Mechanistic insights into seeded growth processes of gold nanoparticles

PAPER www.rsc.org/nanoscale | Nanoscale

Dow

nloa

ded

on 0

1 A

pril

2011

Publ

ishe

d on

29

Sept

embe

r 20

10 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C0N

R00

541J

View Online

Mechanistic insights into seeded growth processes of gold nanoparticles†

J€org Polte,a Martin Herder,a Robert Erler,a Simone Rolf,a Anna Fischer,b Christian W€urth,a

Andreas F. Th€unemann,a Ralph Kraehnertb and Franziska Emmerling*a

Received 27th July 2010, Accepted 18th August 2010

DOI: 10.1039/c0nr00541j

A facile approach for the synthesis of monodisperse gold nanoparticles with radii in the range of 7 to

20 nm is presented. Starting from monodisperse seeds with radii of 7 nm, produced in the first step, the

addition of a defined amount of additional precursor material permits distinct size regulation and the

realization of predicted nanoparticle sizes. These information were derived from ex- and in situ

investigations by comprehensive small angle X-ray scattering (SAXS), X-ray absorption near edge

structure (XANES) and UV-Vis data to obtain information on the physicochemical mechanisms. The

obtained mechanisms can be transferred to other seeded growth processes. Compared to similar

approaches, the presented synthesis route circumvents the use of different reducing or stabilizing

agents. The size of resulting nanoparticles can be varied over a large size range presented for the first

time without a measurable change in the shape, polydispersity or surface chemistry. Thus, the resulting

nanoparticles are ideal candidates for size dependence investigations.

1. Introduction

Nanoparticles are intensively investigated due to their wide range

of potential applications in fields such as medicine,1 biotech-

nology,2 and catalysis.3 Among these, gold nanoparticles (GNP)

are the most prominent nanoscale materials. GNPs can be

prepared via various synthesis routes including chemical, sono-

chemical or photochemical paths.4 The most common route is

the reduction of a dissolved gold precursor, e.g. (tetrachloroauric

acid, HAuCl4), by a reducing agent such as sodium citrate,

ascorbic acid, sodium boron hydride or block-copolymers. A

further stabilizing agent is typically required to prevent ag-

glomeration or further growth of the particles. In many appli-

cations the properties of metallic nanoparticles are determined

by parameters such as size, shape, composition or crystalline

structure. In principle, it is possible to adjust the behaviour of

nanoparticles in such applications in the desired manner by

controlling one of the listed properties.

Control over the size of GNPs can be achieved by varying the

synthesis parameters, e.g. the concentration of the precursor and

reducing agent or the temperature (see ref 4 and references therein).

A precise size control is often difficult to achieve, since small

changes in the above-mentioned parameters often lead to

pronounced differences in the size distribution. In most cases it is

also difficult to keep the polydispersity low when changing the

reaction parameters in order to obtain larger particles. A prominent

example for challenges faced in controlling particle-size distribution

aBAM Federal Institute for Materials Research and Testing, RichardWillst€atter-Straße 11, 12489 Berlin, Germany. E-mail: [email protected]; Fax: +00 (0)49 30 8104 1137; Tel: +00 (0)49 308104 1133bTechnische Universit€at Berlin, Technische Chemie, Straße des 17. Juni124, 10623 Berlin, Germany

† Electronic supplementary information (ESI) available: Additionaldata. See DOI: 10.1039/c0nr00541j

This journal is ª The Royal Society of Chemistry 2010

is the frequently employed synthesis of gold nanoparticles by

reduction of HAuCl4 with sodium citrate.5 For this GNP synthesis,

Frens et al.5 demonstrated control over particle size by varying the

citrate concentration; hence the synthesis is referred to as ‘‘Frens

method’’.6–8 Based on this method, several approaches of changing

the initial reaction parameters were developed, but none of them

could assure a specific particle size, spherical shape and the same

low polydispersity concurrently.5,6,9,10 Unfortunately, these charac-

teristics are required for applications of GNPs in which the size

effect is utilized, e.g. for well-defined surface plasmon resonances in

chemical and biomedical detection and analysis.11

An alternative approach to overcome such difficulties is

‘‘seeded growth’’, the use of previously prepared small particles,

which act as seeds for a further particle growth. For precise

control of the particle size it is necessary to suppress further

nucleation events during seed-mediated growth and to promote

heterogeneous growth. However, seed-mediated growth is pre-

dominantly used for controlled growth of particles of distinct

elongated shapes.12–16 To the best of our knowledge, an addi-

tional stabilizing and/or reducing agent has been required in the

previously reported studies to control particle size and shape

during seeded growth.4,6–8,10,17–21

In the following, a procedure for a size-controlled, seed-

mediated growth of spherical GNPs is presented. It solely relies

on the reduction of HAuCl4 with sodium citrate, without addi-

tional stabilizing or reducing agents, thus in the following this

approach is referred to as ‘self-seeded’ growth. Hence, side re-

actions inducing further nucleation can be prevented. By

applying this novel procedure, the size of the GNPs can be

adjusted while the spherical shape, as well as the low poly-

dispersity and the surface chemistry is retained. The precision

of size control depends solely on the analytic determination of

the particle sizes, as is demonstrated by using the two comple-

mentary methods SAXS and SEM. The data obtained from

in situ SAXS, XANES and UV-Vis analysis throughout the

synthesis provides information about the underlying growth

Nanoscale, 2010, 2, 2463–2469 | 2463

http://dx.doi.org/10.1039/c0nr00541j

Dow

nloa

ded

on 0

1 A

pril

2011

Publ

ishe

d on

29

Sept

embe

r 20

10 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C0N

R00

541J

View Online

mechanism. The potential of the concept is demonstrated for the

synthesis of monodisperse GNPs with adjustable radii ranging

from 7 to 20 nm.

2. Experimental

2.1 Nanoparticle synthesis

Gold nanoparticles were synthesized according to the proce-

dure described by Turkevich et al.,22 i.e. chemical reduction of

the gold precursor HAuCl4 by dissolved trisodium citrate at

75 �C from aqueous solutions containing 0.25 mmol/L and

2.5 mmol/L of gold precursor and citrate, respectively. Prior to

each experiment 35 mL aqueous (Millipore) solution of gold

precursor (7 mg HAuCl4 � 3H2O, Aldrich) and 35 mL aqueous

solution of trisodium citrate (51.45 mg Na3C6H5O7 � 2 H2O,

Aldrich) were prepared and pre-heated to the reaction

temperature (75 �C). The synthesis was carried out under stir-

ring in a flask immersed in a temperature-controlled water bath,

adding the citrate solution to the gold solution. Self-seeded

growth of GNPs at 65 �C was initiated by adding volumes of

1, 2 or 5 ml of aqueous solutions of HAuCl4 (1 mmol/L) to

10 mL of the seed solution under stirring, after adding the

respective 10-fold amount of citrate to the seeds. The amount of

gold salt nadd needed to obtain a specific gold nanoparticle

radius can be simply determined from stoichiometry as nadd ¼(r3

final � r3initial)/r

3initial � ninitial, where ninitial is the molar amount of

gold in the nanoparticle stock solution, and rinitial and rfinal are

the initial and final radii. Liquid samples were extracted by

pipetting for SAXS/XANES, UV analysis or pH measurement

(5.8 to 6.2). For SEM imaging, ca. 0.1 mL of the final colloid

was dried on a Si wafer.

2.2 Nanoparticle characterization

At different reaction times, 10 ml of the liquid samples were

extracted from the batch of reaction solution and approx. 5 ml

were placed as droplets in an acoustic levitator (Tec5, Oberursel,

Germany) used as sample holder for X-ray analysis. The acoustic

levitator was positioned in the mSpot beamline at BESSY II

synchrotron (Berlin, Germany) as described in ref 23 to measure

time-resolved combined SAXS and XANES (for details see

Supporting Information, S1). The in-house scattering experi-

ments were carried out using the Kratky camera type system

SAXSess (Anton Paar, Graz, Austria).

Scanning electron microscopy (SEM) imaging of desiccated

nanoparticles was performed on a JEOL JSM-7401F instrument,

operated with an acceleration voltage of 4 kV, at a working

distance of 2.8 mm, covering the Si wafer that carried the

nanoparticles with a gold mesh to reduce surface charging. The

size determination was performed using Image J software (http://

rsbweb.nih.gov/ij/). For each sample three SEM images from

non overlapping areas were taken. From these images about

200 particles were measured in size to provide statistical signifi-

cance. UV absorption (UV-Vis) spectra were recorded on an

Avantes AvaSpec-2048TEC-2 (Deuterium halogen light source),

connected to a 10 mm optical path length cuvette holder via fibre

optical cables.

2464 | Nanoscale, 2010, 2, 2463–2469

3. Results and discussion

The concept of size-controlled GNP synthesis extends our

previous work,24 which focussed on understanding the nucle-

ation and growth process of monodisperse gold nanoparticles

synthesized according to the Turkevich method.22 Briefly, it was

shown that the growth mechanism consists of four steps. Starting

with a rapid reduction of the gold precursor and nucleation, the

process is followed by coalescence (resulting in a GNP radius of

about 4 nm) and a diffusional growth step (rGNP ¼ 5 nm). The

last growth step is a rapid and complete consumption of the

remaining gold precursor retaining the number of particles as

well as constant polydispersity, i.e. without any further nucle-

ation. An autocatalytic surface reduction was proposed as the

underlying process.

These previous results suggest that the final size of the GNPs

obtained by applying the synthesis procedure described above is

determined by two factors, the number and size of particles

formed initially due to nucleation and coalescence, as well as the

remaining quantity of precursor material available in the final

step of particle growth. Thus, adding additional amounts of gold

precursor to the colloidal solution during or after the GNP

formations final step should increase the particle size, whereas

the particle number and polydispersity should remain constant.

According to the proposed mechanism the autocatalytic surface

growth will continue until the added gold precursor is completely

consumed and deposited onto the existing particles.

To prove this hypothesis, a coupled SAXS and XANES in situ

experiment was performed using the same setup and data eval-

uation procedure previously reported.24 In this context SAXS

delivers information about the size, shape, and number of

particles. The potential of this method for growth studies is

demonstrated in several recent contributions.12,24–31 XANES

provides information on the oxidation state of the gold atoms in

the solution and thus on the progress of the reduction process.

The proof of principle of such self-seeded growth was tested for

the citrate reduction at 85 �C, a temperature close to the condi-

tions for the typical Turkevich synthesis.22 The standard syn-

thesis for GNPs was carried out until completion. Thereafter, an

additional amount of gold precursor solution (HAuCl4) was

added to induce self-seeded particle growth.

The time-dependent results of the evaluated respective SAXS

and XANES data are provided in Fig. 1. The radii and number of

particles (a) as well as polydispersities (b) were calculated from

fitted SAXS curves. The inset in (b) exemplarily shows the size

distributions of the GNPs at different reaction times, including

seed formation and particle growth processes. Oxidation states

and volume fractions derived from XANES are plotted in (c).

Both SAXS and XANES show the expected phases of formation

and growth of the seeds being completed within ca. 40 min.

By adding an additional amount of 25% of the gold precur-

sor (with respect to the starting solution) to the reaction solu-

tion, a rapid increase in the particle radius from 7.9 nm to

8.5 nm was observed (Fig. 1a) whereas the polydispersity re-

mained constant (Fig. 1b). Correspondingly, the total volume

of the gold particles increases by approx. 25%, suggesting

complete conversion of the added precursor. Whereas the total

number of GNP remained constant, the number of particles per

unit volume decreased due to the dilution resulting from the



Fig. 1 Results of the evaluated SAXS and XANES data as a function of time, recorded for the synthesis of gold seeds (at 85 �C) via citrate method (0 to

45 min) and the addition of further gold precursor solution to the seed solution (45 min). Significant points of the experiments are indicated by different

colours: start (red), first phase (black), before (blue) and after the (green) addition of further precursor material. a) The evaluation of the mean radius and

the normalized number of particles as a function of time. b) Polydispersity of particles as a function of time and the corresponding particle size

distribution of the Schulz-Zimm distribution exemplarily shown for four reaction times (inset). c) Average formal oxidation state of gold derived from

normalized XANES spectra plotted vs. reaction time; volume fraction of particles calculated from SAXS data.

Fig. 2 SAXS data and data evaluation for the stepwise growth of GNP

at 65 �C via addition of further reactants (Na3Ct, HAuCl4): a) SAXS data

as a function of the precursor addition obtained during the reduction of

HAuCl4 with sodium citrate. The SAXS data indicate the formation of

spherical nanoparticles with a narrow size distribution. b) Data evalua-

tion showing examples for measured vs. fitted scattering curves for

different particle radii obtained after addition of corresponding quanti-

ties of precursor ranging from GNP seeds (bottom, 6.8 nm radius) to

a radius of 20.8 nm (top).

Dow

nloa

ded

on 0

1 A

pril

2011

Publ

ishe

d on

29

Sept

embe

r 20

10 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C0N

R00

541J

View Online

addition of more solvent (Fig. 1a). For the same reason the

volume fraction of gold particles increases only by about 13%

and not by 25% as would be expected for a constant total

volume of solvent (Fig. 1c). The XANES results confirm the

rapid reduction as well as the full stoichiometric conversion of

the gold salt (see Fig. 1c). The results indicate that further

nucleation and formation of GNPs can be suppressed by

applying the described synthesis procedure. Consequently, the

final size of the particles can be regulated by the controlled

addition of further precursor material.

To provide further evidence for GNP-accelerated reduction of

the added gold precursor, the same experiment (0.06125 mmol/L

HAuCl4, 2.5 mmol/L sodium citrate) was carried out at 65 �C

with and without nanoparticles being present in the solution (for

corresponding UV-vis results see ESI (S2)†). Whereas complete

reduction of gold precursor requires less than 15 min when GNP

are present, more than 200 min are required in the absence of

already formed GNP. Hence, the consumption of gold precursor

is significantly accelerated in presence of colloidal GNP.

Further experiments using SAXS as the analytical technique

were performed to test whether it is possible to increase the

radius of the GNPs step by step by adding additional amounts of

gold-precursor solution. The seeds were synthesized at 75 �C

employing a slight modification of the standard procedure

described by Turkevich et al.22 to obtain particles with a radius of

approximately 7 nm. The subsequent self-seeded growth was

carried out at 65 �C. At this temperature the reduction rate is

slowed down and therefore the probability of nucleation events is

minimized. Since sodium citrate acts both as reducing and

stabilizing agent, it is important to keep the molar concentration

of sodium citrate constant at 2.5 mmol throughout the growth

process to avoid the aggregation of the gold particles.

Thus, proceeding in a stepwise manner, first 1 mL, then 2 mL

and finally 5 mL of HAuCl4 solution (1 mmol) were added slowly

to 10 mL of the seed solution. For each addition of HAuCl4solution the same volume of citrate solution (5 mmol/L) was

added. This alternating addition was repeated until the particles

reached the desired size. Between two additions, a time span of

approximately 5 min was found to be sufficient to ensure the

complete conversion of the gold salt.


In Fig. 2a the evolution of the scattering curves during the

growth period is displayed (SAXS measured at the synchrotron

setup). The integration time for each scattering curve was 30 s

and every 3 min a fresh sample was taken for SAXS analysis.

Further gold salt was added after the particles reached the

calculated radius. An interval of at least 5 min was allowed

between two additions of aliquots of gold precursor. As evident

from Fig. 2a the shape of the scattering curves persists while the

curves are shifted to smaller q-values with addition of further

reactants. This indicates without mathematical modelling that

the particles grow while retaining their spherical shape and low

polydispersity. Fig. 2b illustrates the good agreement of five

selected scattering curves (black) and their respective mathe-

matical fits (red). The fit assumes spherical particles having

a Schulz size distribution,32 which was found to be a good

approximation in the previous analysis of the growth process of

citrate stabilized GNPs.24

For the stepwise self-seeded growth of GNPs from 6.8 to

20.8 nm radius, the results of fitted scattering curves are dis-

played in Fig. 3. Each vertical blue line corresponds to the

Nanoscale, 2010, 2, 2463–2469 | 2465


Fig. 3 Evaluation of radius, particle number and volume fraction as

a function of time for the self-seeded growth of GNPs at 65 �C, with

repeated injection of additional reactant solution (Na3Ct, HAuCl4). Each

vertical blue line indicates the addition of the respective volume of

precursor material. Dashed lines correspond to the addition of a volume

of 1 mL, dotted lines the addition of 2 mL and solid lines the addition of 5

mL of gold precursor solution. a) Particle radius vs. time: the values for

the radii derived from SAXS (squares) are in excellent agreement with the

radii predicted by the stoichiometry of precursor addition, assuming

complete conversion (crosses). b) The number of particles derived from

SAXS (circles) remains nearly constant. The volume fraction of GNP

normalized to the initial reaction volume (triangles) increases continu-

ously with further addition of precursor solution.

Fig. 4 SEM images from different stages of the self-seeded growth of

GNP at 65 �C: GNP seeds (a), after the addition of 30 mL of precursor

(b), and after the addition of 65 mL of precursor (c). Scale bars: 50 nm.

The insets show the size distribution of the particles (Schulz size distri-

bution) as used for the analysis of the SAXS data.

Dow

nloa

ded

on 0

1 A

pril

2011

Publ

ishe

d on

29

Sept

embe

r 20

10 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C0N

R00

541J

View Online

addition of 1, 2 or 5 mL of 1 mmol L�1 gold precursor solution to

the seed colloid. In Fig. 3a the ‘‘theoretical radius’’ calculated

from stoichiometry and the measured radius (‘‘experimental

radius’’) for each step of the addition is given. Fig. 3b shows the

corresponding number of particles, with the dilution due to the

addition of solvent being corrected for and the volume fraction.

Upon adding HAuCl4 solution to the gold seeds, the particle

radius increases exactly as predicted by the reaction’s stoi-

chiometry (Fig. 3a), whereas the number of particles (Fig. 3b)

and their polydispersity remained constant at a value of 13%

throughout the experiment. A small decrease in the number of

particles of about 5% is observed as the particle radius exceeds

18 nm. This might be due to precipitation of some particles onto

the stirrer or the glass wall (not directly observed), and related to

gradually failing stabilisation of the GNPs via citrate. As clearly

seen in Fig. 3a the calculated theoretical radius agrees very

well with the particle radius derived experimentally from SAXS.

Hence, the repeatedly added gold precursor is completely re-

duced and grown within 3 to 5 min onto the existing particles.

The SAXS data obtained on the synchrotron for all final

samples (Fig. 2, Fig. 3) were confirmed experimentally using

a lab-scale SAXS instrument (SAXSess, Anton Paar GmbH,

Graz, Austria), which allows measurements at lower q-values

and thus a more precise size determination of particles with radii

>15 nm (see Fig. 5).

GNP size and polydispersity deduced from SAXS evaluation

were also confirmed by scanning electron microscopy (SEM).

Fig. 4 shows SEM images of three different steps of the self-

seeded growth process, as well as histograms indicating the

particle size distribution derived from SEM images. The SEM

images confirm that while the particle radius increases from 6.2

to 20.4 nm, the predominantly spherical shape of the particles is

preserved. The formed GNPs retain a low polydispersity of less

than 13%. TEM images of particles synthesized with the method

mentioned above are in accordance with the SEM and SAXS

results (for details see ESI (S4)†).

2466 | Nanoscale, 2010, 2, 2463–2469

The GNP radius and polydispersity derived from SAXS and

SEM results reveals that the particle sizes determined with the

two methods agree to within 4 to 5%. The direct comparison of

particle radii obtained via SAXS at the synchrotron and in the

lab as well as from SEM is displayed in Fig. 5, plotting experi-

mental values vs. theoretically calculated radii. The observation

that the data lies almost perfectly on the angle bisector confirms

the high accuracy of the SAXS analysis. The results illustrate the

capabilities of SAXS as a reliable tool for size determination of

nanoparticles. Furthermore, SAXS allows online monitoring of

nanoparticle growth, especially when in situ experiments are

needed or further sample preparation is undesirable. However it

should be noted that with increasing heterogeneity in the shape

of the particles the information obtained from SAXS is limited

and thus electron microscopy is in general the preferred choice.

The seed-mediated growth process presented here meets the

criteria for a SAXS analysis in particular, since the formed

particles exhibit a spherical shape.



Fig. 5 Comparison of the experimentally determined average GNP radii

obtained from SAXS (synchrotron-based and SAXSess) as well as SEM,

plotted as a function of the calculated theoretical radius. The agreement

between experimental and theoretical radii is indicated by the straight

parity line.

Dow

nloa

ded

on 0

1 A

pril

2011

Publ

ishe

d on

29

Sept

embe

r 20

10 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C0N

R00

541J

View Online

The results derived from SAXS, XANES, and SEM are in

agreement with the previously suggested interpretation of the

growth step as autocatalytic reduction. However, further nucle-

ation may occur but these small gold clusters/particles can

undergo coalescent processes, which would be in accordance

with recent findings in the NaBH4 system, where the growth of

GNP occurs exclusively due to coalescence in the absence of

a stabilizing agent.25 At first glance it is surprising that SAXS

does not indicate a temporary increase in the particle number,

Fig. 6 (a) Color change of the GNP solution due to the addition of furthe

obtained during GNP growth from seeds with radii of 7.2 nm recorded at diffe

absorbance at 600 nm as a function of the reaction time. (c) The wavelength

process. The inset depicts the corresponding absorbance at the maximum.


but in fact this is not contradictory. When one considers that

signal intensity in SAXS analysis scales proportional to the sixths

power of the particle radius,33 it is practically impossible to detect

gold clusters with radii smaller that a 1 nm in the presence of

GNPs with a radius larger than 7 nm, since the intensity ratio

between new formed cluster and GNP seeds would be in the

order of 1 to 105.

However, a further nucleation process should be apparent

from the optical properties of the GNPs (see Fig. 6), since the

UV-Vis spectra of particles with radii of about 1 nm differ clearly

from the seeds.34 After addition of HAuCl4 to the seeds, the

solution colour turned from wine red to violet within seconds.

Subsequently, it returns to wine red within 5 to 10 min (see

Fig. 6a), which is in agreement with the time span of a self-seeded

growth step. Thus, the seed-mediated GNP growth was moni-

tored using UV-Vis spectroscopy with a time resolution of 10 s.

The resulting spectra, the absorbance values and the wavelength

shift of the GNPs plasmon resonance are displayed in Fig. 6b

and c for GNPs growing 3 monolayers with a thickness of 0.3 nm

(the thickness of a monolayer is based on the dimension of the

unit cell) onto the seed particles thus from a starting radius of

7.2 nm to 8.1 nm determined with SAXS. It can be seen that the

maximum in the absorption spectrum shifts from 518 nm to

530 nm within 15 s after adding an aliquot of gold precursor

(black and red curve in Fig. 6b), i.e. the colour changes from wine

red to violet. Within the next 5 to 15 min the location of the

plasmon resonance band returns to the initial position of 518 nm

(blue curve in Fig. 6b). Fig. 6c displays in a time-resolved manner

the corresponding shift of the band as well as the absorbance at

r precursor and the subsequent self-seeded growth. (b) UV-Vis spectra

rent time intervals after adding an aliquot of gold salt. The inset shows the

shift at the absorbance maximum with respect to the time of the reaction

Nanoscale, 2010, 2, 2463–2469 | 2467


Dow

nloa

ded

on 0

1 A

pril

2011

Publ

ishe

d on

29

Sept

embe

r 20

10 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C0N

R00

541J

View Online

the peak maximum position (inset in Fig. 6c). Hence, the UV-Vis

data confirm the results from SAXS analysis that the seeded

growth process requires about 5–10 min for completion.

Unfortunately, obtaining quantitative information from UV-

Vis is only possible for GNPs larger than 5 nm in diameter.34

Thus the spectra of small gold clusters can only provide quali-

tative information. The initially observed shift towards higher

wavelength in the absorption spectrum can hardly be explained

by the existence of small gold cluster/particles in the vicinity of

the larger GNP seeds. This change in the absorption matches

neither the UV-Vis spectra measured in the beginning of the

standard citrate synthesis when predominantly very small

particles are present (radius z 1 nm) nor other GNP synthesises

in which the final particle size is small.25

Apart from the presence of additional small GNPs, a plausible

explanation for the observed reversible change of the plasmon

resonance band towards higher wavelengths might be the

temporary decrease in the pH of the reactant solution upon

adding the acidic precursor solution. It has been reported that

decreasing the pH of GNP colloids leads to a shift of the plasmon

band towards higher wavelengths.35,36 In fact, upon adding the

acidic precursor solution (pH ¼ 4), the solutions pH should

decrease. After the complete consumption of the gold precursor,

the pH would return to its initial value. However, it has been also

reported that sodium citrate acts as buffer, so that only small

changes of the pH are observed throughout the reduction/growth

process. A sufficiently accurate experimental determination of

the present pH value did not succeed. That colour changes result

from temporary aggregation of large GNPs, as possibly sug-

gested by the shift of the plasmon resonance towards higher

wavelengths, can be excluded on the basis of the SAXS data: no

aggregates were observed, although even small numbers of

aggregates would produce a strong increase of the SAXS inten-

sity at low q-values. To support the interpretation of the

temporary colour change being pH related, but not due to

aggregation, the addition of HAuCl4 solution to a GNP colloid

was repeated at room temperature complemented by simulta-

neously UV-vis, SAXS and pH analysis using indicator paper.

The described colour change and the shift in the UV-vis spectrum

towards higher wavelengths is reproducibly observed, whereas

SAXS curves did not show any changes upon adding HAuCl4,

solution (see ESI (S3)†). Hence, the particle size distribution did

not change upon adding gold precursor. Thus, the temporary

shift of the GNP plasmon resonance cannot be explained by

aggregation of the present GNP.

The presented results on self-seeded GNP growth are in

agreement with previous publications on seeded growth in which

the addition rate of further metal precursor was discussed as

crucial factor.8,17,37 Jana et al.8 showed in detail for a common

method for seeded growth of gold nanoparticles (GNP seeds

prepared via citrate method, whereas for further growth ascorbic

acid is used as a reducing agent) that a heterogeneous growth and

thus a precise size control is only possible when adding small

amounts of gold precursor step by step to the seed solution, thus

avoiding further nucleation and preserving the spherical shape

and low polydispersity.

A seed mediated growth in which the addition rate had no

influence for a heterogeneous seeded growth was described by

Niu et al.38 They described a ‘one pot’ GNP seeded growth

2468 | Nanoscale, 2010, 2, 2463–2469

synthesis with citrate-reduced and -stabilized seeds, where the

success of the heterogeneous growth shows that it is independent

of the addition rate. The further growth of the seeds was carried

out using 2-mercaptosuccinic acid (MSA) which acts as reducing

and stabilizing agent. The authors presume a surface enhanced

reduction of HAuCl4, based on the gold affinity of the mercapto

groups that was derived because the MSA doesn’t reduce the

HAuCl4 in the absence of the seeds. Their findings are in

accordance with our investigations since the surface reduction in

their case is much stronger than the reduction in solution. The

authors state that the UV-Vis spectra for the growth processes do

not show any aggregation.

As a consequence of this study, and the comparison to the

recent literature, it appears that successful seeded growth

procedures (heterogeneous growth) are primarily governed by

the addition rate, which is depending or influenced by mainly two

factors. The first factor is the relation of the surface reduction

rate and the reduction rate in solution separate from any other

particle. The second influencing factor is the need for the pres-

ervation of the nanoparticle stabilization in each step of addition.

4. Conclusions

The present study shows that from a comprehensive under-

standing of the mechanism of particle growth based on time-

resolved in situ studies, it is possible to derive a synthesis strategy

for size-controlled self-seeded growth of gold nanoparticles.

Analysis by SAXS, SEM and UV-Vis confirm that the radius of

GNP can be precisely adjusted between 7 and 20 nm in GNP

synthesis from HAuCl4 and citrate, by adding incremental

amounts of reactants in the absence of additional stabilizing or

reducing agents. These GNP are ideal candidates for size

dependence investigations. Moreover, the mechanistic results

imply that rapid reduction of added precursor in presence of seed

particles is a crucial factor for a heterogeneous seeded growth.

Comprehensive investigation of related nanoparticle-synthesis

procedures with the employed analytical tools will show whether

the presented self-seeded growth procedure and synthesis

strategy also applies to the formation of other noble-metal

nanoparticles.

Acknowledgements

R.K. acknowledges generous funding from the BMBF within the

frame of the NanoFutur program (FKZ 03X5517A).

Notes and references

1 J. A. Copland, M. Eghtedari, V. L. Popov, N. Kotov, N. Mamedova,M. Motamedi and A. A. Oraevsky, Mol. Imaging Biol., 2004, 6, 341–349.

2 J. J. Storhoff, R. Elghanian, R. C. Mucic, C. A. Mirkin andR. L. Letsinger, J. Am. Chem. Soc., 1998, 120, 1959–1964.

3 G. C. Bond, C. Louis and D. T. Thompson, Catalysis by Gold,Imperial College Press 2006.

4 M. C. Daniel and D. Astruc, Chem. Rev., 2004, 104, 293–346.5 G. Frens, Nature-Physical Science, 1973, 241, 20–22.6 S. D. Perrault and W. C. W. Chan, J. Am. Chem. Soc., 2009, 131,

17042.7 K. R. Brown, D. G. Walter and M. J. Natan, Chem. Mater., 2000, 12,

306–313.8 N. R. Jana, L. Gearheart and C. J. Murphy, Chem. Mater., 2001, 13,

2313–2322.



Dow

nloa

ded

on 0

1 A

pril

2011

Publ

ishe

d on

29

Sept

embe

r 20

10 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C0N

R00

541J

View Online

9 G. Frens, Phys. Lett. A, 1973, 44, 208–210.10 X. H. Ji, X. N. Song, J. Li, Y. B. Bai, W. S. Yang and X. G. Peng,

J. Am. Chem. Soc., 2007, 129, 13939–13948.11 A. M. Schwartzberg and J. Z. Zhang, J. Phys. Chem. C, 2008, 112,

10323–10337.12 A. Henkel, O. Schubert, A. Plech and C. Sonnichsen, J. Phys. Chem.

C, 2009, 113, 10390–10394.13 H. A. Keul, M. Moller and M. R. Bockstaller, Langmuir, 2007, 23,

10307–10315.14 P. H. Qiu and C. B. Mao, J. Nanopart. Res., 2009, 11, 885–894.15 T. K. Sau and C. J. Murphy, Langmuir, 2004, 20, 6414–6420.16 Y. Xia, Y. J. Xiong, B. Lim and S. E. Skrabalak, Angew. Chem., Int.

Ed., 2009, 48, 60–103.17 N. R. Jana, L. Gearheart and C. J. Murphy, Langmuir, 2001, 17,

6782–6786.18 N. R. Jana, L. Gearheart and C. J. Murphy, Adv. Mater., 2001, 13,

1389–1393.19 J. L. Niu, T. Zhu and Z. F. Liu, Nanotechnology, 2007, 18.20 J. Rodriguez-Fernandez, J. Perez-Juste, F. J. G. de Abajo and

L. M. Liz-Marzan, Langmuir, 2006, 22, 7007–7010.21 T. Zhu, K. Vasilev, M. Kreiter, S. Mittler and W. Knoll, Langmuir,

2003, 19, 9518–9525.22 B. V. Enustun and J. Turkevich, J. Am. Chem. Soc., 1963, 85, 3317.23 O. Paris, C. H. Li, S. Siegel, G. Weseloh, F. Emmerling,

H. Riesemeier, A. Erko and P. Fratzl, J. Appl. Crystallogr., 2006,40, s466–S470.

24 J. Polte, T. T. Ahner, F. Delissen, S. Sokolov, F. Emmerling,A. F. Thunemann and R. Kraehnert, J. Am. Chem. Soc., 2010, 132,1296–1301.


25 J. Polte, R. Erler, A. F. Thunemann, S. Sokolov, T. T. Ahner,K. Rademann, F. Emmerling and R. Kraehnert, ACS Nano, 2010,4, 1076–1082.

26 M. Eichelbaum, K. Rademann, A. Hoell, D. M. Tatchev,W. Weigel, R. Stosser and G. Pacchioni, Nanotechnology, 2008,19, 135701.

27 B. Abecassis, F. Testard, O. Spalla and P. Barboux, Nano Lett., 2007,7, 1723–1727.

28 J. Polte, F. Emmerling, M. Radtke, U. Reinholz, H. Riesemeier andA. F. Thunemann, Langmuir, 2010.

29 T. Morita, E. Tanaka, Y. Inagaki, H. Hotta, R. Shingai,Y. Hatakeyama, K. Nishikawa, H. Murai, H. Nakano andK. Hino, J. Phys. Chem. C, 2010, 114, 3804–3810.

30 G. Sankar and W. Bras, Catal. Today, 2009, 145, 195–203.31 B. Abecassis, F. Testard and O. Spalla, Phys. Rev. Lett., 2008, 100,

115504.32 M. Kotlarchyk, R. B. Stephens and J. S. Huang, J. Phys. Chem., 1988,

92, 1533–1538.33 O. Glatter and O. Kratky, Small Angle X-ray Scattering, Academic

Press, London, 1982.34 W. Haiss, N. T. K. Thanh, J. Aveyard and D. G. Fernig, Anal. Chem.,

2007, 79, 4215–4221.35 Y. Shiraishi, D. Arakawa and N. Toshima, Eur. Phys. J. E, 2002, 8,

377–383.36 A. J. Viudez, R. Madueno, T. Pineda and M. Blazquez, J. Phys.

Chem. B, 2006, 110, 17840–17847.37 S. M. Humphrey, M. E. Grass, S. E. Habas, K. Niesz, G. A. Somorjai

and T. D. Tilley, Nano Lett., 2007, 7, 785–790.38 J. L. Niu, T. Zhu and Z. F. Liu, Nanotechnology, 2007, 18, 7.

Nanoscale, 2010, 2, 2463–2469 | 2469


Mechanistic insights into seeded growth processes of gold nanoparticles

Documents