Absorption spectrum of the trimer silver cluster Ag3 2+ and metal nanoparticles in microemulsion

Article abstract of DOI:10.1016/j.cplett.2004.08.080

Classical quantum simulation suggests the formation of Ag₃ ²⁺ [Chem. Phys. Lett. 389 (2004) 150]. We present the effect of water pool radii on the transient absorption spectra of initial silver clusters. Experimental evidence for the

Full text of DOI:10.1016/j.cplett.2004.08.080

Chemical Physics Letters 396 (2004) 415–419
www.elsevier.com/locate/cplett
Absorption spectrum of the trimer silver cluster Ag32þ
and metal nanoparticles in microemulsion
*
Sudhir Kapoor , Ravi Joshi, Tulsi Mukherjee
Radiation Chemistry and Chemical Dynamics Division, Bhabha Atomic Research Centre, MOD LABS, Mumbai 400 085, India
Received 1 June 2004; in final form 18 August 2004
Abstract
Classical quantum simulation suggests the formation of Ag32þ [Chem. Phys. Lett. 389 (2004) 150]. We present the effect of water
pool radii on the transient absorption spectra of initial silver clusters. Experimental evidence for the formation of Ag23þ in micro-
emulsion system is reported for the first time. The surface plasmon absorption band of stable silver clusters obtained by bombard-
ment of electron pulses in water-in-oil microemulsions showed absorption maximum at 405 nm. The transmission electron
micrograph studies showed the presence of silver nanoparticles having an average size of 10 nm.
ꢀ 2004 Elsevier B.V. All rights reserved.
1. Introduction
coalesce and reform. It requires a change in the surfac-
tant film or curvature of droplets. This process is ther-
modynamically unfavourable and has not been
governed by diffusion-controlled process. Thus, inter-
change of nuclei becomes more difficult as the nuclei
grow.
Pileni and co-workers [4] have used extensively in-
verse micelles and microemulsions for the preparation
of nanoparticles. It was found that if the organic med-
ium was isooctane, spherical size of nanoparticles was
obtained, while replacing isooctane by cyclohexane lead
to the formation of small silver clusters.
Information regarding the kinetics of particle forma-
tion is of critical importance to know the mechanism of
the growth process of the particles. Pulse radiolysis is an
important technique by which the initial processes can
be probed easily. In addition, the role played by the
water pool radii in microemulsion in the growth kinetics
of the metal particles is also important to control the
size and distribution of particles. So far only little infor-
mation is available in the literature on this aspect [5]. In
the present study, the reduction of silver ions was stud-
ied in water/sodium dodecyl sulfate/cyclohexane/1-
pentanol by the pulse radiolysis technique. Formation
A variety of methods to prepare metal nanoparticles
have been reported in the literature. At present the
search of methods, which can give monodisperse nano-
particles, is actively pursued [1–3]. Microemulsions are
currently being explored for making ultrafine particles
of metals and semiconductors for many reasons. The
main reason for using microemulsion in the preparation
of nanoparticles is that they promise a medium of con-
trolling particle size as well as their distributions. This
is because of the reason that they consist of water drop-
lets of nanometer size, which are dispersed in a continu-
ous medium. Microemulsions are used as micro/
nanoreactors because they can exchange the content of
their water pools by a collision process. It has been sug-
gested that the rate of exchange/communication between
two droplets is very fast. It occurs only when an ener-
getic collision between the two droplets is able to estab-
lish a water channel between the droplets or the droplets
*
Corresponding author. Fax: +91 22 25505151.
E-mail addresses: sudhirk@apsara.barc.ernet.in, sk0112@rediffmail.
com (S. Kapoor).
0009-2614/$ - see front matter ꢀ 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.cplett.2004.08.080

416
S. Kapoor et al. / Chemical Physics Letters 396 (2004) 415–419
of metal nanoparticles was characterized by UV–Vis
absorption and transmission electron microscopy.
desiccator before putting them in a specimen holder.
TEM characterization was carried out using a JEOL
JEM-2000FX electron microscope. Particle sizes were
measured from the TEM micrographs. The particle size
was calculated by taking at least 100 particles. Absorp-
tion measurements were carried out on a Jasco V-530
spectrophotometer. The UV–Visible absorption spectra
were recorded at room temperature using a 1 cm quartz
cuvette.
2. Experimental
2.1. Materials
The surfactant sodium dodecyl sulfate (SDS),
obtained from SISCO, India, and spectrograde cyclo-
hexane and 1-pentanol obtained from Spectrochem,
India, were used as received. All other chemicals were
of Analytical Reagent grade or equivalent.
3. Results and discussion
Microemulsions were prepared by adding nanopure
water to SDS, followed by addition of cyclohexane
and 1-pentanol. All microemulsions were bubbled with
N2 prior to irradiation. On irradiation of the microemul-
sion the hydrated electrons are produced via the follow-
ing reactions [8]
2.2. Pulse radiolysis
The pulse radiolysis set-up consists of an electron lin-
ear accelerator (Forward Industries, England) capable
of giving single pulses of 50 ns, 500 ns or 2 ls of 7 MeV
electrons. The pulse irradiates the sample contained in a
1 cm · 1 cm Suprasil quartz cuvette kept at a distance of
approximately 12 cm from the electron beam window,
where the beam diameter is approximately 1 cm. Further
details of the LINAC can be seen elsewhere [6]. An aer-
ated 10ꢀ2 mol dmꢀ3 KSCN solution was used for dosime-
C6H12 , C6Hþ12 þ eꢀ
ð1Þ
ð2Þ
þ
H2Oðwater poolsÞ , eꢀaq; H; OH; H3O
Å
Å
Å
The H and OH radicals produced are scavenged by 1-
pentanol. However, the yield of hydrated electrons in
the microemulsion is lower than that in pure water. It in-
creases as w0 increases (w0 = [H2O]/[SDS]). The details
of the system are given elsewhere [9]. The rate constants
for the reaction of Ag+ ions with eaꢀq were measured at
different w0 to see the effect of water droplet size on
the reactivity of silver ions. The reaction of Ag+ with
eꢀaq was monitored at 700 nm. It was observed that the
bimolecular rate constant for the reaction of eꢀaq with
Ag+ increased as w0 increased. The results are compiled
in Table 1. The variation in rate constants can be ex-
plained on the basis of ionic strength of the solution.
At lower w0, the local concentration of Ag+ is high
and therefore the ionic strength is high. As w0 increases,
both the local concentration of Ag+ and the ionic
strength decrease. Hence the bimolecular rate constant
for the reaction of eꢀaq with Ag+ increases.
try and ðSCNÞꢀ radical was monitored at 475 nm. The
2
absorbed dose per pulse was calculated assuming
Åꢀ
Ge½ðSCNÞ ꢁ ¼ 2:6 ꢂ 10ꢀ4 m2 Jꢀ1 at 475 nm [7]. The dose
2
employed in the present study, unless otherwise stated,
was typically 67 Gy per pulse.
Microemulsions were prepared by adding de-ionized
nanopure water (conductivity 0.06 lS cmꢀ1) obtained
from Barnstead (USA) water system, to SDS, followed
by the addition of cyclohexane and 1-pentanol. Solu-
tions were optically clear after vigorous shaking. All
microemulsions were thoroughly purged with high pur-
ity N2 (>99.99%) prior to irradiation.
In the present study microemulsions of representative
w0 (w0 = [water]/[surfactant]) values 16, 36 and 60 were
used. Hydrated electrons ðeꢀaqÞ and other radicals, such
Å
as HÅ and OH, were generated in the water pools of
microemulsion by the electron pulse. In the microemul-
Å
3.1. Reduction of Ag+ in solutions of different w0
sion, the rate constants of HÅ and OH with the 1-pent-
anol and surfactant are sufficiently high, and hence,
scavenging ofꢀthese radicals occurs. The reaction of surf-
actant with eaq is 4–5 orders of magnitude slower than
Fig. 1 shows the time-resolved absorption spectra of
the transient species obtained at four different times in
N2-bubbled microemulsion solution (w0 = 36) contain-
Å
that with HÅ and OH. Transmission electron micro-
graphs were taken using a JEOL JEM-2000FX electron
microscope.
Table 1
Effect of w0 on the size of silver nanoparticles and bimolecular rate
constants for the reaction of eꢀaq with Ag+
2.3. Characterization
w0
k2 · 10ꢀ8 (molꢀ1 dm3 sꢀ1
)
Size (nm)
Samples for transmission electron microscopy (TEM)
were prepared by putting a drop of the colloidal solution
on a copper grid coated with a thin amorphous carbon
film. Samples were dried and kept under vacuum in a
16
36
60
0.52
0.82
25.2
8
10
20
2
2
2

S. Kapoor et al. / Chemical Physics Letters 396 (2004) 415–419
417
the results obtained in microemulsion with that reported
in aqueous solution [1,10] it can be inferred that the ob-
served two paths are due to the following reactions (Eqs.
3 and 4). This is followed by the decay of the transients
formed.
0.08
0.06
0.04
0.02
0.00
Agþ2 þ Agþ ! Ag23þ
Agþ2 þ Agþ2 ! Ag24þ
ð3Þ
ð4Þ
The spectra in Fig. 1 are not spectra of single species,
but superimpositions of the absorption bands of at least
four species present in different concentrations. Hence,
due to the overlapping of the bands the growth and de-
cay kinetics of the transients could not be determined
quantitatively. Similar results were obtained in micro-
emulsion with w0 = 16.
300
350
λ (nm)
400
Fig. 1. Time-resolved transient absorption spectrum obtained in N2-
bubbled microemulsion solution (w0 = 36) containing 6.0 · 10ꢀ3
mol dmꢀ3 AgClO4. (s) 0.6 ls; (d) 0.8 ls; (j) 2 ls; (I) 16 ls after
the electron pulse. Dose = 40 Gy.
Fig. 2 shows the time-resolved absorption spectra of
the transient species obtained at four different times in
N2-bubbled microemulsion solution (w0 = 60) contain-
ing 2.0 · 10ꢀ4 mol dmꢀ3 Ag+. On comparing the results
obtained in Fig. 2 with that of Fig. 1 it can be seen that
the processes remain similar on changing the w0 value.
The significant difference was observed in the decay
kinetics of the higher silver aggregates. In this micro-
emulsion, the use of 1-pentanol as a co-surfactant results
in an increase in the viscosity of the interfacial region
and in the water core. Therefore, at low 1-pentanol/
cyclohexane ratio the microemulsion system becomes
more rigid [11]. Due to the rigidity of the matrix the ex-
change rate between water droplets decreases. This in
turn should affect the aggregation processes of initial
clusters formed in the droplets. On comparing Figs. 1
and 2 it can be seen that the peaks of the initial clusters
of silver are well-resolved in w0 = 36 as compared to that
ing 6.0 · 10ꢀ3 mol dmꢀ3 Ag+. Time-resolved absorption
spectra were measured up to 410 nm as the contribution
of electron tail increases enormouꢀsly at higher wave-
lengths. It was observed that the eaq absorption at 700
nm decreases substantially on increasing the concentra-
tion of Ag+. Concomitantly the increase in the rate of
355 nm absorption band due to the transient formation
of Ag0 was observed. It can be seen from Fig. 1 that with
time the absorption due to the formation of other tran-
sient intermediates build up at lower wavelengths. It can
be seen from Fig. 1 that as soon as Ag0 is formed it re-
acts with Ag+ to give Ag2þ that shows absorption maxi-
ma at 320 and 340 nm. It is important to mention here
that in aqueous solution the absorption maximum for
Agþ2 is reported at 310 and 320 nm. It appears that in
microemulsion the absorption band of Ag0 remains sim-
ilar to that reported in aqueous solution while that of
Agþ2 shifts towards higher wavelength [1]. The observed
red shift in the absorption maximum could be due to the
change in the polarity of the surroundings. However, as
SDS is negatively charged there also exists a possibility
that Agþ2 is bound to the surfactant layer. Hence, it
may affect its energy levels. Indeed we have observed
the red shift in the absorption spectrum of Agþ2 . It is
known that spectrum of silver particles shifts towards
red on adsorption of organic molecules. This suggests
that the presence of the palisade layer affects the energy
levels of silver clusters. This further shows that the surf-
actant layer provides some complexation to the silver
clusters.
0.04
0.03
0.02
0.01
0.00
280
320
360
400
Very recently, Janata [10] has shown that in aqueous
solution Agþ2 reacts with Ag+ to give Ag23þ, which has
absorption maximum at 260 and 310 nm. It can be seen
clearly in Fig. 1 that 2 ls after the electron pulse a clear
absorption band is observed at 310 nm. On comparing
λ (nm)
Fig. 2. Time-resolved transient absorption spectrum obtained in N2-
bubbled microemulsion solution (w0 = 60) containing 2.0 · 10ꢀ4
AgClO4. (j) 1.3 ls; (d) 2.6 ls; (m) 5 ls; (.) 10 ls; (r) 16 ls after the
electron pulse. Dose = 40 Gy.
M

418
S. Kapoor et al. / Chemical Physics Letters 396 (2004) 415–419
in w0 = 60. As droplet size or w0 increases it acts like a
continuum, that is, bulk water. Due to this the diffusion
of the species is faster. This could be the reason for
observing a distinct absorption band of Ag23þ in
w0 = 36 and 16 (results not shown) as compared to that
in w0 = 60. Very recently, Dubois et al. [12], using theo-
retical calculations, have confirmed the existence of
Ag23þ despite the fact that it has large charge repulsion
and only one valence electron. This could be one of
the reasons for not getting any shift in the transient
nm, which probably correspond to larger aggregates of
Agn, due to out-of-plane dipole resonance. It is impor-
tant to mention here that the silver particles showed al-
most similar stability under inert conditions or when
exposed to air. A representative TEM image of silver
particles is shown in Fig. 3b. The corresponding diffrac-
tion pattern is shown in Fig. 3c, which corresponds to
fcc structure of Ag. It can be seen that the particles were
spherical in shape having an average diameter of around
10 nm.
spectrum of Ag23þ
.
3.4. Alloy of Cd and Ag
3.2. Stabilization of silver clusters in the presence of low
silver ions
One-electron reduction potentials of Cd2+/Cd+ and
Ag+/Ag0 are close to ꢀ1.8 V vs. NHE [1]. Therefore,
there exists a possibility that the two may form alloy
rather than core–shell type of particles. Kinetic traces
obtained for N2-bubbled microemulsion solution
(w0 = 36) at 360 nm show that electron transfer from
Cd+ to Ag+ takes place under the conditions where
selectively Cd2+ is reduced by eaꢀq (Fig. 4a).
For the formation of alloy it is necessary that the
reduction of both metal ions takes place simultaneously.
Therefore, the equal concentration of metal ions was ta-
ken for making the mixed particles. On bombardment
with trains of 500 ns electron pulses in N2-bubbled
microemulsion solution (w0 = 36) containing 1.0 · 10ꢀ3
mol dmꢀ3 Cd2+ and 1.0 · 10ꢀ3 mol dmꢀ3 Ag+, black
precipitate was observed. TEM studies revealed the
presence of particles of the size of around 25 nm (Fig.
4b). On comparing with Fig. 3b it can be seen that par-
ticle size has increased in the case of mixed particles.
Diffraction pattern of a particular particle reveals the
presence of two different types of metals (Fig. 4b). It is
known that studies of TEM and diffraction methods
can be used to determine the change in lattice parame-
ters that results from the alloying process [13]. However,
the mechanism of alloying of small particles is still not
The time-resolved absorption spectrum obtained on
reduction of silver ions on pulse irradiation by 2 ls elec-
tron pulse, dose 140 Gy, of N2 -bubbled microemulsion
solution (w0 = 60) containing 2.5 · 10ꢀ5 mol dmꢀ3 Ag+.
The concentration of eꢀaq produced by a 2 ls electron
pulse was 10ꢀ5 mol dmꢀ3. Thus, eaꢀq has reacted with
all or nearly all of the Ag+ ions present in the solution.
It was observed that the stabilization of silver clusters do
not take place in the absence of excess silver ions and no
increase in optical absorbance due to the formation of
other clusters of silver was observed within the time win-
dow of our experimental conditions.
3.3. Formation and stabilization of silver nanoparticles
Long-lived clusters were prepared by reducing Ag+
ions by trains of 500 ns electron pulses in N2-bubbled
microemulsion solution containing 1.0 · 10ꢀ3 mol dmꢀ3
Ag+ at different w0. Fig. 3a shows a representative opti-
cal absorption spectra obtained after irradiation by elec-
tron pulses in w0 = 36. None of the bands observed at
short time scales could be observed at these slow time
scales. However, we do observe a broad band at 410
Fig. 3. (a) Absorption spectrum of silver metal sol obtained, on irradiation (dose = 0.7 kGy), in N2-bubbled microemulsion (w0 = 36) solution
containing 1.0 · 10ꢀ3 M AgClO4. (b) TEM picture of Ag nanoparticles. Other conditions are the same as in Fig. 3a. (c) Diffraction pattern of silver
nanoparticles.

S. Kapoor et al. / Chemical Physics Letters 396 (2004) 415–419
419
Fig. 4. (a) Kinetic traces obtained for N2-bubbled microemulsion solution (w0 = 36, dose = 46 Gy) at 360 nm containing (1) CdSO4 (1 · 10ꢀ2
mol dmꢀ3) only (2) CdSO4 (1 · 10ꢀ2 mol dmꢀ3) and AgClO4 (4.9 · 10ꢀ4 mol dmꢀ3). (b) TEM picture of Ag/Cd nanoparticles prepared by irradiating
N2-bubbled microemulsion solution (w0 = 36) containing 1.0 · 10ꢀ3 M CdSO4 and 1.0 · 10ꢀ3 M AgClO4. Dose = 12 kGy. (c) Diffraction pattern of
Ag/Cd nanoparticles.
clear. Kinetic considerations indicate that, for the two
metals to mix, the diffusion coefficients need to be many
orders of magnitude larger than that the bulk materials
[13]. It is suggested that, in addition to the collisional en-
ergy and the striking coefficient, the rates of nucleation
and growth were determined mainly by the collisions be-
tween the atoms [2–4]. In electron beam irradiation the
reduction rate was so large that most of the ions are re-
duced before the formation of the nuclei and therefore
the probability of the effective collisions between the
atoms is higher. Due to this the probability of alloying
increases and the resultant particles would have more
monodispersity. Since the difference in the mean diame-
ter of Ag nanoparticles (10 nm) and mixed particles
(25 nm) was large enough to distinguish, the nearly
mono-dispersion of bimetallic systems suggested that
they were not physical mixtures of the individual metal-
lic nanoparticles and bimetallic nanoparticles were really
formed. Thus, it appears that in microemulsion, one can
get alloy particles of Cd and Ag.
diffusion of the water droplets can be exploited for mak-
ing alloy particles.
References
[1] A. Henglein, J. Phys. Chem. 97 (1993) 5457.
[2] M. Mostafavi, J.L. Marignier, J. Amblard, J. Belloni, Radiat.
Phys. Chem. 34 (1989) 605.
[3] W. Wong, S. Efrima, O. Regev, Langmuir 14 (1998) 602.
[4] (a) C. Petit, P. Lixon, M.P. Pileni, J. Phys. Chem. 97 (1993)
12974;
(b) M.P. Pileni, Langmuir 13 (1977);
(c) M.P. Pileni, J. Phys. Chem. 97 (1993) 6961.
[5] K. Kurihara, J. Kizling, P. Stenius, J.H. Fendler, J. Am. Chem.
Soc. 105 (1983) 2574.
[6] T. Mukherjee, in: S. Ahmad (Ed.), Atomic, Molecular and Cluster
Physics, Narosa, New Delhi, 1997, p. 299.
[7] G.V. Buxton, C.R. Stuart, J. Chem. Soc., Faraday Trans. 91
(1995) 279.
[8] (a) E.S. Pietek, J.L. Gebicki, J. Kroh, J. Colloid Interface Sci. 176
(1995) 264;
(b) J.L. Gebicki, L. Gebicka, J. Phys. Chem. B 101 (1997)
10828.
[9] S. Adhikari, R. Joshi, C. Gopinathan, J. Colloid Interface Sci.
191 (1997) 268.
4. Conclusion
[10] E. Janata, J. Phys. Chem. 107 (2003) 7334.
[11] P. Lianos, S. Modes, J. Phys. Chem. 91 (1987) 6088.
[12] V. Dubois, M. Seijo, P. Archirel, Chem. Phys. Lett. 389 (2004)
150.
It has been shown that initial silver aggregates get
well-resolved in microemulsion droplets. The studied
microemulsion also provides a medium (microreactor)
to stabilize small silver particles. In addition, the slow
[13] H. Yasuda, H. Mori, Phys. Rev. Lett. 69 (1992) 3747.