DOI: 10.2478/s11532-007-0018-8
Research article
CEJC 5(2) 2007 590–604
Growth kinetics of CdSe nanoparticles synthesized
in reverse micelles using bis(trimethylsilyl)
selenium precursor
Saim M. Emin1,2 , Ceco D. Dushkin1∗ ,
Seiichiro Nakabayashi2, Eiki Adachi3
1
Laboratory of Nanoparticle Science and Technology,
Department of General and Inorganic Chemistry,
Faculty of Chemistry, University of Sofia,
Sofia 1164, Bulgaria
2
3
Department of Chemistry, Faculty of Science,
Saitama University,
Saitama 338-8570, Japan
Advanced Research Center, Nihon l’Oreal R&D Center,
Takatsu-ku, Kawasaki, Kanagawa 213-0012, Japan
Received 9 May 2006; accepted 27 November 2006
Abstract:
We first focus on the kinetics of nanoparticle growth in a microemulsion synthesis of
CdSe semiconductor nanocrystals. The process consists of a fast initial stage of typical time constant
of the order of 103 s followed by a slow stage of time constant of the order of 104 s. Growth proceeds
similarly to that described for the hot-matrix synthesis of CdSe, underlining the generality of the twostage growth mechanism, irrespective of the matrix type and synthesis conditions. However, the time
constant of each stage in the microemulsion synthesis is much larger than in the hot-matrix one. Also,
the ratio between the fast and slow time constant is appreciably bigger. We also prove that larger
size reverse micelles, obtained by increasing the water:surfactant ratio, generally lead to larger CdSe
nanoparticles. Bis(trimethylsilyl) selenium is the crucial precursor for the CdSe nanoparticle synthesis.
An intermediate stage of the chemical reaction limiting the bis(trimethylsilyl) selenium production is
described theoretically.
c Versita Warsaw and Springer-Verlag Berlin Heidelberg. All rights reserved.
Keywords: CdSe nanoparticles, kinetics of growth, AOT microemulsion, bis(trimethylsilyl)selenium, NMR
spectra, FTIR spectra
∗
E-mail: [email protected]
Unauthenticated
Download Date | 6/17/17 12:17 AM
S.M. Emin et al. / Central European Journal of Chemistry 5(2) 2007 590–604
1
591
Introduction
Two different methods are in general use for the synthesis of semiconductor nanocrystals,
exploiting different types of matrix: the microemulsion method uses a micellar solution
of surfactant at ambient temperature and the hot-matrix method uses a molten soap at
a high temperature.
The microemulsion method exploits as nanocavities the water cores of reverse surfactant micelles in hydrocarbon [1] such as water/AOT/heptane where AOT is the surfactant sodium bis(2-ethylhexyl) sulfosuccinate [2]. The micelle size depends on the molar
concentration ratio w = [H2 O]/[AOT], where [H2 O] and [AOT] are the water and surfactant concentrations, respectively. At w=0 dry reverse micelles form in the solution.
At 0 < w ≤ 10 swollen micelles form containing water preferentially bound to the polar
heads of surfactant molecules. This range is the most suitable for nanoparticle synthesis.
At w>10 microemulsion droplets form with interiors rich in free water.
Historically, the first microemulsion syntheses were those of CdS nanoparticles [3, 4];
later, CdSe nanoparticles were synthesized [5]. Recently the microemulsion method has
been applied for ZnS [6, 7] and CdS/ZnS core-shell nanoparticles [8, 9]. The main advantages of the microemulsion method are the low temperature, the large volumes of
nanoparticle suspensions obtained and the slow reaction kinetics [6, 7]. The disadvantage
is the poorer crystallinity of semiconductor cores that may detract from the optical properties of the nanoparticles. It is also the case that the reaction kinetics of this process is
poorly understood.
The hot-matrix method uses a concentrated amphiphile like trioctylphosphine oxide
(TOPO) heated up to 300-350 ◦ C [10, 11]. From the very beginning this method is applied mainly to the synthesis of CdSe nanoparticles although one can also synthesize CdS
[10, 12]. A composite matrix of liquid paraffin and stearic acid has recently been used
to replace TOPO, making the process environmentally safer and lowering the synthesis
temperature to 250 ◦ C [12]. The main advantage of the hot-matrix method is the superior
crystallinity and monodisperse size distribution of the semiconductor nanoparticles. The
kinetics of nanocrystal growth is also well understood [13]. However, nothing is known
about the types of aggregate formed in the heated matrix;it is likely that they are virtual structures with the amphiphiles engulfing the nanoparticles rather than preexisting
aggregates in TOPO. This conjecture is supported by the identification of surfactantlike molecules of cadmium stearate (the ligands) formed in situ during the nanoparticle
synthesis in liquid paraffin [12].
Four reactions have been found suitable for the preparation of CdSe nanocrystals
in a liquid medium using various precursors for synthesis at ambient temperature [14]:
(i) between cadmium sulphate and sodium selenide; (ii) between dimethylcadmium and
hydrogen selenide; (iii) between a cadmium salt (cadmium chloride or cadmium perchlorate [5]) and bis(trimethylsilyl)selenium, Se(TMS)2 ; (iv) between dimethylcadmium and
Se(TMS)2 . Since the synthesis of dimethylcadmium turned out to be a rather difficult
task [15], we here use reaction (iii), involving Se(TMS)2 , for the production of CdSe
Unauthenticated
Download Date | 6/17/17 12:17 AM
592
S.M. Emin et al. / Central European Journal of Chemistry 5(2) 2007 590–604
nanocrystals.
The purpose of our paper is to reveal new substantial features of the growth kinetics of CdSe nanoparticles synthesized by the microemulsion method and compare them
with those of the hot-matrix method. We also describe the synthesis of selenium-based
precursor Se(TMS)2 , required for this study.
The CdSe nanoparticle synthesis follows the recipe from ref. [5]. One of the new points
of our research here is the calculation of the nanoparticle size from the absorption spectra,
which has not previously been done for nanoparticles produced by the microemulsion
method. This enables one to plot the particle diameter as a function of time and to
fit it using the two-stage growth model from ref. [13] originally derived for the hotmatrix synthesis. This model also describes the microemulsion synthesis surprisingly
well, providing larger time constants and comparable values for other parameters. We also
examined the effect of water content on the nanoparticle size in the range of w = 4 − 10.
As might be expected, the CdSe nanoparticles increase in size with increasing w. However,
the nanoparticle sizes remain always smaller than those of the reverse micelles.
The synthesis of Se(TMS)2 follows a route described in the literature [16, 17]. However, we built for this purpose a special reaction cell (four-neck flask), which enables
precise temperature and atmosphere control throughout the synthesis. Photo-images
taken from the cell bottom at different reaction times reveal new aspects of the reaction
kinetics of Se(TMS)2 production. For example, both Li2 Se2 and Li2 Se co-exist during the
dissolution of selenium pearls. We propose a simple model for the kinetics based on the
dissolution of selenium pearls as the rate limiting step.
2
Materials and methods
Elemental Se pearls, lithium triethylborohydride (1 M in THF), chlorotrimethylsilane and
deuterated chloroform (CCl3 D) were purchased from Aldrich. The compounds d-benzene
(C6 D6 ) and tetrahydrofuran (THF) were purchased from Wako Chemicals. The other
chemicals CdCl2 , sodium bis(2-ethylhexyl) sulfosuccinate (AOT) and heptane were obtained from Merck. The air-sensitive compounds were manipulated under an argon gas
atmosphere maintained in a glove bag. The specialized synthetic apparatus, whose main
part is the four-neck flask with a glass jacket and a flat bottom window, was made in
our laboratory. The jacket surrounding the flask allows the circulation of liquid at different temperatures to cool or heat the reaction mixture. The flat window facilitates
photography of the selenium pearls at different reaction times.
Se(TMS)2 provides the precursor for synthesis of CdSe semiconductor nanoparticles.
The reaction was maintained in water-in-oil (w/o) microemulsion (swollen reverse micelles). The dry reverse micelles were prepared as a 0.05 M solution of AOT in benzene.
A 0.1 M aqueous solution of Cd2+ ions was then injected into the AOT solution to make
the reverse micelles with entrapped water. The ratio w = [H2 O]/[AOT] was adjusted for
each sample by adding the appropriate amount of water. Reverse micelles of different
sizes, w=4, 6, 8 or 10, were chosen to host the nanoparticles. The amount of injected
Unauthenticated
Download Date | 6/17/17 12:17 AM
S.M. Emin et al. / Central European Journal of Chemistry 5(2) 2007 590–604
593
organometallic precursor Se(TMS)2 was fixed at 2 μl for each sample of 15 ml of microemulsion. The Cd2+ ions in the water droplets react with the Se2− ones to form CdSe
nuclei, which further grow as the nanoparticles. The mean nanoparticle diameter was
calculated from the absorbance spectra using the empirical formula of Peng [18]. This
formula is devised for CdSe nanocrystals synthesized by pyrolysis in a hot-amphiphile
matrix. Although these conditions are rather different from the microemulsion synthesis
utilized here, the formula gives plausible results for the nanoparticle diameters that are
comparable with those calculated from the absorption peak using the attenuated quantum confinement model [13, 19]. The dispersion of size distribution was calculated from
the experimental data for the absorbance as described in ref. [13]. For comparison, the
CdSe nanoparticle size was also determined from photographs taken by the Hitachi 7500
Transmission Electron Microscope (TEM).
AOT micelle sizes were measured by Dynamic Light Scattering (DLS) using Autosizer
4700C (Malvern Ins., UK) with an argon laser (Innova 70, Coherent) operating at a wavelength of λ0 =488 nm (vertical plane polarized light). The scattering angle (90◦) defines
the direction of the scattering vector q = (4nπ/λ0 ) sin(θ/2), where n is the refractive
index of the sample. The temperature in these experiments was maintained at 25.0±0.1
◦
C.
UV-visible spectra of CdSe nanoparticles in suspension were obtained using a JASCO
V-560 spectrophotometer. Fourier-Transformed Infrared (FTIR) spectra of Se(TMS)2
were measured with a JASCO spectrophotometer FT/IR-660.
A Bruker DRX-400 NMR spectrometer operating at 400 MHz was used to obtain
NMR spectra of Se(TMS)2 . The formation of hydrolyzed product was observed (in 1 H
NMR spectra) in the appearance of two singlets with chemical shifts at 0.1 and 0.2 ppm,
and integrated relative intensity 1:9. Tetramethylsilane was used as a reference standard,
which has similar structure to that of the possible hydrolyzed product. Taking into
account the intensity of 1 H signals, we concluded that this compound was trimethylsilanol.
The compound hydrolyzes due to the weak Si-Se chemical bond.
3
Results and discussion
3.1 Synthesis of Se(TMS)2
3.1.1 Reaction pathway
In the classical method for synthesis of Se(TMS)2 [17], the final reaction which provides
the compound is:
Li2 Se + 2Me3 SiCl → Me3 SiSeSiMe3 + 2LiCl
(1)
where Me stands for the methyl group in Se(TMS)2 (Me3 SiSeSiMe3 ). To get the dilithium
selenide, Li2 Se, elemental selenium is reduced using lithium triethylborohydride,
Li(C2 H5 )3 BH [16]
Se + 2Li(C2 H5 )3 BH → Li2 Se + 2(C2 H5 )3 B + H2
(2)
Unauthenticated
Download Date | 6/17/17 12:17 AM
594
S.M. Emin et al. / Central European Journal of Chemistry 5(2) 2007 590–604
For this purpose 1.0 equiv of selenium shot is mixed with 2.1 equiv of Li(C2 H5 )3 BH
dissolved in tetrahydrofuran (THF) at a concentration of 1.0 M. A milky white dispersion
of Li2 Se is obtained. Subsequent addition of 2.1 eqiuv chlorotrimethylsilane, (CH3 )3 ClSi,
to the Li2 Se leads to the Se(TMS)2 .
However, dilithium diselenide (Li2 Se2 ) can also form in excess Se [16]
2Se + 2Li(C2 H5 )3 BH → Li2 Se2 + 2(C2 H5 )3 B + H2
(3)
For this purpose one mixes 1 equivalent Se with 1.05 equivalents of Li(C2 H5 )3 BH.
The resultant dispersion is of brown-red color. Then the addition of 1.05 equiv of
(CH3 )3 ClSi and stirring results in the formation of bis(trimethyldisilyl)diselenide rather
than of Se(TMS)2 . The compound Li(C2 H5 )3 B, which is a byproduct in these reactions,
is removed as a co-distillate with THF and workup solvents.
3.1.2 Synthesis procedure
The synthesis of bis(trimethylsilyl) selenium was carried out using a vacuum Schlenk
technique in an inert atmosphere. The four-neck flask was charged with 8 selenium
pearls (0.1925 g in total) and then argon gas was blown through the flask. The solution
was then cooled to 0 ◦ C by circulating a cooled liquid through the flask jacket. Then
lithium triethylborohydride (50.2 ml) was injected into the solution, resulting in hydrogen
gas evolution – cf. reactions (2) and (3). The reaction is exothermic. The reaction
temperature was kept at 25 ◦ C in order to follow the kinetics of dissolution of Se pearls.
Photographs were taken through the flat-bottom flask window with a digital photocamera
(Sony). After the completion of Se dissolution the mixture was again cooled to 0 ◦ C. Then
chlorotrimethylsilane was added in a single portion (7.41 ml). The resulting mixture was
stirred for 2 h at a room temperature in order to obtain the desired compound Se(TMS)2 .
The low boiling volatiles, accompanying the Se(TMS)2 , were removed by distillation at
20 torr and 30-60 ◦ C. The product Se(TMS)2 itself was distilled at 5 torr (boiling point 46
at 5.3 torr [? ]) as a colorless oil-like liquid. Due to the high reactivity of the compound
with oxygen, Se(TMS)2 was sealed in a glass ampoule and stored at -20 ◦ C for later use.
3.1.3 Kinetics of Se dissolution
A typical image of the dissolution of Se pearls, after addition of Li(C2 H5 )3 BH, is shown
in the inset to Fig. 1. The brown colored spots (seen as the dark area) are the clouds
of Li2 Se2 , which are surrounded by a milky-white colored dispersion of Li2 Se (the bright
area). A series of images were taken over a time interval of 30 min; the change of the
brown area is not significant at longer times. The expanding spots serve as a measure
k2
k1
for the kinetics of the reaction chain Se −→
Li2 Se2 −→
Li2 Se. The first reaction is
the dissolution of selenium in the solvent of lithium triethylborohydride to form Li2 Se2
with a rate constant k1 . The second reaction is the transition of Li2 Se2 to Li2 Se with a
rate constant k2 . Obviously, the second reaction is a slower process than the first one,
i.e. k1 >> k2 . This is due to the diffusion necessary to bring the solute (Li2 Se2 ) to a
Unauthenticated
Download Date | 6/17/17 12:17 AM
S.M. Emin et al. / Central European Journal of Chemistry 5(2) 2007 590–604
595
fresh solvent. After completion of the reaction (upon stirring) the entire solution becomes
milky white, which confirms the complete transformation of Li2 Se2 into Li2 Se.
To calculate the area of brown spots at different times, we cut the respective paper
from the photographs and carefully weighed it on an analytical balance with a precision
of 10−5 g. The total area of the clouds S(t) at time zero and time t is a measure of the
amount of Li2 Se2 formed. Figure 1 depicts the function S(t) for our experiment.
4
Experiment
Theory
2
S (cm )
3
2
1
0
0
5
10
15
20
25
30
t (min)
Fig. 1 Time dependence of the total projected area, S, of the Li2 Se2 clouds surrounding
the selenium pearls during their dissolution. The theoretical line is drawn by Eq. (6).
The inset shows the dissolution of Se pearls at 10 min after the beginning of process (the
photograph is taken from the bottom of the reaction flask).
In view of the above considerations, we propose the following model for the transition
of Li2 Se2 to Li2 Se. Let us consider first a single Se pearl, placed on the bottom of reaction
flask, which is dissolving in the solvent. The cloud of Li2 Se2 around it can be represented
as a hemisphere of radius R, whose volume is V = (2/3)πR3 . The volume V is assumed
increasing with a constant volumetric rate Q, which is typical for many natural processes.
The quantity Q can in principle be related to the real dissolution rate of the selenium
pearl measured by weighing of the pearl at different time intervals. Then the kinetic
equation for the volume increase is
dV
=Q
(4)
dt
The integration of Eq. (4) leads to the following equation for V :
V (t) = V0 + Qt
(5)
where V0 is the initial volume at zero time, co-inciding with the volume of Se pearl.
The increase of V increases also the surface area of the cloud, 2πR2 , which is related to
the cross-sectional area S = πR2 . By replacing V in Eq. (5) we obtain for the radius
R = [3(V0 + Qt)/2π]1/3 and for the spot area
3
(V0 + Qt)
S(t) = π
2π
2/3
(6)
Unauthenticated
Download Date | 6/17/17 12:17 AM
596
S.M. Emin et al. / Central European Journal of Chemistry 5(2) 2007 590–604
Equation (6) predicts a power law for the increase of the cross-sectional area of a
single cloud with time. Then the total area of clouds created by all Se pearls in the
suspension turns out to be described by the same equation but with V0 meaning the total
initial volume of pearls.
Indeed, the fit of the experimental data for S with Eq. (6) shows a very good coincidence at fitting parameters Q=0.092 cm3 /s and V0 =0.047389 cm3 . Assuming the pearls
are spheres of equal size, we get for the average pearl diameter 2.24 mm (close to the real
size of Se pearls).
3.1.4 NMR spectra
We intend to confirm the existence of two compounds, Li2 Se2 and Li2 Se, formed during
the synthesis. In Fig. 2 the 7 Li NMR spectra of these compounds are presented. The
materials are dissolved in THF with a small amount of added C6 D6 . The positions of
the two peaks clearly show the formation of two products: Li2 Se2 (with a chemical shift
at 0.36 ppm) and Li2 Se (with a chemical shift at –0.15 ppm).
(a)
(b)
(c)
Fig. 2 7 Li NMR spectra at 298 K of: a) Li2 Se2 ; b) Li2 Se (both in THF). c) 77 Se spectrum
of Li2 Se2 at 298 K in mixture of solvents d-benzene and terahydrofuran (0.6:0.2 ml).
77
Se is a NMR active isotope having a nuclear spin quantum number I=1/2; highresolution NMR spectroscopy is therefore possible for this isotope. Since 77 Se is approximately three times more sensitive than 13 C, the resonance range is very wide: from minus
several hundreds up to 1000 ppm. The reported 77 Se chemical shift is at –527 ppm for
2−
the K+
species in water [20]. The chemical shift of certain species can be strongly
2 Se
affected by the solvent. For example, the 77 Se shift of Na+ HSe− can vary from -529 ppm
in water to -447 ppm in dimethylformamide, both measured at 300 K. The measured
value for Li2 Se2 in our solvent system (deuterated benzene+THF) is –445 ppm. The
Unauthenticated
Download Date | 6/17/17 12:17 AM
S.M. Emin et al. / Central European Journal of Chemistry 5(2) 2007 590–604
597
solubility of Li2 Se in THF is very poor and, hence, 77 Se NMR was not recorded for a
mixture of d-benzene+THF. The stability of the species in water is under question due
to the possible formation of HSe− .
3.2 Optical spectroscopy
3.2.1 FTIR spectra of Se(TMS)2
The organoselenium compound, Se(TMS)2 , is characterized by FT/IR spectroscopy
(Fig. 3a). The spectrum is recorded in THF solvent, which has a weak absorption region
around 500 to 800 cm−1 (Fig. 3b). The asymmetrical characteristic vibration for Si-Se-Si
bond in Se(TMS)2 also occurs in this region. The infrared bands could be classified as
follows (Fig. 3a): Symmetric stretching νs of the Si-C-H bond at 855 cm−1 ; asymmetric
stretching νa of the Si-C-H bond at 1338 cm−1 ; symmetrical deformation δs of the SiMe3
group at 636 cm−1 ; asymmetric stretching νa of the Si-Se-Si bond at 756 cm−1 ; νa CH3
at 2916 cm−1 ; νs CH3 at 2980 cm−1 ; δs CH3 at 1252 cm−1 . The estimated Se-Si bond in
Se(TMS)2 is roughly 35 kcal/mol weaker than the C-Si bond in tertamethylsilane SiMe4 .
The positions of the peaks are in good coincidence with the literature [21].
100
T (%)
80
60
40
20
(a)
0
3200
2800
2400
2000
1600
1200
800
400
-1
k (cm )
100
T (%)
80
60
40
20
(b)
0
3200
2800
2400
2000
1600
1200
800
400
-1
k (cm )
Fig. 3 IR spectra of: (a) bis(trimethylsilyl) selenium in THF (110 μm inner space thickness of the KBr cell); (b) tetrahydrofuran (THF) (20 μm inner space thickness of the
KBr cell).
Unauthenticated
Download Date | 6/17/17 12:17 AM
598
S.M. Emin et al. / Central European Journal of Chemistry 5(2) 2007 590–604
3.2.2 Absorbance of CdSe nanoparticles
The growth of CdSe nanoparticles takes about 80 min. Figure 4 presents the changes in
the absorption spectra due to the growth of semiconductor nanoparticles during synthesis
in reverse micelles of w=8 (4a) and w=10 (4b). Each spectrum contains a well-pronounced
excitonic peak that is characteristic for the microemulsion synthesis of nanoparticles. The
peak wavelength (λmax ) is related to the nanoparticle size (quantum confinement effect).
Therefore, it is possible to associate the spectral shift to longer wavelength with increasing
growth time with an increase in nanoparticle sizes.
3.3 Growth of CdSe nanoparticles
3.3.1 Time dependence
The nanoparticle diameter in Fig. 5a increases rapidly at the beginning but later it
reaches a plateau. A number of authors have discussed the particle size increase after the
initial nucleation; some of them attribute this increase to a sort of coagulation, because
the Ostwald ripening should be very slow [7, 22]. If we plot log(d − d0 ) versus log(t) the
gradient of the straight line can give us information about the growth rate (here d0 is
the diameter at zero time). In our case the value of 0.45 is comparable with the Ostwald
ripening value 0.33 [23].
We found that the nanoparticle growth in microemulsion is well described by the twostage theoretical model proven for CdSe nanoparticles in the hot-matrix synthesis [13].
During the fast initial process (Fig. 5a) the nanoparticles grow due to a reaction-driven
mechanism. The semiconductor precursors in the vicinity of the nanoparticle adsorb on
the nanoparticle surface where they react to build the crystal lattice of CdSe. As material
is exhausted from the immediate particle vicinity, a slow diffusion starts to play a role
as reactants must be brought in from farther away. Such a diffusion-limited mechanism
determines the plateau-like growth at longer times. This is the first proof that such
kinetics may operate also in the microemulsion synthesis. However, here the respective
time constants τr of reaction growth and τd of diffusion growth are much larger than those
in the hot-matrix synthesis – see Table 1. Also, the Damköhler number (Da = τγ − τd )
remains 10 times larger than that observed in the hot-matrix method.
Table 1 Comparison of the parameters of growth for various syntheses of CdSe nanoparticles.
Synthesis
τr (s)
τd (s)
Da
w=8
w=10
hot-matrix [13]
2300
1100
20
23000
11000
2000
0.1
0.1
0.01
Unauthenticated
Download Date | 6/17/17 12:17 AM
S.M. Emin et al. / Central European Journal of Chemistry 5(2) 2007 590–604
599
3
w=8
1 min
5 min
28 min
35 min
55 min
60 min
80 min
A/A*
2
1
(a)
0
400
500
600
O(nm)
3
1 min
3 min
7 min
9 min
13 min
16 min
19 min
26 min
30 min
45 min
52 min
70 min
80 min
w=10
A/A*
2
1
(b)
0
400
500
600
O(nm)
3
A/A*
2
1
0
300
w=4
w=6
w=8
w=10
400
(c)
500
600
700
O (nm)
Fig. 4 Absorbance spectra of CdSe nanoparticles in the system water/AOT/benzene as a
function of time at: (a) w=8 and (b) w=10. The growth time increases from left to right.
(c) Visible spectra of CdSe particles synthesized at different w ratio increasing from left
to right. The inset depicts the final suspension of CdSe nanoparticles in heptane, which
show green photoluminescence (PL) under UV-light illumination. In this case (w=4) the
nanoparticles diameter is 1.93 nm.
Unauthenticated
Download Date | 6/17/17 12:17 AM
600
S.M. Emin et al. / Central European Journal of Chemistry 5(2) 2007 590–604
(a)
d (A)
26
°
24
22
20
w=8
3.2
w=10
V (A)
3.0
°
2.8
2.6
2.4
2.2
(b)
2.0
0
50
100
150
t (min)
1.0
(c)
0.8
1 min
5 min
28 min
N/Nmax
55 min
0.6
80 min
0.4
0.2
0.0
0
10
o
20
30
R (A)
Fig. 5 Kinetics of growth of CdSe nanoparticles (the data points are extracted from
Fig. 4a,b). (a) The mean diameter d is plotted versus time. The inset shows TEM image
of the final CdSe nanoparticles, which are readily seen as the black spots (w=10; scale
bar 20 nm). The average diameter obtained from this image (d=3.12 nm) corresponds
roughly to that calculated from the optical spectra (d=2.65 nm). (b) Plot of the dispersion
σ versus the time. σ is determined from the peaks in (c) by taking its width at 0.6
height from the bottom and dividing by 2. (c) Evolution of the size distribution of CdSe
nanoparticles at w=8. The peaks are calculated as described in ref. [13].
Unauthenticated
Download Date | 6/17/17 12:17 AM
S.M. Emin et al. / Central European Journal of Chemistry 5(2) 2007 590–604
601
The other parameters of the data fit in Fig. 5a remain rather close to the values used
for CdSe in the hot-matrix synthesis [13]. For example, the initial particle diameter is
20.8 Å (w=8) and 21.8 Å (w=10). These are plausible values for the sizes of initial nuclei
preceding the CdSe nanoparticles. The thickness of the subsurface layer surrounding a
nanoparticle (δ) is 19 Å, which is essentially the same as for CdSe in TOPO matrix. In
our case, however, this should be considered as a part of the water core of a micelle rather
than as the thickness of the AOT capping layer.
Regarding the dispersion of size distribution σ ∼2.2-3.2 Å plotted in Fig. 5b, it
remains appreciably larger than that from the hot-matrix synthesis (σ ∼1.2-2 Å). This
means that the nanoparticles from the microemulsion syntheses ought to be more polydisperse. The change in σ with time follows the general two-stage mechanism [13]. Initially,
it decreases with the reaction-driven growth (focusing of size-distribution), but later it
remains almost constant, again in accordance with the theoretical predictions.
Based on the above considerations for d and σ, we reconstructed the nanoparticle
size distribution in Fig. 5c for the case w=8. It is clearly seen that the peak position of
nanoparticle size shifts toward larger sizes, and that the size distribution narrows with
increasing growth time.
3.3.2 Dependence on micelle size
Apart from the growth time, the main factors that can affect the nanoparticle sizes are
the concentration of the organoselenium compound, the concentration of Cd2+ ions and
the sizes of micelles (w ratio). The latter is known from the literature [2]. Indeed, this
effect can be seen in Fig. 5a,b. Increasing the ratio w to 10 leads to the growth of bigger
nanoparticles with larger dispersion σ. Also, they grow faster than the nanoparticles at
w = 8 (Table 1). This trend can be seen for the final nanoparticle sizes. Figure 6 shows
the change in the final absorption spectra by varying the w ratio. The linear relationship d(w) = 1.4342 + 0.1255w (d in nm) is obtained from the plot of the semiconductor
nanoparticle size versus w ratio.
The diameters of the reverse micelles dw are also plotted versus w in Fig. 6. The water
pools are appreciably bigger than the CdSe nanoparticles. The relationship between water
pool diameter and w is dw (w) = 1.6534 + 0.3649w (dw in nm), which seems comparable
with the ones derived in ref. [2, 24] for similar systems. The results from Fig. 6 suggest
that the semiconductor nanoparticle occupies only a part of the water volume of micelles
that provide precursors for the nanocrystal growth. This is consistent with the continuous
model for the two-stage reaction kinetics outlined above. The capping of the nanoparticle
cores with the AOT shells can only occur later after water is somehow removed from the
suspension – for example, by evaporation. Ostwald ripening can thus occur only in
the long term – over a period of several days [7]. Nevertheless, the final nanoparticle
suspension is stable with PL for months.
The synthesis of Se-containing semiconductor nanoparticles illustrates the applications of pre-synthesized Se(TMS)2 . Apart from CdSe, ZnSe nanoparticles can also be
prepared using Se(TMS)2 as described in ref. [25]. The approach utilized by us here
Unauthenticated
Download Date | 6/17/17 12:17 AM
602
S.M. Emin et al. / Central European Journal of Chemistry 5(2) 2007 590–604
8
Micelles
Nanoparticles
d (nm)
6
4
2
0
0
5
w
10
Fig. 6 Comparison of the diameters of the reverse micelles measured by DLS and those
of CdSe nanoparticles calculated from the Visible spectra in Fig. 4c. The error bars give
the statistical error from three independent measurements. The solid lines are fits of the
data.
for Se(TMS)2 , is suitable also for the synthesis of sulphur containing precursor such as
bis(trimethylsilyl) sulphur, which has been applied to the preparation of S-containing
semiconductor nanoparticles [26].
4
Conclusions
CdSe nanoparticles are synthesized in water/AOT/heptane microemulsion. The nanoparticle size is calculated as a function of time from the absorbance spectra at two different
ratios of the water to surfactant: w=8 and w=10. The kinetic curves are then fitted
with a theoretical model which considers two subsequent growth processes: a fast process (limited by the reaction building the nanoparticle surface) and a slow one (limited
by the diffusion transport of precursor from the bulk to the subsurface layer). This is
similar to the model used for CdSe nanoparticles synthesized by the hot-matrix method
with some important differences: In the microemulsion method the nanoparticles grow
much slower than in the hot-matrix one. Also, the ratio between the time constants
of fast and slow processes is much bigger in the microemulsion synthesis. At bigger water/surfactant ratio (w=10) the nanoparticles also grow bigger and faster than in the case
of smaller ratio (w=8). The final nanoparticle size can also be increased by increasing
the water/surfactant ratio, because the reverse micelles hosting the nanoparticles are also
bigger.
One of the precursors in the nanoparticle synthesis is bis-trimethylsilylselenium synthesized also by us here. In this synthesis we visualized and quantitatively explained the
dissolution of selenium pearls via a transition through a dilithium diselenide intermediate
into dilithium selenide.
Unauthenticated
Download Date | 6/17/17 12:17 AM
S.M. Emin et al. / Central European Journal of Chemistry 5(2) 2007 590–604
603
Acknowledgment
The authors thank to Prof. M. Saito from Saitama University for the valuable discussion
on 7 Li and 77 Se NMR spectra and to Mr. A. Abadjimarinov from the Institute of Physical
Chemistry, Bulgarian Academy of Sciences, for the glass masterwork. This research is
being supported by grant VUH-09/05 of the Bulgarian Ministry of Education and Science
and by COST Action D43.
References
[1] M.P. Pileni (Ed.): Structure and Reactivity of Reverse Micelles, Elsevier, Amsterdam, 1989.
[2] M.P. Pileni: “Reverse micelles as microreactors”, J. Phys. Chem., Vol. 97, (1993),
pp. 6961–6973.
[3] H. Weller, H.M. Schmidt, U. Koch, A. Fojtik, S. Baral, A. Henglein, W. Kunath and
K. Weiss: “Photochemistry of colloidal semiconductors. Onset of light absorption as
a function of size of extremely small CdS nanoparticles”, Chem. Phys. Lett., Vol.
124, (1986), pp. 557–560.
[4] P. Lianos and J.K. Thomas: “Cadmium sulfide of small dimensions produced in
inverted micelles”, Chem. Phys. Lett., Vol. 125, (1986), pp. 299–302.
[5] M.L. Steigerwald, A.P. Alivisatos, J.M. Gibson, T.D. Harris, R. Kortan, A.J. Muller,
A.M. Thayer, T.M. Duncan, D.C. Douglass and L.E. Brus: “Surface derivatization
and isolation of semiconductor cluster molecules”, J. Am. Chem. Soc., Vol. 110,
(1988), pp. 3046–3050.
[6] P. Calandra, M. Goffredi and V. Turco Liveri: “Study of the growth of ZnS nanoparticles in water/AOT/n-heptane microemulsions by UV-absorption spectroscopy”, Colloid Surfaces A, Vol. 160, (1999), pp. 9–13.
[7] G. Ilieva, K. Papazova, C. Dushkin and E. Adachi: “Aging effects in the microemulsion synthesis of ZnS nanoparticles”, Ann. Univ. Sofia Fac. Chim., Vol. 96, (2004),
pp. 155–164.
[8] L. Qi, J. Ma, H. Cheng and Z. Zhao: “Synthesis and characterization of mixed
CdS-ZnS nanoparticles in reverse micelles”, Colloid Surfaces A, Vol. 111, (1996), pp.
195–202.
[9] A.R. Loukanov, C.D. Dushkin, K.I. Papazova, A.V. Kirov, M.V. Abrashev and E.
Adachi: “Photoluminescence depending on the ZnS shell thickness of CdS/ZnS coreshell semiconductor nanoparticles”, Colloid Surfaces A, Vol. 245, (2004), pp. 9–14.
[10] C.B. Murray, D.J. Norris and M.G. Bawendi: “Synthesis and characterization of
nearly monodisperse CdE (E=S, Se, Te) semiconductor nanocrystallites”, J. Am.
Chem. Soc., Vol. 115, (1993), pp. 8706–9715.
[11] J.E. Bowen Katari, V.L. Colvin and A.P. Alivisatos: “X-ray photoelectron spectroscopy of CdSe nanocrystals with applications to studies of the nanocrystal surface”, J. Phys. Chem., Vol. 98, (1994), pp. 4109–4117.
Unauthenticated
Download Date | 6/17/17 12:17 AM
604
S.M. Emin et al. / Central European Journal of Chemistry 5(2) 2007 590–604
[12] G.G. Yordanov, C.D. Dushkin, G.D. Gicheva, B.H. Bochev and E. Adachi: “Synthesis of high-quality semiconductor nanoparticles in a composite hot-matrix”, Colloid
Polymer Sci., Vol. 284, (2005), pp. 229–232.
[13] C. D. Dushkin, S. Saita, K. Yoshie and Y. Yamaguchi: “The kinetics of growth of
semiconductor nanocrystals in a hot amphiphile matrix”, Adv. Colloid Interface Sci.,
Vol. 88, (2000), pp. 37–78.
[14] S. M. Stuczynski, J. G. Brennan and M. L. Steigerwald: “Formation of metalchalcogen bonds by the reaction of metal alkyls with silyl chalcogenides”, Inorg.
Chem., Vol. 28, (1989), pp. 4431–4432.
[15] G. Yordanov, C. Dushkin, B. Bochev, G. Gicheva and E. Adachi: “Preparation of
dimethyl cadmium precursor for the hot-matrix synthesis of CdSe nanoparticles”,
Ann. Univ. Sofia, Fac. Chim., Vol. 98–99, (2006), pp. 131–139 (English)..
[16] J. A. Gladysz, J. L. Hornby and J. E. Garbe: “A convenient one-flask synthesis of
dialkyl selenides and diselenides via lithium triethylborohydride reduction of Sex ”,
J. Org. Chem., Vol. 43, (1978), pp. 1204–1208.
[17] M. R. Detty and M. D. Seidler: “Bis(trialkylsilyl) chalcogenides. 1. Preparation and
reduction of group 6A oxides”, J. Org. Chem., Vol. 47, (1982), pp. 1354–1356.
[18] W. W. Yu, L. Qu, W. Guo and X. Peng: “Experimental determination of the size
dependent extinction coefficients of high quality CdTe, CdSe and CdS nanocrystals”,
Chem. Mater., Vol. 15, (2003), pp. 2854–2851.
[19] C. Dushkin, K. Papazova, N. Dushkina and E. Adadchi: “Attenuated quantum confinement of the exciton in semiconductor nanoparticles”, Colloid Polymer Sci., Vol.
284, (2005), pp. 80–85.
[20] H. Duddeck: “Selenium-77 Nuclear Magnetic Resonance Spectroscopy”, In: Progress
in NMR Spectroscopy, Vol. 127, Elsevier, Oxford, 1995, pp. 1–323.
[21] K. Hamada and H. Morishita: “Raman, infrared and 1 H-NMR spectra of hexamethyldisilylchalcogenides”, Spectr. Lett., Vol. 19, (1986), pp. 815–826.
[22] K. Suzuki, K. Harada and M. Shioi: “Growth mechanism of cadmium sulfide ultrafine
particles in water-in-oil microemulsion”, J. Chem. Eng. Jpn., Vol. 29, (1996), pp.
264–276.
[23] A. Bhakta and E. Ruckenstein: “Ostwald ripening: a stochastic approach”, J. Chem.
Phys., Vol. 103, (1995), pp. 7120–7131.
[24] I. Stoychev: Microemulsion synthesis of CdS nanoparticles, Thesis (Master), University of Sofia, Sofia, 2003 (in Bulgarian).
[25] S. Emin, T. Fujita, M. Nawa, S. Nakabayashi, E. Adachi and C. Dushkin: “Semiconductor core/polymer shell nanoparticles”, In: E. Balabanova and I. Dragieva (Eds.):
Nanoscience and Nanotechnology, Vol. 3, Heron Press, Sofia, 2003, pp. 165–168.
[26] S. Emin, C. Papazova, C. Dushkin, D. Mladenova and E. Adachi: “Synthesis of
water dispersible semiconductor nanoparticles using bis(trimethyldisilyl)sulfide”, In:
E. Balabanova and I. Dragieva (Eds.): Nanoscience and Nanotechnology, Vol. 4,
Heron Press, Sofia, 2004, pp. 58–61.
Unauthenticated
Download Date | 6/17/17 12:17 AM