Proceedings of Spie: Cds Nanoparticles and The Sol-Gel Process
Proceedings of Spie: Cds Nanoparticles and The Sol-Gel Process
SPIEDigitalLibrary.org/conference-proceedings-of-spie
ABSTRACT
We present two approaches for the sol-gel synthesis of transparent solids in which CdS nanoparticles are homogeneously
dispersed. In any case CdS nanocrystals are obtained by precipitation either in reverse micelles or in the presence of a
complexing thiol. The grafting of the 4-fluorothiophenol at the surface of the particles leads to highly concentrated colloids in
acetone.
In a first approach, CdS/silica nanocomposites are prepared introducing this concentrated CdS colloid in a silica sol which
contains 3-mercaptopropyltrimethoxysilane to ensure the homogeneity of the CdS dispersion. The luminescence properties of
the materials are improved by using methyltriethoxysilane instead of tetraethoxysilane as the alkoxide precursor of the silica
matrix. Further increase of the luminescence efficiency is also observed if a single Mn2+ ion is present inside the CdS
particle.
In a second approach, pure CdS nanocristalline materials (bulk or thin films) are obtained from the sol-gel processing of
the concentrated colloids stabilized by 4-fluorothiophenol. The mechanisms driving the aggregation of the particles and the
gelation of the system are explicited. Thin films as well as monoliths can be produced. The general principles of this method
are not restricted to chalcogenide materials, and should therefore enlarge the applications of the sol-gel process to other non-
oxide materials.
1. INTRODUCTION
For about fifteen years, semiconductor nanoparticles have aroused great interest in the scientific community because of
their physical properties related to quantum confinement, and the potential applications resulting from their high nonlinear
optical 1 Researches were first achieved on glasses in which CdSxSe 1 -x nanoparticles were precipitated by high
temperature treatments. But it appeared a few years ago that new development of the subject requires a better control of the
parameters which determine the properties of the nanoparticles : chemical composition, average size and size distribution, and
surface state of the nanoparticles. The sol-gel process and the colloid chemistry offer good opportunities for the synthesis of
transparent materials embedding semiconductor nanocrystals. Most of the work in this field has been achieved on 11-VT
semiconductors, and especially on CdS and CdSe for which the quantum size effect is well understood and the colloidal
synthesis well controlled. Nevertheless, some work has still to be done in order to understand and control the optical
properties of the materials. Most problems arise indeed from the major role played by the surface of the particles on
recombination pathways of excited carriers.
In the first part of this paper, we describe our procedure concerning the synthesis of CdS/silica nanocomposite materials.
The idea is to develop a process which would give enough flexibility to study the influence of various chemical parameters
on the optical properties of the final material. We will here discuss the influence of the chemical structure of the sol-gel host
matrix and of the incorporation of Mn2+ ions in the CdS network.
In the second part of this paper, we show when using highly concentrated CdS colloids stabilized in acetone by 4-
fluorothiophenol, it is possible to elaborate a transparent solid material by assembling the CdS nanoparticles through a sol-
gel process. In this case, the nanoparticles are not only the optically active system, but also the building blocks for the
synthesis of the whole solid.
Most of the works published on the synthesis of sol-gel matrices doped with semiconductor nanoparticles use either the
synthesis of the particles directly in a sol containing functionalized alkoxides or precipitation within a porous silica network2.
In the last few years, we have developed a process in which the synthesis of the nanoparticles is completely separated from the
sol-gel medium3. This allows to take advantage of the different techniques developed in colloid chemistry to better control the
parameters which have a direct impact on the properties of the final composite materials : chemical composition, average size
and size distribution of the particles, electronic states of their surface and concentration in the final matrix.
The general scheme is described figure 1 . On the one hand, the particles are obtained through usual techniques of colloid
chemistry. In the case of 11-VT semiconductors, concentrated colloids are obtained through the grafting at their surface of a
molecule whose the chemical nature determines the solvent of dispersion. On the other hand, a so! is prepared by hydrolysis
of silicon alkoxides whose further polymerization and drying give the final glassy matrix.
The incorporation of the particles in the so! is directly achieved by adding the CdS colloidal solution to the silica so!.
provided that the colloid does not flocculate, i.e. the particles are still stable in the solvent used for the sol-gel synthesis. In
order to prevent the irreversible flocculation of CdS colloidal particles in the silica so!, we use a functionalized alkoxide F-R-
Si(OEt)3, where R is an alky! spacer and F is a complexing function which can react with the surface of CdS crystallites.
This alkoxide thus allows the stabilization of the particles through their grafting to the silica network.
xeroael
(bulk or thin film)
Figure 1. General scheme for the synthesis of 11- VI nanoparticles in so/-gel !natrices
The colloidal synthesis of Il-VI nanoparticles has been the subject of intensive research. In most cases, the precipitation is
achieved by mixing the metal and chalcogenide precursors (salts or organometallics) in the presence of a surface complexing
agent46. This complexation limits the size and prevents the aggregation of the particles. Many molecules have thus been
used such as polyvinyl alcohol, thiols, phosphine/phosphates ... Another technique consists in achieving the precipitation
within the water pools of inverted micelles7'8. The particles are then stabilized in the water droplets because of their
hydrophobic surface. hut also because of a surface complexation caused by the polar head of the surfactant. This process was
successfully applied for the synthesis of Il-VT semiconductor crystallites, for which the size can be varied from 1 to 10 nm,
playing either on the size of the water pools or the concentration of the reactants.
359
Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 26 May 2021
Terms of Use: https://www.spiedigitallibrary.org/terms-of-use
Most techniques do not allow the direct obtention of concentrated colloidal solutions (i.e. a CdS concentration more than a
few lO mo1.l). Higher concentrations require a strong passivation of the particles to prevent their agglomeration.
Nevertheless, using a strong complexing agent directly during the synthesis of the particles limits their growth so that only
small nanoparticles can be obtained (less than 2 nm in diameter). The idea is then to achieve the synthesis of diluted colloids
in the presence of a slightly complexing agent (or in inverted micelles) and, after the synthesis, to graft a much more
complexing agent at the surface of the particles, thus allowing their concentration without flocculation.
Among the best results obtained, the use of phenyithiol as a strong complexing agent led to highly concentrated colloids
in pyridine8. Other standard solvents only allow dispersion with a much lower concentration. This can be explained by the
ability of pyridine molecules to complex surface cadmium ions which are not bounded to thiolates.
In our work, we found that the use of 4-fluoro,phenylthiol (FPhSH) instead of phenylthiol greatly improves the dispersion
of the particles in many organic solvents9,lU. Dispersion can indeed be achieved in acetone, tetrahydrofuran or
dimethylformamid at very high concentration (at least 5 mol.l, i.e. a volumic fraction of about 15%). Details on the
preparations of such colloids, adapted from previous works by other authors4'8, are given elsewhere9' 10. In a first synthesis,
the particles are obtained by direct precipitation in the presence of FPhSH. This leads to small polymeric CdSSPhF clusters,
similar to those described by Dance et a11 . In a second route, the particles are synthesized in inverted micelles and FPhSH is
subsequently grafted at their surface. The size of the particles obtained in this way can be tuned from 1 to 10 nm.
Different alkoxides can be used for the synthesis of silica matrices, but we focused on the TEOS (tetraethoxysilane) which
finally gives purely inorganic matrices, and MTEOS (methyltriethoxysilane) which gives mixed organic/inorganic matrices.
As it was said before, 10 molar% MPTMOS (mercaptopropyltrimethoxysilane) was added as the functionalized alkoxide
which prevents the colloid flocculation. Hydrolysis is performed by the addition of 3 molar equivalent of pH 2.5 water
acidified with HC1, and 3 other equivalents of ethanol. The colloid is added after 12 hours of aging, and the further evolution
of the sol gives a gel, which is dried at 60°C for 72 hours to give the final composite materials (bulk or thin film).
2-4 Luminescence
Luminescence experiments give information on the recombination of the carriers (electrons and holes), excited after the
absorption of light by the semiconductor nanoparticles. Figures 2a and 2b show the luminescence spectra of the same CdS
nanoparticles (2.5 nm in diameter) in a TEOS and MTEOS matrix respectively. In the TEOS sample, the luminescence
spectrum essentially consists in a broad band centered around 600 nm. This band is often observed in CdS nanoparticles, and
is attributed to recombination from deep surface traps. In the case of MTEOS samples, the same band is observed, but it is
associated with a much sharper band peaking near the absorption threshold. This signal is attributed to direct electron-hole
recombination. Another difference between the two samples concerns the quantum efficiency, which is about 0.1% in TEOS
and 1 % in MTEOS. The reasons for the difference between the two matrices is not well understood yet, but it might come
from a change of interactions between the surface of the particles and the host matrix. The presence of hydrophobic methyl
groups at the pore surface in the MTEOS matrix make the surface traps shallower than in TEOS and restrict non radiative
recombination pathways of the excited charge carriers.
Further improvement of the luminescence quantum yield was also obtained by studying the properties of CdS
nanoparticles containing some Mn2+ ions in solid solution. ZnS:Mn and CdS:Mn micronic powders are indeed well-known
phosphors, due to an efficient energy transfer between excited carriers and the Mn2+ d electrons. The synthesis of such
nanoparticles was achieved under conditions similar to those used in the case of pure CdS particles, but we have taken into
account the differences of chemical behavior between the Cd2+ and the Mn2+ ions in solution, which made their
coprecipitation as a solid solution difficult. A great care was also taken to accurately determine the effective concentration of
Mn2 in the CdS particle and its spatial location12"3.
Figure 2-c shows the luminescence spectrum of a sample in which most particles contain about 1 Mn2+ ion. The
luminescence hand is peaking at about 590 nm, which corresponds to the energy of the
4Tl--->6A1 transition energy between the 3d electronic configurations of the Mn ions in CdS crystals. The quantum
efficiency was determined roughly equal to 10%, which is much more than it can be observed in pure CdS. This seems to be
the result of an efficient energy transfer between the excited carriers trapped at the surface of the nanoparticles, and the Mn 3d
electronic configurations. In that case. the surface of the particle acts as a coactivator of the luminescence of the Mn2+ ions.
360
3
=
= = 3
C
=
=
C-' =
=
=
=
In the previous section, CdS nanoparticles were considered as optically active systems, whose properties were investigated
after their incorporation in sol-gel silica matrices. We now change our point of view and consider these particles not only as
distinct objects with interesting optical properties, but also as elementary entities similar to the silica clusters or particles
used for the sol-gel synthesis of transparent matrices.
Starting from a colloidal solution, the further evolution into a gel requires to find successive reactions which would
ensure the bonding between the particles in a controlled way, similarly to the hydrolysis/condensation reactions in the silica
polymerization. The evolution of the colloid into a gel, instead of a precipitate. requires the formation of aggregates with an
open structure. Gelation will only occur when the number and/or the reactivity of the surface sites allowing the bonding
between the particles is low, thus preventing the formation of dense aggregates. This may be induced by the addition to the
colloid of a chemical agent which slowly increases the reactivity of the particles through the elimination of the stabilizing
molecules.
361
C'D
XEROGEL
GEL
QcDO0 4
The remarkable ability of 4-fluorothiophenol to stabilize CdS nanoparticles in poorly complexing solvents such as
acetone or tetrahydrofuran gives us a unique opportunity to observe the sol-gel transition in these colloids. Thiols are indeed
easily oxidized by various agents such as oxygen, hydrogen peroxyde or sodium periodate to give either dithiol or sulfonate
derivatives. As these compounds are no longer bounded to the surface of the particles, they leave reactive sites and thus permit
aggregation. Since the surface of the particles can be slowly activated, random aggregation occurs, leading to lacunar
aggregates, and thus making the sol-gel transition possible.
Gelation was then studied in controlled conditions by using addition of H202 to deaerated CdS colloid9' lO For example.
stiff gels were obtained within a few minutes for 2 nm cristallites occupying a volume fraction of 5%, and treated with 0.3
equivalent of hydrogen peroxide (relative to the thiol concentration). The time evolution of the system was followed by 19F
NMR (figure 4) and small angle X-ray scattering (figure 5).
362
>
=
0,1
Q(nm 1)
Figure 5: Temporal evolution of Small Angle X-Ray Scattering (S.A.X.S) curves for a cadmium sulfide sol (the
diameter of tile CdS particles is about 2.5 inn)
363
Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 26 May 2021
Terms of Use: https://www.spiedigitallibrary.org/terms-of-use
3-2 CdS xerogels
The drying of CdS gels into transparent xerogels has been demonstrated10. Monoliths with dimensions of a few
millimeters were obtained, but the drying step has to cope with the same problems as the ones encountered for silica
monoliths. On the contrary (and as usual for the sol-gel process), thin films with a thickness up to 2 .tm were easily
deposited on glass slides or silicon wafers by spin coating. Characterization by ellipsometry (UVISEL, Jobin-Yvon) evidenced
a refractive index lower than the bulk value (about 2.4), due to the residual organic thiolates and the porosity associated with
the deposition of spheres.
Preliminary experiments on the optical properties of CdS gels and xerogels have shown that the quantum confinement is
still observed, like in the initial colloids. Temperature treatments of CdS thin films have also shown that the nanostructure
remains up to 300°C, but disappears at higher temperature due to the coalescence of the nanoparticles.
n k
0.42
'0.38
'0.34
0.30
'0.26
'0.22
'0.18
'0.14
0.10
'0.06
320 380 440 500 560 620 680 740 800
Wavelength (nm)
Figure 6 : Real (,z) and imaginaiy parts (k) of the refractive index ofa C'dS thin film measured by ellipsometry. (the
size ofthe CdS nanoparticles is about 4 nnz)
4. CONCLUSION
We have presented two approaches for the sol-gel synthesis of transparent solids in which CdS nanoparticles are
homogeneously dispersed. In both cases, the CdS particles are produced by classical techniques of colloid chemistry and final
concentrated solutions are obtained by surface complexation with 4-fluorophenylthiol. A first category of materials consists in
the dispersion of the nanoparticles in sol-gel silica matrices. We have shown that the luminescence of the particles can be
controlled to some extent by changing the chemical nature of the silica matrix (use of hybrid organic/inorganic silica
matrices), or by offering an alternative pathway for the recombination of surface trapped carriers (presence of a single Mn2+
ion inside the particles). A second group of materials was obtained by considering the CdS nanoparticles not only as the
optically active units, but also as the building blocks for the whole solid. The sol-gel synthesis of transparent chalcogenide
materials was then demonstrated, starting from concentrated CdS colloids which are slowly destabilized by the progessive
depassivation of the particles.
Those results tend to show the considerable potentiality of the colloid chemistry and the sol-gel process for the synthesis
of original materials, since the principles given here should go beyond the simple case of CdS and other chalcogenides.
364
1. C. Flytzanis, F. Hache, M.C. Klein, D. Ricard, P. Roussignol, "Progress in Optics XXJX", edited by E. Wolf, Elsevier
Science Publisher B.V., 1991.
2. see for exemple H. Schmidt, "Sol-gel Optics", edited by L.C. Klein, Kiuwer Academic Publishers, 20, 45 1, 1994.
M. Nogami, ibid, 15, 329.
3. T. Gacoin, C. Train, F. Chaput, J.P. Boilot, P. Aubert, M. Gandais, Y. Wang, A. Lecomte, SPIE Proc. 1758, 565
1992.
T. Gacoin, L. Malier, G. Counio, S. Esnouf, J.P. Boilot, L. Audinet, C. Ricolleau, M. Gandais, Mat. Res. Soc. Proc.
435, 643, 1996.
4. N. Herron, Y. Wang, H. Eckert, J. Am. Chem. Soc. 112, 1322, 1990.
Y. Nosaka, N. Otha, T. Fukuyama, N. Fujii, J. Colloid Interface Sci. 155, 23, 1993.
5. A. Fojtik, H. Weller, U. Koch, A. Henglein, Ber. Bunsen-Ges. Phys. Chem 88, 969, 1984.
6. C.B. Murray, D.J. Norris, M.G. Bawendi, J. Am. Chem. Soc. 115, 8706, 1993.
7. P. Lianos, J.K. Thomas, Chem. Phys. Lett. 125, 299, 1986.
8. M.L. Steigerwald, A.P. Allivisatos, J.M. Gibson, T.D. Harris, R. Kortan, A.J. Muller, A.M. Thayer, T.M. Duncan,
D.C. Douglass, L.E. Brus, J. Am. Chem. Soc. 110, 3046, 1988.
9. T. Gacoin, L. Malier, J.P. Boilot, J. Mat. Chem. 7, 6, 859, 1997.
10. T. Gacoin, L. Malier, J.P. Boilot, Chem. Mat., 9,1502, 1997.
11. G.S. Lee, K.J. Fisher, A.M. Vassallo, J.V. Hanna, I.G. Dance, Inorg. Chem. 32, 66, 1993.
12. G. Counio, S. Esnouf, T. Gacoin, J.P. Boilot, J. Chem. Phys. 100, 20021, 1996.
13. G. Counio, T. Gacoin, J.P. Boilot, "Proceedings of the 4' symposium on quantum confinment,
jjth international congress of the Electrochenzical Society", Montreal, in press, 1997.
14. D.W. Schaefer, D. Keefer, in Fractals in Physics, edited by L. Pietronero and E. Tosatti, Elsevier Science Publisher
B.V., 39, 1986.
R. Vacher, T. Woigner, J. Pelous, E. Courtens, Phys. Rev. B 37, 11, 6500, 1988.
365