Supplemental material
TiO2@Silica nanoparticles in a Random Laser: Strong relationship of silica shell
thickness on scattering medium properties and Random laser performance.
Ernesto Jimenez-Villar1,5*, Valdeci Mestre2, Paulo C. de Oliveira2, Wagner M. Faustino3, D. S. Silva4,
Gilberto F. de Sá5
1Instituto
de Ciencia Molecular, Universitat de València. C/ Catedrático José Beltrán Nº 2, 46980 Paterna Valencia, Spain
2Departamento
de Física, Universidade Federal da Paraíba, João Pessoa, 58051-970, Paraíba, Brazil
3Departamento
de Química, Universidade Federal da Paraíba, João Pessoa, 58051-970, Paraíba, Brazil
4
Instituto de Química, Universidade Estadual de Campinas, Campinas, 13083-970, SP, Brazil
5Departamento
de Química Fundamental, Universidade Federal de Pernambuco, Recife, 50670-901, Pernambuco, Brazil
Materials
R6G laser dye (C28H31N2O3Cl), with molecular weight 479.02 g/mol, supplied by Fluka:
Ethanol alcohol (C2H5OH) with spectroscopic grade purity, supplied by Alphatec: Tetraethyl-ortho-silicate (TEOS), supplied by Sigma-Aldrich. The titanium dioxide (TiO2
nanoparticles 410 nm) of rutile crystal structure was acquired from DuPont Inc (R900).
Chemical synthesis of silica coating on TiO2 nanoparticles
In the first stage, 2 g of TiO2 NPs were diluted in 250 ml of absolute ethanol. The solution of
TiO2 nanoparticles was then divided into four portions: one 125 ml (sample γ), another 100
ml and two equal portions of 12.5 ml. Three portions - one of 125 ml and two of 12.5ml were placed in a bath at 5 °C and TEOS, previously diluted in ethanol at 10%, was added. For
the sample γ, the 10% diluted solution of TEOS was added in 110 portions of 100 μl during
the course of 1 hour. For the others two (12.5 ml) samples, the 10% diluted solution of TEOS
was added in 10 and 50 portions of 10 μl corresponding to α and β samples, respectively. The
solutions were stirred during the TEOS addition. Previously, before adding the TEOS, the
solutions of TiO2 NPs were placed in an ultrasound bath for 20 minutes to disperse the
particles. The proportions of TEOS added in the TiO2 suspensions, regarding the amount
added in the γ sample, were 9% and 45% for the samples α and β, respectively.
The TEOS hydrolysis and subsequent condensation of silica on the TiO2 surface was
provoked, taking advantage of the catalytic effect in the nearby surroundings of TiO2 NPs.1,2
In addition, the possible accumulation of water molecules around the nanoparticles would
favor TEOS hydrolysis. Probably, the electric potential associated with the surface of the
nanoparticles themselves, in conjunction with the higher dielectric permittivity of the water,
should promote the accumulation and polarization of water molecules in the vicinity of TiO2
NPs.
3
Self superficial hydrolysis of TiO2 nanoparticles
4
would also favor the TEOS
hydrolysis. The superficial irregularities of the TiO2 nanoparticles must cause intensification
of superficial electric potential in the regions with a smaller curvature radius. This
phenomenon should increase the medium polarization, water accumulation and catalytic
effects in these regions, leading to silica coating irregularities. On the other hand, the silica
shell volume was not proportional to amount of TEOS added. Figure 1S shows the silica shell
volume (in a.u.) as a function of the added TEOS percentage. It is clear that the increased
silica shell volume increases with additional amounts of TEOS. This could be caused by the
enhanced catalytic effect of the silica surface, by which the hydrolyzed TEOS proportion
should increase with the silica surface area. The complete coating of TiO2 surface by silica
shell requires a critical TEOS amount (silica thickness), above which the TiO2 surface is
entirely protected by the silica shell, avoiding the nanoparticle photodegradation.
Supplementar Fig. 1S. Influence of the added TEOS amount on the silica shell volume of
TiO2@(α,β,γ)Silica nanoparticles determined by ED-XRF.
Characterization and sample preparation:
Transmission electron microscopy (TEM) and electron-energy-loss spectroscopy (EELS)
were done using a Carl Zeiss Libra 120 kV transmission electron microscope. The
commercial carbon-coated Cu TEM grids were immersed in the suspensions of
TiO2@(β,γ)Silica nanoparticles that had previously been diluted 100-fold and then left to dry,
before being introduced into the microscope. The stoichiometric ratio (Ti/Si) was determined
by Energy Dispersive X-Ray fluorescence (ED-XRF) using an X-ray spectrometer SIEMENS
D5000. The sample was prepared in three steps: precipitation, washing and deposition. The
films prepared for ED-XRF analysis were deposited onto a copper substrate (chemically
different from the materials under study). The samples were prepared by drop casting of
TiO2@(α,β,γ)Silica suspensions in order to obtain the mass percentage ratios (Ti/Si). Table IS
shows the mass percentage ratio (Ti/Si) (determined by ED-XRF) and the average SST of the
three TiO2@(α,β,γ)Silica samples. The silica shell thicknesses were calculated considering
the typical density of the silica obtained by the TEOS hydrolysis
5
(between 2.0 and 2.2
g/cm3).
TABLE IS. Mass percentage ratio (Ti/Si)% and calculated silica shell thicknesses (2.1 g/cm3
silica density).
Samples
α
β
γ
Mass (Ti/Si)%
(96.5/3.5)
90/8
70/30
Silica Thickness (nm)
4.4
10.4
39.2
Experimental setup for random laser and scattering mean free path measurement:
Supplementary figure 2a shows a schematic diagram of the random laser (RL) experimental
setup. The pumping source of the random laser was the second harmonic of a Q-switched
Nd:YAG Continuum Minilite II (25 mJ, λ = 532 nm, with a pulse width of ~6 ns, repetition
rate up to 15 Hz and spot size of 3 mm). The pump laser beam was incident upon the sample
at 15 degrees. The laser power was regulated through neutral density filters (NDF), a
polarizer and a half wave plate. The samples were accommodated in a 2 mm path length
quartz cuvette. The emission spectra were collected through a multimode optical fiber (200
μm), coupled to a spectrometer HR4000 UV-VIS (Ocean Optics) with a 0.36 nm spectral
resolution (FWHM). The collection angle was ~60 degrees with respect to the sample
surface, that is, 45 degrees with respect to the incident pumping beam. The liquid samples
were placed in an ultrasound bath for about 10 minutes before recording the spectrum, in
order to obtain the same dispersion of nanoparticles (initial conditions) in all measurements.
Supplementar Fig. 2S. (a) Schematic diagram of the RL experimental setup; NDF neutral
density filter; (b) Schematic diagram of the experimental setup for the determination of
transmitted intensity as a function of slab thickness. L1 and L2, lens; F+F, cell consisting in
two optical flats (quartz), mounted on a translation stage; PH1 and PH2, pinholes; OP, optical
fiber to collect the light in the spectrometer.
Figure 2Sb shows a schematic diagram of the scattering mean free path measurement. The
transmitted intensity It was determined as a function of slab thickness: the laser beam (second
harmonic of Q-switched Nd:YAG Continuum Minilite II, λ = 532 nm, 1 μJ (1mJ attenuated
103 times by neutral density filters), with a pulse width of ~6 ns, repetition rate 10 Hz) was
passed through a positive lens L1 (200 mm focal length) so as to obtain the focus with its
waist near the pinhole PH1 (600μm diameter). The cell consisted of two optical flats F in
wedge form, in this way; the slab thickness depends on the incidence point of the cell.
Another pinhole, PH2 (1200μm diameter), was positioned 80 mm away to PH1 in order to
reduce the diffuse light. Yet another lens, L2 (50 mm focal length), allowed focalization of the
coherent light on the optical fiber (200μm).
TABLE IIS. Silica shell thickness calculated from ED-XRF measurements, ls determined by
transmission experiment and ls calculated by Mie theory for the TiO2 and TiO2@(α,β,γ)Silica
systems.
Samples
γ
β
α
TiO2
SST (nm)
39.2
10.4
4.4
0
ls (μm) Measured
20.1±0.4
27.5±0.4
31.3±0.8
52±4
28
31
40
ls (μm) Calculated 20
Table IIS shows the SST and ls values (determined by transmission experiment and
calculated by Mie theory) for the TiO2 and TiO2@(α,β,γ)Silica samples. The ls value
calculated by Mie theory for a TiO2 particles (0.41μm diameter) suspension (in ethanol) at
[5.6x1010 NPs/ml] is ~40 μm (lower than TiO2 (~52 μm) and higher than TiO2@αSilica (~31
μm). Therefore, ls behavior could be associated with the OCS improving and light coupling
enhancement (LCE) with the TiO2 scattering cores provided by the silica shell. The
theoretical treatment of the LCE effect by Mie theory for these TiO2@(αβγ)Silica NPs could
be very complex. Especially, if we consider the silica shell irregularities, which would mean
that, the refractive index is not constant on the silica shell. Therefore, we have made some
approximations in order to understand the ls values experimentally obtained. The ls value
measured for TiO2@αSilica (~ 31μm) could be theoretically obtained (Mie theory), if we
consider that TiO2@αSilica Nps have a mean refraction index and diameter of 2.33 and 0.425
μm, respectively. This calculation has been made by taking account the average refraction
index of a TiO2 particle (0.41 μm) with a silica shell of 7 nm thickness. For TiO2@βSilica
system (ls = 27.5 μm), it is required nanoparticles with a refraction index of 2.15 and a
diameter of 0.44 μm. These values of refraction index and diameter were calculated by
considering a silica shell of 15 nm thickness. For the ls calculus of TiO2@γSilica system, it
was considered TiO2 nanoparticles with 0.41μm diameter but at the double of volume
concentration. The medium refractive index was considered 1.4, which would be the average
refraction index between ethanol and the irregular silica shell (~50 nm). These
approximations for the TiO2@γSilica system have been made because the light should be
dispersed by the TiO2 cores (0.41μm diameter), however, the NP scattering cross section
should be equivalent to particles with a higher diameter (0.51 μm).
For the previous
calculations, the considerate SST values have been higher than those determined from EDXRF measurements. Notice that the LCE effect should come from outside of the silica rough
shell.
Figure 3S shows the intensity increase and the bandwidth narrowing of RL emission spectra
with the increase of pumping energy fluence. The emission intensity increases around 3
orders for pumping fluencies from 0.6 mJ/cm2 to 60 mJ/cm2 (below and above RL threshold,
respectively).
Supplementar Fig. 3S. Typical emission spectra of RL for the four samples; TiO2 (red),
TiO2@αSilica (black), TiO2@βSilica (green) and TiO2@γSilica (magenta) at pumping
energy fluencies; (a) below (0.6 mJ/cm2), (b) around (4.8 mJ/cm2) and (c) above (60 mJ/cm2)
RL threshold.
Figure 4S shows a photograph of the suspensions of TiO2 and TiO2@αβγSilica nanoparticles
in an ethanol solution of R6G. The solution containing TiO2 nanoparticles presents a duller
appearance and is more turbid than those of TiO2@(α,β,γ)Silica solutions, where the
rhodamine pigment is more pronounced with the increase of SST.
Supplementar Fig. 4S. Photograph of ethanol solutions of R6G containing nanoparticles of:
(a) TiO2, (b) TiO2@αSilica, (c) TiO2@βSilica and (d) TiO2@γSilica.
Photodegradation study
The photocatalytic pathway involves a reaction on the TiO2 surface following several steps:
1) photogeneration of electron–hole pairs by exciting the semiconductor with >3.2 eV light,
2) separation of electrons and holes by traps existing on the TiO2 surface and 3) a redox
process induced by the separated electrons and holes with the adsorbates present on the
surface.
The exponential decrease of RL intensity, for the TiO2 and TiO2@αβSilica systems, indicates
that the photodegradation is proportional to its derivative, in regards to the photodegradation
rate. This means that the charge transfer,
6
and therefore the redox reaction,
7
will cause a
greater charge transfer in the next laser shot. Thus, one might think that the high
concentrations of charges created in the TiO2 core of nanoparticles (TiO2 and
TiO2@αβSilica) at high pumping fluencies must react with the proper surface of TiO2 core
(devoid of silica coating), reducing Ti4+ and oxidizing O2-. This process results in oxygen
vacancies, 8 which act as traps for photoelectrons. These electrons, trapped near the surface,
act as a source of electron transfer coming from these superficial traps, increasing the
efficiency of the redox process.
9
Additionally, the creation of oxygen vacancies in TiO2
causes a progressive decrease of gap on the TiO2 surfaces, which is reflected in the
progressive increase in the creation of electron-hole pairs. This photo-darkening effect is
observed in films of TiO2 exposed to successive irradiation of laser pulses. 8,10
The nanoparticles photodegradation percentage, measured by the RL emission signal after readdition of R6G, shows a linear decrease with SST (figure 4c). This is most likely caused by
the increase of the TiO2 surface coated with the silica shell. If this linear behavior is
extrapolated, the 100% of RL initial signal would be reached for a SST of ~24 nm. The
spaces (devoid of silica shell) most likely originate from an irregular coating process. The
silica shell obtained by this method requires a critical thickness, above which the TiO2
surface is protected entirely with the silica shell, avoiding the photodegradation of
nanoparticles and R6G molecules
In the TiO2@γSilica system, the R6G photodegradation can be explained as follows:11

Through the ethanol radical CH3 ĊHOH reaction with R6G ground state molecules,
12
the free radical CH3 ĊHOH is produced by energy transfer from the R6G molecules in a higher
triplet state, which is produced by two sequential single-photon absorptions. 13
References
1
Rafi-ud-din, Qu Xuanhui, Li Ping, Lin Zhang, Wan Qi, M. Zubair Iqbal, M. Yasir
Rafique, M. Hassan Farooq, and Islam-ud-din, J. Phys. Chem. C 2012, 116 (22), 11924–
11938
2
Kamal Abderrafi, Ernesto Jimenez, T. Ben, S.I. Molina, R. Ibáñez, V. Chirvony, J.P.
Martínez-Pastor, J. Nanoscience and Nanotechnology 2012, 12 (8), 6774-6778
3
A. Luzar, S. Svetina, and B. Zeks, J. Chem. Phys. 1985, 82, 5146
4
Philippe Leroy, Christophe Tournassat, Mohamed Bizi, Journal of Colloid and Interface
Science 2011, 356, (2), 442-453
5
Basudeb Karmakar; Goutam De; Dibyendu Ganguli, Journal of Non-Crystalline Solids
2000, 272, 119-126.
6
Amy L. Linsebigler, Guangquan Lu, and John T. Yates, Jr. Chem. Rev. 1995, 95, 735-758.
7
Serpone, N., Pelizzetti, E., Eds.; Wiley: New York Chapter 7, van Damme, H. Supports in
Photocatalysis 1989.
8
M. Tsukamoto, N. Abe, Y. Soga, M. Yoshida, H. Nakano, M. Fujita, J. Akedo, Appl Phys
A 2008, 93, 193–196.
9
Heinz Ceriscber and Adam Heller, The Journal of Physical Chemistry 1991, 95 (13),
5261.
10
Tsukamoto Masahiro, Shinonaga Togo, Takahashi masanari, Fujita Masayuki,· Abe
Nubuyuki, Transactions of JWRI 2011, 40 (1), 21-23.
11
E. Jimenez-Villar, Valdeci Mestre, Paulo C. de Oliveira, Gilberto F. de Sá, Nanoscale
2013, 5, 12512-12517.
12
Adrian Dunne, Michael F. Quinn, J. Chem. Soc. Faraday Trans. 1 1976, 72, 2289-2295.
13
Yamashita, M.; Kashiwagi, H., IEEE Journal of Quantum electronics 1976, 12 (1), 90-95.