Anisotropy in periodic mesoporous silica and organosilica films studied by generalized ellipsometry

Article abstract of DOI:10.1063/1.2140472

The dielectric functions of a series of periodic mesoporous silica as well as periodic mesoporous organosilica thin films were measured using generalized variable angle spectroscopic ellipsometry over the spectral range 300-1400 nm. Ellipsometry results indicate that following template removal, both types of films possess uniaxial anisotropy, with the optic axis perpendicular to the plane of the film. This anisotropy is apparently caused by the structural distortion of the channels, oriented primarily parallel to the substrate plane. We also find that the birefringence increases as a function of porosity.

Full text of DOI:10.1063/1.2140472

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Anisotropy in periodic mesoporous silica and organosilica films studied by generalized
ellipsometry
F. C. Peiris, B. D. Hatton, G. A. Ozin, and D. D. Perovic
Citation: Applied Physics Letters 87, 241902 (2005); doi: 10.1063/1.2140472
View online: http://dx.doi.org/10.1063/1.2140472
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/87/24?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Evolution of optical properties of tin film from solid to liquid studied by spectroscopic ellipsometry and ab initio
calculation
Appl. Phys. Lett. 104, 121907 (2014); 10.1063/1.4869722
Stress-induced anisotropy of graphitelike amorphous carbon
J. Appl. Phys. 101, 053522 (2007); 10.1063/1.2562551
Skeletal silica characterization in porous-silica low-dielectric-constant films by infrared spectroscopic ellipsometry
J. Appl. Phys. 97, 113504 (2005); 10.1063/1.1903108
Ordinary optical dielectric functions of anisotropic hexagonal GaN film determined by variable angle
spectroscopic ellipsometry
J. Appl. Phys. 88, 3463 (2000); 10.1063/1.1289224
In situ layer characterization by spectroscopic ellipsometry at high temperatures
AIP Conf. Proc. 449, 336 (1998); 10.1063/1.56816
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.119.67.237 On: Wed, 03 Sep 2014 03:09:10

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
APPLIED PHYSICS LETTERS 87, 241902 ͑2005͒
Anisotropy in periodic mesoporous silica and organosilica films studied by
generalized ellipsometry
F. C. Peiris,a͒ B. D. Hatton,b͒ and G. A. Ozinc͒
Materials Chemistry Research Group, Chemistry Department, University of Toronto, 80 St. George Street,
Toronto, Ontario M5S 3H6, Canada
D. D. Perovic
Department of Materials Science and Engineering, University of Toronto, 184 College Street, Toronto,
Ontario M5S 3H4, Canada
͑Received 9 June 2005; accepted 11 October 2005; published online 5 December 2005͒
The dielectric functions of a series of periodic mesoporous silica as well as periodic mesoporous
organosilica thin films were measured using generalized variable angle spectroscopic ellipsometry
over the spectral range 300–1400 nm. Ellipsometry results indicate that following template
removal, both types of films possess uniaxial anisotropy, with the optic axis perpendicular to the
plane of the film. This anisotropy is apparently caused by the structural distortion of the channels,
oriented primarily parallel to the substrate plane. We also find that the birefringence increases as a
function of porosity. © 2005 American Institute of Physics. ͓DOI: 10.1063/1.2140472͔
In order to sustain the ultralarge scale integration ͑ULSI͒
of semiconductor circuits, intense research is dedicated to
exploring low dielectric constant materials that can provide
low resistance-capacitance time delays in various intercon-
nect structures.1–3 Silica ͑SiO2͒ has been a longtime industry
standard, but with a static dielectric constant of roughly 4.4
is probably twice too large to maintain the ULSI. Periodic
mesoporous films are an exciting class of materials with po-
tential to be used in microelectronics, catalysis, and chemical
sensing applications. Especially, due to their low dielectric
constants, robust mechanical properties, and thermal stabil-
ity, they are attractive candidates to replace SiO2 in intercon-
nect structures. Periodic mesoporous silica ͑PMS͒, first re-
ported in 1992, are synthesized using surfactant as a self-
assembled template, to produce monodispersed, hexagonally
ordered channels with diameters tunable with Angstrom di-
mensional control.4 The surfactant template is subsequently
removed from the as-synthesized material to obtain ordered
mesopores within a silica matrix framework. Extending on
this work, in 1999, we5 and two other groups6,7 indepen-
dently reported the synthesis of periodic mesoporous organo-
silica ͑PMO͒ in which a variety of bridging organic groups
are incorporated into the silica framework without blocking
or occupying the pore volume. The motivation to introduce
different organic groups is based on the ability to tune the
optical, electronic, catalytic, hydrophobicity, hardness, and
density properties of the structure.
to understand both their dispersion and anisotropy more
comprehensively.
Spectroscopic ellipsometry is an efficient method to de-
termine the dielectric functions as well as the thicknesses of
thin
films
without
involving
Kramers-Kronig
transformation.10 Furthermore, this method can determine
the optical anisotropy, and shed light on issues such as mo-
lecular orientation in thin films.11,12 In the present study we
have therefore used ellipsometry to investigate the dielectric
functions for a series of PMS and PMO films spin coated on
silicon substrates in order to better understand their optical
properties.
PMS and PMO films were synthesized using an
evaporation-induced self-assembly technique,13 and details
can be found in Refs. 14 and 9. The PMS films were synthe-
sized from tetramethoxysilane ͑TMOS͒, and the PMO films
were synthesized using
a cyclic three-ring precursor
͓͑͑EtO͒2SiCH2͔3͒.8 Synthesis of the films first involved mix-
ing HCl ͑10−3 M͒ with CTACl surfactant ͑25 wt %͒ and an-
hydrous EtOH. Once mixed homogeneously, the TMOS or
three-ring precursor was added. The CTACl/precursor molar
ratios ͑R͒ were varied to produce two-dimensional hexagonal
mesostructures with increasing volume fraction of surfactant
template, which is used to control the final porosity. After the
addition of the precursor, the solutions were stirred for
30–40 min before spin coating onto Si wafer at
2000–3000 rpm, to deposit films of thickness 500–800 nm.
Finally, in order to remove the surfactant, the films were
treated by solvent extraction ͑EtOH/HCl͒, plasma oxidation,
and calcination ͑400 °C in air for the PMS films, and
300 °C in nitrogen for the PMO films͒ for 5 h, using a
1 °C/min ramp.
In the past few years, there have been many studies re-
ported on periodic mesoporous films, especially in terms of
synthesis and their structural properties.8,9 Very few of these
studies have investigated the dispersion of the optical prop-
erties as a function of the porosity, and none of these have
pursued the question of the degree of anisotropy with respect
to the optical constants of these films. If these structures are
to be used in optoelectronic applications, it is important
The lattice parameters of all of the samples were ob-
tained by x-ray diffraction ͑XRD͒ experiments using a
double crystal diffractometer with Cu K␣1 radiation ͑Bruker
D5000͒. Ellipsometric spectra were obtained using a J.A.
Woollam rotating analyzer ellipsometer between 300 and
1400 nm spectral range. For each sample, ellipsometry spec-
tra were obtained at the incidence angles of 70 and 75 deg.
The hexagonal mesostructure is shown schematically in
Fig. 1͑a͒, where the chemical precursors ͑i͒ and ͑ii͒ are
a͒
On leave from: Department of Physics, Kenyon College, Gambier, OH
43022.
b͒
Also at: Department of Materials Science and Engineering, University of
Toronto, 184 College Street, Toronto, Ontario M5S 3H4, Canada.
Electronic mail: gozin@alchemy.chem.utoronto.ca
c͒
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
0003-6951/2005/87͑24͒/241902/3/$22.50 87, 241902-1 © 2005 American Institute of Physics
129.119.67.237 On: Wed, 03 Sep 2014 03:09:10

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
241902-2 Peiris et al.
Appl. Phys. Lett. 87, 241902 ͑2005͒
FIG. 1. ͑a͒ Schematic illustration of ͑i͒ PMO and ͑ii͒ PMS chemical pre-
cursors and the hexagonal mesostructure, ͑b͒ XRD patterns for as-
synthesized and calcined silica films having R=0.079, ͑c͒ SEM ͑and TEM
inset͒ of film cross section, and ͑d͒ d spacing and porosity of PMS films
with respect to surfactant/Si molar ratio.
FIG. 2. Experimental and modeled ⌿ and ⌬ spectra for the PMO sample
with surfactant/Si molar ratio ͑R͒=0.206. The continuous and dashed lines
represent best-fit results obtained by modeling the PMO layer as an isotropic
and an anisotropic material, respectively.
shown for the three-ring PMO and PMS materials, respec-
tively. Figure 1͑c͒ shows a cross-sectional view of the film
obtained through scanning electron microscopy ͑SEM͒, and
the orientation of the channels in the xy plane ͓transmission
electron microscopy ͑TEM͒ insert of the aligned channels͔.
The high degree of order produces sharp peaks in XRD ͓Fig.
1͑b͔͒, which shift to smaller d spacing due to condensation
contraction after high temperature calcination. Finally, the
change in d spacing ͑and porosity͒ due to increasing surfac-
tant template volume fraction ͑surfactant/Si molar ratio͒ is
shown in Fig. 1͑d͒. The surfactant/Si molar ratio increases
the final surfactant volume fraction in the film, and so can be
used to increase the porosity,15 and decrease the channel d
spacing.16
samples measured in this study, where both rps and rsp were
zero. However, in order to model the experimental spectra
accurately, the periodic mesoporous films had do be consid-
ered anisotropic.
The experimental ⌿ and ⌬, along with the values ob-
tained from our model are shown in Fig. 2 for a representa-
tive sample from this study. For clarity, we have shown only
the spectra obtained at one incidence angle ͑i.e., 75 deg͒. In
the figure, the continuous line showing a very low value of
the error function ͓mean squared error ͑MSE͒=1.1͔, repre-
sents a fit based on uniaxial anisotropy. Keeping the same
layered-model, but by relaxing the anisotropy, one obtains a
fit with a MSE value that is four times higher, represented by
the dashed line. Similarly, the rest of the samples also had
lower MSE values when the mesoporous layer was repre-
sented using an anisotropy model. These results indicate that
the anisotropy model is far more suited in representing the
films studied in this work. It must be noted that the uniaxial
anisotropy model could not be further improved by introduc-
ing a biaxial anisotropic model, which is also confirmed by
the fact that both ⌿ and ⌬ spectra did not show any notable
changes when the experiment was performed for different
azimuthal angle.
The n values derived from the earlier scheme are dis-
played in Figs. 3 and 4 for PMS and PMO films, respec-
tively. Both types of periodic mesoporous films are nega-
tively birefringent; higher n for light propagating with its
electric field vector oscillating along the plane of the film,
compared to light propagating with its electric field perpen-
dicular to the plane of the film. The birefringence for these
films is very small ͑roughly an order of magnitude͒ in com-
parison to polymers that exhibit uniaxial anisotropy.1,18 In
the case of these polymers, it is well established that the
bonding electrons in the chain are mainly localized parallel
to the polymer backbone resulting in a situation where nx
=ny Ͼnz. However, polarized Raman experiments22 per-
formed on both types of mesoporous films studied in this
work reveal that there is no preferential bond ordering within
the channel walls of the film, eliminating the possibility of
bond orientation giving rise to anisotropy. Since we were
For isotropic samples, the two parameters, ⌿ and ⌬
measured by ellipsometry at each wavelength are related to
the ratio of reflection coefficients by
rpp
␳ =
= tan͑⌿͒ei⌬,
rss
where rpp and rss are the complex reflection coefficients for
light polarized parallel ͑p͒, and perpendicular ͑s͒ to the plane
of incidence, respectively. For anisotropic samples, all four
components in the Jones matrix, which relates the p and the
s states of the incident and reflected waves, must be evalu-
ated using generalized ellipsometry.17 Unlike conventional
ellipsometry, where only a single ␳ is measured, in general-
ized ellipsometry, one measures two additional ␳’s ͑i.e.,
rps/rpp and rsp/rss͒ to determine all four matrix elements in
the Jones matrix defined as follows:
rpp rps
J =
,
ͩ ͪ
rsp rss
where rps and rsp are the Fresnel coefficients for mixed po-
larizations, and contains ⌿ and ⌬ values associated with
each of them. However, the measurement of zero off-
diagonal terms in the Jones matrix does not automatically
indicate an isotropic film. This is because an anisotropic
sample with its optic axis directed in one of the principle axis
of the experiment will also have its off-diagonal terms
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
zero.18 This was indeed the case we observed for all of the
able to model the ellipsometry spectra of the as-synthesized
129.119.67.237 On: Wed, 03 Sep 2014 03:09:10

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
241902-3 Peiris et al.
Appl. Phys. Lett. 87, 241902 ͑2005͒
FIG. 3. Dispersion of the index of refraction for template-free PMS films
with different surfactant/Si molar ratios ͑R͒. Ordinary ͑nx =ny͒ and extraor-
dinary ͑nz͒ indices of refraction are represented by solid and dashed lines,
respectively.
FIG. 4. Dispersion of the index of refraction for template-free PMO films
with different surfactant/three-ring precursor molar ratios ͑R͒. Ordinary
͑nx =ny͒ and extraordinary ͑nz͒ indices of refraction are represented by solid
and dashed lines, respectively.
removal procedure, and is caused by the lattice contraction.
The authors acknowledge the National Sciences and En-
gineering Research Council of Canada ͑NSERC͒ for the fi-
nancial support, and thank Professor Eugenia Kumacheva for
providing access to the ellipsometer.
samples ͑i.e., before the removal of the surfactant͒ using
only an isotropic model, it was clear that the surfactant re-
moval resulted in the anisotropy. It has been previously
found that there is an anisotropic contraction of mesoporous
films along z axis ͓see Fig. 1͑c͔͒ due to constraint in the xy
plane.20–22 We found that all three different methods of sur-
factant removal ͑i.e., solvent extraction, plasma oxidation,
and calcination͒ produced an anisotropy although they varied
slightly with respect to the degree of anisotropy, with the
calcination and the solvent extraction giving the highest and
the lowest anisotropy, respectively. There seems to be a di-
rect correlation between the degree of anisotropy and the
contraction of the lattice, as indicated by our XRD results;
the calcination and the solvent-extraction procedures intro-
duce the highest and the lowest contraction of the ͑100͒ lat-
tice spacing, respectively. In addition, as the surfactant/
precurcor ͑R͒ is increased, and the resulting porosity
increases in these structures, there is a slight enhancement in
the birefringence ͑see Figs. 3 and 4͒ because the lattice is
able to contract more in the z axis.
These results therefore clearly indicate that the aniso-
tropy is induced because of the contraction of the lattice.
While first-principle calculations that would identify the
changes to the bonding structure—as a result of the lattice
contraction—are beyond the scope of this paper, one could
speculate that this distortion probably results in more bond-
ing electrons parallel to the channels ͓i.e., x and y directions
in Fig. 1͑a͔͒ compared to perpendicular direction ͑i.e., z
axis͒, and therefore causes n in the x and y directions to be
slightly higher in comparison to the n in the z direction.
We have investigated the optical properties of PMS and
PMO thin films using spectroscopic ellipsometry. The ex-
perimental ellipsometry spectra were best modeled by repre-
senting the mesoporous films as uniaxially anisotropic with
its optic axis directed perpendicular to the plane of the film.
We find that both types of mesoporous films are negative
birefringent, where the index of refraction perpendicular to
the film is smaller than the index of refraction parallel to the
film. This anisotropy is a direct consequence of the surfactant
1J. J. Senkevich and S. B. Desu, Appl. Phys. Lett. 72, 258 ͑1998͒.
2K. Postava and T. Yamaguchi, J. Appl. Phys. 89, 2189 ͑2001͒.
3H. J. Lee, C. L. Soles, D. W. Liu, B. J. Bauer, E. K. Lin, W. Wu, and A.
Grill, J. Appl. Phys. 95, 2355 ͑2004͒.
4C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli, and J. S. Beck,
Nature ͑London͒ 359, 710 ͑1992͒.
5T. Asefa, M. J. MacLachlan, N. Coombs, and G. A. Ozin, Nature
͑London͒ 402, 867 ͑1999͒.
6S. Inagaki, S. Guan, Y. Fukushima, T. Oshuna, and O. Terasaki, J. Am.
Chem. Soc. 121, 9611 ͑1999͒.
7B. J. Melde, B. T. Holland, C. F. Blanford, and A. Stein, Chem. Mater. 11,
3302 ͑1999͒.
8K. Landskron, B. D. Hatton, D. D. Perovic, and G. A. Ozin, Science 302,
266 ͑2003͒.
9B. D. Hatton, K. Landskron, W. Whitnall, D. D. Perovic, and G. A. Ozin,
Adv. Funct. Mater. 15, 823 ͑2005͒.
10P. Lautenschlager, S. Logothetiidis, L. Vina, and M. Cardona, Phys. Rev.
B 32, 3811 ͑1985͒; M. R. Buckley, F. C. Peiris, O. Maksimov, M. Muñoz,
and M. C. Tamargo, Appl. Phys. Lett. 81, 5156 ͑2002͒.
11M. Losurdo, M. M. Giangregorio, P. Capezzuto, G. Bruno, F. Babudri, D.
Colangiuli, G. M. Farinola, F. Naso, and A. Irene, Macromolecules 36,
4492 ͑2003͒.
12B. P. Lyons and A. P. Monkman, J. Appl. Phys. 96, 4735 ͑2004͒.
13Y. Lu, R. Ganguli, C. A. Drewien, M. T. Anderson, C. J. Brinker, W.
Gong, Y. Guo, H. Soyez, B. Dunn, M. H. Huang, and J. I. Zink, Nature
͑London͒ 389, 364 ͑1997͒.
14B. D. Hatton, D. D. Perovic, and G. A. Ozin ͑unpublished͒.
15A. R. Balkenende, F. K. de Theije, and J. C. K. Kriege, Adv. Mater.
͑Weinheim, Ger.͒ 15, 139 ͑2003͒.
16S. Besson, T. Gacoin, C. Ricolleau, C. Jacquiod, and J. P. Boilot, J. Mater.
Chem. 13, 404 ͑2003͒.
17M. Schubert, Phys. Rev. B 53, 4265 ͑1996͒.
18M. Losurdo, G. Bruno, and E. A. Irene, J. Appl. Phys. 94, 4923 ͑2003͒.
19O. Dag and G. A. Ozin, Adv. Mater. ͑Weinheim, Ger.͒ 13, 1182 ͑2001͒.
20H. Yang, N. Coombs, I. Sokolov, and G. A. Ozin, Nature ͑London͒ 381,
589 ͑1996͒.
21M. Klotz, P. A. Albouy, A. Ayral, C. Menager, D. Grosso, A. Van der Lee,
V. Cabuil, F. Babonneau, and C. Guizard, Chem. Mater. 12, 1721 ͑2000͒.
22D. Grosso, F. Babonneau, P. A. Albouy, H. Amenitsch, A. R. Balkenende,
A. Brunet-Bruneau, and J. Rivory, Chem. Mater. 14, 931 ͑2002͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.119.67.237 On: Wed, 03 Sep 2014 03:09:10