SCIENCE
Volume: 312
Issue: 5779
Pages: 1504-1508
Published: JUN 9 2006
Converting Ceria Polyhedral Nanoparticles into Single-Crystal Nanospheres
Xiangdong Feng1, *, Dean C. Sayle 2, Zhong Lin Wang 3, *, M. Sharon Paras4, Brian Santora1, Anthony
C. Sutorik4, Thi X. T. Sayle2, Yi Yang1, Yong Ding3, Xudong Wang3, Yie-Shein Her1
1 Ferro
2 Cranfield
3 School
Corporation, 7500 E. Pleasant Vally Road, Independence, Ohio USA
University, Defense Academy of the United Kingdom, Shrivenham, Swindon SN6 8LA, UK
of Materials Science and Engineering, Georgia Institute of Technology, Atlanta, Georgia
30332-0245 USA.
4 Nanocerox
Inc., 712 State Circle, Ann Arbor, MI 48108 USA
* Corresponding authors: E-mail: shawn.feng@jhresearchusa.com (XDF), Current address: James
Hardie, 10901 Elm Ave. Fontana, CA 92337 USA; zhong.wang@mse.gatech.edu (ZLW).
Ceria (CeO2) nanoparticles are one of the key abrasive materials for chemical-mechanical
planarization (CMP) of advanced integrated circuits. However, CeO2 nanoparticles
synthesized by all existing techniques are faceted with irregular faceted-shapes, and they
scratch the silicon wafers with increased defect concentrations. Here, we show for the first
time an innovative approach for large-scale synthesis of spherical, single-crystal, CeO2
nanoparticles. Our synthetic strategy involves doping the CeO2 system with titanium, using
flame temperatures that facilitate crystallization of the CeO2, yet retains the TiO2 in a molten
state. In conjunction with Molecular Dynamics simulation, we show that under these
conditions, the inner CeO2 core evolves in a single-crystal spherical-shape without faceting,
because, throughout the crystallization, it is completely encapsulated by a molten 1-2 nm
shell of TiO2, which, in liquid state, minimizes the surface energy. The single-crystal
nanospheres reduce CMP defects by 80% and increase the silica removal rate by 50%,
which will facilitate precise and reliable mass-manufacturing of chips for nanoelectronics at a
precision of sub-nanometers. The principle demonstrated here could be applied to other
oxide systems.
1
Oxide nanoparticles have a wide range of applications spanning catalysts 1, 2, fuel cell 3,
chemical-mechanical planarization of advanced integrated circuits 4, and phosphor/luminescence 5 and
many more. Ceria nanoparticles, for example, are a key abrasive nanomaterial for chemicalmechanical planarization (CMP) of advanced integrated circuits. CMP using ceria, for example, took
over 60% of the $1 billion market for nanomaterials in 2005 6. As the size of the integrated circuit
architecture decreases rapidly in today’s climate for chip miniaturization and nanorization, defects
must be reduced by at least 50% to make chip mass-manufacturing viable for each next-generation.
Crystalline ceria nanoparticles can be synthesized via a wealth of different methods, including
room temperature solution precipitation 7, 8, microwave-assisted hydrothermal route 9, hydrothermal
crystallization 10, microemulsion 11, mechanochemical processing 12, thermal decomposition 13, aerosol
pyrolysis
14
, sol-gel method
15
, thermal hydrolysis
16
, and solvothermal synthesis
17
. But ceria
nanoparticles synthesized by all existing techniques are faceted and possess sharp edges, corners and
apexes, which are more prone to scratch the silicon wafers and limit the CMP rates. Clearly, for
superior performance, spherical nanoparticles would be ideal. However, this is a grant challenge to the
existing synthesis approaches, which usually produce faceted and/or irregular shaped nanoparticles
that are enclosed by the low energy surfaces, while spherical single-crystal nanoparticles must have
energetically unfavorable high index surfaces. As of today, there is no breakthrough technique that
can produce spherical ceramic particles.
Here, we describe an innovative synthetic strategy for rational design and synthesis of singlecrystal ceria nanospheres, which was achieved by doping with Ti(IV) and crystallizing under flame
conditions. This is the first known example where oxide/ceramic nanoparticles pertain spherical
shapes yet preserve a single-crystalline lattice structure. The nanparticles are shown to facilitate a 50%
increase in the silica removal rate and 80% reduction in defects for CMP in comparison to the faceted
nanoparticles. This technology will allow precise and reliable large-scale fabrication of chips for
microelectronics and nanoelectronics, and establishes a foundation for the synthesis of spherical
single-crystal, inorganic nanoparticles for industrial and scientific applications.
The nanospheres of CeO2 and Ce1-xTixO2 (0 ≤ x ≤ 0.25) reported here were produced via a
liquid-phase flame spray pyrolysis (L-FSP) of solutions of cerium and titanium precursors dissolved
in a flammable solvent (i.e., an alcohol) 18 (see Fig. S1 in 19). The alcohol solution of metal precursors
is pumped 150g/min into an atomization device that sprays a fine mist into the combustion chamber.
The mist is directly ignited by pilot torches in the line of the spray, leading to instantaneous
combustion of the metal precursors and generation of the desired metal oxide as a nanoparticulate
"smoke". Depending on experimental conditions, combustion occurs in the range of 1200-2500 °C.
2
Temperatures are rapidly quenched moving downstream from the initial combustion zone, leading to a
measure of kinetic control over the chemical reaction, crystallization, phase evolution, and particle
growth. Due to the short particle residence time, atomic diffusion and particle growth end rapidly
after initial particle formation, tending to quench out oxide nano-powders. The nanospheres are
collected as the product stream (powder plus CO2 and water vapor) is drawn from the combustion
chamber by exhaust fans into powder collectors. The apparatus used for this study can generate
nanoparticles at a rate of about up to 300 g/h.
Pure ceria nanoparticles without doping were synthesized through L-FSP of a solution made
from dissolving cerium carbonate into propionic acid 20. The non-doped CeO2 nanoparticles show
faceted and sometime irregular shapes (Fig. 1A). The as-synthesized nanoparticles have a large size
variation. Transmission electron microscopy (TEM) image displays that each nanoparticle is a singlecrystal (Fig. 1B). Electron diffraction pattern acquired from dozens of nanoparticles indicate that they
are cubic ceria phase with the fluorite (CaF2) type structure (Fig. 1C). High-resolution TEM clearly
show that the nanoparticles are dominated by {111} and {001} type of facets and some of them
exhibit octahedral (OT) and truncated octahedral (TO) shape (Fig. 1D). For cubic structured ceria,
{111} and {001} are the dominant facets possibly due to low surface energy. The surfaces of the
nanoparticles are atomically sharp, abrupt and clean without visible contaminant (see Fig. S2 in (19)).
In comparison to image simulation, the (001) surface is likely terminated with oxygen (see the
simulated image in Fig. S2 in 19).
Ceria nanospheres doped with varied amount of titanium were synthesized through L-FSP of
a solution made from dissolving cerium carbonate into propionic acid plus stoichiometric amounts of
titanium (IV) (triethanolaminato) isopropoxide 21. The cerium oxide containing 6.25 atom% of Ti are
dominated by spherical shape (Figs. 1E), and each particle is single-crystal (Fig. 1F). High-resolution
TEM shows that a nano particle is single-crystal spherical-like with visible {111} and {001} as well
as high index facets (Fig. 1G), and the particle shape is close to be an ideal sphere. The smaller size
particles are single-crystal and more spherical (Fig. 1H). For larger size nanospheres, the facets are a
little more pronounced and the {110} type of facets can also be identified (Fig. 1I). The crystal lattice
extends to the surface of the nanosphere and the surfaces are fairly clean and atomically sharp.
For the ceria nanosphere doped with 12.5 atom% Ti, the particles are dominated by spherical
shape and single-crystal (Fig.2A). High resolution TEM clearly shows that a large particle is nearspherical and its surface is covered with a uniform amorphous thin layer (1-2 nm) (Fig. 2B). The areas
of the {111} and {001} types of facets are largely reduced with an increased formation of high index
3
planes, which is required for forming spherical shape. A smaller size particle displays similar
characteristic as the bigger ones (Fig. 2C).
To identify the nature of the surface amorphous layer, chemical maps of Ce and Ti were
acquired from the 12.5 atom% Ti doped nanospheres using a scanning TEM (STEM) with a probe
size of ~1 nm. The morphology and distribution of the nanospheres is displayed in the bright-field
STEM image (Fig. 2D), and corresponding maps for Ce and Ti acquired using the X-ray signal
emitted from the sample as the probe was scanned across the sample are shown in Fig. 2E and 2F,
respectively. The clear evidence is that Ti atoms are distributed at the surface of the particles, and
some precipitated TiO2 particles have also been found. The data clearly show that the surface
amorphous layer is TiO2.
To complement our experiments and help rationalize the sphericalization of CeO2
nanoparticles when doped with Ti(IV), we performed Molecular Dynamics simulations using the
DL_POLY code 22 and potential parameters taken from references 23, 24. Most atomistic simulations
start by proposing a structural model, which is then simulated using, for example, Molecular
Dynamics (MD). However, this has the disadvantage of requiring the simulator to choose or design,
perhaps erroneously, the structural model. Clearly, the best way of capturing realistic atomistic
models that are not influenced artificially or pre-determined by the simulator, is to simulate the
nanoparticle synthesis process itself. Here, our experimental synthetic strategy involved crystallizing
nanoparticles in flame, which we have endeavoured to simulate. In particular, atomistic models for
our nanoparticles were generated by constructing an amorphous/molten precursor, which was then
crystallized, at high temperatures, and quenched to room temperature.
At the beginning of the simulation, two identical ‘cubes’ of CeO2, comprising 15,972 atoms
(5324 Ce, 10648 O), were generated. One cube was doped with 25 atom% Ti (1330 atoms - located at
the surface), while the other nanoparticle remained ‘pure’. Both cubes were first amorphosized and
then recrystallized at high temperature using MD simulation. Finally, the two nanoparticles were
cooled (quenched) to room temperature; snapshots, showing the crystallization of the Ti-doped
system, are presented in Fig. 3(A). Further simulation details can be found in reference 25.
Fig. 3(B-H) shows the (room temperature) structure of the nanoparticles at the end of the
simulation. The undoped nanoparticle has an octahedral shape/morphology with {111} surfaces,
truncated with {100} facets and includes sharp edges and corners. Conversely, the Ti-doped CeO2
nanoparticle is effectively spherical. Both structures are in accord with our experimentally synthesized
nanoparticles (Figs. 1 and 2).
4
The crystallization process, starting from the amorphous precursor to the final crystalline
structure, was then studied by generating and then analyzing a movie of the crystallization using
molecular graphics. In the pure system, a single, fluorite-structured (CaF2), crystalline ‘seed’ was
observed to evolve ‘naturally’ within the amorphous sea of ions (Ce, O) during the MD simulation. Ce
and O surrounding this seed then started to condense onto its surface propagating the crystallization
front, which traverses through the nanoparticle. Moreover, as the crystallization front impinged the
surface, the crystallization was observed to evolve energetically stable {111} surfaces resulting in a
fluorite-structured single crystal of CeO2 with {111} and {100} facets (Fig. 3(B-D)). The simulations
also revealed that energy was liberated during this process, which reflects the latent heat of
crystallization.
Similar to the pure CeO2 nanoparticle, the Ti-doped system also evolved a fluorite-structured
crystalline seed, Fig. 3A, and the amorphous sea of ions (Ti, Ce and O) started to condense onto its
surface. The crystallization front traversed through the nanoparticle emanating spherically from the
crystalline seed, fig. 3A, and continued until it consumed the entire CeO2 core. However, in contrast
to the pure system, as the crystallization front impinged on the outer surfaces of the CeO 2 it did not
start to evolve {111} and {100} because it is enveloped by the amorphous TiO2 shell. Moreover, as
the crystallization front moves out to the surface (TiO2 region), the TiO2 does not crystallize – rather it
remains amorphous. This enables the nanoparticle to retain its sphericity, which is driven by the
system minimizing its surface energy to facilitate an energetically more stable nanoparticle. This is in
contrast to the facetted surfaces, including sharp edges and corners, associated with the pure CeO2
nanoparticles.
By combining our experimental and simulation data, we have been able to rationallize the
evolution of spherical nanoparticles: Specifically, the energy difference between CeO2 surfaces is
significant and therefore its morphology is dominated by low index (energetically stable) surfaces
such as {111} (Fig. 1), which results in sharp edges and corners. However, to manufacture spherical
single-crystal CeO2 nanoparticles, higher index (less stable) surfaces must be formed. Our synthetic
strategy of generating the nanoparticles in flame at about 2500oC facilitates the crystallization of the
cerium oxide (CeO2, m.p. 2400°C), while maintaining the TiO2 in a molten state (TiO2 mp 1843oC). In
particular, the CeO2 occupies the core region of the nanoparticle and is encapsulated by a TiO2 shell;
limited Ti(IV) is incorporated into the lattice. Here, in contrast to the pure CeO2 system, the CeO2 core
crystallizes while enveloped by a TiO2 shell. Crucially, because the CeO2 surfaces are covered with
amorphous TiO2 as they evolve, the energy difference between different surfaces is no longer
significant. Accordingly, the CeO2 evolves a variety of surfaces, which results in a spherical
5
nanoparticle. Therefore, it is the flame temperature that facilitates sphericalization of the Ti-doped
CeO2 nanoparticles.
The single-crystal Ce1-xTixO2, (0 ≤ x ≤ 0.25) nanospheres have a key application as an
abrasive material for chemical-mechanical planarization (CMP) of advanced integrated circuits
because of its special spherical and crystalline structure characteristics. The primary function of CMP
is to smooth a nominally macroscopically flat wafer at the feature (or micro-level), i.e., planarize
features. Therefore, two of the important parameters to evaluate a CMP slurry are the uniform
material removal rate across the wafer and the defectivity of the wafer. A good CMP slurry is
expected to achieve good removal rate with low defectivity. Cerium oxide particles have been utilized
in the oxide or silicon wafer polishing due to its unique selectivity and good oxide removal rate. The
cerium oxide particles have always been crystalline and cubic in structure. It is believe that the
relatively sharp edges of the cubic shaped ceria particles (Fig. 1) may gouge and/or scratch wafer
surfaces being planarized, which is particularly detrimental as the size of the integrated circuit
architecture decreases rapidly in today’s climate for chip miniaturization and nanorization.
Four CMP slurries were prepared by mixing 1.0 wt% of the spherical single-crystalline Ce1xTixO2
powders with a chemical additive, proline (see Supplementary Materials Table 1). The CMP
efficiency is characterized by the removal rate of thermal oxide, SiO2 (26), which is shown in Fig. 4a
and Table S1 in (19). The rate for the non-doped CeO2 nanoparticles of slurry A is at 195 nm/minute.
The SiO2 removal rates gradually increased to 224 and 301 nm/minute as the Ti content in the
nanoparticles Ce1-xTixO2 increased to x = 0.0625 and 0.125, respectively. This increase in CMP rate
can be understood that the reactivity of the nanoparticles increased as the Ti doping is increased. The
reactivity of Ce0.875Ti0.125 reached the maximum, and it is why the CMP rate decreased back to 295
nm/minute as the Ti content further increased to Ce0.75Ti0.25O2.
The defects generated during CMP process are detrimental to the chip production yield. Any
defect such as scratches can produce defective chips and only a minimal number of defects can be
tolerated in the production of advanced integrated circuits since the CMP process is utilized many
times during manufacturing each chip. Each next generation of advanced chips requires at least 50%
reduction in defect rate in order to make the chip mass-manufacturing viable in accommodating the
reduced feature sizes and increased number of inter-connects in each chip. Fig. 4B showed clearly that
the adders and especially the scratches (a more critical parameter than adders) are significantly
reduced as the nanoparticles become more spherical and contain higher Ti contents. This can be easily
understood in terms of the lack of sharp edges and corners as well as the ability to roll freely on the
wafer surface during polishing as the nanoparticles become more spherical.
6
In summary, we report for the first time the synthesis of spherical, single-crystal Ti-doped
CeO2 nanoparticles at large-scale (100-200 grams per hour). This was achieved by synthesizing Tidoped CeO2 nanoparticles in flame temperatures that facilitate crystallization of the CeO 2 yet retain
the TiO2 in a molten state. The (molten) titanium oxide shell encapsulates the inner ceria core and
accommodates a spherical morphology, which minimizes its surface energy. Moreover, the
nanoparticles become more perfect (spherical) with increasing TiO2 content. In particular, as the CeO2
core crystallizes, the relative areas of {111} and {001} are reduced at the expense of increased
formation of higher index planes, which are stabilized by the amorphous TiO2 shell. High index
planes are a pre-requisite to forming spherical shapes. For CMP, spherical, single-crystal Ti-doped
CeO2 nanparticles are shown to facilitate an 80% reduction in defects and a 50% increase in the silica
removal rate during the chemical mechanical planarization of advanced integrated circuits. This
technology will impact the development of high-quality and high precision microelectronics and
nanoelectronics over large-size wafers. The methodology demonstrated here for synthesizing spherical
single-crystal inorganic particles might offer unlimited possibilities in a broad range of fields such as
chemical mechanical planarization, photonics, magnetics, catalysis, and inorganic pigment where
spherical nanoscale inorganic particles are pivotal.
7
Figure captions
Figure 1. (A-D) Microstructure of CeO2 nanoparticles without Ti doping. (A) SEM image, (B) lowmagnification TEM image, and (C, D) high-resolution TEM images of the ceria nanocrystals with
faceted shapes. (E-I) Microstructure of CeO2 nanospheres doped with 6 at.% of Ti. (E) SEM image,
(F) low-magnification TEM image, and (G-I) high-resolution TEM images of the ceria nanocrystals,
showing their spherical shape and single-crystal structure. The inset in (F) is an electron diffraction
pattern recorded from the area, proving the cubic ceria structure of the sample. The inset in (I) is a
Fast Fourier Transform of the high-resolution TEM image.
Figure 2. Microstructure of CeO2 nanospheres doped with 12 at.% of Ti. (A) low-magnification TEM
image, and (B, C) high-resolution TEM images of the ceria nanocrystals, showing their spherical
shape and single-crystal structure as well as the amorphous layer on the surface. The inset in (A) is an
electron diffraction pattern recorded from the area, proving the cubic ceria structure of the sample. (D,
E, F) Scanning transmission electron microscopy images of the nanocrystals doped with 12 at.% of Ti,
showing the morphology, corresponding Ce distribution map and Ti distribution map. The
nanospheres are dominated by Ce, and the Ti is distributed at the surfaces. The bright clusters seen in
(F) are precipitated TiO2 particles.
Figure 3. (A): snapshots, taken during crystallization (Ti-doped CeO2), showing the initial amorphous
precursor (left), evolution and growth of the seed (circled), to the fully crystalline nanoparticle with
amorphous shell (right); (B,C,D): non-doped nanoparticle, (E,F,G) Ti-doped CeO2 nanosphere, (H)
enlarged segment of the Ti-Doped nanosphere. (B,E) Sphere model representation of the atom
positions, (C,F) side view with smaller spheres to view through the nanoparticle, (D,G) surface
rendered model. Cerium is colored white; Ti(IV) is blue and oxygen, red. The nanoparticles are about
7-8 nm in diameter. Note all images show actual atom positions and are not schematics.
Figure 4. Chemical-mechanical planarization performance of single-crystal nanospheres of
Ce1-xTixO2, (x = 0 – 0.25). (A) A comparison of the thermal oxide, SiO2, removal rates among the four
CMP slurries, and (B) A comparison of the defect rates in terms of the total number of the adders that
adhere to the pattern wafers after CMP testing, and the number of scratches observed.
8
1. Bera, P. & Hegde, S.M. Characterization and catalytic properties of combustion synthesized Au/CeO 2
catalyst. Catalysis Lett. 79, 75-81 (2002).
2. Valenzuela, R. X. et al. Nanostructured ceria-based catalysts for oxydehydrogenation of ethane with CO 2.
Topics In Catalysis 15, 181-188 (2001).
3. Kang, Z.C. & Wang, Z.L. Novel oxides for cycling hydrogen productioin from methane and water by
temperature swing. Adv. Mater. 15, 521-526 (2003).
4. Feng, X. et al. CeO2 particles for chemical mechanical planarization. MRS Symp. Proc. 767, 173 (2003).
5. Yu, X. J., Xie, P.B. & Su, Q.D. Size-dependent optical properties of nanocrystalline CeO 2:Er obtained by
combustion synthesis. Phys. Chem. Chem. Phys. 3, 5266-5269 (2001).
6 Singh, R. K. & Bajaj, R. Advances in Chemical--Mechanical Planarization. MRS Bull. 27 743-747 (2002);
CMP is a process that is used in the semiconductor industry to isolate and connect individual transistors on a
chip. The CMP process has been the fastest-growing semiconductor operation in the last decade, and its
future growth is expected to be equally explosive because of the introduction of copper-based interconnects
in advanced microprocessors and other novel applications of CMP for next-generation nanoscale devices.
The CMP slurries typically contain particle-based abrasives, which constitutes nearly 60% of the total $1
billion worldwide market for nanopowders in 2005.
7. Chen, H. & Chang, H. Synthesis of nanocrystalline cerium oxide particles by the precipitation method.
Ceramics International 31, 795-802 (2005).
8. Yuan, W. Bi, H. & Wei, C. Huanan Ligong Daxue Xuebao, Structure characterization of nanosized cerium
oxide particles prepared by homogeneous precipitation, Huanan Ligong Daxue Xuebao, Ziran Kexueban,
32(4), 52 (2004).
9. Bondioli, F. et al. Synthesis and characterization of praseodymium-doped ceria powders by a microwaveassisted hydrothermal (MH) route. J. Mater. Chem. 15 1061-1066 (2005).
10. Shuk, P. & Greenblatt, M. Hydrothermal synthesis and properties of mixed conductors based on
Ce1−xPrxO2−δ solid solutions. Solid Stae Ionics 116, 217-223 (1999).
11. He, Y. & Yang B. Synthesis of nanometer sized CeO2 particles by the coupling route of microemulsion with
homogeneous precipitation, Huaxue Tonbao, Xi’an Jiaotong University, 66(2), 120 (2003)
12. Tsuzuki, T., Robinson, J. S. & McCormick, P.G. J. UV-shielding nanoparticles synthesized by
mechanochemical processing, J. Australasian Ceramic Society, 38(1), 15 (2002)
13. Wang, Y. et al. Low-temperature synthesis of praseodymium-doped ceria nanopowders. J. Am. Ceram. Soc.
85 3105-3107 (2002).
14. Lopez-Navarrete, E. et al. Low-temperature preparation and structural characterization of Pr-doped ceria
solid solutions. J. Mater. Res. 17 797 (2002); Kodas, T. T. & Hampden-Smith, M. J. Aerosol Processing of
Materials (Wiley-VCH, New York, 1999).
9
15. Hartridge, A. & Bhattacharya, A. K. Preparation and analysis of zirconia doped ceria nanocrystal
dispersions. J. Phys. Chem. Solids 63 441-448 (2002).
16. Hirano, M. et al. Preparation and spherical agglomeration of crystalline cerium(IV) oxide nanoparticles by
thermal hydrolysis. J. Am. Ceram Soc. 83 1287-1289 (2000).
17. Yabe, S. Solvothermal synthesis of cerium dioxide, KOSE Corporation, Tokyo, Japan. Kidorui, 34 124
(1999)
18. Bickmore, C. et al. Ultrafine spinel powders by flame spray pyrolysis of a magnesium aluminum double
alkoxide. J. Am. Ceram. Soc. 79 1419-1423 (1996); Sutorik, A.C. & Baliat, M. S. Solid solution Behavior of
CexZr1-xO2 nanopowders prepared by flame spray pyrolysis of solvent-borne precursors. Mater. Sci. Forum
386-388 371-376 (2002).
19. See supplementary materials provided.
20. Preparation of a ceria stock solution: 3100 g of Propionic Acid (Food Grade, Univar) was heated in a 12-L
round bottom flask. Once refluxing, a total of 740 g Cerium Carbonate (99.9995%, PIDC) was added. The
solution was turbid due to insoluble cerium sources and impurities. Once the solution cooled to room
temperature, 620 g of deionized water (DI) was added, and the solution was vacuum filtered through a 0.2 μ
filter. This stock solution (9-10 wt% CeO2 as determined by mass loss on heating) was then diluted with a
mixture of methanol:water (5.2:1 weight ratio) to a final solution of about 5 wt% ceria before flame prolysis.
The L-FSP operating parameters are: pumping rate at 120g/minutes, the atomizer oxygen flow was at 2 cubic
foot per minutes, supplemental oxygen flow at 4.1 cubic foot per minutes.
21. Preparation of Titanium-doped Ceria, Ce1-xTixO2, with x = 0.0625, 0.125 and 0.25: Stoichiometric amounts
of Titanium (IV) (triethanolaminato) isopropoxide (TYZOR-TE) from Dupont were added to the ceria stock
solution prepared above (19) and the combined solution was diluted to 4-5 wt% oxide with methanol:water
(5.2:1 weight ratio). The L-FSP operatingparameters are the same as (19).
22. DL_POLY is a package of Molecular Simulation routines written by W. Smith and T.R. Forester Copyright
by the council for the Central Laboratory of the Research Councils, Daresbury Laboratory, Daresbury,
Warrington, UK, 1996, http://www.dl.ac.uk/TCSC/Software/DLPOLY
23. Sayle, T.X.T., Parker, S. C. and Catlow, C. R. A., The role of oxygen vacancies on ceria surfaces in the
oxidation of carbon monoxide. Surface Science 316 (3): 329-336 ( 1994).
24. Sayle, D. C., Catlow, C. R. A., Perrin, M. A. and Nortier, P., Computer simulation study of the defect
chemistry of rutile TiO2. Journal of Physics and Chemistry of Solids, 56 (6), 799-805 (1995).
25. Sayle, T. X. T., Parker, S. C. and Sayle, D. C., Shape of CeO2 nanoparticles using simulated amorphisation
and recrystallisation. Chemical Communications, 21, 2438-2439 (2004). For the Ti-doped nanoparticle,
constant volume MD simulation was performed for 7000ps at 3750K. The nanoparticle was then quenched:
MD simulation was performed for 400ps at 273K. Each simulation required about 100 hours using 96
processors of a SunFire Galaxy-class supercomputer.
10
26. The SiO2 film layer for CMP testing was a 1000 nm thermal oxide film on a silicon wafer. The wafers were
polished using a Strasbaugh 6EC polisher, a Rodel IC1000 pad with Suba IV backing at a down pressure of
4.5 psi, and a table and head rotation speed of 130 rpm, and slurry flow rate of 150 ml/min. The defect study
was performed on patterned MIT-mask silica wafers by polishing the wafer for 60 seconds using the above
mentioned polishing conditions. The defects were examined under A Applied Materials WF736 defect
inspection station. The polishing slurry mixture was adjusted to pH 4 using nitric acid and was then subjected
to high shear mixing for 30 minutes before polishing.
11