Hydrolysis of tetravalent cerium for a simple route to nanocrystalline cerium dioxide: An in situ spectroscopic study of nanocrystal evolution

Article abstract of DOI:10.1002/chem.201204101

Despite the rapid developments in recent nanocrystal research and their expanding applications, the evolution mechanism of nanocrystals remains veiled for the most part due to the lack of appropriate analytical techniques. Here we demonstrate one promising multi-spectroscopic approach for the in situ investigation of nanocrystal evolution. That is, the formation of nanocrystalline cerium dioxide (NC-CeO₂) has been probed by dynamic light scattering (DLS), X-ray absorption spectroscopy (XAS) and high-energy X-ray scattering (HEXS). The obtained results indicate that the fine colloidal particles of NC-CeO₂ are formed in an acidic aqueous solution simply through the hydrolysis of the initial precursor of small oligomer Ce ^IV species. This information on how NC-CeO₂ evolves is fundamental to simplifying and alleviating the synthetic strategy for NC-CeO₂ production. Nanocrystal evolution: A multi-spectroscopic approach has revealed that the hydrolysis of tetravalent cerium (Ce ^IV) produces a colloidal solution of fine cerium dioxide (CeO ₂) nanocrystals (see figure). A simple synthetic concept of nanocrystalline metal oxides is proposed based on the obtained results. Copyright

Full text of DOI:10.1002/chem.201204101

FULL PAPER
DOI: 10.1002/chem.201204101
Hydrolysis of Tetravalent Cerium for a Simple Route to Nanocrystalline
Cerium Dioxide: An In Situ Spectroscopic Study of Nanocrystal Evolution
Atsushi Ikeda-Ohno,*[a, b, c] Christoph Hennig,[a] Stephan Weiss,[a] Tsuyoshi Yaita,[d] and
Gert Bernhard[a]
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

Abstract: Despite the rapid develop-
ments in recent nanocrystal research
and their expanding applications, the
evolution mechanism of nanocrystals
remains veiled for the most part due to
the lack of appropriate analytical tech-
niques. Here we demonstrate one
mation of nanocrystalline cerium diox-
ide (NC-CeO2) has been probed by dy-
namic light scattering (DLS), X-ray ab-
sorption spectroscopy (XAS) and high-
energy X-ray scattering (HEXS). The
obtained results indicate that the fine
colloidal particles of NC-CeO2 are
formed in an acidic aqueous solution
simply through the hydrolysis of the in-
itial precursor of small oligomer CeIV
species. This information on how NC-
CeO2 evolves is fundamental to simpli-
fying and alleviating the synthetic strat-
egy for NC-CeO2 production.
Keywords: cerium
· hydrolysis ·
promising
multi-spectroscopic
ap-
lanthanides · nanomaterials · X-ray
spectroscopy
proach for the in situ investigation of
nanocrystal evolution. That is, the for-
Introduction
erable attention across a range of technological and medical
applications.[6,7]
The hydrolysis of metal cations is a fundamental reaction
not only for the aqueous solution chemistry of metal cations,
but also for wet chemical processes of metals. For instance,
the sol–gel process, which is often employed to produce
metal oxide compounds, is based on the polymerization and
aggregation initiated by hydrolysis.[1] More recently, sol–gel
processes are further being developed for the fabrication of
metal nanocrystals (NCs).[2] Metal NCs, whose size is gener-
ally less than 100 nm at least in one dimension,[3] exhibit a
unique “intermediate” character between bulk crystals and
atoms.[4] This often enhances their chemical activities, such
as catalytic or redox activity, as compared with those of the
chemically identical bulk materials.[5] Owing to these re-
markable properties, metal NCs are steadily drawing consid-
One significant on-going challenge in metal NCs research
is to control their shape and size, because the physical/chem-
ical properties of metal NCs largely depend on these fac-
tors.[7] The establishment of synthetic strategy to tailor metal
NCs with desirable shape and size requires the elucidation
of their fundamental formation mechanism, that is, the evo-
lution from initial metal precursors into NCs, as the mor-
phology and size distribution of the resultant NCs are deter-
mined predominantly by their evolution route.[7,8] The for-
mation mechanism of metal NCs, however, remains undis-
covered for the most part due to the lack of appropriate ex-
perimental techniques. That is, the most frequently used
tools to characterize NCs thus far are electron microscopy
(e.g., SEM and TEM) and X-ray diffraction. These methods
are powerful enough to visualize the shape, size, and lattice
images of NCs, as well as to acquire their lattice parameters.
The intrinsic limitation of these methods is, however, that
they are applicable exclusively to solid state samples, but
neither to solution samples nor to wet ones. Given the fact
that major syntheses to fabricate metal NCs, including the
sol–gel process, are performed in solution, a new analytical
approach to observe in situ the evolution of metal NCs in
solution is imperative for further expanding the research of
metal NCs as well as their applications.
Based on these backgrounds, the aim of this paper is to
demonstrate one promising analytical strategy for probing
the evolution of metal NCs in solution. That is, we herein
combine dynamic light scattering (DLS), X-ray absorption
spectroscopy (XAS), and high energy X-ray scattering
(HEXS) based on pair-distribution function (PDF) analy-
sis,[9] for in situ monitoring of the formation of nanocrystal-
line cerium dioxide (referred to as “NC-CeO2”, hereafter)
triggered by the hydrolysis of tetravalent cerium (CeIV). The
initial purpose of this study was to systematically investigate
the hydrolysis behavior of CeIV. The hydrolysis of CeIV is a
very primary step of many wet syntheses for fabricating NC-
CeO2,[10] although all the reported synthetic methods require
additional processes (e.g., heating, adding organic solvents,
or calcination) subsequent to the initial hydrolysis step to fi-
nally yield NC-CeO2. However, the present study reveals,
unexpectedly, that NC-CeO2 is already produced at the
[a] Dr. A. Ikeda-Ohno, Dr. C. Hennig, S. Weiss, Prof. Dr. G. Bernhard
Institute of Resource Ecology
Helmholtz-Zentrum Dresden-Rossendorf
P.O.Box 510119, 01314 Dresden (Germany)
Fax : (+49)351-260-3553
[b] Dr. A. Ikeda-Ohno
School of Civil and Environmental Engineering
The University of New South Wales
UNSW, Sydney, NSW 2052 (Australia)
Fax : (+61)2-9385-5033
E-mail: a.ikeda@unsw.edu.au
aikeda@lax.kuramae.ne.jp
[c] Dr. A. Ikeda-Ohno
Institute for Environmental Research
Australian Nuclear Science and
Technology Organisation
Locked Bag 2001, Kirrawee DC
NSW 2232 (Australia)
Fax : (+61)2-9717-3111
E-mail: atsushi@ansto.gov.au
[d] Dr. T. Yaita
Quantum Beam Science Directorate
Japan Atomic Energy Agency
Kouto 1-1-1, Sayo-cho, Sayo-gun
Hyogo-ken 679-5148 (Japan)
Fax : (+81)791-58-0311
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201204101.
&
2
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Hydrolysis of Tetravalent Cerium
FULL PAPER
early stage of the hydrolysis of CeIV. As a consequence, we
have succeeded in observing the whole evolution process
from the precursor species into NC-CeO2.
critical pH” hereafter. The same behavior was observed
when either the Ce concentration was increased up to 1.0m,
or an NH3 solution was used for titrant.
Because of the flexibility between the tri- and tetravalent
oxidation states, cerium is known to be the only rare earth
element (REE) forming a stable pure stoichiometric dioxide
compound (MIVO2) with a fluorite-type structure, whereas
most others (except Pr and Tb) form a sesquioxide com-
pound (MIII2O3).[11] Owing to this chemical specificity along
with the highest natural abundance of Ce among all
REEs,[12] the application of NC-CeO2 has spread out over a
variety of fields, including the catalysts for harmful gas
treatment[13] and the water–gas shift reaction,[14] electrodes
for solid oxide fuel cells (SOFC),[15] polishing materials for
advanced integrated circuits,[16] sunscreen cosmetics,[17] and a
medical use as an artificial superoxide dismutase.[6] These
versatile applications of NC-CeO2 still require a simpler and
more efficient synthetic strategy. This study provides a pri-
mary concept to potentially simplify the NC-CeO2 synthesis
merely based on hydrolysis. This is also a representative ex-
ample which demonstrates the importance of fundamental
understanding of the formation mechanism of metal NCs
for further development of NCs research.
Dynamic light scattering: DLS is a powerful technique to
determine the size distribution of particles dispersed in solu-
tion.[18] The CeIV solutions shown in Figure 1 have been
studied first by DLS to investigate the presence of colloidal
species in the solutions. No DLS signals were detected for
the solutions below the critical pH (i.e., pH<0.4), meaning
that the scattered light intensities were equal to those of a
pure solvent (i.e., aqueous HNO3). On the other hand, the
solutions at the critical pH or higher (i.e., pHꢀ0.5) showed
detectable DLS signals, suggesting a significant increase of
scattered light intensities as compared with that of the pure
solvent and, hence, the formation of colloidal species in the
solutions exceeding the critical pH. The tetravalent cations
of ThIV[19] and UIV,[20] ionic radii of which are similar to that
of CeIV,[21] are known to produce colloidal solution being
composed of their oxyhydroxide compounds. Considering
the chemical similarity among these tetravalent cations, the
CeIV particles formed above the critical pH are likely to be
the hydrolytic colloid of CeIV or possibly the mixture of CeIV
and reduced CeIII. Figure 2(a) shows the particle size distri-
butions of the colloidal species formed at different pH
values. There is a broadening trend in the particle size distri-
bution with increasing pH. The local pH excursion caused
by the base titration process possibly produces aggregated
species with larger particle sizes. This is possibly the cause
of the observed broadening trend in particle size distribution
as the pH increases. The morphology of colloidal particles is
also influenced by metal concentration and the type of basic
reagents employed for pH adjustment. That is, the lower
concentration of Ce in the initial solution results in the nar-
rower particle size distribution of the resultant colloids, and
the pH adjustment with NH3 provides finer and more homo-
geneous colloid than that with NaOH (Sections 1 and 2 in
the Supporting Information). The particle size and homoge-
neity of the Ce colloid also vary with aging. As shown in
Figure 2(b), the freshly formed colloid exhibits a wide parti-
cle size distribution and, thus, is very heterogeneous (the
top-most histogram in Figure 2(b): “1 hour”). With aging,
the particle size distribution gradually narrows and the
mean hydrodynamic diameter of the particles becomes
smaller. After approximately one month of aging, the parti-
cle size of the Ce colloid becomes significantly homogene-
ous (the bottom-most histogram in Figure 2(b): “672 h”)
with a finer particle size (mean values of the dominant hy-
drodynamic diameter=(3.0ꢁ0.7) nm when the pH is adjust-
ed with NH3, and (4.2ꢁ0.8) nm with NaOH (Sections 1 and
2 and Figure S5 in the Supporting Information). The DLS
instrument employed in this study has a minimum detection
limit of <1 nm for particle size estimation, according to the
specification document. Therefore, the particle size of 3–
4 nm diameter is closed to the minimum detection limit, but
it is still detectable. In general, DLS intensity is proportional
to the sixth power of a particle diameter.[18] This means that
Results
Visual observation of hydrolysis behavior of CeIV: Figure 1
shows the color change of a 0.1m CeIV solution in 1.0m
HNO3 titrated with NaOH. The initial color of the solution
(“pH 0.0” in Figure 1) had a strong and clear yellow color.
Figure 1. Color change of 0.1m CeIV in 1.0m HNO3 as it is being titrated
with NaOH. The photo was taken immediately after sample preparation.
With increasing pH, the yellow color began to fade at
around pH 0.5 and became almost colorless above pH 0.7.
The solution remained transparent until the production of
hydrolytic precipitate above pH 3.0. The precipitate initially
had a pale-yellow color (e.g., “pH 3.0” in Figure 1). It then
turned into purple (e.g., “pH 8.0” in Figure 1) when the sol-
ution became basic (pH>7.0). It should be noted that, as
far as pH 0.5 (i.e., from pH 0.0 (initial) up to 0.5), the pH
value of the sample solution increased in proportion to the
amount of NaOH added. However, at around pH 0.5, at
which the solution color began to fade, the pH value
became transiently unchanged regardless of the addition of
NaOH. Continuous addition of NaOH resulted in the re-in-
crease of pH in due course. After that, the pH value rose
proportionally again. We refer this pH value of 0.5 as “the
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
3
&
ÞÞ
These are not the final page numbers!

Dr. A. Ikeda-Ohno et al.
To summarize the DLS results, the apparently transparent
CeIV solutions with pH 0.5–2.0 in Figure 1 are, in fact, a col-
loidal solution of fine particles. The hydrodynamic diameter
and size distribution of these fine particles are influenced by
the Ce concentration, pH, the type of basic reagent, and
aging.
X-ray absorption spectroscopy: To acquire the information
on Ce oxidation states and local structural arrangement of
Ce species, the CeIV samples in Figure 1 were further investi-
gated by XAS including X-ray absorption near-edge struc-
ture (XANES) and extended X-ray absorption fine structure
(EXAFS) analyses. The analysis of Ce K-edge XANES
spectra in conjunction with spectrofluorometry suggests that
the oxidation state of Ce remains mostly tetravalent at any
pH value (Section 1–1 in the Supporting Information). This
means that the change of solution color observed in Figure 1
is likely to be attributed not to the reduction from CeIV to
colorless CeIII, but to the change in CeIV speciation.
Figure 3 shows the k3-weighted Ce K-edge EXAFS spec-
tra (left, labeled (a)) for 0.1m CeIV solutions at different pH
values along with reference data of a 0.05m CeIV solution in
2.0m HClO4 and CeO2 powder, and their corresponding
Fourier transforms (FTs) (right, labeled (b)). Our recent
study combining EXAFS with DFT calculations has re-
vealed that the primary aqua species of CeIV in an aqueous
perchloric acid (grey colored spectra at the top most in
Figure 3) is not a simple mononuclear hydrate complex, but
an oxo- or/and hydroxo-bridging dinuclear complex(es).[22]
Such smaller oligomer complexes can be readily identified
on EXAFS-FT data by the presence of a distinguishable
peak at around R+D=3.5–3.7 ꢁ, which is derived from the
[22]
ꢂ
ꢂ
Ce Obridging Ce arrangement.
This characteristic peak at
around R+D=3.5–3.7 ꢁ appeared also in the spectrum for
1.0m HNO3 without an addition of basic reagent (“0.0” with
a deep blue color in Figure 3). Coordination of nitrate ions
ꢂ
(NO3 ) may also give an EXAFS-FT peak at around R+
D=3.6–3.7 ꢁ,[23]
potentially
overlapping
with
the
IV
ꢂ
ꢂ
Ce Obridging Ce peak. However, tetravalent actinides (An ),
which are often referred to as analogues of CeIV, are coordi-
ꢂ
nated by two NO3 at best in lower nitrate concentrations
like 1.0m,[23c,24] exhibiting small and indistinctive EXAFS-FT
ꢂ
peaks related with NO3 coordination.[23c] Taking into ac-
count this fact along with the observed spectral identity be-
tween the grey and deep-blue colored spectra in Figure 3
(see Section 1–3 in the Supporting Information), we con-
clude that the oxo/hydroxo-bridging dinuclear complexes
are the dominant form of CeIV in 1.0m HNO3 at pH 0.0,
albeit the present EXAFS results do not completely exclude
the co-existence of higher oligomer complexes at this stage.
These results indicate that these small oligomer complexes,
not a mononuclear hydrate complex, are the precursors that
finally evolve into NC-CeO2 through hydrolysis, as will be
discussed in the following sections. When the pH is in-
creased from “0.0” to “0.3” (deep- and normal-blue-colored
data in Figure 3), Peak 3 in Figure 3(b) is split into two with
an additional peak at a shorter R range of R+D=3.0 ꢁ.
Figure 2. (a) Particle size distribution of Ce colloid formed in 1.0m HNO3
at different pH values adjusted with NaOH; [Ce]=0.1m. DLS data were
collected 48 h after sample preparation. (b) Time variation of particle
size distribution of Ce colloid formed in 1.0m HNO3 at pH 0.5; [Ce]=
0.1m. NH3 was employed for pH adjustment.
the presence of a small amount of larger particles may
lower the sensitivity for smaller particles. However, the
error of DLS signals becomes significantly smaller with time
(Figure S5 in the Supporting Information), suggesting that
there is little contribution of larger particles to the detection
of 3–4 nm particles. Hence, we conclude that the obtained
hydrodynamic diameter of 3–4 nm is above the limit of min-
imum detection.
&
4
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Hydrolysis of Tetravalent Cerium
FULL PAPER
Figure 3. (a) k3-weighted Ce K-edge EXAFS spectra for 0.1m CeIV in 1.0m HNO3 with different pH values and reference spectra for 0.05m CeIV in 2.0m
HClO4 (grey) and CeO2 (red): solid lines, experimental data; dotted lines, theoretical curve fits. NaOH was used for pH adjustment. (b) Fourier trans-
forms of EXAFS spectra in (a): the experimental data is given as solid lines; theoretical curve fits are given as dotted lines; k range for Fourier transfor-
mation, k=2.0–15.0 ꢁꢂ1; phase shifts (D) are not corrected. Peak 1=Ce Ooxo/Ohyd, Peak 2=Ce O
ACHTNUGTRENNU(G H2O), Peak 3=Ce Ce, Peak 4=Ce OoxoACHTUNGTREN(NUGN CeO2) (see
the main text for more details). The magnitude of the data for CeO2 powder (red) is reduced by one third relative to the y axes for facilitating compari-
ꢂ
ꢂ
ꢂ
ꢂ
sons of the spectra.
The Ce Ce distance in CeIV oligomers is expected to shorten
as they grow from dimers into larger oligomers.[22,25] The
thermodynamic data suggests the presence of only di-, tri-,
and hexanuclear complexes for soluble CeIV oligomer spe-
cies in an aqueous solution.[26] This has been further support-
ed by DFT calculations.[25] Given these facts, the EXAFS-
FT peak at R+D=3.0 ꢁ for the blue-colored spectra “0.3”
in Figure 3(b) possibly implies the partial evolution of CeIV
species from smaller to larger oligomer complexes. As a
matter of fact, the EXAFS curve fitting for the data “0.3”
further increase in pH leads to a drastic change in EXAFS
spectra. That is, at pH 0.4 and over, EXAFS oscillation in a
higher k range (k>8 ꢁꢂ1 in Figure 3(a)) becomes more em-
phatic, producing an intense peak at R+D=3.5 ꢁ (Peak 3
in Figure 3(b)) on EXAFS-FT spectra. The position of this
ꢂ
ꢂ
intense peak is consistent with that for the nearest Ce Ce
arrangement in CeO2 (red-colored data in Figure 3(b)). Be-
sides, the EXAFS oscillation cycle for the samples exceeding
the critical pH (pH 0.5–12.0) coincides with that of CeO2 es-
pecially in the higher k range (Figure 3(a)). Therefore, the
EXAFS-FT peak at around R+D=3.5 ꢁ for the samples
ꢂ
calculates a shorter Ce Ce distance of 3.28 ꢁ (Table 1),
ꢂ
which may be comparable with 3.35 ꢁ found in the trinu-
“pH 0.4–12.0” can be reasonably assigned to the Ce Ce ar-
clear CeIV complex optimized by a DFT calculation.[22]
A
rangement of CeO2 or similar structure. The calculated Ce
ꢂ
Table 1. Summary of EXAFS structural parameters obtained from theoretical curve fitting. A complete list of the obtained parameters is provided in
the Supporting Information (Table S1).
Peak 1 or 4 in Figure 3 (b)
Peak 2 in Figure 3 (b)
Peak 3 in Figure 3 (b)
Peak 3 in Figure 3 (b)
ꢂ
ꢂ
Ce O
ꢂ
ꢂ
Sample
Form
Ce O
(H2O)
Ce Ce1
Ce Ce2
CN[a]
R [ꢁ][b]
CN[a]
R [ꢁ][b]
CN[a]
R [ꢁ][b]
CN[a]
0.5
R [ꢁ][b]
[c]
2.0 m HClO4
liquid
1.0
0.7
1.8
4.0
5.5
8.2
8.3
8.2
8.0
7.7
8[d]
2.03
2.02
2.08
2.35
2.29
2.28
2.28
2.28
2.30
2.31
2.35
7.1
7.4
6.9
3.5
3.4
2.46
2.43
2.44
2.47
2.46
0.8
1.1
1.1
5.3
5.8
7.9
8.8
11.6
11.7
11.7
12[d]
4.05
4.05
4.05
3.83
3.81
3.82
3.81
3.81
3.82
3.82
3.83
1.0 m HNO3 (pH 0.0)
1.0 m HNO3 (pH 0.3)
1.0 m HNO3 (pH 0.4)
1.0 m HNO3 (pH 0.5)
1.0 m HNO3 (pH 0.8)
1.0 m HNO3 (pH 1.0)
1.0 m HNO3 (pH 2.0)
1.0 m HNO3 (pH 5.0)
1.0 m HNO3 (pH 12.0)
CeO2
liquid
liquid
liquid
liquid
liquid
liquid
liquid
precipitate
precipitate
powder
3.28
[a] CN: coordination number; error ꢃ ꢁ15%. [b] R: interatomic distance; error ꢃ ꢁ0.01 ꢁ. [c] Values taken from Ref. [22]. [d] Fixed values.
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
5
&
ÞÞ
These are not the final page numbers!

Dr. A. Ikeda-Ohno et al.
Ce distances for the spectra “pH 0.4–12.0” range from 3.81
to 3.83 ꢁ (Table 1), being in agreement with that found in
the crystal structure of CeO2 (3.83 ꢁ).[27] The EXAFS-FT
peaks observed at R+D=1.5–2.0 ꢁ (Peaks 1 and 2 in Fig-
ure 3(b)) reflect the primary coordination sphere of CeIV.
The present CeIV samples are composed of H2O, HNO3, and
samples. For making the discussion in the following sections
clearer, the term “colloid/colloidal” hereafter refers to any
particulate forming a colloidal suspension, regardless of the
chemical structure of the particulate. The term “crystalline”
refers to a compound having a long-range structural order
(<1 nm), whereas the term “amorphous” is used for a com-
pound without the long-range structural order.
Figure 4 shows the two-dimensional images of X-ray scat-
tering patterns from the CeIV solution without pH adjust-
ment (pH 0.0, (a) on the left) and the solution at pH 1.0 ((b)
on the right). The colloid-free solution at pH 0.0 exhibits a
featureless scattering image, whose scattering intensity is de-
ꢂ
NaOH. Taking into account the fact that NO3 are unlikely
to interact directly with CeIV, the coordinative ligands poten-
tially found in the primary coordination sphere are water
ꢂ
ꢂ ꢂ
molecules, hydroxide ions (OH ) and oxo-oxygens ( O ),
all of which are O-coordinating ligands. The peaks 1 and 2
for “HClO4” in Figure 3(b), grey-colored data) correspond
to the bridging oxo/hydroxo oxygen (Ooxo/Ohyd) and water
molecules, respectively.[22] Peak 2 is observable from pH 0.0
to the critical pH of 0.5, indicating the presence of hydrate
species. Peak 4 in Figure 3(b) corresponds to the nearest
ꢂ
Ce Ooxo arrangement in CeO2 (red-colored spectra). Peak 4
is dominant also for the precipitate samples (pH “5.0” and
“12.0” in Figure 3(b)). This suggests that the local coordina-
tion arrangement of CeIV in the precipitate formed above
pH 2.0 is similar to that in CeO2. The structural parameters
obtained for these precipitate samples are also well consis-
tent with those for the reference CeO2 sample (Table 1 and
Table S1 in the Supporting Information). For the solution
samples (pH “0.0”–“2.0”), Peak 1 becomes distinguishable
with a slight fluctuation of peak position as an increase in
pH, being dominant in the primary coordination sphere
above the critical pH. Since Peak 1 appears in a shorter dis-
tance than Peak 2 relating to water molecules, Peak 1 can be
attributed to stronger O ligands than water molecules, such
as Ooxo/Ohyd or OHꢂ. This suggests that, with increasing pH,
hydrated water molecules are transformed into oxo/hydroxo
ligands, which potentially bridge Ce4+ cations through oxo-
lation and olation reactions[28] finally to form colloidal spe-
cies. The solution samples above the critical pH display
Peaks 1, 3, and 4 (pH “0.8”–“2.0” in Figure 3(b)), indicating
that the colloidal species in these samples are a mixture of
CeO2-like species and other CeIV oligomers.
Figure 4. Two-dimensional X-ray scattering images (colored circular
image), integrated scattering curves (inset graph with a white line) and
sample appearance (inset photo) of 0.5m CeIV in 1.0m HNO3 at pH 0.0
(a) and at pH 1.0 (b). NaOH was employed for pH adjustment. Data
were collected 14 days after the sample preparation.
creased monotonously from the scattering center outward
(i.e., toward higher scattering angles, inset graph of Fig-
ure 4 (a)). This is a typical scattering pattern often called
“diffuse scattering”,[9] representing the absence of long-
range order in the sample (e.g., liquid or glass). On the
other hand, the solution at pH 1.0, which contains the CeIV
colloid, shows clear ring patterns producing Bragg peaks on
its integrated scattering curve (inset graph of Figure 4 (b)).
This is a clear indication that the particles of the CeIV col-
loid have a crystalline structure, not an amorphous one.
Since (oxy)hydroxide compounds are generally amor-
phous[19] and do not exhibit Bragg peaks, the CeIV colloid
formed is most likely a crystalline oxide compound of CeIV.
The collected HEXS data were further treated to extract
the total scattering structure function (Stotal(Q))[9] derived
only from the CeIV colloid, and finally to produce its pair-
distribution functions (PDF; g(R)),[9] which represent the
radial distribution of the interatomic distances of atomic
pairs present in the colloid. Figure 5(a) shows the Stotal(Q)
of the CeIV colloid formed at different pH values, along with
a reference data of CeO2 powder. Although the colloid
formed at a lower pH of 0.5 displays weaker peak intensity,
the position of each peak is almost identical for all the col-
loid samples, regardless of pH. Besides, the peak positions
observed for the CeIV colloid are in good agreement with
those for CeO2 (red-colored data in Figure 5(a)). The PDFs
shown in Figure 5(b) further indicate the structural similari-
The present EXAFS results clearly demonstrate the struc-
tural transition of CeIV species from the initial oligomer
complexes to CeO2-like precipitate through the potential
formation of small oligomers. However, it is difficult to fully
identify the CeIV colloid forming above the critical pH only
by EXAFS, as presumable species (oligomers like hexam-
er,[29] hydroxide species,[30] and CeO2[27]) are expected to
ꢂ
have similar Ce Ce distances and, hence, they all would ex-
hibit Peak 3 at similar position in Figure 3(b). Therefore, ad-
ditional information is required to further characterize the
CeIV colloid.
High energy X-ray scattering: HEXS combined with PDF
analysis has been recently demonstrated to be a powerful
tool for the structural analysis of chemical species in solu-
tion and colloidal systems,[31] being complementary to
EXAFS.[31c,32] To acquire additional structural information
particularly on the CeIV colloid, HEXS data were collected
for selected transparent solutions, colloidal and precipitate
&
6
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Hydrolysis of Tetravalent Cerium
FULL PAPER
and Figure S12 in the Supporting Information). This also
supports the identification of CeO2 in the colloid particles,
as CeO2 is known to be hardly soluble even in a strong acid,
whereas other hydroxide or hydrous oxide compounds are
readily soluble in an acidic solution.
ꢂ
The Ce Ce distances in the colloid structure can be esti-
mated directly from the obtained PDFs. Table 2 summarizes
ꢂ
the Ce Ce distances in the colloid samples prepared in dif-
ferent conditions, along with those for reference CeO2
ꢂ
powder. The Ce Ce distances for CeO2 powder listed in
Table 2 agree well with those of the reported crystal struc-
ture.[27] The CeIV colloid samples exhibit similar Ce Ce dis-
ꢂ
tances as those found for the CeO2 powder within the aver-
age distance error of ꢄ0.02 ꢁ, regardless of the type of
basic reagent used for pH adjustment, the pH, or the con-
centration of Ce. This indicates that, unless precipitated, the
hydrolysis of CeIV results in the formation of NC-CeO2 at
any condition above the critical pH (Section 1–5 in the Sup-
porting Information). The PDF of the CeIV colloid exhibits
an obvious cut-off in correlation peaks at around 2.5–2.6 nm
(Figure S15 in the Supporting Information), suggesting that
the colloid particles keep their structural order up to 2.5–
2.6 nm.
The crystallite (domain) size of the CeIV colloid can be es-
timated from the peak width of their extracted scattering
curves (in 2q) based on the Scherrer equation.[33] As shown
in Figure 6, the intensity of each scattering peak grows
larger as an increase in Ce concentration, although their
scattering patterns are identical. The domain size was esti-
Figure 5. (a) Total scattering structure functions (Stotal(Q)) of CeIV colloid
([Ce]=0.1m) formed at different pH values and a reference compound
of CeO2 powder. (b) Pair-distribution functions (PDFs, g(R)) derived
from Stotal(Q) in (a): dotted lines; the positions corresponding to the Ce
ꢂ
Ce pairs found in the CeO2 structure.[27] NaOH was employed for pH ad-
justment. Data were collected 14 days after the sample preparation.
ty between the CeIV colloid particles and crystalline CeO2.
That is, all the colloid samples display clear peaks at the po-
ꢂ
sitions corresponding to the Ce Ce distances of the CeO2
Figure 6. X-ray scattering curves (in 2q) of CeIV colloid formed at pH 2.0
with different Ce concentrations: Energy of incident X-ray; 87.1 keV
(l=0.142 ꢁ), pink colored data; NH3 was employed for pH adjustment,
orange colored data; NaOH was employed for pH adjustment. Black
lines with numbers at the bottom denote the hkl indices of CeO2.[27a]
Data were collected 14 days after the sample preparation.
structure (dotted lines in Figure 5(b)). These results reveal
that the colloidal CeIV species formed above the critical pH
is actually the nano-sized particles of crystalline CeO2 (i.e.,
NC-CeO2). The CeIV colloid once formed is no longer solu-
ble even if the solution is re-acidified to pH 0.0 (Section 1–5
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
7
&
ÞÞ
These are not the final page numbers!

Dr. A. Ikeda-Ohno et al.
IV
ꢂ
Table 2. Summary of Ce Ce distances for Ce colloid and CeO2 powder. The values are obtained by direct reading of x coordinates of the correspond-
ꢂ
ing peak tops on PDFs. The error in Ce Ce distances is estimated to be ꢁ0.001 nm based on 95% height of peak maximum.
[b]
[Ce] pH
[m]
Basic
Form
RCe [nm]
5
jR N(CeO2)ꢂRi javerage
ꢂ
Ce
reagent[a]
1
2
3
4
6
7
8
9
0.5
1.0
2.0
0.383
0.383
0.383
0.537
0.537
0.539
0.664
0.662
0.664
0.767
0.767
0.763
0.854
0.855
0.857
1.016
1.014
1.014
1.153
1.151
1.154
1.211
1.212
1.212
1.386
1.384
1.384
0.0024
0.0017
0.0023
0.1
0.1
NH3
colloid
0.5
1.0
2.0
colloid
colloid
colloid
precipitate
colloid
0.384
0.384
0.383
0.381
0.383
0.557
0.544
0.549
0.540
0.546
0.656
0.661
0.662
0.663
0.661
0.770
0.769
0.767
0.767
0.767
0.857
0.856
0.854
0.856
0.854
1.012
1.016
1.015
1.015
1.015
1.162
1.155
1.149
1.151
1.149
1.246
1.214
1.213
1.209
1.212
1.376
1.382
1.381
1.383
1.381
0.0078
0.0014
0.0019
0.0021
0.0017
NaOH
5.0
0.0[c]
0.5
1.0
2.0
2.0
NaOH
NaOH
colloid
colloid
powder
0.383
0.383
0.384
0.545
0.549
0.543
0.660
0.662
0.663
0.770
0.767
0.768
0.856
0.854
0.855
1.017
1.015
1.013
1.154
1.149
1.152
1.215
1.212
1.213
1.382
1.381
1.380
0.0022
0.0019
CeO2
CeO2
crystal structure[d] 0.3826 0.5411 0.6627 0.7652 0.8556 1.0123 1.1478 1.2099 1.3795
for each peak (i=1–9), RiA(CeO2)=RCe for reference CeO2 powder (listed on the
(CeO2)ꢂRi javerage” is a measure indicating the conformity of each sample with CeO2. Sample
iA
[a] Type of basic reagent used for pH adjustment. [b] Ri =RCe
CHTUNGTRENNUNG
ꢂ
Ce
ꢂ
Ce
second bottom line on the table). Therefore, the value “jR
iACHTUNGTRENNUNG
“[Ce]=0.1m, pH 0.5 with NaOH” (the 4th line from the top) indicates a large “jR (CeO2)ꢂRi javerage” value of 0.078 ꢁ, mainly due to a large noise level
N
on its spectrum (light-green-colored data in Figure 5 (b)). [c] The sample was first adjusted at pH 2.0 to produce CeIV colloid, and re-acidified to pH 0.0
by using HNO3. [d] Values taken from Ref. [27b].
mated from the full widths of half maximum for three distin-
guishable scattering peaks attributed to (220), (311) and
(331) reflections. Table 3 summarizes the domain sizes esti-
mated from their scattering peaks. Interestingly, the domain
size in the CeIV colloid particles is little influenced by the
formation conditions (i.e., type of basic reagents for pH ad-
justment, pH value, or Ce concentration), and is calculated
to be almost constant (2.6 nm assuming a cubic shape, or
2.9 nm assuming a spherical shape). This domain size is well
consistent with the cut-off distance of correlation peaks in
PDF (2.5–2.6 nm) and also in line with those of the reported
NC-CeO2 synthesized by hydrolysis-related methods (2–
3 nm).[34] The derived domain size of 2–3 nm is smaller than
those estimated by DLS (approximately 4–8 nm, Figure S5
in the Supporting Information). The particle size derived
from DLS corresponds to the hydrodynamic diameter,
which includes the interfacial double layer surrounding the
particle, and also reflects agglomerated particles. Hence, it is
theoretically larger than the physical diameter of an individ-
ual particle (i.e., domain size). Besides, the domain particles
would agglomerate to each other to form larger particles.
The particle size estimation by DLS involves these factors,
calculating a larger particle size than a single domain size.
Another possible scenario could be that DLS is sensitive to
a different part of the size distribution, whereas the domain
size determination by HEXS is sensitive to only small parti-
cle sizes. This is also a possible reason why the domain size
derived from HEXS is different from the particle size esti-
mated by DLS. The type of countercations (i.e., Na+ or
Table 3. Domain size of CeIV colloid formed in different conditions. The
listed values are the average of those derived from the full widths of half
maximum (FWHM) or integral widths of (220), (311), and (331) reflec-
tion peaks in Figure 6. A detailed description of the domain size estima-
tion based on the Scherrer equation[33] is given in Section 1–7 of the Sup-
porting Information.
+
Cubic shape[a]
Spherical shape[b]
NH4 in the present case) also affects the thickness of the
interfacial double layer and agglomeration properties. In
fact, the NC-CeO2 prepared by adding NaOH exhibits a
wider particle size distribution than that prepared by adding
NH3 (Figure S4 in the Supporting Information). Given the
fact that the domain size of the resultant NC-CeO2 is un-
changed regardless of their formation conditions (Table 3),
the type of basic reagent for pH adjustment (i.e., counterca-
tions) has an influence mainly on the thickness of the inter-
facial double layer and/or the agglomeration properties of
the domain particles, but not on the domain size itself. The
NC-CeO2 with several-nanometer size is composed of just
hundreds of atoms, bearing internal structural strain more
easily than the rigid and orderly crystal structure in CeO2
powder. This is probably the cause of the 0.02 ꢁ difference
[Ce]
[m]
pH
Basic
Domain
size [nm]
St. dev.
(ꢁ) [nm]
0.21
0.10
0.08
Domain
size [nm]
St. dev.
(ꢁ) [nm]
0.08
0.19
0.17
reagent[c]
0.5
1.0
2.0
2.63
2.70
2.43
2.91
3.01
2.72
0.1
0.1
NH3
1.0
2.0
2.59
2.61
0.21
0.24
2.87
2.84
0.13
0.14
NaOH
0.5
1.0
2.0
2.0
NaOH
NaOH
2.66
2.61
0.26
0.15
2.92
2.82
0.21
0.04
[a] Domain size (DhklACHTUNGTRENNUNG
(cubic) ꢁꢂ1)=0.94l/b1/2cosq, in which l=0.142 ꢁ
and b1/2 =FWHM of (hkl) peak. [b] Domain size (Dhkl(spherical) ꢁꢂ1)=
1.07l/bicosq, in which l=0.142 ꢁ and bi =integral width of (hkl) peak.
[c] Basic reagent employed for pH adjustment.
&
8
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Hydrolysis of Tetravalent Cerium
FULL PAPER
CeIV suggests the formation of hydrolytic CeIV species of up
to hexamer.[26] A recent DFT study by Aubriet et al.[25] has
further suggested that, among the possible hydrolytic CeIV
oligomers, trimer and hexamer tend to be stabilized during
the hydrolysis in particular. It should be noted that tetrava-
lent actinides (AnIV), the chemical analogues of CeIV, can
form stable hexanuclear complexes in aqueous solution only
in the presence of strongly coordinating functionalized li-
gands, such as carboxylic ligands, which can bridge and link
adjacent MIV centers to reinforce and stabilize the hexanu-
ꢂ
in Ce Ce distances between NC-CeO2 and a reference
CeO2 powder (jR
iA
ing curve of NC-CeO2 precisely reproduces the hkl families
of CeO2 without anomalous reflections (Figure S13 in the
Supporting Information), suggesting that the NC-CeO2 has a
high crystallinity and is characterized as a fluorite-type
CeO2 structure.
Transmission electron microscopy: The CeIV colloid (i.e.,
NC-CeO2) formed above the critical pH was further investi-
gated by TEM to characterize its dried state. Figure 7 shows
ꢂ
clear core framework.[35] The present system contains NO3
as a potentially coordinating ligand. However, the coordina-
ꢂ
tion of NO3 towards CeIV is expected to be much weaker
ꢂ
than that of carboxylic ligands and, hence, NO3 is probably
unable to serve as a bridging ligand to stabilize the hexanu-
clear core structure. Taking into account these facts, the
dominant precursor species to finally evolve into NC-CeO2
are considered to be dimer and trimer species. Of course,
the potential presence of other oligomer species cannot be
completely excluded, particularly as transition species from
dimer/trimer to NC-CeO2. However, their lifetime must be
very short and, therefore, it is difficult to prove their exis-
tence by the present experimental capability. Recently de-
veloped time-resolved spectroscopic techniques[36] may be
able to gain more insight into the dynamics of the formation
of these minor oligomer species.
Figure 7. A high-resolution transmission electron micrograph of the dried
CeIV colloid (left) and fast Fourier transformations of Areas (i) and (ii)
(right). The sample was prepared by drying a CeIV colloid solution (1.0m
CeIV in 1.0m HNO3, pH 2.0 adjusted with NH3) in the air.
When a basic reagent is added into the system, the follow-
ing three fundamental reactions will occur with respect to
the hydrolysis of CeIV:
a high-resolution TEM image of the dried CeIV colloid parti-
cles formed at pH 2.0 (left) and its Fourier transformation
(FT) images (right). The FT image of Area (i) is well consis-
Ceðaq:Þ þ OHꢂ ! CeꢂOH
ð1Þ
ð2Þ
ð3Þ
tent with two independent CeO
3.12 ꢁ, dihedral angle=70.58), whereas the FT image of
Area (ii) is in good agreement with CeO2A(200) and CeO2-
(022) reflections (d200 =2.71 ꢁ, 022 =1.91 ꢁ, dihedral
2ACHTUNGTERNN(UGN 111) reflections (d111 =
Ceðaq:ÞꢂOH2 þ CeꢂOH ! CeꢂOHꢂCe þ H2O
CeꢂOH þ CeꢂOH ! CeꢂOꢂCe þ H2O
CTHUNGTRENNUNG
A
d
angle=908). These results demonstrate that the NC-CeO2
formed in the solution phase retains its crystalline structure
even when it is dried up. The individual domains seem to be
agglomerated during the dying process. The size of individu-
al particles of the dried NC-CeO2 is estimated to be 2–3 nm,
being in line with the HEXS and DLS results, as well as the
size of the reported NC-CeO2 particles.[34] To generate a lat-
tice image on TEM, several atoms are needed to be lying in
the direction of electron beam. Therefore, the clear lattice
image shown in Figure 7 suggests that the NC-CeO2 pre-
pared in this study is highly crystalline.
Both olation and oxolation reactions (Equations 2 and 3,
respectively) can promote the transformation of the initial
dimer species into trimer and other oligomer species, as
these polynuclear complexes are bridged by oxo- and hy-
droxo ligands.[22,25,29] Here we note that the oxolation reac-
tion (Equation 3) consumes OHꢂ to form an oxo-bridge,
whereas the stoichiometric amount of OHꢂ is unchanged in
the olation reaction (Equation 2). It has been mentioned in
the section “Results: Visual observation of hydrolysis be-
havior of CeIV” that the pH value of the CeIV solution in-
creases proportionally to the volume of the basic reagent
added from the initial pH (pH 0.0) until the critical pH of
0.5. This increase in pH is stopped at the critical pH, at
which the formation of NC-CeO2 begins, and the pH value
becomes constant transiently regardless of the addition of
basic reagent. These facts clearly indicate that the olation
reaction is dominant in the region of pH 0.0–0.5 to increase
hydroxo-bridging species, whereas the oxolation reaction is
accelerated suddenly at the critical pH of 0.5 to convert the
hydroxo-bridging species to the oxo-bridging ones, which
further evolve into CeO2.
Discussion
Evolution mechanism of nanocrystalline cerium dioxide:
The EXAFS results have demonstrated that the initial pre-
cursor species, which finally evolve into NC-CeO2, is a
smaller oligomer species of CeIV, such as a dinuclear com-
plex. An increase in pH will transform such a smaller
oligomer species into larger ones before the formation of
NC-CeO2. The thermodynamical data on the hydrolysis of
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
9
&
ÞÞ
These are not the final page numbers!

Dr. A. Ikeda-Ohno et al.
The HEXS results indicate that the NC-CeO2 particles
formed above the critical pH have a constant domain size of
2–3 nm. This size is consistent with the smallest size of 2–
3 nm for the reported NC-CeO2 prepared by different meth-
ods.[10b,37] The domain particles do not seem to grow larger,
even if solution conditions are modified (i.e., Ce concentra-
tion, pH, or type of basic reagent). Assuming that the
domain particle of NC-CeO2 has a polyhedral shape with an
average diagonal length of 2.6 nm based on the HEXS re-
size of 2–3 nm (Figure 2(b)). The aggregation of the domain
particles can be also minimized by adjusting pH at around
the critical pH (pH 0.5 in the present condition), which can
be readily identified by the observation of pH variation
during the titration process.
Simple preparation of nanocrystalline cerium dioxide: Thus
far, the synthesis of NC-CeO2 has required involving “not
gentle” chemical processes, such as calcination or use of or-
ganic solvents/emulsion, to produce homogeneous and fine
particles. Based on the results obtained in this study, the
synthetic process for NC-CeO2 can be simplified and allevi-
ated. Firstly, the formation of CeO2 does not necessarily re-
quire a highly basic condition or calcination. The strong ola-
tion/oxolation tendency of CeIV is sufficient to produce
CeO2 even in an acidic condition without heating. Besides,
NC-CeO2 powder is obtained simply by evaporating the sol-
ution of the CeIV colloid formed above the critical pH. Sec-
ondly, NC-CeO2 seems to possess an intrinsic nature to be
stabilized with a specific domain size (2–3 nm in the present
conditions). Therefore, the effort for adding organic solvent/
emulsion or physical/chemical processes (e.g., pyrolysis, me-
chanochemical- or sonochemical process) to homogenize the
particle size distribution can be saved. Thirdly, the homoge-
neousness of NC-CeO2 particles can be controlled by opti-
mizing pH values, Ce concentration, and the type of basic
reagents for pH adjustment. The particle size distribution of
NC-CeO2 colloid also tends to become homogeneous with
aging. This aging effect may be accelerated by gentle heat-
ing. The size of NC-CeO2 particles can be possibly control-
led simply by selecting the type of basic reagents (i.e., coun-
tercations), which play an important role in stabilizing nano-
crystals with a specific size in solution.[38]
ꢂ
ꢂ
sults, the number of the Ce Ooxo Ce unit on one diagonal
line of the polyhedron is calculated to be 6–7 (Section 1–6 in
the Supporting Information). Tetravalent plutonium (PuIV),
whose ionic radius is very close to that of CeIV (hexagonal:
PuIV =0.86 ꢁ; CeIV =0.87 ꢁ, octahedral: PuIV =0.96 ꢁ,
CeIV =0.97 ꢁ),[21] is known to form a stable nano-sized oxide
cluster of [Pu38O56]40+ in an aqueous solution.[31d] The struc-
ture of this PuIV cluster is based on the fluorite-type PuO2
ꢂ
ꢂ
arrangement, in which three Pu Ooxo Pu units are lying on
a diagonal line. Given the fact that CeO2 is also based on
the fluorite structure, the present NC-CeO2 is highly likely
to form a similar but slightly larger oxide cluster. Assuming
that the domain particle of the CeIV colloid is forming a
cubic polyhedral cluster (Figure S13 in the Supporting Infor-
mation), the cluster possesses a huge specific surface. That
is, of the 146 Ce atoms, 102 are the end member atoms con-
stituting the cluster surface. This suggests that 69% of Ce
atoms in the CeIV cluster are still surrounded by OHꢂ and/
or water molecules in solution. Hence, the EXAFS-FT Peak
1 observed for the CeIV colloid samples in Figure 3(b)
ꢂ
(pH “0.5”–“2.0”) is probably attributed to the Ce OACTHNUTRGNEUNG
(OHꢂ)
bonding of the end-member Ce atoms on the cluster surface,
ꢂ
ꢂ
ꢂ
as the Ce O
O
(OH ) distance can be shorter than that of Ce
oxo in the rigid crystal structure of CeO2.[22]
DLS results suggest that the hydrodynamic diameter of
the CeIV colloid and its size distribution are controlled by
several factors. As discussed in the “Results: High energy
X-ray scattering” section, the difference in hydrodynamic di-
ameter could be attributed to the different thickness of the
interfacial double layer around the domain NC-CeO2 parti-
cles, as well as the agglomeration property of the particles.
The observed particle size distribution can be interpreted by
the agglomeration of domain particles. That is, the domain
NC-CeO2 particles agglomerate to form larger clusters.[31d,34a]
The surface of the domain particles is more hydroxidized as
an increase in pH, prompting the condensation of the
domain particles through the olation reaction. This reasona-
bly explains the observed broadening trend in particle size
distribution with increasing pH (Figure 2(a)). A higher
metal concentration increases the number of domain parti-
cles in the system, augmenting the opportunity for the
domain particles to be agglomerated. This is also in line
with the broadening trend of particle size distribution with
increasing Ce concentration (Figure S3 in the Supporting In-
formation). However, these agglomerated clusters are not
stable thermodynamically for a long period of time, as they
are disassembled gradually with time and the particle size
distribution finally becomes homogeneous with the domain
Conclusion
The present study has demonstrated that a combination of
DLS, XAS, and HEXS is a very powerful and comprehen-
sive approach to study in situ the evolution of nanocrystal-
line materials in solution. This multiple spectroscopic ap-
proach has been applied to the formation of NC-CeO2 in-
duced by the hydrolysis of CeIV. The obtained results have
proven that, with increasing pH, the initial precursor CeIV
species of small oligomer complexes, such as a dinuclear
complex, grow into larger ones through olation/oxolation re-
actions, finally producing a colloid being composed of CeO2
with a uniform domain size of 2–3 nm. In the solution condi-
tions studied, the NC-CeO2 colloid is already formed at a
critical pH of 0.5, which is very acidic. The formation of
NC-CeO2 in an acidic condition, as well as the presence of
the initial hydroxo/oxo-bridging dinuclear complexes, stems
mainly from the strong olation/oxolation tendency of CeIV.
Based on these findings, the synthetic route to NC-CeO2 can
be simplified and alleviated. That is, a simple evaporation of
the CeIV colloidal solution, which is prepared by a careful
pH adjustment with a selected basic reagent, will result in
&
10
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Hydrolysis of Tetravalent Cerium
FULL PAPER
99 mA with top-up mode. A liquid N2-cooled SiACTHNUTRGENUG(N 311) double crystal was
the production of fine NC-CeO2 powder. Besides, this
simple synthetic method can be potentially applied to the
production of other nanocrystalline metal oxides, such as
TiO2 from TiIV,[39] or ZrO2 from ZrIV,[40] as these oxides can
be prepared through the hydrolysis of the corresponding tet-
ravalent cations. The key to establishing the hydrolysis-
based simple route to nanocrystalline metal oxides is to
identify the critical pH condition at which the nanocrystals
begin to form. To this end, the present multi-spectroscopic
analytical approach serves as a promising tool for optimizing
the synthesis conditions to obtain nanocrystalline metal
oxides, as it is a versatile approach applicable to any other
solution or colloidal system.
employed to monochromatize white X-rays from the synchrotron ring.
All spectra were collected in transmission mode by using ionization
chambers filled with an Ar/N2 mixture (Ar/N2 =5:5) at ambient tempera-
ture and pressure. Prior to the measurement of each sample, the energy
of incident X-rays was corrected by measuring the first inflection point of
the Ce K-edge spectrum (defined as 40.447 keV) for reference CeO2
powder. For each sample, a single scan from 40.07 to 41.92 keV was com-
pleted in 60 seconds, and the scan was repeated 20 times. The obtained
20 spectra were then averaged for data analysis. The pH adjustment of
sample solutions was carried out in a N2-filled glove box immediately
before the measurement. The sample solutions were enclosed in a doubly
sealed plastic cuvette with an optical path length of 10 mm in the same
N2-filled glove box, and transferred to the beamline directly. The whole
measurement process, including pH adjustment, sample sealing, sample
transport, and collecting XAS spectra was completed within 30 min. Pre-
cipitate samples were measured as a “wet” form, in which the precipitate
was gathered at the bottom of the sample cuvette by spontaneous sedi-
mentation, without removing solution phase or evaporation. EXAFS
data analysis was performed according to the standard procedure[43] by
using a dedicated program WinXAS (version 3.2).[44] EXAFS structural
parameters were obtained by theoretical curve fitting both in k space
(i.e., EXAFS oscillation spectra) and in R space (i.e., Fourier-trans-
formed spectra). The threshold energy, Ek=0, was defined as the first in-
flection point of each spectrum. Theoretical phase and amplitude re-
quired for the curve fitting were calculated by a program code FEFF
8.20[45] based on the reported crystal structures of CeO2,[27b] CeIV dinu-
clear complexes[46] and DFT-optimized CeIV oligomer complexes.[22] All
significant single-scattering- (SS) and multiple-scattering (MS) paths,
whose relative amplitude was calculated to be more than 15%, were
taken into account on the curve fitting. The amplitude reduction factor,
Experimental Section
Sample preparation: A weighted amount of CeACHTUNGTRNEUNG(NO3)3·6H2O (Rare Met-
allic Co. Ltd., or Sigma–Aldrich Co., 99.99%) was dissolved into 1.0m
HNO3 (100 mL) to give a Ce concentration of 0.1, 0.5, or 1m. The solu-
tion was then electrolyzed at 1.9 V[22] to oxidize CeIII to CeIV by using a
potentiostat (ALS Electrochemical Analyzer Model 600C or Metrohm
Autolab PGSTAT12/30/302) with
a three electrode system (Pt-plate
working and counter electrodes, and a Ag/AgCl reference electrode).
After the electrolysis, the Ce oxidation state in the solution was con-
firmed to be tetravalent by fluorescence spectroscopy (Figure S1 in the
Supporting Information) and Ce K-edge XANES spectroscopy (Fig-
ure S2 in the Supporting Information). The resultant solution was em-
ployed as a stock solution for subsequent pH titration experiments. That
is, the stock solution was sequentially titrated by a NaOH or NH3 solu-
tion to prepare a series of CeIV solution samples with different pH
values. To minimize the local pH excursion effects, which potentially pro-
duce larger aggregate species, the titrand solution was stirred vigorously
and the basic concentration and titrant volume added were minimized in
the course of pH adjustment. At each pH, a small portion of the titrated
solution was collected for spectroscopic measurements. The whole titra-
tion process involved less than 5% decrease in solution volume. There-
fore, we consider that a suite of the collected solution samples has a con-
stant Ce concentration. The pH measurement was performed by using a
pH meter (HORIBA F-52 or inoLab WTW-pH720) calibrated with four
different pH buffer solutions (pH 1.68, 4.01, 6.86, and 9.18 at 298 K). All
sample preparation and sealing were performed in an inert glove box
filled with N2 to avoid the penetration of atmospheric oxygen or carbon
dioxide into the sample solutions. Additionally, deionized water used for
sample preparation was degassed and deoxygenated by purging N2 prior
to its use.
2
S0 , was fixed at 0.9, and the shifts in the threshold energy, DEk=0, were
constrained to be the same value for all shells, to reduce the free parame-
ters on the curve fit procedure as much as possible.
High energy X-ray scattering: HEXS measurements were performed on
beamline ID15B at the ESRF by using a 87 keV (l=0.142 ꢁ) monochro-
matic X-ray beam selected by a single SiACHTUNTRGNEUNG(511) bent Laue crystal. The
cross section of the beam was adjusted to 0.1ꢂ0.1 mm2 by tungsten slits
in front of the sample. Two-dimensional scattering images were recorded
at room temperature on a mar345 image plate detector (Marresearch
GmbH) in transmission geometry. The distance between the sample and
the detector was 300 mm. Samples were sealed in a polycarbonate capil-
lary (2.0 mm in diameter) with epoxy and packaged in a N2-filled sample
container having a Kapton window. The scattering data of an empty ca-
pillary and Ce-free blank solutions (i.e., aqueous HNO3 with NaOH or
NH3) were also collected for background subtraction. Data reduction and
PDF analysis were carried out on a software package IGOR Pro (version
6.0) with the dedicated macro commands developed by S. Skanthakumar
and co-workers at Argonne National Laboratory, according to the same
procedure described previously by the same group.[31,47] Fourier transfor-
mation from Q space (ꢁꢂ1) into R-space (ꢁ) was performed in the range
Dynamic light scattering: Sample solutions enclosed in a Pyrex glass cuv-
ette were irradiated by an argon ion laser LEXEL 95 (LEXEL Laser,
400 mW, 514.5 nm wavelength) at a constant temperature of 298 K. The
intensity of scattered light from the samples was detected at a right angle
to the incident light by using a BI-90 Particle Size Analyzer (Brookhaven
Instruments Co.) with a 120 second duration. The measurement was re-
peated ten times for each sample and the collected data were averaged.
The autocorrelation functions, C(t), were calculated from the fluctuations
of the scattered light intensity to estimate the hydrodynamic radii (i.e.,
Stokes radii) of the particles in the samples.[18] The Cumulant expansion
and the CONTIN analysis were employed to derive the particle size in-
formation based on the obtained C(t).[41] For the measurements of aging
effect, the samples were triply sealed in a glass cuvette and stored in a
N2-filled inert glove box between the measurements, to avoid possible
penetration of air into the samples.
of Q=1.0–21.0 ꢁꢂ1
.
Transmission electron microscopy: A Titan 80-300 electron microscope
(FEI) operating at 300 kV was employed for acquiring TEM images in
bright field. The instrument is equipped with a field emission gun, a Su-
perTWIN-a lens, and an image corrector providing a point resolution of
0.10 nm. Samples for TEM measurement were prepared by dropping a 5
or 10 mL of a sample solution on a 400-mesh copper grid coated with
carbon film (10 to 15 mm in thickness, Plano GmbH), wicking away the
solution with a filter paper, and drying in a desiccator for one day.
Acknowledgements
X-ray absorption spectroscopy: X-ray absorption spectra, including both
XANES and EXAFS regions, were collected at Ce K-edge (40.447 keV)
in quick-scanning mode at the JAEA Quantum Dynamics Beamline
BL11XU,[42] SPring-8, under the ring operating conditions of 8 GeV and
The authors thank H. Shiwaku and Y. Tomisugi for technical support for
XAS measurement on the beamline BL11XU at SPring-8, V. Honkimꢃki
and T. Buslaps for HEXS measurement on the beamline ID15B at the
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
11
&
ÞÞ
These are not the final page numbers!

Dr. A. Ikeda-Ohno et al.
ESRF and relevant data analysis, W. Kraus for advice on the Scherrer
equation, S. Skanthakumar for providing the IGOR macros dedicated for
HEXS data analysis, and A. Mꢄcklich for TEM measurement at HZDR.
We also acknowledge the allocation of beam times and assistance for
HEXS experiments at the beamline ID15B, ESRF, under Proposal Code
HD488. We also greatly appreciate the constructive comments from an
anonymous reviewer.
Hennig, A. Rossberg, H. Funke, A. C. Scheinost, G. Bernhard, T.
Yaita, Inorg. Chem. 2008, 47, 8294–8305; c) A. Ikeda-Ohno, C.
Hennig, S. Tsushima, A. C. Scheinost, G. Bernhard, T. Yaita, Inorg.
Chem. 2009, 48, 7201–7210.
[24] G. Johansson, M. Magini, H. Ohtaki, J. Solution Chem. 1991, 20,
775–792.
[25] F. Aubriet, J.-J. Gaumet, W. A. de Jong, G. S. Groenewold, A. K.
Gianotto, M. E. McIlwain, M. J. Van Stipdonk, C. M. Leavitt, J.
Phys. Chem. A 2009, 113, 6239–6252.
[26] C. F. Baes, Jr., R. E. Mesmer, The Hydrolysis of Cations, Wiley, New
York, 1976, chap. 7, pp. 138–146.
[27] a) G. Wang, Q. Mu, C. Ting, W. Yude, J. Alloys Compd. 2010, 493,
202–207; b) R. W. G. Wyckoff, Crystal Structure, Vol. 1, 2nd ed., In-
terscience, New York, 1963, pp. 239–444.
[28] M. Henry, J. P. Jolivet, J. Livage, Struct. Bonding 1992, 77, 153–206.
[29] a) V. Mereacre, A. M. Ako, M. N. Akhtar, A. Lindemann, C. E.
Anson, A. K. Powell, Helv. Chim. Acta 2009, 92, 2507–2524; b) R.
Das, R. Sarma, J. B. Baruah, Inorg. Chem. Commun. 2010, 13, 793–
795.
[1] L. L. Hench, J. K. West, Chem. Rev. 1990, 90, 33–72.
[2] X. Wang, J. Zhuang, Q. Peng, Y. Li, Nature 2005, 437, 121–124.
[3] B. D. Fahlman, Materials Chemistry, Vol. 1, Springer, Mount Pleas-
ant, MI, 2007.
[4] W. P. Halperin, Rev. Mod. Phys. 1986, 58, 533–606.
[5] a) J. S. Bradley, Clusters Colloids, VCH, Weinheim, 1994; b) A.
Tschope, W. Liu, M. Flytzani-Stephanopoulos, J. Y. Ying, J. Catal.
1995, 157, 42–50.
[6] A. Karakoti, S. Singh, J. M. Dowding, S. Seal, W. T. Self, Chem. Soc.
Rev. 2010, 39, 4422–4432, and references therein.
[7] Y. Xia, Y. Xiong, B. Lim, S. E. Skrabalak, Angew. Chem. 2009, 121,
62–108; Angew. Chem. Int. Ed. 2009, 48, 60–103, and references
therein.
[8] a) Y. Xiong, I. Washio, J. Chen, M. Sadilek, Y. Xia, Angew. Chem.
2007, 119, 5005–5009; Angew. Chem. Int. Ed. 2007, 46, 4917–4921;
b) Y. Xiong, Y. Xia, Adv. Mater. 2007, 19, 3385–3391.
[9] T. Egami, S. J. L. Billinge in Underneath the Bragg Peaks—Structural
Analysis of Complex Materials (Pergamon Materials Series), Vol. 7
(Ed.: R. W. Cahn), Elsevier, Oxford, 2003.
[30] M. Balasubramanian, C. A. Melendres, A. N. Mansour, Thin Solid
Films 1999, 347, 178–183.
[31] a) J. Neuefeind, S. Skanthakumar, L. Soderholm, Inorg. Chem. 2004,
43, 2422–2426; b) R. E. Wilson, S. Skanthakumar, G. Sigmon, P. C.
Burns, L. Soderholm, Inorg. Chem. 2007, 46, 2368–2372; c) S. Skan-
thakumar, M. R. Antonio, R. E. Wilson, L. Soderholm, Inorg.
Chem. 2007, 46, 3485–3491; d) L. Soderholm, P. M. Almond, S.
Skanthakumar, R. E. Wilson, P. C. Burns, Angew. Chem. 2008, 120,
304–308.
[32] a) C. Hennig, A. Ikeda, K. Schmeide, V. Brendler, H. Moll, S.
Tsushima, A. C. Scheinost, S. Skanthakumar, R. Wilson, L. Soder-
holm, K. Servaes, C. Gçrrler-Walrand, R. Van Deun, Radiochim.
Acta 2008, 96, 607–611; b) C. Hennig, S. Weiss, D. Banerjee, E.
Brendler, V. Honkimꢃki, G. Cuello, A. Ikeda-Ohno, A. C. Scheinost,
H. Zꢃnker, Geochim. Cosmochim. Acta 2013, 103, 197–212.
[33] a) P. Scherrer, Gçttinger Nachrichten Gesell. 1918, 2, 98–100;
b) A. L. Patterson, Phys. Rev. 1939, 56, 978–982.
[34] a) Z. L. Wang, X. Feng, J. Phys. Chem. B 2003, 107, 13563–13566;
b) F. Zhang, Q. Jin, S.-W. Chan, J. Appl. Phys. 2004, 95, 4319–4326.
[35] a) S. Takao, K. Takao, W. Kraus, F. Emmerling, A. C. Scheinost, G.
Bernhard, C. Hennig, Eur. J. Inorg. Chem. 2009, 4771–4775; b) K.
Takao, S. Takao, A. C. Scheinost, G. Bernhard, C. Hennig, Inorg.
Chem. 2012, 51, 1336–1344; c) C. Hennig, S. Takao, K. Takao, S.
Weiss, W. Kraus, F. Emmerling, A. C. Scheinost, Dalton Trans. 2012,
41, 12818–12823.
[10] a) N.-C. Wu, E.-W. Shi, Y.-Q. Zheng, W.-J. Li, J. Am. Ceram. Soc.
2002, 85, 2462–2468; b) H. Wang, J.-J. Zhu, J.-M. Zhu, X.-H. Liao,
S. Xu, T. Ding, H.-Y. Chen, Phys. Chem. Chem. Phys. 2002, 4, 3794–
3799; c) H. Zou, Y. S. Lin, N. Rane, T. He, Ind. Eng. Chem. Res.
2004, 43, 3019–3025; d) M. L. Dos Santos, R. C. Lima, C. S. Riccar-
di, R. L. Tranquilin, P. R. Bueno, J. A. Varela, E. Longo, Mater. Lett.
2008, 62, 4509–4511.
[11] L. Eyring in Handbook on the Physics and Chemistry of Rare
Earths, Vol. 3 (Eds.: K. A. Gschneidner, Jr., L. Eyring,) North-Hol-
land, Amsterdam, 1979, chap. 27, pp. 337–399.
[12] Rare Earths: Science Technology, and Applications III (Eds.: R. G.
Bautista, C. O. Bounds, T. W. Ellis, B. T. Kilbourn), The Minerals,
Metals & Materials Society, Pennsylvania, 1997.
[13] A. Trovarelli, Catal. Rev. Sci. Eng. 1996, 38, 439–520.
[14] a) Q. Fu, A. Weber, M. Flytzani-Stephanopoulos, Catal. Lett. 2001,
77, 87–95; b) Q. Fu, H. Saltsburg, M. Flytzani-Stephanopoulos, Sci-
ence 2003, 301, 935–938; c) S. Carrettin, P. Concepciꢅn, A. Corma,
J. M. Lꢅpez Nieto, V. F. Puntes, Angew. Chem. 2004, 116, 2592–
2594; Angew. Chem. Int. Ed. 2004, 43, 2538–2540.
[15] a) R. J. Gorte, S. Park, J. M. Vohs, Nature 2000, 404, 265–267; b) T.
Hibino, A. Hashimoto, T. Inoue, J. Takuno, S. Yoshida, M. Sano, Sci-
ence 2000, 288, 2031–2033; c) B. C. H. Steele, A. Heinzel, Nature
2001, 414, 345–352.
[16] X. Feng, D. C. Sayle, Z. L. Wang, M. S. Paras, B. Santora, A. C. Su-
torik, T. X. T. Sayle, Y. Yang, Y. Ding, X. Wang, Y.-X. Her, Science
2006, 312, 1504–1508.
[17] a) S. Yabe, T. Sato, J. Solid State Chem. 2003, 171, 7–11; b) N. M.
Zholobak, V. K. Ivanov, A. B. Shcherbakov, A. S. Shaporev, O. S.
Polezhaeva, A. Ye. Baranchikov, N. Ya. Spivak, Yu. D. Tretyakov, J.
Photochem. Photobiol. B 2011, 102, 32–38.
[18] W. Brown, Dynamic Light Scattering: The Method and Some Appli-
cations, Oxford University Press, New York, 1993.
[36] S. Pascarelli, O. Mathon, Phys. Chem. Chem. Phys. 2010, 12, 5535–
5546.
[37] a) T. Masui, K. Fujiwara, K. Machida, G. Adachi, Chem. Mater.
1997, 9, 2197–2204; b) M. Inoue, M. Kimura, T. Inui, Chem.
Commun. 1999, 957–958; c) T. Yu, J. Joo, Y. I. Park, T. Hyeon,
Angew. Chem. 2005, 117, 7577–7580; Angew. Chem. Int. Ed. 2005,
44, 7411–7414; d) Y. Liu, S. He, M. Uehara, H. Maeda, Chem. Lett.
2007, 36, 764–765; e) Z. Huo, C. Chen, X. Liu, D. Chu, H. Li, Q.
Peng, Y. Li, Chem. Commun. 2008, 3741–3743.
[38] J. D. Aiken III, R. G. Finke, J. Mol. Catal. A 1999, 145, 1–44.
´
[39] a) E. Matijevic, M. Budnik, L. Meites, J. Colloid Interface Sci. 1977,
61, 302–311; b) C. Charbonneau, R. Gauvin, G. P. Demopoulos, J.
Cryst. Growth 2009, 312, 86–94.
[40] a) H. Cçlfen, H. Schnablegger, A. Fischer, F. C. Jentoft, G. Wein-
berg, R. Schlçgl, Langmuir 2002, 18, 3500–3509; b) N. M. P. Ricar-
do, F. C. F. de FranÅa, C. Price, N. M. P. S. Ricardo, Colloids Surf. A
2004, 232, 113–118.
[41] I. Dreissig, S. Weiss, C. Hennig, G. Bernhard, H. Zꢃnker, Geochim.
Cosmochim. Acta 2011, 75, 352–367.
[42] H. Shiwaku, T. Mitsui, K. Tozawa, K. Kiriyama, T. Harami, T. Mo-
chizuki, AIP Conf. Proc. 2004, 705, 659–662.
[19] T. Bundschuh, R. Knopp, R. Mꢄller, J. I. Kim, V. Neck, T. Fanghꢃ-
nel, Radiochim. Acta 2000, 88, 625–629.
[20] K. Opel, S. Weib, S. Hꢄbener, H. Zꢃnker, G. Bernhard, Radiochim.
Acta 2007, 95, 143–149.
[21] R. D. Shannon, Acta Crystallogr. Sect. A 1976, 32, 751–767.
[22] A. Ikeda-Ohno, S. Tsushima, C. Hennig, T. Yaita, G. Bernhard,
Dalton Trans. 2012, 41, 7190–7192.
[23] a) P. G. Allen, D. K. Veirs, S. D. Conradson, C. A. Smith, S. F.
Marsh, Inorg. Chem. 1996, 35, 2841–2845; b) A. Ikeda-Ohno, C.
[43] R. Prins, D. E. Koningsberger, X-ray Absorption: Principles, Appli-
cations, Techniques for EXAFS, SEXAFS, and XANES, Wiley, New
York, 1988.
[44] T. Ressler, J. Synchrotron Radiat. 1998, 5, 118–122.
&
12
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Hydrolysis of Tetravalent Cerium
FULL PAPER
[45] A. L. Ankudinov, B. Ravel, J. J. Rehr, S. D. Conradson, Phys. Rev. B
1998, 58, 7565–7576.
[47] S. Skanthakumar, L. Soderholm, Mater. Res. Soc. Symp. Proc. 2006,
893, 411–416.
[46] a) O. Lindgren, Acta Chem. Scand. Ser. A 1977, 31, 163–166; b) N.
Guillou, J. P. Auffrꢆdic, D. Louꢇr, J. Solid State Chem. 1994, 112,
45–52.
Received: November 16, 2012
Revised: March 1, 2013
Published online: && &&, 2013
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
13
&
ÞÞ
These are not the final page numbers!

Dr. A. Ikeda-Ohno et al.
Nanocrystal evolution: A multi-spec-
troscopic approach has revealed that
the hydrolysis of tetravalent cerium
(CeIV) produces a colloidal solution of
fine cerium dioxide (CeO2) nanocrys-
tals (see figure). A simple synthetic
concept of nanocrystalline metal
oxides is proposed based on the
obtained results.
Nanomaterials
A. Ikeda-Ohno,* C. Hennig, S. Weiss,
T. Yaita, G. Bernhard . . . . . . . &&&& — &&&&
Hydrolysis of Tetravalent Cerium for a
Simple Route to Nanocrystalline
Cerium Dioxide: An In Situ Spectro-
scopic Study of Nanocrystal Evolution
CeO2 nanocrystals
Simple hydrolysis of aqueous tetravalent cerium (CeIV),
that is, without additional physical or chemical processes,
gives nanosized crystalline cerium dioxide (CeO2) parti-
cles. Multispectroscopic analysis shows that the transfor-
mation of small oligomeric CeIV complexes into nanosized
cerium dioxide proceeds through olation/oxolation reac-
tions. The study outlines a simple and conceptually new
way of preparing metal oxide nanocrystals and a strategy
for analyzing nanocrystal evolution in situ. For more
details, see the Full Paper by A. Ikeda-Ohno et al. on
page &&ff.
&
14
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!