View
1
Download
0
Category
Preview:
Citation preview
S1
Supplementary information for
Quantum Size Effect in TiO2 Nanoparticles
Prepared by Finely Controlled Metal Assembly on
Dendrimer Templates
Norifusa Satoh, Toshio Nakashima, Kenta Kamikura, Kimihisa Yamamoto*
Department of Chemistry, Faculty of Science and Technology, Keio University, Yokohama 223-8522, Japan
*e-mail: yamamoto@chem.keio.ac.jp
Supplementary information for
Quantum Size Effect in TiO2 Nanoparticles Prepared by Finely Controlled Metal Assembly on Dendrimer Templates
S2
Supplementary Note
The following relationship between the bandgap of a semiconductor particle (Einfinite) and
the radius (R) is described by Brus, using the effective mass approximation1:
!
Einfinite " Eg +h2# 2
2R2
1
µ$1.8e
2
%R+ smaller terms (S1)
where ħ is Planck’s constant; e is the charge on the electron; Eg is the bandgap in the bulk; µ is the
reduced mass of the electron me and hole mh in the bulk; and ε is the dielectric constant of the bulk
semiconductor. Although this description is widely used in the study of TiO2 nanoparticles, the
failure of the Brus equation is generally known especially in the ultrasmall region; for example, in
the case of CdS.2, 3, 4 The main reason is the conventional infinite depth well model of the Brus
equation. Additionally, the electronic band structure of such small clusters is affected by their
surface state and related relaxation in the atomic positions.
Assuming an apparent reduced mass of the electron and hole µ´ and an apparent dielectric
constant ε´ in the infinite depth well model, we can replace the Nosaka equation4 for the excitation
energy Efinite using a finite depth well model with the following expression looking like a Brus
equation:
!
E finite ~ Eg +h2" 2
2R2
1
# µ $1.8e
2
%#R (S2)
In the case of TiO2, the normally used value of 1.63 mo as the µ, experimentally estimated from the
Brus equation using E* (∆E = 0.15 eV) at R = 1.2 by Kormann et al.5 would be the above µ´ to
describe the realistic finite potential well model in the rigorous meaning. As Kormann et al. also
mentioned, the real µ is believed to be less than 1 mo in the bulk because mh is much smaller than me
N. Satoh, T. Nakashima, K. Kamikura & K. Yamamoto*
* yamamoto@chem.keio.ac.jp
S3
(e.g., me ~ 5–100 mo, mh ~ 0.01–0.8 mo)6, 7. By contrast, the ε in TiO2 is large enough to permit
ignoring the failure due to the infinite depth well model because of the small second term of the
Brus equation (cf. ε = 5.7 for CdS). For these reasons, the Brus equation for TiO2, generally
adopted, would correspond to the experimental data. Namely, the equation (S2) (= equation (1) in
the main text) is not the Brus equation based on the infinite depth well model but is equivalent to
the Nosaka equation based on the finite depth well model in physical meaning.
Additionally, this kind of semi-empirical prediction would essentially include the
information about the surface and the atomic character of the nanoparticles under the same
condition. Zunger and coworkers succeed in treating these problems in the calculation of the
electronic structure, using the semi-empirical pseudopotential method.8, 9, 10 Experimentally, it has
been shown that the surface modification of nanoparticles, e.g., CdSe and CdS, has little effect on
the absorption spectra without the electronic interaction with the aromatic rings of the
surface-capping molecules.11, 12
In contrast, Jorne and Fauchet et al. have observed the dramatic effect of the surface Si-H
and Si=O bonds on the electronic structure, such as absorption and luminescence, in Si nanocrystals
(1nm < 2R < 4 nm);13 Si is an indirect bandgap semiconductor. The phenomena are confined with
the theoretical computation by Zdetsis et al. and Brus et al.14, 15, 16 The H- and O-passivation
remarkably changes the chemical structure from almost pure Si in a bulk to a 1:1 ratio in the range
of 1.1–1.4 nm15, 16 and finally to the molecular structure, such as SiH4 and SiO2 (or Si(OH)4). In
the ultra-small range, the effect of the newly formed chemical bonds needs to be considered in the
electronic structure.
Although TiO2 is also classified as an indirect bandgap semiconductor (see S 23–25), the
Supplementary information for
Quantum Size Effect in TiO2 Nanoparticles Prepared by Finely Controlled Metal Assembly on Dendrimer Templates
S4
chemical structure changes less than that of Si; from TiO2 in the bulk to Ti(OR)4 in the final
molecular structure, where R is H or Si of the substrate SiO2 under the present study condition.17, 18
In both of the structures, a Ti atom is always surrounded by O atoms. Additionally, these atoms
similarly contribute to the electronic transitions; O of the valence band to Ti of the conduction band
in the bulk, and O of the ligands to Ti of the metal in the molecular structure. Therefore, we can
expect to observe the size dependency of an indirect bandgap semiconductor TiO2 in the ultra-small
region, based not on the surface effect but on the Q-size effect combined with a simple electrostatic
continuum, equation (1) in the main text.
In the theoretical field, numerous quantum mechanical calculations for free TiO2 clusters
have been performed. 19 However, these calculations neglect the chemical character of the Ti(IV)
ion, which tends to form a TiO4 tetrahedral or a TiO6 octahedral structure. Although these
calculations may successfully explain the experimental data for these free TiO2 clusters in a very
specific environment such as a high vacuum chamber20, these structures are too unstable in surface
energy and structure in an air atmosphere21. By contrast, Doren et al. have recently performed a
theoretical calculation for TiO6 octahedral chains. 22 They revealed that the Ti(3d) lowest
unoccupied state is delocalized in the 3Ti chain and that the lowest optical energy transition is an
O(2p) → Tibulk(3d). The theoretical calculation for Q-size TiO2 particles is beyond the scope of
our present study, but it will be needed to understand the detailed electronic structure in individual
particles.
N. Satoh, T. Nakashima, K. Kamikura & K. Yamamoto*
* yamamoto@chem.keio.ac.jp
S5
Supplementary Figure 1 1:1 complexation between phenylazomethine and
Ti(acac) Cl3. a, Changes in UV-vis spectra during addition of Ti(acac)Cl3 (in steps of 0.2
eq. until 1.0 eq.): b, titration curve, based on the spectral changes at 315 nm; c, scheme
of complexation of Ti(acac)Cl3 and phenylazomethine. These spectra changes are similar
to those of SnCl2 and FeCl3. Thus, the results indicate the complexation of the imine
ligand with Ti(acac)Cl3.
Supplementary information for
Quantum Size Effect in TiO2 Nanoparticles Prepared by Finely Controlled Metal Assembly on Dendrimer Templates
S6
Supplementary Figure 2 Changes in 1H-NMR spectra of phenylazomethine during
addition of Ti(acac)Cl3. a, phenylazomethine; b, c, d, e, phenylazomethine:
Ti(acac)Cl3 = 5:1 (b), 2:1 (c), 1:1 (d), 1:2 (e). With the addition of Ti(acac)Cl3, the
chemical shift moved upfield. This means the formation of a complex of
phenylazomethine with Ti(acac)Cl3. The shift is saturated at the ratio of 1:1. A similar
upfield shift is also observed during the complexation of phenylazomethine with SnCl2.
N. Satoh, T. Nakashima, K. Kamikura & K. Yamamoto*
* yamamoto@chem.keio.ac.jp
S7
Supplementary Figure 3 Changes in UV-vis spectra during addition of Ti(acac)Cl3.
a, DPA G3 (1 × 10–5 M) while adding 14 drops of 3 μl dehydrated acetonitrile solution of
Ti(acac)Cl3 (3 × 10–2 M); b, DPA G2 (3 × 10–5 M) while adding 6 drops of 3 μl dehydrated
acetonitrile solution of Ti(acac)Cl3 (1 × 10–2 M).
a
b
Supplementary information for
Quantum Size Effect in TiO2 Nanoparticles Prepared by Finely Controlled Metal Assembly on Dendrimer Templates
S8
Supplementary Figure 4 AFM image (50 nm × 50 nm) of [Ti(acac)Cl3]30@DPA G4.
a, top view; b, cross-sectional view.
N. Satoh, T. Nakashima, K. Kamikura & K. Yamamoto*
* yamamoto@chem.keio.ac.jp
S9
Supplementary Figure 5 Molecular modeling of [Ti(acac)Cl3]30@DPA G4. a, Top
view; b, side view. The size is in good agreement with that obtained in the AFM image
(Supplementary S5). An assembly of 30 equivalent Ti(acac)Cl3s increased the size from
2.3 × 2.5 × 2.9 nm to 3.0 × 3.9 × 4.0 nm; the size for this assembly is greater than that
reported using SnCl2.
Supplementary information for
Quantum Size Effect in TiO2 Nanoparticles Prepared by Finely Controlled Metal Assembly on Dendrimer Templates
S10
450455460465470 395400405410
N1s
Ti2p3/2
Ti2p1/2
[Ti(acac)Cl3]30@DPA G4
[Ti(acac)Cl3]14@DPA G4
[Ti(acac)Cl3]6@DPA G4
[Ti(acac)Cl3]2@DPA G4
Ti(acac)Cl3
DPA G4
Binding energy (eV)
Supplementary Figure 6 XPS spectra of Ti-assembled DPA G4. The binding
energy of Ti(2p3/2) showed an upfield shift as complexing occurred with the imine groups on
the outer layer of DPA G4. The shift shows that the imine groups on the inner layer are
stronger bases than those on the outer layer. Ti(acac)Cl3 showed the lowest binding
energy, because it exists in a dimeric structure in the solid form; the chlorine atoms, which
bind two titanium atoms, acts as stronger donors than the N atoms of DPA G4.
N. Satoh, T. Nakashima, K. Kamikura & K. Yamamoto*
* yamamoto@chem.keio.ac.jp
S11
50 nm
Eg
A1gru
ru
300 400 500 600200 700Raman shift (cm-1)
a
b
Supplementary Figure 7 Hydrolysis of Ti-assembled DPA G4 under bulk
conditions. a, TEM image of hydrolyzed TiO2 under bulk conditions. b, Raman
spectrum of TiO2 obtained from hydrolysis of Ti-assembled DPA G4 in 10 %-HCl solution.
Supplementary information for
Quantum Size Effect in TiO2 Nanoparticles Prepared by Finely Controlled Metal Assembly on Dendrimer Templates
S12
Supplementary Figure 8 1H-NMR spectra of hydrolyzed DPA G4. DPA G4 was
hydrolyzed by the addition of a 10% HCl solution, forming small organic compounds, such
as benzophenone, diaminobenzophenone and a complex between diaminobenzophenone
and HCl.
N. Satoh, T. Nakashima, K. Kamikura & K. Yamamoto*
* yamamoto@chem.keio.ac.jp
S13
Supplementary Figure 9 Thermogravimetric curves of hydrolyzed DPA G4.
Benzophenone, diaminobenzophenone and the complex between diaminobenzophenone
diminished at about 180, 330, and 630 ºC, respectively, with a heating rate of 20 ºC/min.
under an air atmosphere. Although the ionic bonds of the complex are strong, the final
decomposition of the complex under HCl treatment starts at about 500 ºC.
Supplementary information for
Quantum Size Effect in TiO2 Nanoparticles Prepared by Finely Controlled Metal Assembly on Dendrimer Templates
S14
Supplementary Figure 10 Observed carbon remaining on substrate after
annealing of hydrolyzed Ti-assembled DPA G4. a, Changes in XPS spectra of
Ti-assembled DPA G4 hydrolyzed by vapor of HCl solution on a mica substrate after
annealing at 500 ºC; b, decay in the C1s peak area. The C(1s) peak area was determined
by subtracting the mica substrate background and normalizing to the Ti(2p) peak area.
N. Satoh, T. Nakashima, K. Kamikura & K. Yamamoto*
* yamamoto@chem.keio.ac.jp
S15
a
b
200 300 400 500 600 700Raman shift (cm-1)
50 nm
Eg
B1g
A1g
B1gan
an
an an
Supplementary Figure 11 Thermolysis of Ti-assembled DPA G4 under bulk
conditions. a, TEM image of thermolyzed TiO2 under bulk condition. b, Raman
spectrum of TiO2 obtained from thermolysis of solid Ti-assembled DPA G4.
Supplementary information for
Quantum Size Effect in TiO2 Nanoparticles Prepared by Finely Controlled Metal Assembly on Dendrimer Templates
S16
Supplementary Figure 12 Thermogravimetric curves of DPA G4 and Ti-assembling
DPA G4. Under a N2 atmosphere, DPA G4 has a high thermostability (Td-10% = 530 ºC)
and carbon remains on pan at 1000 ºC. Under an air atmosphere, however, DPA G4 was
largely removed at 800 ºC at the heating rate of 20 ºC/min. The amount of Ti-assembled
DPA G4 that remained on the pan was matched with the weight of Ti added as TiO2. The
final decomposition of the Ti-assembled DPA G4 starts about 500 ºC.
N. Satoh, T. Nakashima, K. Kamikura & K. Yamamoto*
* yamamoto@chem.keio.ac.jp
S17
Supplementary Figure 13 Observed carbon remaining on substrate after
annealing Ti-assembled DPA G4. a, Changes in XPS spectra of Ti-assembled DPA G4
on mica substrate after annealing at 500 ºC; b, decay in the C1s peak area. The C(1s)
peak area was determined by subtracting the mica substrate background, and normalizing
to the Ti(2p) peak area.
Supplementary information for
Quantum Size Effect in TiO2 Nanoparticles Prepared by Finely Controlled Metal Assembly on Dendrimer Templates
S18
Supplementary Figure 14 XPS spectra of the obtained TiO2. a. rutile particle and
hydrolyzed Q-size TiO2. b. anatase particle and thermolyzed Q-size TiO2. The y-axis is
Intensity (A.U.). In bulk, Eb (Ti2p3/2) is 458.9 and 458.7 eV for rutile and anatase,
respectively. The Eb of rutile is higher than that of anatase in bulk, which would also be
correct on a nanoscale as long as the TiO6 octahedrons exist. As a size-dependency, the
Eb of Q-size TiO2 shifted upfield with a reduction in the size. The difference in the
synthesis methods, hydrolysis and thermolysis, provide the Eb of Q-size TiO2 at different
peak positions: the Eb of the hydrolyzed one is higher than that of the thermolyzed one by
0.1-2 eV. These results also indicate the Q-size effect and crystal form existence.
N. Satoh, T. Nakashima, K. Kamikura & K. Yamamoto*
* yamamoto@chem.keio.ac.jp
S19
Supplementary Figure 15 Relation of observed size and bandgap energy of TiO2
molded on DPA G4. a. hydrolyzed TiO2 and the equation (S2) (= equation (1) in the main
text) for rutile. b. thermolyzed TiO2 and the equation (S2) for anatase. The equation (S2)
is plotted using with Eg = 3.0, ɛ = 173 (rutile) for the hydrolyzed samples and Eg = 3.2, ɛ = 31
(anatase) for the thermolyzed samples.
a
b
Supplementary information for
Quantum Size Effect in TiO2 Nanoparticles Prepared by Finely Controlled Metal Assembly on Dendrimer Templates
S20
Supplementary Figure 16 Molecular modeling structure, highlighted in
Supplementary Table 1. a, 7-rutile (O-center); b, 15-rutile (Ti-center); c, 31-rutile
(Ti-center); d, 7-anatase (O-center); e, 15-anatase (O-center); f, 34-anatase
(Ti-center).
N. Satoh, T. Nakashima, K. Kamikura & K. Yamamoto*
* yamamoto@chem.keio.ac.jp
S21
Supplementary Table 1 The number of TiO62– octahedral units within the TiO2
cluster models and the minimum radius R (Å) of the cluster models, approximated
as a spherical shape with a Ti or O atom as the center. The number of octahedral
units, included in a sphere with a radius increasing in steps of 0.01 Å, was counted in each
case of the Ti-center and O-center. Since the number quantumly increased with the
radius, we regarded the minimum radius as 6, 14, and 30 octahedrons as the radius of
these (highlighted).
number R (Ti-center, Å) R (O-center, Å) 3 4.62 3.97 5 ↓ 4.92
6TiO2 rutile
7 5.4 5.04 11 5.44
13 ↓ 6.44
15 6.16 ↓
14TiO2 rutile
17 6.46 26 7.26 27 7.12 ↓
30 ↓ 7.49 30TiO2 rutile
31 7.45
3 3.96 5 4.76 ↓
7 ↓ 5.46 6TiO2
anatase
9 5.7
9 5.7
13 6.04 15
↓ 6.16
14TiO2 anatase
17 6.6
29 7.25 7.65 33 ↓ 8.16
30TiO2 anatase
34 7.83
Supplementary information for
Quantum Size Effect in TiO2 Nanoparticles Prepared by Finely Controlled Metal Assembly on Dendrimer Templates
S22
Supplementary Figure 18 Crystalline phase formation from TiO62– octahedra
nucleus. The nucleus structure of the rutile and anatase phases differs at the point
where the 3rd octahedron is added to the dimer, giving a linear chain and zigzag chain,
respectively. These chains three-dimensionally grow to form TiO2 clusters and/or bulk
crystals.
N. Satoh, T. Nakashima, K. Kamikura & K. Yamamoto*
* yamamoto@chem.keio.ac.jp
S23
Supplementary Note
Titanium oxide is classified as an indirect bandgap semiconductor; a weak indirect
transition is observed from a lower energy point (3.0 eV for rutile and 3.2 eV for anatase) than a
strong direct transition bandgap (3.3 eV for rutile23 and 3.4 eV for anatase24, 25) in bulk.
Nanoparticles of indirect semiconductors would be considered to change the transition nature to
direct one by decreasing in size, because the structure finally reaches the molecular size17, 18 having
a direct transition property.
Serpone et al. inferred that the blue shift of TiO2 nanoparticles should be based not on the
quantum size effect but on the direct transition from the observation of the size-independent blue
shift using the size-uncontrolled samples (2R = 2.1 ± 1.1 nm, 13.3 ± 4.5 nm, and 26.7 ± 9.0 nm).26
However, this insistence is unacceptable because of the experimental error due to the enormous
large size-distribution. Additionally, they assigned the absorption spectra of the anatase particles
using the band calculation for the rutile single crystal27.
For direct transition in the templated TiO2, four regression lines could be obtained in the
Tauc plots of (αhν)2 versus energy (Supplementary Figure 19; estimated error in energies < 0.01 eV,
coefficients of determination > 95 %). All the obtained direct bandgap energies (E*direct) the cross
points of these lines and the x-axis are larger than the indirect bandgap (E* in the main text),
suggesting that the Q-size TiO2 would retain the nature of the indirect bandgap semiconductor.
Additionally, the differences between E*direct and E* in the 6, 14, and 30 TiO2 are almost the same
in each group, which agrees with the difference in the bulk (Supplementary Table 2; error in
energies ≤ 0.02 eV). This means that the size dependency of E*direct parallelly shifts to the upper
field of E* and that the Q-size effect (∆E) also affects these direct transitions comparably;
Supplementary information for
Quantum Size Effect in TiO2 Nanoparticles Prepared by Finely Controlled Metal Assembly on Dendrimer Templates
S24
!
"E = E *#Eg = E *direct #Eg#direct =h2$ 2
2R2
1
% µ #1.8e
2
&%R (S3)
!
E *direct "E* = Eg"direct " Eg (S4)
Here, Eg-direct is the bandgap energy of each direct transition in the bulk.
Comparing the hydrolyzed and the thermolyzed samples, we can observe the different
manner in the direct transition from the differences between E*direct and E*. The different manner
indicates that the band structure of anatase and rutile would be different because of the different
crystal structure and space group. The band structure of rutile is observed using the single crystal
and calculated sufficiently to discuss the electronic transition.27, To the best of our knowledge, the
experimental data24, 25 and theoretical calculation28 for anatase are not sufficient to assign the
electronic transition. The calculated values in the rutile band structure seem to compare with the
hydrolyzed samples (for rutile) better than the thermolyzed ones (for anatase). These results also
support the belief that the Q-size TiO2 would retain the crystal structure and nature in the range of 1
nm < 2R < 2 nm. hν
According to our conjecture, the crystal structure of TiO2 nanoparticles, such as rutile and
anatase, would lose their nature when the particle size reaches less than 0.68 nm (discussed in the
main text). At this point, the indirect bandgap, E*(0.68) is estimated to be 4.91 eV. On the other
hand, the electronic transition band of the molecular unit, such as TiO4 tetrahedral and TiO6
octahedral coordinations, can be assigned in terms of ligand-to-metal charge transfer (LMCT)
transition: O2–Ti4+ + hν → O–Ti3+. The energy of the molecular LMCT transition, ELMCT, is
calculated from a phenomenological relationship to be 5.39–5.96 eV, corresponding to the TiO6 and
TiO4 units.29 Thus, the ELMCT is higher than the E*(0.68) by 0.48–1.05 eV. The lowest indirect
N. Satoh, T. Nakashima, K. Kamikura & K. Yamamoto*
* yamamoto@chem.keio.ac.jp
S25
transition at 0.68 nm, E*direct(0.68), however, would exist in the upper field of the E*(0.68) by ca.
0.30–0.46 eV, which almost reaches the ELMCT of the TiO6 units. If the size dependence of E*direct
progressed to the dimer size, E*direct could reach the ELMCT of the TiO4 units. These conjectures
could well explain the shift of the indirect transition to the direct one at the point that would lose the
crystal nature. Needless to say, the possibility for the transition shift to happen in the range of
0.68 nm < 2R < 1 nm may be considerable.
0
2
4
6
8
10
3.5 4 4.50
10
20
30
40
50
60
Energy (eV)
b
c
d
a
b
a
c
d
b
ab
c
a
d
c
bb
a
c
A B
0
1
2
3
4
5
6
0
10
20
30
40
50
60
70
80
0
5
10
15
20
25
30
3.5 4 4.5 5
Energy (eV)
d d
c
d
5
b
a
0
2
4
6
3.5 4 4.5Energy (eV)
b
a
0
0.5
1
b
a
a
Supplementary Figure 19 Tauc plots for direct bandgap of 6TiO2 (green), 14TiO2
(red), 30TiO2 (blue). A, hydrolyzed TiO2. B, thermolyzed TiO2. The direct bandgap is
categorized into four groups, a, b, c, d.
Supplementary Table 2 Direct bandgaps of 6TiO2, 14TiO2, 30TiO2, and bulk (and
the difference with the indirect bandgap)
experimental* experimental‡group 30TiO2 14TiO2 6TiO2 bulk bulk transition 30TiO2 14TiO2 6TiO2 bulka 3.78 (0.44) 4.00 (0.45) 4.26 (0.45) 3.3 (0.3) 3.45 (0.45) X1 X1 3.77 (0.31) 3.93 (0.29) 4.18 (0.30) b 3.94 (0.61) 4.15 (0.60) 4.42 (0.60) 3.4 (0.4) 3.59 (0.59) X2 X1 3.97 (0.51) 4.15 (0.51) 4.39 (0.51)c 4.38 (1.05) 4.61 (1.07) 4.92 (1.10) 4.07 (1.07) 4.05 (1.05) 5 1 4.08 (0.62) 4.28 (0.63) 4.50 (0.62)d 4.45 (1.12) 4.73 (1.12) 4.95 (1.13) 4.15 (1.15) 4.3 (1.3) 2 1 4.15 (0.69) 4.36 (0.71) 4.58 (0.70)
* Reference 23. † Reference 27. ‡ Summarized in reference 25.
AnataseRutile
3.40 (0.20),3.63–3.67 (0.43–0.47),
and 3.80 (0.60)
hydrolysis calculated† thermolysis
Supplementary information for
Quantum Size Effect in TiO2 Nanoparticles Prepared by Finely Controlled Metal Assembly on Dendrimer Templates
S26
Supplementary Materials
DPA G2-4 and phenylazomethine were prepared according to a previous method30. Ti(acac)Cl3
was synthesized following a literature method31. The acetonitrile-d3 and chloroform-d used in the
NMR measurements (ACROS, 99% D) were dehydrated using activated molecular sieves (Wako
Pure Chemical Industries, Ltd., 4A/16) for titration. All other chemicals were purchased from
Kantoh Kagaku Co. and used as received.
Supplementary General Methods
The NMR spectra were recorded using a JEOL JMN400 FT-NMR spectrometer (400 MHz) in
CDCl3/CD3CN (v/v = 1/1, with tetramethylsilane as the internal standard) solution. UV-vis
spectra were recorded using a Shimadzu UV-3100PC spectrometer with a closed quartz cell (optical
path length: 1cm). AFM was performed using a SII SPA400 instrument under ambient conditions
with the tapping mode of imaging (DFM). Si probes having a spring constant of 42 N/m (SII
SI-DF40P for DFM) were used at a resonance frequency of 300 kHz. A 20-µm scanner
(SPA400-PZT (FS-20A), 970P3202) was used. The TEM images were obtained at 120 kV with a
JEOL JEM-2010 instrument. The molecular modeling of the Ti-assembled DPA G4 was
performed using Fujitsu CAChe 5.0 Molecular Mechanics software with MM3 parameters
according to a previous method32. The XPS spectra were measured using a JEOL JPS-9000MC
photoelectron spectrometer without the plasma etching process. The binding energy calibration is
based on the measurement of the mica O(1s) signal, Eb = 531.61 eV33. The thermogravimetric
measurements were performed using a Rigaku TG8120 TG-DTA/MS system. The OWG spectra
were obtained using a System Instruments Co., Ltd., SIS-5000 spectrophotometer. The structures
of the TiO2 clusters were drawn using the CrystalMaker software CrystalMaker 7.
N. Satoh, T. Nakashima, K. Kamikura & K. Yamamoto*
* yamamoto@chem.keio.ac.jp
S27
Supplementary References
1. Kayanuma, Y. Quantum-size effects of interacting electrons and holes in semiconductor
microcrystals with spherical shape. Phys. Rev. B 38, 9797–9805 (1988).
2. Lippens, P. E. & Lannoo, M. Calculation of the band gap for small CdS and ZnS crystallites.
Phys. Rev. B, 39, 10935–10942 (1989).
3. Rama Krishna, M. V. & Friesner, R. A. Exciton spectra of semiconductor clusters. Phys. Rev.
Lett. 67, 629–632 (1991).
4. Nosaka, Y. Finite depth spherical well model for excited states of ultrasmall semiconductor
particles: an application. J. Phys. Chem. 95, 5054–5058 (1991).
5. Kormann, C., Bahnemann, D. W. & Hoffmann, M. R. Preparation and characterization of
quantum-size titanium dioxide. J. Phys. Chem. 92, 5196–5201 (1988).
6. Enright, B. & Fitzmaurice, D. Spectroscopic determination of electron and hole effective
masses in a nanocrystalline semiconductor film. J. Phys. Chem. 100, 1027-1035 (1996).
7. Kashiski, J. J., Gomez-Jahn, L. A., Faran, K. J., Gracewski, S. M. & Miller, R. J. D.
Picosecond dynamics of surface electron transfer processes: Surface restricted transient grating
studies of the n-TiO2/H2O interface. J. Chem. Phys. 90 1253–1269 (1989).
8. Wang L.-W. & Zunger, A. Pseudopotential calculations of nanoscale CdSe quantum dots.
Phys. Rev. B 53, 9579–9582 (1996).
9. Franceschetti A. & Zunger, A. Direct pseudopotential calculation of exciton coulomb and
exchange energies in semiconductor quantum dots. Phys. Rev. Lett. 78, 915–918 (1997).
10. Fu, H. & Zunger, A. InP quantum dots: Electronic structure, surface effects, and the
redshifted emission. Phys. Rev. B 56, 1496–1508 (1997).
11. Steigerwald, M. L. & Brus, L. E. Semiconductor crystallites: a class of large molecules. Ace.
Chem. Res. 23, 183-188 (1990).
12. Torimoto, T. et al. Characterization of ultrasmall CdS nanoparticles prepared by the
size-selective photoetching technique. J. Phys. Chem. B, 105, 6838–6845 (2001).
Supplementary information for
Quantum Size Effect in TiO2 Nanoparticles Prepared by Finely Controlled Metal Assembly on Dendrimer Templates
S28
13. Wolkin, M. V. et al. Electronic states and luminescence in porous silicon quantum dots: the
role of oxygen.
14. Garoufalis C. S., Zdetsis, A. D. & Grimme, S. High level ab initio calculations of the optical
gap of small silicon quantum dots. Phys. Rev. Lett. 87, 276402 (2001).
15. Zhou, Z., Brus, L. & Friesner, R. Electronic structure and luminescence of 1.1- and 1.4-nm
silicon nanocrystals: oxide shell versus hydrogen passivation. Nano Letters, 3, 163–167 (2003).
16. Zhou, Z., Friesner, R. A. & Brus, L. Electronic structure of 1 to 2 nm diameter silicon
core/shell nanocrystals: surface chemistry, optical spectra, charge transfer, and doping. J. Am.
Chem. Soc. 125, 15599–15607 (2003).
17. Marchese, L. et al. Structure–functionality relationships of grafted Ti-MCM41 silicas.
Spectroscopic and catalytic studies. Phys. Chem. Chem. Phys., 1, 585–592 (1999).
18. Hu, Y. et al. Characterization of the local structures of Ti-MCM-41 and their photocatalytic
reactivity for the decomposition of NO into N2 and O2. J. Phys. Chem. B, 110, 1680–1685 (2006).
19. Lundqvist, M. J., Nilsing, M., Persson, P. & Lunell, S. DFT study of bare and dye-sensitized
TiO2 clusters and nanocrystals. Int. J. Quantum. Chem. 106, 3214–3234 (2006).
20. Zhai, H.-J. & Wang, L.-S. Probing the electronic structure and band gap evolution of titanium
oxide clusters (TiO2)n– (n = 1-10) using photoelectron spectroscopy. J. Am. Chem. Soc. 129,
3022–3026 (2007).
21. Zhanpeisov, N. U. & Anpo, M. Theoretical ab initio study of the intrinsic band gap in
semiconductor oxides based on modified titanium dioxide. Theor. Chem. Acc. 114, 235–241
(2005).
22. Zimmerman, A. M., Doren, D. J. & Lobo, R. F. Electronic and Geometric Properties of
ETS-10: QM/MM Studies of Cluster Models. J. Phys. Chem. B, 110, 8959–8964 (2006).
23. Frova, A., Boddy, P. J. & Chen, Y. S. Electromodulation of the optical constants of rutile in
the uv. Phys. Rev. 157, 700–708 (1967).
24. Jiménez González A. E. & Gelover Santiago, S. Structural and optoelectronic characterization
of TiO2 films prepared using the sol–gel technique. Semicond. Sci. Technol. 22 709–716 (2007).
N. Satoh, T. Nakashima, K. Kamikura & K. Yamamoto*
* yamamoto@chem.keio.ac.jp
S29
25. Gao, Y. et al. Room temperature deposition of a TiO2 thin film from aqueous peroxotitanate
solution. J. Mater. Chem. 13, 608–613 (2003).
26. Serpone, N., Lawless, D. & Khairutdinov, R. Size effects on the photophysical properties of
colloidal anatase TiO2 particles: Size quantization versus direct transitions in this indirect
semiconductor? J. Phys. Chem. 99, 16646–16654 (1995).
27. Daude, N., Gout, C. & Jouanin, C. Electronic band structure of titanium dioxide. Phys. Rev.
B 15, 3229–3235 (1977).
28. Mikami, M., Nakamura, S., Kitao O. & Arakawa, H. Lattice dynamics and dielectric
properties of TiO2 anatase: A first-principles study. Phys. Rev. B, 66, 155213 (2002)
29. Borello, E. et al. Quantum-size effects in the titanosilicate molecular sieve. Appl. Phys. Lett.
71, 2319–2321 (1997).
30. Takanashi, K., Chiba, H., Higuchi, M. & Yamamoto K. Efficient synthesis of
poly(phenylazomethine) dendrons allowing access to higher generation dendrimers. Org. Lett., 6,
1709-1712 (2004).
31. Serpone, N., Bird, P. H., Somogyvari, A. & Bickley, D. G. Five-coordinate titanium(IV)
complexes. Infrared spectral studies on X3Ti(diketonato) and XY2Ti(diketonato) complexes and
the crystal and molecular structure of Di-µ-chloro-tetrachlorobis(2,4-pentanedionato)dititanium(IV).
Inorg. Chem. 16, 2381-2386 (1977).
32. Higuchi, M., Tsuruta, M., Chiba, H., Shiki, S. & Yamamoto, K. Control of stepwise radial
complexation in dendritic polyphenylazomethines. J. Am. Chem. Soc. 125, 9988-9997 (2003)
33. Liu, Z. H. & Brown N. M. D. XPS characterization of mica surfaces processed using a
radio-frequency (rf) argon plasma. J. Phys. D; Appl. Phys. 31 1771–1781 (1998).
Recommended