New insights into ETS-10 and titanate quantum wire: A comprehensive characterization

Article abstract of DOI:10.1021/ja903638v

The titanate quantum wires in ETS-10 crystals remain intact during ion exchange of the pristine cations (Na⁺_0.47 + K ⁺_0.53) with Mⁿ⁺ ions (Mn⁺) Na ⁺, K⁺, Mg²⁺

Full text of DOI:10.1021/ja903638v

Published on Web 08/20/2009
New Insights into ETS-10 and Titanate Quantum Wire: A
Comprehensive Characterization
Nak Cheon Jeong,† Young Ju Lee,† Jung-Hyun Park,‡ Hyunjin Lim,§
Chae-Ho Shin,‡ Hyeonsik Cheong,§ and Kyung Byung Yoon*,†
Center for Nano Materials, Departments of Chemistry and Physics, Program of Integrated
Biotechnology, Sogang UniVersity, Seoul 121-742, Korea, and Department of Chemical
Engineering, Chungbuk National UniVersity, Cheongju, Chungbuk 361-763, Korea
Received May 11, 2009; E-mail: yoonkb@sogang.ac.kr
Abstract: The titanate quantum wires in ETS-10 crystals remain intact during ion exchange of the pristine
cations (Na+
+ K+0.53) with Mn+ ions (Mn+ ) Na+, K+, Mg2+, Ca2+, Sr2+, Ba2+, Pb2+, Cd2+, Zn2+) and
0.47
during reverse exchange of the newly exchanged cations with Na+. The binding energies of O(1s) and
Ti(2p) decrease as the electronegativity of the cation decreases, and they are inversely proportional to the
negative partial charge of the framework oxygen [-δ(Of)]. At least five different oxygen species were
identified, and their binding energies (526.1-531.9 eV) indicate that the titanate-forming oxides are much
more basic than those of aluminosilicate zeolites (530.2-533.3 eV), which explains the vulnerability of the
quantum wire to acids and oxidants. The chemical shifts of the five NMR-spectroscopically nonequivalent
Si sites, δ(IA), δ(IB), δ(IIA), δ(IIB), and δ(III), shift downfield as -δ(Of) increases, with slopes of 2.5, 18.6,
133.5, 216.3, and 93.8 ppm/[-δ(Of)], respectively. The nonuniform responses of the chemical shifts to -δ(Of)
arise from the phenomenon that the cations in the 12-membered-ring channels shift to the interiors of the cages
surrounded by four seven-membered-ring windows. On the basis of the above, we assign δ(IA), δ(IB), δ(IIA),
and δ(IIB) to the chemical shifts arising from Si(12,12), Si(12,7), Si(7,12), and Si(7,7) atoms, respectively. The
frequency of the longitudinal stretching vibration of the titanate quantum wire increases linearly and the bandwidth
decreases nonlinearly with increasing -δ(Of), indicating that the titanate quantum wire resembles a metallic
carbon nanotube. As the degree of hydration increases, the vibrational frequency shifts linearly to higher
frequencies while the bandwidth decreases. We identified another normal mode of vibration of the quantum
wire, which vibrates in the region of 274-280 cm-1. In the dehydrated state, the band-gap energy and the first
absorption maximum shift to lower energies as -δ(Of) increases, indicating the oxide-to-titanium(IV) charge-
transfer nature of the transitions.
removal of harmful heavy-metal ions7-9 and photocatalytic
removal of organic pollutants10 and has also been tested as a
Introduction
ETS-10 is a highly intriguing nanoporous titanosilicate1 that
has shown the potential to work as an effective catalyst for
various reactions, such as dehydration of tert-butyl alcohol to
isobutene,2 aldol condensation of acetone,3 oxidative dehydro-
genation of cyclohexanol,4 production of propylene carbonate
from carbon dioxide and propylene oxide,5 and aromatization
of n-hexane.6 It is also an effective material for the selective
material for solar cells.11 However, the most interesting feature
that makes ETS-10 so unique is that it contains periodically
positioned quantum wires [-O-Ti(O-)4-O-] with a diameter
(d) of ∼0.67 nm running along the two perpendicular directions
in the crystal (Chart 1A).1,12-17
(6) Waghmode, S. B.; Vetrivel, R.; Gohinath, C. S.; Sivasanker, S. J.
Phys. Chem. B 2004, 108, 11541–11548.
(7) (a) Lv, L.; Tsoi, G.; Zhao, X. S. Ind. Eng. Chem. Res. 2004, 43, 7900–
7906. (b) Zhao, X. S.; Lee, J. L.; Chia, P. A. Langmuir 2003, 19,
1977–1979. (c) Lv, L.; Wang, K.; Zhao, X. S. J. Colloid Interface
Sci. 2007, 305, 218–225.
† Department of Chemistry, Sogang University.
‡ Chungbuk National University.
§ Department of Physics, Sogang University.
(1) (a) Kuznicki, S. M. U.S. Patent 4,853,202, 1989. (b) Anderson, M. W.;
Terasaki, O.; Ohsuna, T.; Philippou, A.; MacKay, S. P.; Ferreira, A.;
Rocha, J.; Lidin, S. Nature 1994, 367, 347–351. (c) Anderson, M. W.;
Terasaki, O.; Ohsuna, T.; Malley, P. J. O.; Philippou, A.; MacKay,
S. P.; Ferreira, A.; Rocha, J.; Lidin, S. Philos. Mag. B 1995, 71, 813–
841.
(8) (a) Choi, J. H.; Kim, S. D.; Noh, S. H.; Oh, S. J.; Kim, W. J.
Microporous Mesoporous Mater. 2006, 87, 163–169. (b) Choi, J. H.;
Kim, S. D.; Kwon, Y. J.; Kim, W. J. Microporous Mesoporous Mater.
2006, 96, 157–167.
(9) Lopes, C. B.; Otero, M.; Coimbra, J.; Pereira, E.; Rocha, J.; Lin, Z.;
Duarte, A. Microporous Mesoporous Mater. 2007, 103, 325–332.
(10) Luca, V.; Osborne, M.; Sizgek, D.; Griffith, C.; Araujo, P. Z. Chem.
Mater. 2006, 18, 6132–6138.
(2) Philippou, A.; Naderi, M.; Rocha, J.; Anderson, M. W. Catal. Lett.
1998, 53, 221–224.
(3) Philippou, A.; Anderson, M. W. J. Catal. 2000, 189, 395–400.
(4) Valente, A.; Lin, Z.; Branda˜o, P.; Portugal, I.; Anderson, M. W.; Rocha,
J. J. Catal. 2001, 200, 99–105.
(11) At´ıenzar, P.; Valencia, S.; Corma, A.; García, H. ChemPhysChem 2007,
8, 1115–1119.
(12) Jeong, N. C.; Lee, M. H.; Yoon, K. B. Angew. Chem., Int. Ed. 2007,
46, 5868–5872.
(5) Doskocil, E. J. J. Phys. Chem. B 2005, 109, 2315–2320.
9
13080 J. AM. CHEM. SOC. 2009, 131, 13080–13092
10.1021/ja903638v CCC: $40.75  2009 American Chemical Society

New Insights into ETS-10 and Titanate Quantum Wire
A R T I C L E S
2-
Chart 1. (A) Illustration of 3D Networks of Titanate (TiO3
Quantum Wires (red) and SiO2 Channels (gray) in a
)
Scheme 1
Truncated-Bipyramidal ETS-10 crystal with Polymorph B and (B,
C) Single Titanate Quantum Wires Charge-Balanced with (B)
Hydrated and (C) Naked Cations
balancing cations to compensate for the negative charges of the
titanate quantum wires. In the related aluminosilicate zeolites,
the (negative) charge density of the framework23 has often been
represented by the binding energy of an O 1s electron,24 despite
the fact that it is a core electron. However, the Sanderson partial
(negative) charge of the framework oxygen, -δ(Of), has often
been used instead of the O 1s binding energy because the
binding energy of an O 1s electron cannot be routinely obtained
and it has been established that the two values [the O 1s binding
energy and -δ(Of)] are linearly correlated.24,25
The negative charge density of the framework can be
conveniently tuned by varying both the type of charge-balancing
cation and the degree of dehydration of the zeolite because the
framework oxygen atoms interact with the cations through a
ligand-to-metal charge-transfer (LMCT) interaction (Scheme 1)
and the degree of the LMCT interaction increases as the degree
of dehydration increases.25,26
Thus, in a dry zeolite, the framework charge density decreases
as the acceptor strength or the electronegativity of the charge-
balancing cation increases and also as the degree of dehydration
increases, since dehydration of the zeolite leads to the removal
of water shells from the cations, thereby increasing the degree
of direct LMCT from the framework to the cation in the ground
state.
Among the known quantum wires,18,19 the TiO3 quantum
2-
wire is an anionic counterpart of anatase nanowire18 and can
be regarded as a one-dimensional (1D) extreme of three-
dimensional (3D) bulk titanates, which are widely used in
industry as capacitors, etc.20 Therefore, acquiring the properties
of the TiO3 quantum wire also means acquiring the 1D-
extreme properties of titanates. Each TiO3 quantum wire is
surrounded by insulating nanoporous silica with a pore size of
8 Å × 5 Å, and the wires are supramolecularly organized within
the crystalline solid from its birth. It exhibits the quantum
confinement effect along the chain direction even at length scales
longer than 50 nm.12 The estimated reduced mass of an exciton
along its chain direction (µz, 0.0006me)12 is much smaller than
those of InSb (0.014me)21 and single-walled carbon nanotubes
(SW-CNTs, 0.019me).22 Such an unusually small µz further
2-
2-
Normally, a mixture of Na+ and K+ exists within the pristine
ETS-10 crystals for charge neutralization. It has been believed
that the charge-balancing cations are evenly distributed along
the quantum wire in the vicinity of each titanium center (Chart
1B).27 They can be exchanged with other cations through ion
exchange in aqueous solution.7,9,15,28-30 In the hydrated state,
the cations are also expected to be at least partially surrounded
by water molecules, as illustrated in Chart 1B. In the dehydrated
state, however, they should be directly coordinated by the
framework oxygen atoms, as illustrated in Chart 1C. Accord-
2-
suggests the possibility that the carrier mobility along the TiO3
molecular wire, despite the wire’s oxide nature, can be much
greater than that in single-walled carbon nanotubes, indicating
the possibility that the quantum wire is metallic.
The unit cell composition of ETS-10 is Mn+(2/n)Ti1Si5O13.
Thus, like aluminosilicate zeolites, ETS-10 also carries charge-
(13) (a) Borello, E.; Lamberti, C.; Bordiga, S.; Zecchina, A.; Area´n, C. O.
Appl. Phys. Lett. 1997, 71, 2319–2321. (b) Llabre´s i Xamena, F. X.;
Damin, A.; Bordiga, S.; Zecchina, A. Chem. Commun. 2003, 1514–
1515. (c) Damin, A.; Llabre´s i Xamena, F. X.; Lamberti, C.; Civalleri,
B.; Zicovich-Wilson, C. M.; Zecchina, A. J. Phys. Chem. B 2004,
108, 1328–1336. (d) Bordiga, S.; Palomino, G. T.; Zecchina, A.;
Ranghino, G.; Giamello, E.; Lamberti, C. J. Chem. Phys. 2000, 112,
3859–3867.
(23) It is more often called the framework electron donor strength or
framework basicity. These terms eventually represent the richness of
the electron density in the framework. Among these, the terms electron
donor strength and framework basicity should rather be used when
one deals with the interaction between the framework and its
counterparts such as electron acceptors and acids. Therefore, in this
work, we use negative framework charge density to represent the
richness of electron density.
(14) Zimmerman, A. M.; Doren, D. J.; Lobo, R. F. J. Phys. Chem. B 2006,
110, 8959–8964.
(24) Huang, H.; Adnot, A.; Kaliaguine, S. J. Am. Chem. Soc. 1992, 114,
10005–10010.
(15) Lamberti, C. Microporous Mesoporous Mater. 1999, 30, 155–163.
(16) Wang, X.; Jacobson, A. J. Chem. Commun. 1999, 973–974.
(17) Yilmaz, B.; Warzywoda, J.; Sacco, A., Jr. Nanotechnology 2006, 17,
4092–4099.
(25) (a) Mortier, W. J.; Schoonheydt, R. A. Prog. Solid State Chem. 1985,
16, 1–125. (b) Heidler, R.; Janssens, G. O. A.; Mortier, W. J.;
Schoonheydt, R. A. J. Phys. Chem. 1996, 100, 19728–19734. (c)
Genechten, K. A. V.; Mortier, W. J. Zeolites 1988, 8, 273–283.
(26) (a) Yoon, K. B. In Handbook of Zeolite Science and Technology;
Auerbach, S. M., Carrado, K. A., Dutta, P. K., Eds.; Marcel Dekker:
New York, 2003; Chapter 13. (b) Park, Y. S.; Um, S. Y.; Yoon, K. B.
J. Am. Chem. Soc. 1999, 121, 3193–3200. (c) Choi, S. Y.; Park, Y. S.;
Hong, S. B.; Yoon, K. B. J. Am. Chem. Soc. 1996, 118, 9377–9386.
(27) (a) Anderson, M. W.; Agger, J. R.; Luigi, D.-P.; Baggaley, A. K.;
Rocha, J. Phys. Chem. Chem. Phys. 1999, 1, 2287–2282. (b) Grillo,
M. E.; Carrazza, J. J. Phys. Chem. 1996, 100, 12261–12264.
(28) Krisnandi, Y. K.; Lachowski, E. E.; Howe, R. F. Chem. Mater. 2006,
18, 928–933.
(18) (a) Liu, C.; Yang, S. ACS Nano 2009, 3, 1025–1031. (b) Venkataraman,
L.; Lieber, C. M. Phys. ReV. Lett. 1999, 83, 5334–5337. (c) Huang,
Y.; Duan, X.; Cui, Y.; Lauhon, L. J.; Kim, K.-H.; Lieber, C. M. Science
2001, 294, 1313–1317. (d) Duan, X.; Huang, Y.; Cui, Y.; Wang, J.;
Lieber, C. M. Nature 2001, 409, 66–69.
(19) (a) Xia, Y.; Yang, P.; Sun, Y.; Wu, Y.; Mayers, B.; Gates, B.; Yin,
Y.; Kim, F.; Yan, H. AdV. Mater. 2003, 15, 353–389. (b) Golden,
J. H.; DiSalvo, F. J.; Fre´chet, J. M. J.; Silcox, J.; Thomas, M.; Elman,
J. Science 1996, 273, 782–784. (c) Yu, H.; Li, J.; Loomis, R. A.; Wang,
L.-W.; Buhro, W. E. Nat. Mater. 2003, 2, 517–520.
(20) Bavykin, D. V.; Friedrich, J. M.; Walsh, F. C. AdV. Mater. 2006, 18,
2807–2824.
(29) Waghmode, S. B.; Vetrivel, R.; Hegde, S. G.; Gopinath, C. S.;
Sivasanker, S. J. Phys. Chem. B 2003, 107, 8517–8523.
(21) Burs, L. E. J. Chem. Phys. 1984, 80, 4403–4409.
(22) Pedersen, T. G. Phys. ReV. B 2003, 67, 073401.
(30) Pavel, C. C.; Zibrowius, B.; Lo¨ffler, E.; Schmidt, W. Phys. Chem.
Chem. Phys. 2007, 9, 3440–3446.
9
J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009 13081

A R T I C L E S
Jeong et al.
this, 0.5 g of Mn+-ETS-10 was introduced into the saturated NaCl
solution (45 mL), and the solution was shaken for 1 day at room
temperature. The reverse ion exchange was repeated two more
times. Despite this, the degree of reverse exchange was not
complete, as summarized in Table SI 2 in the Supporting Informa-
tion.
ingly, the degree of intimate LMCT interaction in Scheme 1
increases as the degree of dehydration of the zeolite increases,
and it reaches a maximum in the fully dehydrated state.
With the above background in mind, we have addressed the
following questions: Would the electronegativity of the charge-
balancing cation and the degree of hydration of the molecular
sieve also affect the charge density of the TiO32-quantum wires
and the silicon atoms that surround them? If yes, how would
the charge density on the titanate quantum wires affect their
band-gap energy (Eg), vibrational frequency, and vibrational
bandwidth? Since ion exchange of Na+ and K+ of the pristine
Dehydration of Mn+-ETS-10 Samples. Since the M-ETS-10
samples were not stable above 250 °C (Figure SI 1 in the Supporting
Information), they were dried at 250 °C for 4 h under vacuum
(<10-6 Torr). For the measurements of magic-angle-spinning (MAS)
solid-state 29Si NMR spectra, the dried Mn+-ETS-10 samples were
treated with octadecytrichlorosilane in a glovebox according to the
procedure described in our previous report.33
ETS-10 with NH4 and H+ readily decomposes the titanate
+
Preparation of Partially Hydrated Mn+-ETS-10. Dry Mn+
-
quantum wire28,31 while ion exchange of the cations with Cs+
can be done reversibly without decomposing it,30 are there other
cations that could be reversibly exchanged into ETS-10 without
decomposing the quantum wires?
ETS-10 powder (∼200 mg) was spread onto a light, shallow, plastic
container (40 mm × 40 mm) placed on top of a microbalance, and
the weight was quickly measured after a desired period of time.
After the weight measurement, the powder was then quickly
transferred into a flat, round, fused silica cell (i.d. ) 19 mm), and
Raman and UV-vis spectra were measured immediately. After the
spectral measurements, the powder was spread back onto the
weighing container and allowed to pick up more moisture from
the atmosphere. After a desired period of exposure to the atmo-
sphere, the weight increase was measured using a microbalance,
and the powder was subsequently transferred into a fused silica
cell for the measurements of Raman and UV-vis spectra.
3D Bulk Titanates. Li2TiO3, BaTiO3, SrTiO3, and PbTiO3 were
purchased from Aldrich. CaTiO3 was prepared by thoroughly
grinding a 1:1 mixture of CaCO3 (1 g, Aldrich) and TiO2 (0.8 g,
Aldrich) at room temperature and subsequently heating at 1000 °C
for 16 h.
We now report the ion exchange behavior of ETS-10, the
effect of the electronegativity and degree of dehydration of the
cation on the charge densities of the Ti, O, and Si atoms, and
the effect of negative charge density in the titanate quantum
wire on its Eg, vibrational frequency, and vibrational bandwidth.
We also report the XPS spectrum of Ti, the detailed assignment
of 29Si NMR peaks to the silicon sites, and the valuable
information that the cations in the 12-memberd-ring (12MR)
channels move to the cages surrounded by four seven-membered-
ring (7MR) windows in the dehydrated state and that the
framework oxides in ETS-10 are much more basic than those
in zeolites.
Instrumentation. Scanning electron microscopy (SEM) images
were obtained using a field-effect scanning electron microscope
(Hitachi S-4300) operating at an acceleration voltage of 20 kV.
Elemental analyses of various Mn+-ETS-10 samples for Mn+, Si,
and Ti were carried out by analyzing the energy-dispersive X-ray
(EDX) spectra of the samples using a Horiba EMAX 6853-H EDX
spectrometer. Comparison of some of our elemental analysis results
with those obtained from chemical analysis showed good matching,
supporting the accuracy of the EDX analyses in this report. Powder
X-ray diffraction (XRD) patterns were obtained using a Rigaku
D/MAX-2500/pc diffractometer. Raman spectra of the samples were
recorded on a homemade setup equipped with an Ar+ ion laser
(Spectra-Physics Stabilite 2017) as an excitation beam source, a
spectrometer (Horiba Jobin Yvon TRIAX 550), and a CCD detector
(Horiba Jobin Yvon Symphony) cooled to -133 °C. The wave-
length of the excitation beam was 514.5 nm. Diffuse-reflectance
UV-vis spectra of the samples were recorded on a Varian Cary
5000 UV-vis-NIR spectrophotometer equipped with an integrating
sphere. Barium sulfate was used as the reference. The diffuse-
reflectance spectra were converted into the Kubelka-Munk (K-M)
formalism. X-ray photoelectron spectroscopy (XPS) of Mn+-ETS-
10 samples (for Mn+ ) Na+, K+, Ca2+, Sr2+, Ba2+) were recorded
on a VG Scientifics ESCALAB 250 XPS spectrometer at the Korea
Basic Science Institute located at Pusan National University. The
Brunauer-Emmett-Teller (BET) surface area, micropore volume,
and total pore volume were obtained from argon adsorption at 87.3
K using a volumetric method with a Micromeritics ASAP 2020
analyzer. Micropore volume was determined using the t-plot
method. Prior to adsorption of argon, the samples were evacuated
under a vacuum of 3 × 10-5 Torr at 250 °C. The 29Si MAS NMR
spectra were measured on a Bruker Avance 500 spectrometer
operating at a spinning rate of 10 kHz and a 29Si frequency of 99.351
Experimental Section
Synthesis of ETS-10. ETS-10 crystals with the sizes of ∼500
nm were prepared according to the previously reported procedure.32
The composition of pristine ETS-10 is Na0.94K1.06TiSi5O13. Thus,
the cations in pristine ETS-10 are a mixture of Na+ and K+ in the
ratio of 0.94:1.06. Hereafter, the pristine cations are represented
by the expression Na+ + K+
.
0.47
0.53
Preparation of Mn+-Exchanged ETS-10 Crystals (Mn+-ETS-
10, Mn+ ) Na+, K+, Mg2+, Ca2+, Sr2+, Ba2+, Pb2+, Cd2+
,
Zn2+). ETS-10 samples exchanged with ten different metal ions
(Mn+-ETS-10) were prepared by exchanging Na+ and K+ cations
in pristine ETS-10 with the desired metal ion using the correspond-
ing salt [NaCl, KCl, MgCl2 CaCl2, SrCl2, BaCl2, Pb(CH3CO2)2,
Cd(CH3CO2)2, or Zn(CH3CO2)2]. The ion exchange was carried out
by introducing 1 g of pristine ETS-10 powder into 45 mL of a 2 M
aqueous solution of the salt at room temperature and then stirring
for 1 h at room temperature. The ion exchange was repeated two
more times to achieve the highest possible degree of ion exchange.
Although the ion exchange did not go to completion, we denote
the ion-exchanged ETS-10 as Mn+-ETS-10, where Mn+ is the major
cation (Na+, Na+ + K+, K+, Mg2+, Ca2+, Sr2+, Ba2+, Pb2+, Cd2+
,
or Zn2+). The compositions of cations in each Mn+-ETS-10 are listed
in Table SI 1 in the Supporting Information. For the ion exchange
of Na+ with K+ ion in pristine ETS-10, the K+ content in ETS-10
was first decreased to ∼20%, after which ion exchange of the 80%
Na+ ion with K+ ion was carried out. Likewise, the Na+ content
was first decreased to 20% before the ion exchange of Na+ with
K+ was carried out.
Preparation of Na+-Reverse-Exchanged ETS-10. To check the
stability of the titanate quantum wires during and after ion exchange,
we reverse-exchanged each Mn+ in Mn+-ETS-10 with Na+ ion. For
π
kHz using a /2 rad pulse length of 8 µs. A long recycle of 30 s
was used to avoid relaxation effects in the signal intensities, and
(31) (a) Lv, L.; Lee, F. Y.; Zhou, J.; Su, F.; Zhao, X. S. Microporous
Mesoporous Mater. 2006, 96, 270–275. (b) Lv, L.; Zhou, J. K.; Su,
F.; Zhao, X. S. J. Phys. Chem. C 2007, 111, 773–778.
(32) Jeong, N. C.; Lee, Y. J.; Yoon, K. B. Microporous Mesoporous Mater.
2008, 115, 308–313.
(33) (a) Jeong, N. C.; Kim, H. S.; Yoon, K. B. Langmuir 2005, 21, 6038–
6047. (b) Jeong, N. C.; Kim, H. S.; Yoon, K. B. J. Phys. Chem. C
2007, 111, 10298–10312.
9
13082 J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009

New Insights into ETS-10 and Titanate Quantum Wire
A R T I C L E S
2500 pulse transients were typically accumulated. The 29Si chemical
shifts are referenced to tetramethylsilane (TMS).
Results and Discussion
Quality of the ETS-10 Crystals Used in This Study. ETS-10
has often been produced with seriously damaged (short)
TiO32-quantum wires.17,34 Therefore, even if the XRD pattern
of a newly prepared sample matches the standard diffraction
pattern of ETS-10, this does not guarantee that high-quality
TiO32-molecular wires (i.e., with long lengths and no defects)
were produced. The presence and quality (length) of the
TiO32-quantum wires in ETS-10 have often been analyzed by
the longitudinal (along-the-wire) stretching vibration band of
the TiO32-molecular wire and its bandwidth in the Raman
spectrum.13b,17,28,30,31,34,35 In the hydrated pristine ETS-10 (with
Na+0.47 + K+0.53 as the countercations), when the lengths of the
TiO32- quantum wires are longer than ∼50 nm and the quality
of the titanate quantum wires is good, the longitudinal stretching
vibration band remains constant at ∼724 cm-1 with a bandwidth
Figure 1. Plots of (A) the degrees of ion exchange of the pristine cations
(Na+0.47 + K+0.53) with Mn+ [or (Na+0.8 + K+0.2) for the ion exchange with
K+ or (Na+0.2 + K+0.8) for the ion exchange with Na+] and (B) the degrees
of reverse ion exchange of Mn+ with Na+ with respect to the number of
ion exchange cycles. The Mn+ are listed in (A).
(fwhm) of ∼26 cm-1
.
12 As the quality of the ETS-10 decreases
2-
(i.e., as the lengths of TiO3 quantum wires become shorter
than ∼50 nm), the longitudinal stretching vibrational band is
blue-shifted and the bandwidth (fwhm) increases.13b,17,28,30,31,34,35
In this respect, the observation that the frequency and bandwidth
of our pristine ETS-10 were 724 and 24 cm-1, respectively,
indicates that the titanate quantum wires in the ETS-10 crystals
used in this study were produced with lengths of >50 nm with
few defects.
showed that between the two pristine cations, Na+ and K+, the
former (Na+) is more readily exchanged by other cations than
the latter (K+).
Reverse Ion Exchange and Special Affinity for Pb2+. Reverse
ion exchange of the newly introduced Mn+ with Na+ was also
carried out three times for each Mn+-ETS-10. The degrees of
reverse ion exchange (Figure 1B and Table SI 2 in the
Supporting Information) were higher than 77% in all cases
except Mn+ ) Pb2+, for which the degree of reverse ion
exchange was only 8%. This indicates that the ion exchange of
the pristine metal ions (Na+ and K+) with Pb2+ is almost
irreversible. The exchange of Pb2+ with the most preferred
cation, Zn2+, also did not occur readily. Thus, even after six
ion exchange cycles with Zn2+, the degree of ion exchange of
Pb2+ with Zn2+ was only 20%. Ion exchange of the pristine
cations with Pb2+ does not cause structural collapse (see the
next section). Since Zn2+ is more preferred than Pb2+, the
difficulty of exchanging Pb2+ with not only Na+ but also Zn2+
indicates that the back-exchange of Pb2+ with other cations is
kinetically hindered.
Ion Exchange of Pristine ETS-10 with Several Metal Ions
(Mn+ ) Na+, K+, Mg+, Ca2+, Sr2+, Ba2+, Pb2+, Cd2+, Zn2+
)
and Reverse Exchange of Mn+ with Na+. The Sanderson
electronegativities of the cations are as follows: K+ ) 0.445,
Na+ ) 0.560, Ba2+ ) 0.651, Sr2+ ) 0.721, Ca2+ ) 0.950, Mg2+
) 1.318, Pb2+ ) 1.920, Cd2+ ) 1.970, and Zn2+ ) 2.223.36
Thus, the Sanderson elctronegativity increases in the order K+
< Na+ < Ba2+ < Sr2+ < Ca2+ < Mg2+ < Pb2+ < Cd2+ < Zn2+
.
Nondestructive Forward Ion Exchange and High Affinity
for Pb2+, Cd2+, and Zn2+. The ion exchange seemed to reach
equilibrium within 1 h at room temperature. The degree of ion
exchange after the first cycle increased in the order Ba2+ (49)
< Sr2+ (54) < [Ca2+ (63) ≈ Na+ (64) ≈ K+ (65) ≈ Mg2+ (67)]
< [Zn2+ (76) ≈ Cd2+ (77) ≈ Pb2+ (79)] (Figure 1A and Table
SI 1 in the Supporting Information). Consistent with the
literature reports,7,8a Pb2+ was the most preferred cation after
the first cycle. In general, ETS-10 showed higher preferences
for Zn2+, Cd2+, and Pb2+ than for alkali and alkaline-earth metal
ions. The degrees of ion exchange after the third cycle ranged
between 80 and 94%, with specific values of 80% (Na+, K+,
Among the cations studied in this work, the alkaline-earth
metal ions (Ba2+, Sr2+, Ca2+, and Mg2+) were most readily
replaced by Na+. This fact is consistent with the phenomenon
that these cations were most reluctantly exchanged into ETS-
10, indicating that ETS-10 has a relatively low affinity for the
alkaline-earth metal ions, particularly the large cations Sr2+ and
Ba2+. The reverse-ion-exchanged ETS-10 samples are denoted
as (Mn+fNa+)ETS-10.
Mg2+), 85-88% (Sr2+, Ba2+, Ca2+), and 90-94% (Pb2+, Cd2+
,
Zn2+). Thus, as noted, the ion exchange did not go to completion
(100%) even after the ion exchange procedure was repeated
three times, despite the fact that each cycle was carried out in
a freshly prepared, highly concentrated (2 M) solution of the
salt. The degrees of ion exchange for Pb2+, Cd2+, and Zn2+ were
90, 93, and 94%, respectively. Although the differences were
small, the order of preference increased as Pb2+ < Cd2+ < Zn2+,
which coincides with the order of the increase in the Sanderson
electronegativities. Analysis of the remaining alkali metal cations
(Table SI 1 in the Supporting Information, fourth column)
Powder XRD Patterns of Mn+-ETS-10 and (Mn+fNa+)
ETS-10. The powder XRD patterns of Mn+-ETS-10 were similar
to that of the pristine ETS-10 in the hydrated state for all of the
Mn+ ions except Pb2+ and Cd2+ (Figure 2A), which showed
minor variations. Thus, when Mn+ ) Pb2+, the following took
place: the intensity of the lowest-angle peak (2θ ) 5.98°)
increased by a factor of ∼2; a new peak at 2θ ) 15.23° appeared
in significant intensity; the intensity of the peak at 2θ ) 17.86°
increased significantly; the peak at 20.05° disappeared; the peaks
at 23.69, 25.73, and 27.09° shifted to slightly lower angles
(23.64, 25.66, and 27.01°, respectively) with accompanying
intensity increases; and a new peak appeared at 26.69°. Similar
changes were observed when Mn+ ) Cd2+, except that even
(34) Southon, P. D.; Howe, R. F. Chem. Mater. 2002, 14, 4209–4218.
(35) Pavel, C. C.; Park, S.-H.; Dreier, A.; Tesche, B.; Schmidt, W. Chem.
Mater. 2006, 18, 3813–3820.
(36) Sanderson, R. T. J. Am. Chem. Soc. 1983, 105, 2259–2261.
9
J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009 13083

A R T I C L E S
Jeong et al.
Figure 2. Powder XRD patterns of Mn+-ETS-10 in the (A) hydrated and
(B) dehydrated states and of (C) (Mn+fNa+)ETS-10 in the hydrated state.
the two small peaks at 12.20 and 12.92° disappeared, but the
peak at 20.09° did not disappear and the peak at 25.76° did not
shift to the lower-angle region. We ascribe the above changes
to the presence of Pb2+ and Cd2+ ions on the regular positions
in ETS-10 even in the hydrated states. Upon dehydration, when
Mn+ ) Pb2+, the intensity of the lowest-angle peak at 2θ )
6.1° increased even further, and the peak at 12.20° disappeared;
also, the intensities of the peaks at 15.23 and 17.86° decreased
a bit (Figure 2B). However, when Mn+ ) Cd2+, no further
changes were observed.
Figure 3. Ar adsorption isotherms of Mn+-ETS-10 for p/p0 in the range
0.0-1.0.
argon BET isotherms of Mn+-ETS-10 samples in the region of
relative pressure (p/po) between 10-6 and 1.0 (Figure 3) to check
whether mesopores arising from structural damage are produced
after ion exchange of the pristine cations in ETS-10 with various
metal ions under the given exchange conditions (also see Figure
SI 2 in the Supporting Information). In fact, this is the first report
to systematically compare the argon BET isotherms of ETS-10
exchanged with various metal ions. The isotherms showed no
sudden increases in adsorption volume and no hysteresis
between adsorption and desorption for p/po ) 0.1-0.8, indicat-
ing that the ion exchange does not cause the formation of
mesopores. This result further confirms that the structure of ETS-
10 is not damaged even partially either during or after ion
exchange of the pristine cations with the above cations or even
after dehydration.
Restoration of the Original Diffraction Patterns after
Reverse Exchange with Na+. Interestingly, except when Mn+
) Pb2+, the powder XRD patterns of the reverse-Na+-exchanged
ETS-10 samples [(Mn+fNa+)ETS-10] returned to that of the
pristine or Na+-ETS-10 (Figure 2C). This result shows that ETS-
10 retains its structural integrity during ion exchange with K+,
Na+, Ba2+, Sr2+, Ca2+, Mg2+, and Zn2+ and reverse ion exchange
with Na+ in the dry or fully hydrated state. The above results
also show that ETS-10 has an exceptionally high affinity for
Pb2+, as if this cation is very tightly encapsulated within the
framework pockets, and as a result, the exchange of Na+ and
K+ with Pb2+ is almost irreversible. Although Pb2+-exchanged
ETS-10 gives additional diffraction peaks, the existence of the
original diffraction peaks of the pristine ETS-10 indicates that
ETS-10 also does not undergo structural damage after exchange
The micropore volumes (in cm3 g-1) ranged from 0.115
(Zn2+) to 0.198 (Ca2+) when measured using the t-plot method
or from 0.141 (Zn2+) to 0.240 (Ca2+) when estimated by the
value at p/po ) 0.12 (Table SI 3 in the Supporting Information).
The total pore volumes at p/po ) 0.99 (in cm3 g-1) ranged from
0.155 (Zn2+) to 0.285 (Ca2+). The differences between the total
and micropore volumes were small and nearly constant
(0.024-0.034). The BET surface areas (in m2 g-1) were 554
28,31
with Pb2+, in contrast to its behavior with NH4 and H+.
+
Argon Adsorption Isotherms of Mn+-ETS-10. No Indication
of Mesopore Formation after Ion Exchange. We also obtained
9
13084 J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009

New Insights into ETS-10 and Titanate Quantum Wire
A R T I C L E S
(Ca2+), 536 (Mg2+), 507 (Ba2+), 500 (Na+), 496 (K+), 480
(Sr2+), 407 (Cd2+), 396 (Na+0.47 + K+0.53), 359 (Pb2+), and 323
(Zn2+). Interestingly, the surface areas of ETS-10 decreased
significantly when the high-affinity cations (Zn2+, Pb2+, and
Cd2+) were the charge-balancing cations. The above results show
that the size and number of cations sensitively affect the
micropore volume and surface area, the cation exchange does
not cause the formation of mesopores, and the micropores and
the external surface of ETS-10 are the two constituents of the
total surface area.
S SNa SK STi SSi5SO - SO
i+j+k+19
i
j
k
1
13
√
M
δ(Of) )
(1)
2.08 S
√
O
where, SM, SNa, SK, STi, SSi, and SO denote the Sanderson
electronegativities of the exchanged metal ion, sodium, potas-
sium, titanium, silicon, and oxygen, respectively.
Effects of Cation Electronegativity and Degree of Hydration
on Chemical Shifts. Figure 4A shows that even in the fully
hydrated state, the three chemical shifts δ(IIA), δ(IIB), and δ(III)
experience gradual downfield shifts (to the left) with decreasing
Sanderson electronegativity of the cation, in the order Sr2+, Ba2+,
29Si MAS NMR Spectra of Mn+-ETS-10 and (Mn+fNa+)
ETS-10: Periodic Localization of Cations at the Cross
Section of Titanate Quantum Wires in the Dehydrated State.
Solid-state MAS 29Si NMR spectra of Mn+-ETS-10 in both the
hydrated (Figure 4A) and dehydrated (Figure 4B) states and
(Mn+fNa+)ETS-10 in the hydrated state (Figure 4C) were
Na+, Na+ + K+0.53, and K+ (from bottom to top). Further-
0.47
more, in the case of Sr2+-ETS-10, a new peak appears at around
-99 ppm. We assign the peak as δ(IIB), which seems to be due
to a very large upfield shift, as the results for the dry state make
clear. Thus, in the dehydrated state, the gradual downfield shifts
of the peaks with decreasing the Sanderson electronegativity
of the cation become more straightforward (Figure 4B). In the
dehydrated state, the very large upfield shift appears not only
for Sr2+-ETS-10 but also for Ba2+-ETS-10. However, they
merge into a pseudo-single peak in the hydrated state upon
exchanging the Sr2+, Ba2+, Na+0.47 + K+0.53, and K+ ions with
Na+ (Figure 4C), indicating that the gradual shifts of δ(IIA),
δ(IIB), and δ(III) do not arise from structural damage of ETS-
10 that might have occurred during ion exchange. Thus, the
above results clearly demonstrate that both the nature (elec-
obtained for Mn+ ) K+, Na+0.47 + K+0.53, Na+, Ba2+, and Sr2+
.
The chemical shifts of the three observed peaks are listed in
Table SI 4 in the Supporting Information. The 29Si NMR spectra
of dehydrated Mn+-ETS-10 could be resolved into the five
chemical shifts δ(IA), δ(IB), δ(IIA), δ(IIB), and δ(III), as indicated
in Figure 4B (bottom). The chemical shifts are tabulated in Table
SI 5 in the Supporting Information. The linear relationships
between the chemical shifts and the calculated -δ(Of) values
are shown in Figure 4D. The -δ(Of) values for Mn+-ETS-10
were obtained from the chemical composition according to eq
1:
Figure 4. 29Si solid-state MAS NMR spectra of Mn+-ETS-10 in the (A) hydrated and (B) dehydrated states and of (C) (Mn+fNa+)ETS-10 in the hydrated
state. (D) Plot of the five chemical shifts described at the bottom of (B) vs -δ(Of) for Mn+ as indicated. (E) Illustration of the positions of the five different
2-
types of Si atoms near a (7MR)4 cage in ETS-10 along a TiO3 quantum wire and the correlation between the Si atoms and the NMR peaks (general
description). (F, G) Illustrations of a (7MR)4 cage in ETS-10 from two different perspectives, showing three different types of Si atoms and four different
types of O atoms protruding into the interior of the cage. (H-J) Illustrations of cation-existing regions for the case of polymorph B as (H) a 3D perspective
or as two-dimensional views along the (I) a or b or (J) c direction.
9
J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009 13085

A R T I C L E S
Jeong et al.
tronegativity) of the cation and the degree of hydration
sensitively affect δ(IIA), δ(IIB), and δ(III).
Si-O-Ti-O-Si unit, regardless of its type.27 In fact, two
Si-O-Ti-O-Si units share a Ti atom, forming the letter X
[(Si-O)2-Ti-(O-Si)2]. As a result, two cations exist in the
vicinity of each Ti center (in opposite positions) regardless of
the type of Si-O-Ti-O-Si unit, giving rise to a uniform
distribution of cations along the titanate wire. However, the
“uniform distribution model” cannot account for the very large
difference in the sensitivities of δ(IA), δ(IB), δ(IIA), δ(IIB), and
δ(III) to the variation of -δ(Of) (Figure 4D), particularly the
fact that the sensitivities of δ(IA) and δ(IB) are negligible and
very small, respectively. We attribute such a large difference
in sensitivities to the nonuniform distribution of countercations
along the titanate (TiO32-) molecular wires in the dehydrated
state.
Migration of Cations from 12MR Channels to (7MR)4
Cages upon Dehydration. We now propose that the above
cations that may exist near Si(12,7)-O-Ti-O-Si(12,7) and
Si(12,12)-O-Ti-O-Si(12,12) in the 12MR in the hydrated
state migrate into the neighboring (7MR)4 cages through the
7MR windows (Figure 4E) upon dehydration.38 Since the cations
can have 3D coordination to the oxygen atoms (yellow) in a
(7MR)4 cage, we believe that the increase in coordination
number is the driving force for the above phenomenon. As a
result, upon dehydration, Si(12,12) (green, marked as IA) and
Si(12,7) (green, marked as IB) atoms existing in 12MR channels
should experience very low degrees of electron density deple-
tion, since the cations do not coordinate to the O atoms adjacent
to the Si(12,12) and Si(12,7) atoms. On the other hand, upon
dehydration, all of the framework Si atoms that form a (7MR)4
cage, namely, Si(7,12) (blue, marked as IIA), Si(7,7) (blue,
marked as IIB), and even Si(4Si) (red, marked III), should
experience much higher degrees of electron density removal
because of the influx of additional cations into the (7MR)4 cage,
which should eventually deplete additional electron density from
the three types of Si atoms through coordination to the oxygen
atoms. Accordingly, Si(7,7), Si(7,12), and Si(4Si) atoms should
be very sensitively affected by the type of countercation, while
Si(12,12) and Si(12,7) atoms in the dehydrated state should not.
Nonuniform Sensitivities of Chemical Shifts to Framework
Oxide Charge Density. Figure 4D describes the following two
very interesting phenomena. First, the chemical shifts of 29Si
shift downfield as -δ(Of) increases. Second, the slopes in these
linear relationships are not uniform: the absolute slope values
{in ppm/[-δ(Of)]} are 2.5 for δ(IA), 18.6 for δ(IB), 133.5 for
δ(IIA), 216.3 for δ(IIB), and 93.8 for δ(III), which increase in
the order δ(IA) e δ(IB) , δ(III) < δ(IIA) < δ(IIB). Anderson
and co-workers assigned δ(IA), δ(IB), δ(IIA), and δ(IIB) as the
chemical shifts of four NMR-spectroscopically nonequivalent
Si atoms connected to three Si atoms and one Ti atom [denoted
as Si(3Si,1Ti)] and δ(III) as the chemical shift of the Si atom
connected to four Si atoms [denoted as Si(4Si)].1b,c They also
identified the four NMR-spectroscopically nonequivalent Si
atoms in Si(3Si,1Ti) and named them as Si(12,12), Si(12,7),
Si(7,12), and Si(7,7) on the basis of their connectivities to the
adjacent 7MR and 12MR systems (Figure 4E). However, the
one-to-one correlation between the four nonequivalent Si(3Si,1Ti)
atoms [Si(12,12), Si(12,7), Si(7,12), and Si(7,7)] and the four
chemical shifts [δ(IA), δ(IB), δ(IIA), and δ(IIB)] has not been
established to date. It will therefore further deepen our
understanding of ETS-10 if there is a way to make a one-to-
one correlation between the four chemical shifts and the four
Si(3Si,1Ti) sites.
It should be recalled that the chemical shift of 29Si shifts
downfield as the charge density on the Si atom increases25c,37
and that the negative charge density of the framework oxide
increases with decreasing electronegativity of the charge-
balancing cation for a given zeolite and with increasing degree
of hydration of the zeolite (see above).24–26 The latter result
arises because increasing the degree of hydration leads to a
decrease in the degree of direct LMCT interaction between the
framework oxide and the cation (Scheme 1), which gives rise
to a decrease in the degree of partial electron transfer from the
framework oxide to the cation in the ground state.
Cation Locations: Uniform versus Nonuniform Cation
Distributions. In ETS-10, there are V-shaped Si-O-Ti-O-Si
units that repeat along the titanate quantum wires at the positions
perpendicular to the quantum wire direction. The two Si atoms
are equivalent, and they are positioned in four different
environments depending on their connectivity to 7MRs and
12MRs (see above). Accordingly, there are four different types
of Si-O-Ti-O-Si units, namely, Si(7,7)-O-Ti-O-Si(7,7),
Si(7,12)-O-Ti-O-Si(7,12), Si(12,7)-O-Ti-O-Si(12,7),
and Si(12,12)-O-Ti-O-Si(12,12). Among these, Si(7,7)-
O-Ti-O-Si(7,7) and Si(7,12)-O-Ti-O-Si(7,12) units are
located in a small cage produced by connecting two perpen-
dicularly running adjacent titanate quantum wires through four
Si(4Si) atoms (Figure 4E). Eventually, each cage becomes
surrounded by two sets of confronting 7MR, windows and the
confronting windows are mirror images of each other. We call
this structure a (7MR)4 cage on the basis of the fact that it is
surrounded by four 7MRs. The other two units, Si(12,7)-O-
Ti-O-Si(12,7) and Si(12,12)-O-Ti-O-Si(12,12), are lo-
cated in 12MR channels (Figure 4E).
Assignment of Four 29Si Chemical Shifts to Four Si(3Si,1Ti)
Atoms. On the basis of the above “nonuniform distribution
model”, we conclude that δ(IA), δ(IB), δ(IIA), and δ(IIB) arise
from Si(12,12), Si(12,7), Si(7,12), and Si(7,7), respectively, as
illustrated in Figure 4E (also see Figure SI 3 in the Supporting
Information). The following further supports the above conclu-
sion. Among the Si(7,7) and Si(7,12) atoms, the latter are placed
closer to the oxygen atoms connected to Si(12,7) atoms, to
which the cations do not bind. Therefore, we assign δ(IIA) to
Si(7,12) and the most sensitive shift, δ(IIB), to Si(7,7). Likewise,
we assign the least sensitive shift, δ(IA), to Si(12,12) and δ(IB)
to Si(12,7).
(38) We also obtained 29Si MAS NMR spectra of Mn+-ETS-10 for Mn+
)
Ca2+, Cd2+, Pb2+, and Zn2+, cations that are more electronegative than
those shown in Figure 4. In these samples, contrary to the results for
K+, Na+, Ba2+, and Sr2+, the two peaks I and II merge into single
peaks with centers appearing between-96 and-98 ppm; this is due
to more pronounced shift upfield (to the right) for peak I than peak II.
This phenomenon indicates that in the case of the more electronegative
cations (Ca2+, Cd2+, Pb2+, Zn2+), the cations that already exist in the
12MR in the hydrated state tend to remain in the same ring even after
dehydration. However, because of the difficulty in resolving the single
peaks into four peaks (IA, IB, IIA, and IIB), we could not extract the
details. Since this work already contains a large number of complicated
results, we will study the effect of the more electronegative cations
on the 29Si chemical shifts more rigorously and report the results later
elsewhere.
It has been proposed that in the case of monovalent cations,
a cation exists in the vicinity of the two O atoms of a
(37) (a) Melchior, M. T.; Vaughan, D. E. W.; Jarman, R. H.; Jacobson,
A. J. Nature 1982, 298, 455–456. (b) Dyer, A. In An Introduction to
Zeolite Molecular SieVes; John Wiley & Sons: New York, 1988;
Chapter 4.
9
13086 J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009

New Insights into ETS-10 and Titanate Quantum Wire
A R T I C L E S
Clustering of Cations in (7MR)4 Cages. A (7MR)4 cage is
composed of four Ti atoms (black), four Si(7,12) atoms (blue,
IIA), four Si(7,7) atoms (blue, IIB), four Si(4Si) atoms (red), 20
O atoms (yellow) that can coordinate to the interior cations,
and two axial O atoms that connect two Ti atoms (gray) (Figure
4F,G). Out of the 20 coordinating O atoms, eight equatorial O
atoms (labeled 1) are equivalent, as are four O atoms that
connect Si(7,12) and Si(7,7) atoms (labeled 2), four O atoms
that connect Si(4Si) and Si(7,12) atoms (labeled 3), and four O
atoms that connect Si(4Si) and Si(7,7) atoms (labeled 4). In
the case of monovalent cations, the cation population in a
(7MR)4 cage is eight, indicating the possibility that the cations
form ionic clusters within the cage as a result of the space
limitation inside the cage. The formation of ionic clusters should
be confirmed.
Periodic Localization of Cations at the Cross Sections of
the Titanate Quantum Wires. On the basis of the above, for
the case of polymorph B, a 3D representation of the regions in
which the cations exist in the dehydrated state is shown in Figure
4H; a 2D representation of these regions along the a or b axis
is given in Figure 4I and one along the c axis in Figure 4J.
These illustrations show that the cations exist only in the regions
where two perpendicular titanate quantum wires cross in the
dehydrated state, or at the cross sections of the titanate quantum
wires. The corresponding illustrations for polymorph A are
shown in Figure SI 4 in the Supporting Information. Thus, the
above results reveal that along the titanate quantum wire, there
are cation-coordinated and cation-absent regions in the dehy-
drated ETS-10. We believe that this finding will help elucidate
more interesting features of ETS-10 in the future.
Usefulness of Surface Coating of Dry ETS-10 with Octadecyl-
trichlorosilane. The groups of Schmidt30 and others1b,c,28,31b,32,34,35
have also reported 29Si NMR spectra of Mn+-ETS-10 with Mn+
) Na+, Na+0.47 + K+0.53, K+, and Cs+. However, these groups
were not able to elucidate the effects of the cation and degree
of hydration on the shift in the chemical shift because of
experimental difficulties in obtaining solid-state 29Si NMR
spectra of dehydrated samples. We attribute our successful
elucidation of the above phenomena in the dehydrated state to
the ability to preserve the fully dehydrated ETS-10 samples as
such even in the ambient environments by coating the surfaces
of the dehydrated samples with octadecyltrichlorosilane.33 This
method has also proved to be useful for obtaining powder XRD
patterns of dehydrated zeolites.33
Figure 5. (A) XPS spectra of Mn+-ETS-10 (with Mn+ as indicated) in the
Ti(2p) region, showing two different types of Ti. (B) Plot of the Ti(2p)
binding energies vs -δ(Of) for various Mn+. (C) XPS spectra of Mn+-ETS-
10 in the O(1s) region, showing five different types of O. (D) Plot of the
O(1s) binding energies vs -δ(Of) for various Mn+
.
Scheme 2
Triad Interaction between the Cation, Oxide, and Ti(IV). The
peaks gradually shifted to the higher-energy region with
increasing electronegativity of the countercation, indicating that
cations are removing electron density from the Ti species. The
plot of the binding energies of the two Ti species versus -δ(Of)
(Figure 5B) reveals that the Ti(2p) binding energies decrease
linearly as -δ(Of) increases, indicating that the electron density
on Ti increases as the (average) negative charge density of the
framework oxygen increases. This phenomenon also shows that
oxide-to-titanium(IV) LMCT occurs from the oxide ions to the
titanium(IV) ions in the titanate quantum wire, even in the
ground state. Overall, the above results show that Mn+, titanate-
forming oxide, and Ti4+ form an acceptor 1-donor-acceptor
2 (A1-D-A2) triad system, as shown in Scheme 2.
XPS Studies of Mn+-ETS-10. XPS spectra of Ti(2p) and O(1s)
were obtained from Mn+-ETS-10 with Mn+ ) K+, Na+, Ba2+,
Sr2+, and Ca2+. The XPS spectrum of Ti(2p) is supposed to
give two peaks due to the two spin-orbit-coupled energy states,
3
1
J ) /2 and /2, with the former appearing in the lower-energy
region and the latter in the higher-energy region. The XPS
spectra for Ti(2p) in Mn+-ETS-10 (Figure 5A) clearly indicate
that there are two peaks in each of these energy regions,
indicating that there are two Ti species, I and II. The binding
energies for each species are listed in Table SI 6 in the
Supporting Information. The higher-energy peak for the 3/2 state
appears at 458.1-458.9 eV, a typical region for Ti4+, while
the lower-energy peak appears at 456.5-457.6 eV, a typical
region for Ti3+. We therefore assign the XPS peaks as Ti3+(3/
2), Ti4+(3/2), Ti3+(1/2), and Ti4+(1/2), respectively. The Ti3+(3/2)/
Ti4+(3/2) intensity ratios were between 1.0:1.5 and 1.0:2.7, and
the Ti3+(1/2)/Ti4+(1/2) intensity ratios were between 1.0:0.6 and
1.0:1.9.
Interestingly, the sensitivity of Ti3+ to the electronegativity
of the cation was higher than that of Ti4+. We believe this
happens because the Ti3+ species are richer in electron density
than Ti4+. It may also occur because of the presence of the Ti3+
centers in the (7MR)4 cage region, where the cation population
9
J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009 13087

A R T I C L E S
Jeong et al.
is high. We believe that the latter explanation is more compel-
ling, since the Ti3+ centers can be stabilized more as a result of
their location in the (7MR)4 cage.
5.76 eV (see Table SI 7 in the Supporting Information for more
details). Again, we could not find any trend for the bulk 3D
titanates. Such an absence of the trend for the bulk 3D titanates
may arise from the different M-O (where M is the counter-
cation) and O-Ti bond lengths, since the degree of charge
transfer is sensitively affected by the distance between the donor
and acceptor. In any case, it is clear that the titanate quantum
wire in ETS-10 is unique in the sense that it shows a systematic
change of the degree of spin-orbit splitting as the electrone-
gativity of the charge-balancing cation changes.
Highly Basic Framework Oxide. The appearance of signifi-
cant amounts of Ti3+ from Mn+-ETS-10 is quite interesting since
the related bulk titanate materials (Na2TiO3, CaTiO3, and
BaTiO3) do not show the Ti3+ peak in the XPS spectra.39-41
We believe that the Ti3+ species are produced during XPS
measurements in the vacuum chamber by X-ray-(photo)induced
O-to-Ti charge transfer, since we did not detect the ESR signal
of Ti3+ from the dehydrated ETS-10 under vacuum. Therefore,
the different behaviors of ETS-10 and bulk titanates indicate
that the oxide species in titanate quantum wires are much more
basic than those of bulk 3D titanates, as will be discussed in
more detail later in this work. The formation of significant
amounts of Ti3+ species may also be possible as a result of the
likely stabilization of the extra electron on Ti with charge-
balancing cations clustered within the (7MR)4 cages. This
interpretation is consistent with the fact that the sensitivity of
Ti3+ to the electronegativity of the cation is higher than that of
Ti4+ (see above).
It is noteworthy that the spin-orbit splitting values (in eV)
of the 3D titanate and 1D ETS-10 titanate quantum wire for
Ti4+ are similar for a given cation: 5.83 and 5.78 for Ba2+, 5.85
and 5.81 for Sr2+, and 5.79 and 5.74 for Ca2+. Thus, the
spin-orbit splitting does not change upon reduction of the
dimensionality from 3D in the bulk to 1D in the titanate quantum
wire in ETS-10. This result contrasts with the case of PbTiO3,
whose spin-orbit splitting value increases markedly (by 0.9
eV) in going from the 3D bulk (5.6 eV)41 to the 1D nanowire
(6.5 eV),44 indicating that the degree of electron depletion from
Ti by Pb2+ increases markedly as the dimensionality decreases.
The lack of such a marked increase in spin-orbit splitting in
going from the 3D bulk titanate to the 1D titanate quantum wire
in ETS-10 may arise from the open nature of the ETS-10
structure, in which the cations next to the titanate quantum wires
do not bind to the oxide ions as tightly as in the case of the 1D
nanowire.
Spin-Orbit Splitting of Ti. The measured spin-orbit splitting
[∆(2p3/2 - 2p1/2)] values for Ti4+ ions were 5.51, 5.55, 5.83,
5.85, and 5.79 eV for Mn+ ) K+, Na+, Ba2+, Sr2+, and Ca2+
,
respectively, and those for Ti3+ ions were 5.40, 5.43, 5.71, 5.83,
and 6.04 eV for Mn+ ) K+, Na+, Ba2+, Sr2+, and Ca2+
,
respectively. Thus, the degree of spin-orbit splitting increases
as the electronegativity of the cation increases and hence as the
charge densities of O and Ti atoms decrease. This result indicates
that the effective nuclear charge of the Ti atom increases as the
electronegativity of the charge-balancing cation increases
because the degree of spin-orbit splitting of an atom increases
as the effective nuclear charge of the atom increases.42 The
above result further indicates that the depletion of electron
density from the oxide by the charge-balancing cation sequen-
tially leads to the depletion of electron density from the Ti atom,
giving rise to the increase in the effective nuclear charge of the
Ti atom, and that the degree of electron density depletion from
the Ti atom increases as the electronegativity of the charge
balancing cation increases. This phenomenon also serves as
direct evidence that Mn+, titanate-forming oxide, and Ti4+ form
a triad complex as shown in Scheme 2.
The available literature values for the spin-orbit splitting of
bulk 3D titanates are 5.8,39 5.6,41 6.0,40 and 5.643 eV for
Na2TiO3, BaTiO3, CaTiO3, and PbTiO3, respectively. As noted,
the degree of spin-orbit splitting does not increase with
increasing electronegativity of the cation, unlike the case of the
titanate quantum wire in ETS-10. We suspected that the absence
of a trend for the bulk 3D titanates could arise from the fact
that the values were obtained from different laboratories, so we
measured XPS spectra of the bulk 3D titanates Li2TiO3, BaTiO3,
SrTiO3, CaTiO3, and PbTiO3 and obtained the corresponding
spin-orbit splitting values, which are 5.74, 5.78, 5.81, 5.74,
XPS spectra of O(1s). The XPS spectra for O(1s) in Mn+
-
ETS-10 (Figure 5C) reveal that there are at least five different
types of O species in terms of binding energy. The binding
energies for each of these species ranged from 526.1 to 531.9
eV, and the values are listed in Table SI 6 in the Supporting
Information. We believe the lower-binding-energy species are
the oxide species that form titanate quantum wires. Interestingly,
the binding energies of O(1s) of the oxygen atoms in the titanate
quantum wires are significantly lower than those of O(1s) of
various metal-ion-exchanged zeolites (530.2-533.3 eV),24
indicating that the oxide ions in the quantum wires are much
more basic than the zeolite framework oxides. The values are
also significantly lower than those of bulk titanates, which range
between 529 and 531 eV.39,41,44,45 This explains why titanate
quantum wires are so much vulnerable to acid,31 even to NH4
+
ion,28,30 which can easily generate H+ after liberation of NH3
even under mild conditions, and also to H2O2, which is a typical
oxidant.13b,35 The plot of the O(1s) binding energies versus
-δ(Of) (Figure 5D) shows that the binding energies of O(1s)
decrease linearly as -δ(Of) increases, despite the fact that O(1s)
is a core-level electron.
Effects of Cation and Degree of Hydration on the Raman
Spectra of Titanate Quantum Wires. Background. As described
earlier (see above), it is of interest to study the effect of various
+
cations other than H+ and NH4 on the longitudinal stretching
vibration of the titanate quantum wire. The aforementioned fact
that the electronegativity of the charge-balancing cation and the
degree of hydration sensitively affect the negative charge density
of the titanate quantum wire further provides us a good
opportunity to elucidate the effect of the cation and the degree
of hydration on the frequency and bandwidth of the longitudinal
stretching vibration of the titanate quantum wire and to analyze
(39) Chu, C. L.; Chung, C. Y.; Zhou, J.; Pu, Y. P.; Lin, P. H. J. Biomed.
Mater. Res. 2005, 75A, 595–602.
(40) Wiff, J. P.; Fuenzalida, V. M.; Za´rate, R. A.; Arias, J. L.; Ferna´ndez,
M. S. J. Phys.: Condens. Matter 2004, 16, S1345–S1350.
(41) (a) Jin, Y.; Zhu, Y.; Yang, X.; Li, C.; Zhou, J. J. Solid State Chem.
2007, 180, 301–306. (b) Lo´pez, M. D. C. B.; Fourlaris, G.; Rand, B.;
Riley, F. L. J. Am. Ceram. Soc. 1999, 82, 1777–1786.
(42) Atkins, P. W. In Physical Chemistry, 6th ed.; Oxford University Press:
Oxford, U.K., 1998; p. 374.
(44) Gu, H.; Hu, Y.; You, J.; Hu, Z.; Yuan, Y.; Zhang, T. J. Appl. Phys.
2007, 101, 024319.
(43) Kim, J.-N.; Shin, K.-S.; Park, B.-O.; Lee, J.-H.; Kim, N.-K.; Cho,
S.-H. Smart Mater. Struct. 2003, 12, 565–570.
(45) Meng, X.-D.; Wang, D.-Z.; Liu, J.-H.; Zhang, S.-Y. Mater. Res. Bull.
2004, 39, 2163–2170.
9
13088 J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009

New Insights into ETS-10 and Titanate Quantum Wire
A R T I C L E S
Figure 6. Raman spectra of Mn+-ETS-10 (Mn+ as indicated) in the (A) hydrated and (B) dehydrated states and of (C) (Mn+fNa+)ETS-10 in the 500-1000
cm-1 region. (D) Major components of the along-the-wire vibrational bands of Mn+-ETS-10 (for five Mn+, as indicated) with varying water content (from
the fully hydrated states to the fully dehydrated states). (E) Plots of the frequencies of along-the-wire vibrational bands with respect to water content for
various Mn+. (F) Plots of the frequencies and bandwidths of the along-the-wire vibrational bands of fully dehydrated Mn+-ETS-10 with respect to -δ(Of).
the results in the context of the effect of the charge density of
the quantum wire on its vibrational frequency and bandwidth.
One of the ultimate aims of current nanotechnology is to
utilize metallic and semiconducting quantum wires as building
blocks for nanoscale electronic devices and other novel
applications.18,19,46,47 However, because of their extremely small
thicknesses (on the order of molecular dimensions), the charge-
transfer (CT) interaction between the environment and the
quantum wire significantly affects the physical properties of the
molecular wire through charge injection from the environment
into the quantum wire.47 One of the important effects is the
change in vibrational frequency of the quantum wire in response
to charge injection from the environment. Thus, in the cases of
SW-CNTs, it has been shown that electron injection into the
nanotubes significantly affects their vibrational frequencies and
bandwidths.47 As the vibrational frequency significantly affects
their performance during applications, the elucidation of the
effects of various environments on the vibrational frequency
and bandwidth is of great importance. However, studies of such
effects have been limited to SW-CNTs.47a Therefore, their
elucidation in the case of titanate quantum wires will help us
deepen our understanding of the fundamental properties of
quantum wires, which is essential for the novel application of
the these wires in future nanodevices. In this regard, we also
studied the effect of the environment (the nature of cation and
the degree of hydration) on the vibrational frequency and
bandwidth of the titanate quantum wires.
Effects of Cation and Degree of Hydration on the
Stretching Vibrations of Titanate Quantum Wires. The Raman
spectrum of hydrated Mn+-ETS-10 changes markedly as Mn+
is changed (Figure 6A). Most notably, the longitudinal vibra-
tional peak of the titanate quantum wire shifts irregularly to
either the lower- or higher-frequency region (for Mn+ ) K+,
Na+, Ba2+, Sr2+, Ca2+, Mg2+, Zn2+) or even disappears (for Mn+
) Pb2+, Cd2+). Upon dehydration, the vibrational band gradually
shifts to the lower-energy region, and its bandwidth (fwhm)
progressively increases with increasing Sanderson electronega-
tivity for Mn+ ) K+, Na+, Ba2+, Sr2+, and Ca2+ (Figure 6B).
When Mn+ ) Mg2+, the vibrational band intensity decreases
significantly, and when Mn+ ) Pb2+, Cd2+, or Zn2+, the
vibrational band nearly or completely disappears. However, upon
reverse-exchange of Mn+ with Na+, the Raman spectra become
identical with that of Na+-ETS-10, except when Mn+ ) Pb2+
(Figure 6C), indicating that the titanate quantum wires were
not damaged during or after ion exchange or after dehydration.
The nonreappearance of the Raman band of the titanate quantum
wire in the case of Mn+ ) Pb2+ arises from the difficulty of
reverse-exchanging Pb2+ with Na+.
(46) (a) Kan, S.; Mokari, T.; Rothenberg, E.; Banin, U. Nat. Mater. 2003,
2, 155–158. (b) Ahrenkiel, S. P.; Miedaner, O. I.; Curtis, C. J.;
Nedeljkovic´, J. M.; Nozic, A. J. Nano Lett. 2003, 3, 833–837. (c)
Ouyang, M.; Huang, J.-L.; Lieber, C. M. Acc. Chem. Res. 2002, 35,
1018–1025.
The longitudinal stretching vibration frequency of the titanate
quantum wire gradually shifts to the lower-frequency region
and the bandwidth gradually increases as the water content
decreases from ∼12 to 0% for Mn+ ) K+, Na+, Ba2+, Sr2+,
and Ca2+ (Figure SI 5 and Table SI 8 in the Supporting
(47) (a) Tsang, J. C.; Freitag, M.; Perebeinos, V.; Liu, J.; Avouris, P. H.
Nat. Nanotechnol. 2007, 2, 725–730. (b) Shim, M.; Ozel, T.; Gaur,
A.; Wang, C. J. Am. Chem. Soc. 2006, 128, 7522–7530. (c) Rao, A. M.;
Eklund, P. C.; Bandow, S.; Thess, A.; Smalley, R. E. Nature 1997,
388, 257–259.
9
J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009 13089

A R T I C L E S
Jeong et al.
Figure 8. (A) Raman spectra of fully dehydrated Mn+-ETS-10 (Mn+ as
indicated) in the 200-500 cm-1 region and (B) the linear relationship
Figure 7. Plots of the along-the-wire vibrational frequencies of Mn+-ETS-
10 (for five Mn+, as indicated) versus the binding energies of (A) Ti(2p)
and (B) O(1s), and plots of the along-the-wire vibrational bandwidths of
between -δ(Of) and the frequency of the band at 274-280 cm-1
.
M
n+-ETS-10 (for five Mn+, as indicated) with respect to the binding energies
are shown in Figure 6F, which reveals a very interesting linear
relationship between the longitudinal vibrational frequency of
the titanate quantum and -δ(Of) and the fact that the bandwidth
decreases nonlinearly as -δ(Of) increases. This behavior of the
titanate quantum wire is similar to that of metallic SW-CNTs
in the sense that the vibrational frequency increases and the
bandwidth decreases with increasing electron density.47 Al-
though the amounts of charge injection into the titanate quantum
wire and SW-CNT cannot be directly compared, the sizes of
the frequency increase (45 cm-1) and bandwidth decrease (54
cm-1) observed for the titanate quantum wire are much larger
than those observed for metallic SW-CNTs (∼10 cm-1 and ∼20
cm-1, respectively).47
The partial negative charge of the framework oxide, -δ(Of),
has successfully represented the negative charge density, electron
donor strength, and basicity of the framework oxide. In many
cases, the -δ(Of) values have greatly helped explain the
observed physical and chemical properties of zeolites, which
are related to differences in the framework charge density,
electron donor strength, and basicity. Nevertheless, -δ(Of) is
still a theoretical value. Therefore, wherever possible, attempts
to correlate the observed physical and chemical properties with
actual experimental values that can represent the charge density,
electron donor strength, and basicity of the framework elements
have to be made. In this work, the Ti(2p) and O(1s) binding
energies are such values. We found that the longitudinal
stretching frequency linearly increases as the binding energies
of Ti3+(3/2), Ti4+(3/2), Ti3+(1/2), and Ti4+(1/2) decrease (Figure 7A)
and as those of OII(1s), OIII(1s), OIV(1s), and OV(1s) decrease
(Figure 7B). The bandwidth decreases nonlinearly as the binding
energies of Ti3+(3/2), Ti4+(3/2), Ti3+(1/2), and Ti4+(1/2) decrease
of (C) Ti(2p) and (D) O(1s). The slopes in (A) are -39.7, -57.7, -26.3,
and -40.5 cm-1/eV from left to right, and those in (B) are -24.4, -35.8,
-45.1, and -74.6 cm-1/eV from left to right.
Information). Since each vibrational spectrum has at least two
components and there are other additional small peaks that
overlap the main component of the vibrational band in the case
of Ca2+-ETS-10, we extracted the main component from each
spectrum by deconvolution and compared only the extracted
main components for a given water content in Figure 6D. This
panel clearly shows that the frequency is progressively red-
shifted and the bandwidth progressively increases as the water
content decreases. In the fully dehydrated states, the vibrational
frequencies are 720.8, 708.8, 692.0, 682.9, and 679.7 cm-1 and
the bandwidths (fwhm) are 33.7, 40.4, 55.9, 58.2, 94.6 cm-1
for K+, Na+, Ba2+, Sr2+, and Ca2+, respectively.
Linear Relationship between the Stretching Vibrational
Frequency and Water Content and the Metallic Nature of
the Titanate Quantum Wires. Interestingly, for each Mn+, the
longitudinal stretching vibration frequency and the water content
show a linear relationship, as shown in Figure 6E. It also shows
that the sensitivity of the peak shift with respect to the water
content increases in the order K+ e Na+ , Ca2+ < Sr2+
e
Ba2+. Thus, the sensitivity is higher for divalent cations than
monovalent ones, despite the fact that the number of divalent
cations is half that of the monovalent ones, and the sensitivity
increases with increasing size of the divalent cation. This also
makes clear that the degree of direct binding of the metal cation
to the oxide framework increases in the same order as the water
content decreases. Plots of the longitudinal vibrational frequency
and the bandwidth versus -δ(Of) for the fully dehydrated states
9
13090 J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009

New Insights into ETS-10 and Titanate Quantum Wire
A R T I C L E S
indicates that the cations, particularly large cations, cannot
be fully hydrated by water molecules even in the fully
hydrated state, presumably because of spatial restrictions.25
Thus, we have demonstrated for the first time that the charge
2-
density of the TiO3 molecular wire in ETS-10 can be varied
by varying the charge-balancing cation and the degree of
hydration and that the longitudinal stretching vibration frequency
of the titanate quantum wire increases linearly while the
bandwidth of the vibrational band decreases nonlinearly as
-δ(Of) increases.
New Vibrational Mode of the Titanate Quantum Wire. In
addition to the longitudinal stretching vibration band of the
titanate quantum wire that appears between 600 and 800 cm-1
,
we also paid attention to the three weaker, lower-energy Raman
bands that appear in the 250-400 cm-1 region (Figure 8A).
These bands have been proposed to arise from Ti-O-Ti
bending vibrations.48 If they indeed arise from the Ti-O-Ti
bending vibrations, then they should also experience shifts to
higher-energy regions with increasing negative charge density
of the titanate quantum wire, as was the case for the longitudinal
stretching vibration, since they also originate from the vibration
of the titanate quantum wire. However, while the bands at 305
and 345 cm-1 remain almost unaltered, the band at 274-280
cm-1 progressively shifts to higher energy as the cation is
changed from Ca2+ to K+ (bottom to top in Figure 8A). The
plot of the frequency with respect to -δ(Of) shows a linear
relationship (Figure 8B). This trend exactly matches that of the
longitudinal stretching vibration of the titanate quantum wire.
Accordingly, we propose that out of the three bands in the
250-400 cm-1 region, only the band at 274-280 cm-1 is
another vibrational normal mode of the titanate quantum wire.
We propose that this band arises from bending of the quantum
wire, since it is significantly lower in energy than the longitu-
dinal stretching vibration of the quantum wire.
Effect of Cation and Degree of Hydration on the UV-Vis
Spectra of the Titanate Quantum Wire. The band-gap energy
of the titanate quantum wire, Eg, has been known to decrease
as the length increases.12 Now it has become clear that for a
given length, we can change the charge density and the
stretching and bending vibration frequencies of the titanate
quantum wire by changing the electronegativity of the cation
and the degree of hydration. In fact, the longitudinal stretching
and bending vibrations of the titanate quantum wire can be
regarded as its longitudinal and transverse optical phonons, since
titanate quantum wires are a part of a crystalline solid. In view
of the fact that Eg is affected by phonons and the notion that an
increase in charge density in a semiconductor material will affect
the energy levels of its electronic bands, it is of great interest
to investigate the relationship between the charge density of
the titanate quantum wire and the electronic (UV-vis) spectra
of Mn+-ETS-10 for various Mn+ in both the hydrated and
dehydrated states.
Figure 9. Diffuse reflectance spectra of Mn+-ETS-10 (Mn+ as indicated)
in the (A) hydrated and (B) dehydrated states, and (C) plots of the lowest-
energy absorption maxima and the band-gap energies vs -δ(Of) for the
dehydrated states. The slopes in (C) are (top) -8.88 and (bottom) -3.92
eV/[-δ(Of)].
In the hydrated states, the spectra remain almost the same,
except for a slight shift of the lowest-energy absorption
maximum to higher energy as the cation is changed according
to the order shown in Figure 9A. As a result, Eg also remains
the same at 4.03 eV.49 This again confirms that hydration of
the cations prevents the direct LMCT from the framework to
the cation. Lamberti15 has also reported that the UV-vis
(Figure 7C) and as those of OII(1s), OIII(1s), OIV(1s), and OV(1s)
decrease (Figure 7D).
The fact that the vibrational frequency of the titanate
quantum wire sensitively varies with changes in the water
content further explains the observed discrepancies among
the values for the Raman band of pristine ETS-10 samples
in the literature,13b,c,28,30–32,34,35 since the amount of adsorbed
water varies from one laboratory to another depending on
the relative humidity and the season. The fact that the
stretching vibration frequencies of the titanate quantum wires
of Mn+-ETS-10 change even in the fully hydrated states
(48) Hong, S. B.; Kim, S. J.; Uh, Y. S. Korean J. Chem. Eng 1996, 13,
419–421.
(49) The Eg values were determined from the inflection points of the UV-
vis spectra of the Mn+-ETS-10 samples (Figure 9B).
9
J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009 13091

A R T I C L E S
Jeong et al.
spectrum of ETS-10 remains the same regardless of the cation
(among K+, Li+, Ca2+, and Ba2+), indicating that the investigated
ETS-10 samples were fully hydrated.
tivity of the cation and/or decreasing the degree of hydration
of the cation. The XPS data for Ti(2p) and O(1s) confirm the
above behavior and also reveal that the oxide ions of the titanate
quantum wires are much more basic than those of zeolites and
the 3D titanates. The longitudinal stretching vibration frequency
of the titanate quantum wire increases linearly while the Raman
bandwidth decreases nonlinearly as the negative charge density
of the titanate quantum wire increases, indicating that the titanate
quantum wire is similar to a metallic SW-CNT. We also found
another normal vibrational mode of the titanate quantum wire
with a vibrational frequency of 274-280 cm-1, which we have
assigned as the bending vibration of the quantum wire. In the
dehydrated state, Eg and the first absorption maximum shift to
lower energy as the negative charge density of the oxide
increases, indicating the oxide-to-titanium(IV) charge-transfer
nature of the transitions. We have also shown that the cations
existing in the 12MR channels migrate to the (7MR)4 cages
upon dehydration of ETS-10. On the basis of this phenomenon,
we have assigned δ(IA), δ(IB), δ(IIA), and δ(IIB) as the chemical
shifts arising from the Si(12,12), Si(12,7), Si(7,12), and Si(7,7)
atoms, respectively.
Negative Linear Relationship between Eg and Partial
Negative Charge of the Framework Oxide. The UV-vis
spectrum of ETS-10 is blue-shifted as dehydration proceeds,
as shown in Figure SI 6 in the Supporting Information. In the
fully dehydrated states, the lowest-energy absorption maximum
becomes distinct and progressively shifts to higher energies
[from 4.34 eV (286 nm) to 4.88 eV (254 nm)] as Mn+ is changed
from K+ to Zn2+, i.e., the Sanderson electronegativity of the
cation increases (Figure 9B). Accordingly, Eg also progressively
shifts from 4.11 to 4.39 eV. The plots of Eg and the energy of
the lowest-energy absorption maximum, Emax, versus -δ(Of)
give negative linear relationships with slopes of -8.9 and -3.9
eV/[-δ(Of)], respectively (Figure 9C).
The fact that Eg and Emax decrease as -δ(Of) increases
indicates that the valence band has strong oxide character and
the conduction band strong Ti(IV) character and that the lowest-
energy absorption arises from oxide-to-Ti4+ LMCT, consistent
with the proposals of Zecchina and co-workers.13a,14,17 However,
in the case of the latter, it is still not clear whether the LMCT
arises from the equatorial or axial oxide ions to Ti(IV) ions. In
any case, the above results clearly show that Eg and the lowest-
energy absorption sensitively change with changes in the
electronegativity of the cation and the degree of hydration. This
work thus reveals that not only the length of the quantum wire12
but also -δ(Of) sensitively affect its Eg and lowest-energy
absorption. Since -δ(Of) is sensitively governed by the nature
of the cation, this result further confirms the A1-D-A2 triad
interaction between the charge-balancing cation, the titanate-
forming oxide, and the titanium ion in the dry state (Scheme
2).
Acknowledgment. We thank the Ministry of Education, Sci-
ence, and Technology (MEST) of the Korean Government and
Sogang University for supporting this work through the Acceleration
Research and Internal University Research Programs. J.-H.P. and
C.-H.S also thank the Carbon Dioxide Reduction & Sequestration
Research Center, a 21st Century Frontier Program supported by
MEST. H.C. and H.L. also thank the Quantum Photonic Science
Research Center supported by MEST.
Supporting Information Available: Various tables, thermal
stability of the titanate quantum wire, Ar adsorption isotherms,
2D illustration (along the a axis) of the positions of the five
different types of Si atoms and the correlation between the Si
atoms and the NMR peaks, illustrations of cation-existing
regions in polymorph A, Raman spectra of Mn+-ETS-10 with
varying degrees of hydration (raw data), and diffuse reflectance
spectra of Mn+-ETS-10 with varying degrees of hydration. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Summary
Overall, this work reports that in ETS-10, the charge-
balancing cation, the titanate-forming oxide, and the titanium
ion interact via acceptor 1-donor-acceptor 2 triad interaction
(Scheme 2). Accordingly, the negative charge density of the
titanate quantum wire can be finely tuned by varying the
electronegativity and degree of hydration of the charge-balancing
cation. For instance, the negative charge density of the titanate
quantum wire can be decreased by increasing the electronega-
JA903638V
9
13092 J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009