Aligned Inclusion of Hemicyanine Dyes into Silica Zeolite Films for Second Harmonic Generation

Article abstract of DOI:10.1021/ja037772q

Silicalite-1 films (thickness = 400 nm) supported on both sides of glass plates (SL/G) were prepared, and hemicyanine dyes (HC-n) with different alkyl chain lengths (n, n = 3, 6, 9, 12, 15, 18, 22, and 24) were included into the silicalite-1 films by dipp

Full text of DOI:10.1021/ja037772q

Published on Web 12/24/2003
Aligned Inclusion of Hemicyanine Dyes into Silica Zeolite
Films for Second Harmonic Generation
Hyun Sung Kim,† Seung Mook Lee,‡ Kwang Ha,† Changsoo Jung,‡ Yun-Jo Lee,†
Yu Sung Chun,† Doseok Kim,‡ Bum Ku Rhee,‡ and Kyung Byung Yoon*,†
Contribution from the Departments of Chemistry and Physics, Sogang UniVersity,
Seoul 121-742, Korea
Received August 6, 2003; E-mail: yoonkb@ccs.sogang.ac.kr
Abstract: Silicalite-1 films (thickness ) 400 nm) supported on both sides of glass plates (SL/G) were
prepared, and hemicyanine dyes (HC-n) with different alkyl chain lengths (n, n ) 3, 6, 9, 12, 15, 18, 22,
and 24) were included into the silicalite-1 films by dipping SL/Gs into each methanol solution of HC-n (1
mM) for 1 d. The included numbers of HC-n per channel (NC) generally decreased with increasing n; that
is, they were 6.4, 23.1, 15.4, 8.2, 5.7, 3.5, 0.9, and 1.2 molecules per channel, respectively. The d33 value
gradually increased with increasing n but decreased when n > 18; that is, they were 1.12, 0.50, 2.25, 3.59,
4.99, 5.30, 1.71, and 2.57 pm V-1, respectively. However, d33/NC progressively increased with increasing
n. The d31 values were ∼100 times smaller than the corresponding d33 values, and the average d33/d31
ratio was 109, which is higher than those of Langmuir-Blodgett (LB) films and poled polymers of nonlinear
optical (NLO) dyes, by ∼2-5 and ∼30-50 times, respectively. The estimated average tilted angle of the
dyes with respect to the channel direction was 7.7°, and the calculated average order parameter was 0.97,
which is ∼480 times higher than the values observed from poled polymers. The degree of uniform alignment
(DUA) generally increased with increasing n. The progressive increase of both DUA and d33/NC with n is
attributed to the increase in the tendency of HC-n to enter hydrophobic silicalite-1 channels with the
hydrophobic alkyl chain first. A more than 134-fold increase in DUA was observed upon increasing n from
6 to 24. The DUA of HC-24 in the silicalite-1 film reached close to 1. Although the observed d33 values
were lower than those of the LB films of NLO dyes due to very small dye densities of the silicalite films, this
methodology bears a great potential to be developed into the methods for preparing practically viable NLO
films.
Introduction
supported on optically transparent substrates. However, the
disadvantages associated with the organic thin films of NLO
Although there are a large variety of dipolar organic nonlinear
optical (NLO) dyes with high second-order hyperpolarizability
constants (â values), most of their crystals are inactive for optical
second harmonic generation (SHG) due to the strong tendency
of the dyes to align centrosymmetrically within the crystals.1,2
Furthermore, the methods developed for crystallizing the NLO
dyes in acentric symmetries have been shown to be successful
only in limited cases.1-3 Efforts have therefore been directed
at utilizing the NLO dyes as self-assembled mono- or
multilayers4-8 and polymer-NLO dye composite films1,2,9-13
dyes1,2 have triggered scientists to search for other novel
conceptual approaches in the use of the dipolar NLO dyes.
(6) Kanis, D. R.; Ratner, M. A.; Marks, T. J. Chem. ReV. 1994, 94, 195-242.
(7) (a) Facchetti, A.; Abbotto, A.; Beverina, L.; van der Boom, M. E.; Dutta,
P.; Evmenenko, G.; Pagani, G. A.; Marks, T. J. Chem. Mater. 2003, 15,
1064-1072. (b) Wang, G.; Zhu, P.; Marks, T. J.; Ketterson, J. B. Appl.
Phys. Lett. 2002, 81, 2169-2171. (c) Milko, E.; van der Boom, M. E.;
Evmenenko, G.; Marks, T. J. AdV. Funct. Mater. 2001, 11, 393-397. (d)
Zhao, Y.-G.; Wu, A.; Lu, H. L.; Chang, S.; Lu, W.-K.; Ho, S. T. Appl.
Phys. Lett. 2001, 79, 587-589. (e) Facchetti, A.; Abbotto, A.; Beverina,
L.; van der Boom, M. E.; Dutta, P.; Evmenenko, G.; Marks, T. J.; Pagani,
G. A. Chem. Mater. 2002, 14, 4996-5005. (f) Wostyn, K.; Binnemans,
K.; Clays, K.; Persoons, A. J. Phys. Chem. B 2001, 105, 5169-5173. (g)
Be´nard, S.; Yu, P.; Audie`re, J. P.; Rivie`re, E.; Cle´ment, R.; Guilhem, J.;
Tchertanov, L.; Nakatani, K. J. Am. Chem. Soc. 2000, 122, 9444-9454.
(h) Schwartz, H.; Mazor, R.; Khodorkovsky, V.; Shapiro, L.; Klung, J. T.;
Kovalev, E.; Meshulam, G.; Berkovic, G.; Kotler, Z.; Efrima, S. J. Phys.
Chem. B 2001, 105, 5914-5921. (i) Lin, S.; Meech, S. R. Langmuir 2000,
16, 2893-2898.
† Department of Chemistry.
‡ Department of Physics.
(1) Nalwa, H. S., Myata, S., Eds. Nonlinear optics of organic molecules and
polymers; CRC: Florida, 1997.
(2) Gunter, P., Ed. Nonlinear optical effects and materials; Springer: Heidel-
berg, 2000.
(3) (a) Eaton, D. F.; Anderson, A. G.; Tam, W.; Wang, Y. J. Am. Chem. Soc.
1987, 109, 1886-1888. (b) Tam, W.; Eaton, D. F.; Calabrese, J. C.;
Williams, I. D.; Wang, Y.; Anderson, A. G. Chem. Mater. 1989, 1, 128-
140. (c) Wang, Y.; Eaton, D. F. Chem. Phys. Lett. 1985, 120, 441-444.
(d) Tomaru, S.; Zembutsu, S.; Kawachi, M.; Kobayashi, M. Chem.
Commun. 1984, 1207-1208. (e) Weissbuch, I.; Lahav, M.; Leiserowitz,
L.; Meredith, G. R.; Vanherzeele, H. Chem. Mater. 1989, 1, 114-118.
(4) Ashwell, G. J.; Hargreaves, R. C.; Baldwin, C. E.; Bahra, G. S.; Brown,
C. R. Nature 1992, 357, 393-395.
(8) (a) Yang, X.; McBranch, D.; Swanson, B.; Li, D. Angew. Chem., Int. Ed.
Engl. 1996, 35, 538-560. (b) Huang, W.; Helvenston, M. Langmuir 1999,
15, 6510-6514.
(9) Dalton, L. R.; Harper, A. W.; Ghosn, R.; Steier, W. H.; Ziari, M.; Fetterman,
H.; Shi, Y.; Mustacich, R. V.; Jen, A. K.-Y.; Shea, K. J. Chem. Mater.
1995, 7, 1060-1081.
(10) Marder, S. R.; Kippelen, B.; Jen, A. K.-Y.; Peyghambarian, N. Nature 1997,
388, 845-851.
(11) van der Boom, M. E. Angew. Chem., Int. Ed. Engl. 1996, 41, 3363-3366.
(12) Samyn, C.; Verbiest, T.; Persoons, A. Macromol. Rapid Commun. 2000,
21, 1-15.
(5) Katz, H. E.; Scheller, G.; Putvinski, T. M.; Schilling, M. L.; Wilson, W.
L.; Chidsey, C. E. D. Science 1991, 254, 1485-1487.
9
10.1021/ja037772q CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 673-682
673

A R T I C L E S
Kim et al.
Along this line, zeolites and the related nanoporous materials
have been examined as the hosts for aligned inclusion of organic
dipolar NLO dyes to explore novel organic-inorganic composite
SHG materials.14-22 Thus, Stucky and co-workers first reported
that para-nitroaniline (PNA), 2-methyl-4-nitroaniline (MNA),
2-amino-4-nitropyridine, and the analogous compounds readily
enter the straight channels of AlPO4-5 [a noncentrosymmetric
(P6cc) zeolite analogue having one-dimensional channels with
a diameter of 0.8 nm], and the dye-incorporating AlPO4-5
powders generate second harmonic (SH) with the intensity far
exceeding that of quartz powders.14,15
to be SHG active, however, aging of the composite under humid
air for several weeks was necessary, indicating the requirement
of the water-assisted secondary reorganization of the included
PNA molecules to give rise to a net bulk dipole moment.
Thus, the previous pioneering works have demonstrated the
potential of zeolites and the related nanoporous materials to be
developed into versatile inorganic hosts for preparation of
practically viable organic-inorganic composite SHG materials.
However, despite the fact that there are a large variety of dipolar
organic NLO dyes with much higher â values (>500 × 10-30
esu), the previous studies have been limited to PNA and the
analogous NLO dyes with relatively low â values (âPNA ) (34.5
( 4) × 10-30 esu)23,24 and with two distinctively different
terminals, to one of which the aforementioned AlPO4-5, Sb-
SL, and ZSM-5 channels happened to show a higher intrinsic
preference. Furthermore, the examined zeolite forms have been
limited to powders and very small single crystals that bear no
practical applicability. It is therefore necessary to develop
methods for including dipolar NLO dyes with higher â values
into the zeolites in uniform orientations so that the previous
pioneering efforts can bear fruit. Also, for practical viability,
efforts should be directed at utilizing uniformly aligned zeolite
films as hosts for the aligned inclusion of NLO dyes rather than
small crystals and powders.
Subsequent studies by Marlow, Caro, and their co-workers
revealed that the SHG activity of the PNA-including AlPO4-5
crystals arose as a result of the spontaneous inclusion of PNA
into the channels of AlPO4-5 with the nitro group first caused
by the intrinsically higher affinity of the AlPO4-5 channels to
the nitro than to the amino group.16 Because the sizes of the
crystals far exceeded (such as 130 µm) the wavelength of the
incident laser beam (1.064 µm), the polarization reversal that
occurred at the center of each crystal did not affect the overall
SHG activities of the dye-loaded AlPO4-5 crystals. They also
found that the AlPO4-5 crystals loaded with 4-nitro-N,N-dimethyl-
aniline17 or (dimethylamino)benzonitrile18 are active for SHG.
The MFI-type structures such as ZSM-519 and Sb-incorporat-
ing silicalite-1 (Sb-SL)20 [centrosymmetric (Pnma) zeolites
having a three-dimensional channel system consisting of straight
0.54 × 0.56 nm channels in one direction and sinusoidal 0.51
× 0.54 nm channels in the other direction perpendicular to the
straight channels] have also been shown to be SHG active upon
PNA loading. However, in the case of Sb-SL loaded with PNA,
the SHG activity disappeared after several exposures to incident
laser beams. Unlike ZSM-5 and Sb-SL which contain Al and
Sb, respectively, in the framework, the closely related pure silica
ZSM-12 [a centrosymmetric (C2/c) silica zeolite having one-
dimensional channels with a diameter of 0.56 × 0.59 nm] did
not show any SHG activity even after inclusion of PNA.21 PNA-
loaded MCM-41 (an amorphous silica having a hexagonal array
of channels with a diameter of 2-8 nm) also showed a SHG
activity that was comparable to that of potassium dihydrogen
phosphate (KDP) powders whose d33 (a tensor component of
We now report a method for incorporating hemicyanine (HC)
(â g 770 × 10-30 esu) with the N-alkyl pyridinium side first
into the channels of uniformly aligned silicalite-1 films and that
the HC-loaded silicalite-1 films show high SHG activities.
Experimental Section
We prepared silicalite-1 films supported on both sides of glass plates
(SL/G) and a series of HCs with different alkyl chain lengths (HC-n,
n ) 3, 6, 9, 12, 15, 18, 22, and 24), on the basis of the hypothesis that
slim, dipolar NLO dyes having a long hydrophobic tail will enter
hydrophobic zeolite channels with the tail part first, leading to aligned
inclusion of the NLO dyes into the channels.
the quadratic nonlinear susceptibility) is ∼3 pm V-1 22
. For them
(13) (a) Saadeh, H.; Yu, D.; Wang, L. M.; Yu, L. P. J. Mater. Chem. 1999, 9,
1865-1873. (b) Steire, W. H.; Chen, A.; Lee, S.-S.; Garner, S.; Zhang,
H.; Chuyanov, V.; Dalton, L. R.; Wang, F.; Ren, A. S.; Zhang, C.;
Todorova, G.; Harper, A.; Fetterman, H. R.; Chen, D.; Udupa, A.;
Bhattacharya, D.; Tsap, B. Chem. Phys. 1999, 245, 487-506. (c) Jiang,
H.; Kakkar, A. K. J. Am. Chem. Soc. 1999, 121, 3657-3665. (d) Dalton,
L. R.; Steier, W. H.; Robinson, B. H.; Zhang, C.; Ren, A.; Garner, S.;
Chen, A.; Londergan, T.; Irwin, L.; Carlson, B.; Fifield, L.; Phelan, G.;
Kincaid, C.; Amend, J.; Jen, A. J. Mater. Chem. 1999, 9, 1905-1920.
(14) Cox, S. D.; Gier, T. E.; Stucky, G. D.; Bierlein, J. J. Am. Chem. Soc. 1988,
110, 2986-2987.
Subsequently, we incorporated the HC-n dyes into SL/Gs and
investigated the effects of n on the included amounts of HC-n in each
channel of SL/G, on the optical SH intensity (I2ω) of the resulting HC-
n-including SL/G (HC-n-SL/G), and on the two tensor components of
the quadratic nonlinear susceptibility, d31 and d33, which were deter-
mined by Maker’s fringe method.25 The reason for choosing silicalite-1
films despite the fact that pure silica zeolites have never been employed
as the hosts for aligned inclusion of NLO dyes is because they readily
grow on glass with the straight channels (b-axis) orienting perpendicular
to the glass plane and the channels are very hydrophobic. The measured
d33 values were then compared with the maximum theoretical values.
For the above, we also independently measured the absorption
maximums (λmax), the molar extinction coefficients (ꢀ), and â values
of HC-n dyes (âHC-n).
(15) (a) Cox, S. D.; Gier, T. E.; Stucky, G. D. Chem. Mater. 1990, 2, 609-
619. (b) Cox, S. D.; Gier, T. E.; Stucky, G. D.; Bierlein, J. Solid State
Ionics 1989, 32, 514-520.
(16) Marlow, F.; Wu¨bbenhorst, M.; Caro, J. J. Phys. Chem. 1994, 98, 12315-
12319.
(17) Caro, J.; Marlow, F.; Hoffmann, K.; Striebel, C.; Kornatowski, J.; Girnus,
I.; Noack, M.; Ko¨lsch, P. Prog. Zeolite Microporous Mater. 1997, 105,
2171-2178.
(18) Marlow, F.; Caro, J.; Werner, L.; Kornatowski, J.; Da¨hne, S. J. Phys. Chem.
1993, 97, 11286-11290.
(19) Werner, L.; Caro, J.; Finger, G.; Kornatowski, J. Zeolites 1992, 12, 658-
663.
(23) Clays, K.; Persoons, A. Phys. ReV. Lett. 1991, 66, 2980-2983.
(24) Sta¨helin, M.; Burland, D. M.; Rice, J. E. Chem. Phys. Lett. 1992, 191,
245-250.
(25) (a) Maker, P. D.; Terhune, R. W.; Nisenoff, M.; Savage, C. M. Phys. ReV.
Lett. 1962, 8, 21-22. (b) Herman, W. N.; Hayden, L. M. J. Opt. Soc. Am.
B 1995, 12, 416-427.
(20) Reck, G.; Marlow, F.; Kornatowski, J.; Hill, W.; Caro, J. J. Phys. Chem.
1996, 100, 1698-1704.
(21) Kinski, I.; Daniels, P.; Deroche, C.; Marler, B.; Gies, H. Microporous
Mesoporous Mater. 2002, 56, 11-25.
(22) Kinski, I.; Gies, H.; Marlow, F. Zeolites 1997, 19, 375-381.
9
674 J. AM. CHEM. SOC. VOL. 126, NO. 2, 2004

Inclusion of Dyes into Silica Zeolite Films
A R T I C L E S
Materials. A series of n-alkyl bromides (CnH2n+1Br) were purchased
from TCI (n ) 3, 6, 9, 12, 15, and 18) and Aldrich (n ) 22) and were
used as received. 1-Tetracosyl bromide (n-C24H49Br) was prepared from
the corresponding alcohol (Aldrich) according to the following
procedure. Triphenylphosphine (95 mg, 0.36 mmol) and carbontetra-
bromide (120 mg, 0.36 mmol) were sequentially introduced into a
dichloromethane solution (50 mL) dissolved with 1-tetracosanol (60
mg, 0.17 mmol), and the mixture was stirred at room temperature for
12 h. The solvent was then removed by evacuation, and the produced
1-tetracosyl bromide was isolated from the reaction mixture by column
chromatography. 4-[4-(Dimethylamino)styryl]pyridin (Aldrich), tetra-
ethyl orthosilicate (TEOS, Acros), sodium aluminate (NaAlO2, Kanto),
and tetrapropylammonium hydroxide (TPAOH, Aldrich) were pur-
chased and used as received. n-Octadecane and n-heptadecane were
purchased from PolyScience Corp. and used as such. 4-Nitroaniline
was purchased from Aldrich and used as such.
Preparation of 4-[4-(Dimethylamino)styryl]-1-n-alkylpyridinium
Bromide (HC-n). The HC-n dyes were prepared by refluxing the
acetonitrile solution (25 mL) of 4-[4-(dimethylamino)styryl]pyridin (1
mmol) and the corresponding 1-bromoalkane (1 mmol) for varying
periods of time until the initially colorless solution turned dark red
(for instance, it required 6 days for HC-18). The solvent from each
reaction mixture was removed under vacuum. Ethyl acetate (100 mL)
was introduced into the reaction flask containing the crude products,
and the heterogeneous mixture was poured on top of a silica gel column.
To remove nonreacted starting compounds, copious amounts of ethyl
acetate were passed through the column until pure ethyl acetate was
collected from the column. The residual red salt was separated from
silica gel by passing copious amounts of methanol through the column,
and by collecting the red methanol solution. The red product was
isolated by evaporation of methanol. The product was redissolved into
a minimum amount of methanol, and ether was added into the solution
until the solution became cloudy. Upon keeping the cloudy solution at
-20 °C overnight, microcrystals of HC-n precipitated at the bottom of
the glass container, and they were collected by filtration. The yields
were generally over 60%. Results of elemental analyses, matrix-assisted
laser desorption ionization time-of-flight (MALDI-TOF) mass spectral
data, 1H and 13C NMR spectra, and Fourier transform infrared (FT-IR)
spectra are listed in the Supporting Information (Tables SI-1-4 and
Figure SI-1). The above analytical results unambiguously confirmed
the identities of HC-n dyes synthesized in this report.
Figure 1. The HRS signal of the methanol solution of PNA (0.9 mM)
(A), the photoluminescence spectrum of the methanol solution of HC-18
(B), and the HRS signal of the methanol solution of HC-22 (0.3 mM)
overlapping with the upper edge of the intense photoluminescence envelope
(C), arising from irradiation of each sample with 1064-nm laser pulses.
SH radiation and subsequently led to a photomultiplier tube (PMT).
Figure 1A shows the HRS signal of PNA. The HC-n dyes produced
very intense photoluminescence with the maximum at ∼600 nm due
to multiphoton absorption of the incident beam by the HC-n dyes as
typically shown in Figure 1B for the case of HC-18. As a result, the
HRS signals from HC-n dyes overlapped with the upper edge of an
intense photoluminescence envelope as shown in Figure 1C. Accord-
ingly, only the SH signals were extracted from the backgrounds by
deconvolution.
Measurements of λmax and E of the Visible Band of HC-n. A series
of 0.1 mM methanol solutions of HC-n were prepared. The UV-vis
spectra of HC-n solutions were recorded using a matched pair of quartz
cuvettes with optical path lengths of 1 cm. The absorbance values of
the 0.02-mM solutions fell between 0.8 and 1. For each HC-n solution,
the above measurement was repeated three times to deduce an average
value of λmax and ꢀ, respectively.
Measurement of âHC-n by Hyper Rayleigh Scattering (HRS). The
âHC-n values for n ) 6, 12, and 22 were determined indirectly by
comparison with that of PNA (âPNA ) (34.5 ( 4) × 10-30 esu, in
methanol at 1064 nm), which has been measured most reliably.23,24
For this, the intensities of the incoherent SH radiations (Iout) of HC-n
(n ) 6, 12, and 22) and PNA were measured at various concentrations
between 0.1 and 0.5 mM for HC-n and between 0.1 and 30 mM for
PNA. The experimental setup is as follows. A rectangular quartz cuvette
(path length ) 1 cm) containing a methanol solution of a NLO dye
was placed along the beam path of a mode-locked Nd:YAG laser beam
(1064 nm, 10-Hz repetition rate, 40-ps pulse width with the intensity
of Iin) 40 cm away from a focusing lens with the focal length of 50
cm. The cuvette was carefully positioned so that the beam can pass
through the rightmost end of its front face to minimize absorption of
the generated SH wave (532 nm) by the HC-n dyes whose absorption
tails exceed 532 nm. The scattered incoherent SH radiation was
collected with a lens facing the right-hand side of the cuvette, and the
collected beam was passed through a monochromator to select only
Preparation of Silicalite-1 Films. The growth of ZSM-5 and
silicalite-1 films on porous substrates such as porous stainless steel,
porous alumina, and porous Vycor glass disk and on nonporous
substrates such as Teflon, silver, and silicon is well documented.26 We
therefore grew silicalite-1 and ZSM-5 films on flat glass by modification
of the procedures and compositions of the synthesis gels described in
the literature. For comparison, we also grew ZSM-5 films with several
different Si/Al ratios on glass plates.
(26) (a) Auerbach, S. M., Carrado, K. A., Dutta, P. K., Eds. Handbook of Zeolite
Science and Technology; Marcel Dekker: New York, 2003; Chapter 17
edited by Nair, S., Tsapatsis, M. (b) Yan, Y.; Davis, M. E.; Gavalas, G. R.
Ind. Eng. Chem. Res. 1995, 34, 1652-1661. (c) Hedlund, J.; Mintova, S.;
Sterte, J. Microporous Mesoporous Mater. 1999, 28, 185-194. (d)
Tsikoyiannis, J. G.; Hagg, W. O. Zeolites 1992, 12, 126-130. (e) den Exter,
M. J.; van Bekkum, H. V.; Rijn, C. J. M.; Kapteijn, F.; Moulijn, J. A.;
Schellevis, H.; Beenakker, C. I. N. Zeolites 1997, 19, 13-20. (f) Wang,
Z.; Yan, Y. Chem. Mater. 2001, 13, 1101-1107. (g) Xomeritakis, G.;
Tsapatsis, M. Chem. Mater. 1999, 11, 875-878. (h) Ha, K.; Lee, Y.-J.;
Chun, Y. S.; Park, Y. S.; Lee, G. S.; Yoon, K. B. AdV. Mater. 2001, 13,
594-596.
9
J. AM. CHEM. SOC. VOL. 126, NO. 2, 2004 675

A R T I C L E S
Kim et al.
The growth of continuous silicalite-1 films on the glass plates was
carried out by immersing five glass plates (25 × 70 × 1 mm3) into the
synthesis gel consisting of TEOS, TPAOH, and water in the mole ratio
of 0.8:0.1:50, followed by heating at 140 °C for 5 h in a Teflon-lined
autoclave. The synthesis gel was prepared by introducing TEOS (28.3
g) into a plastic beaker containing a TPAOH solution (150 mL, 0.11
M) with vigorous stirring. The gel was aged at room temperature for
12 h with vigorous stirring and was transferred to a Teflon-lined
autoclave before immersion of glass plates. The obtained SL/Gs were
thoroughly washed with copious amounts of water and dried in the
atmosphere at room temperature. The resulting SL/Gs were cut into
four pieces with the size of ∼25 × 18 × 1 mm3, and they were calcined
at 450 °C for 12 h to remove TPA template prior to inclusion of NLO
dyes. The thickness of each silicalite-1 film was monitored by taking
the scanning electron microscope (SEM) images of the cross sections.
Films with higher thickness (g∼1 µm) were also prepared either by
increasing the immersion time in the synthesis gel or by re-immersing
the glass plates coated with zeolite films into a fresh gel for a desired
length of time.
deionized water) were sequentially introduced into a 100-mL plastic
beaker. After the beaker was swirled for 5 min, the solution was
neutralized by adding aqueous NaOH solution (10 mL, 5 M). Methanol
(80 mL) was subsequently added to the solution, and the clean
supernatant solution was collected by centrifugation. Independently, a
stock solution of HC-n in methanol was prepared by dissolving a known
amount of HC-n (20-100 µg) in methanol (10 mL). An aliquot (1
mL) of the stock solution of HC-n was added into 100 mL of the above
simulated solution, and the absorbance of the mixed solution was
recorded. More aliquots were added into the simulated solution until
the absorbance of the final solution became close to the one obtained
from the HC-n solution extracted from HC-n-SL/G. We also extracted
HC-22 from five HC-22-SL/Gs to increase the absorbance of the
extracted solution as a means to increase accuracy in the quantitative
analysis.
In estimating the degree of coverage of silicalite-1 film, the areas
occupied by the empty spots and those covered by a-oriented twined
crystals (the twined crystals whose a-axis are oriented perpendicular
to the underlying basal crystal faces) were deducted from the total area.
Those areas covered by second layers of silicalite-1 crystals were
regarded as a single layer on the basis of the assumption that the
channels in the upper and the bottom crystals are not aligned, and, as
a result, inclusion of HC-n molecules is limited to the channels in the
upper layers. For the above, a SEM image with a size of 32 × 21 µm2
taken at a magnification of 4000 was photocopied onto a piece of paper
so that the area of the SEM image became 32 × 21 cm2, and the area
covered by empty spots and a-oriented twined crystals was cut off using
a sharp knife. By comparing the initial and final weights of the paper,
the percentage of the area occupied by the b-oriented silicalite-1 film
(the film whose b-axis is oriented perpendicular to the glass substrate)
was deduced. The average percentage was deduced by repeating the
above procedure for three different spots in a silicalite-1 film.
Subsequently, given that the unit cell dimension of (010) plane is 2.007
× 1.342 nm2 and each unit cell contains two straight channels,27 the
number of straight channels in each SL/G was deduced by dividing
the area of each SL/G (∼25 × 18 mm2) by 2.007 × 1.342 nm2 and
taking into account the fact that a SL/G has two films.
Measurement of SH Intensity of HC-n-SL/G (I2ω) Relative to
That of Quartz (I2ω (quartz)). As an index matching fluid, a drop of
dimethylsulfoxide (DMSO) was dropped onto each side of a HC-n-
SL/G, respectively, and each side of the HC-n-SL/G was covered with
a clean bare glass plate (25 × 18 × 1 mm3). This was necessary to
avoid scattering of the incident laser beam caused by the irregular
thickness of the silicalite-1 films. A reference channel consisting of a
beam splitter and a photodiode was used to compensate for the intensity
fluctuations of the fundamental beam (1064 nm). The polarity of the
fundamental laser beam was adjusted using a half-wave plate before it
hit the sample. The electric field vector of the incident beam was either
parallel (p-polarization) or perpendicular (s-polarization) to the plane
of incidence. Only the p-polarized SH beam was made to enter a PMT
by using a prism, a SH pass filter, and a polarizer. A HC-n-SL/G was
mounted on the rotator coupled to a step motor. The output signals
from the photodiode and PMT were detected as a function of an incident
angle. A 3-mm-thick Y-cut quartz crystal (a piece of quartz plate whose
plane is perpendicular to the crystalline y-axis and the thickness of the
plate is 3 mm; see Figure SI-2 in the Supporting Information)28 was
used as a reference for determining the relative intensities of the SH
signals (I2ω/I2ω (quartz)) generated from the samples. The SH signals
were collected from three different spots of a HC-n-SL/G, and the
average intensity was taken.
ZSM-5 films with several different Si/Al ratios (50, 25, and 17) were
also grown on glass plates by modification of the gel compositions.
The gel compositions were TEOS:NaAlO2:TPAOH:H2O ) 7:0.14:1:
300, 7:0.28:1:300, and 7:0.42:1:300, respectively. The aging period and
reaction temperature were fixed to be 5 h and 180 °C, respectively,
and the reaction periods were 4, 4.3, and 5 h, respectively. The Si/Al
ratios of the ZSM-5 films were determined with energy-dispersive X-ray
analysis (EDX) of the films.
Inclusion of HC-n into SL/G and ZSM-5 Films. Into each vial
(25 mL capacity) containing a methanol solution of different HC-n
(10 mL, 1 mM) were added two SL/Gs, and the vial was kept at room
temperature for a desired period of time such as 1 day, 1 week, or 3
weeks. After equilibration, the SL-Gs were removed from the solution,
washed with copious amounts of fresh methanol, and dried in the air.
For comparison of the SHG activities, SL/Gs from the same batch were
used so that the characteristic factors of the film, such as the thickness,
the degree of the coverage of the glass with silicalite film, the orientation
of the film, and the morphology of the film, were as similar as possible.
Furthermore, we took out a small portion from each SL/G and analyzed
the characteristic factors of the films with SEM to ensure that the above
factors are indeed similar from one sample to another and to discard
any SL/Gs having significantly different factors from the average values.
Similarly, HC-18 was included into ZSM-5 films with Si/Al ratios
of 50 and 25, respectively. Because of the poor film condition of ZSM-5
films with a Si/Al ratio of 17, inclusion of HC-18 into the films was
not carried out.
Quantitative Analysis of the Included Amounts of HC-n in HC-
n-SL/G. The HC-n dyes incorporated in SL/Gs were quantitatively
analyzed according to the following procedure. A dilute aqueous
solution of hydrofluoric acid (1 mL, a mixture of 0.2 mL of 49% HF
and 0.8 mL of distilled deionized water) was introduced into a 50-mL
plastic beaker containing a HC-n-SL/G. After gentle swirling of the
mixture for 5 min at room temperature, the solution was neutralized
by adding aqueous NaOH solution (1 mL, 5 M). Methanol (8 mL) was
subsequently added into the neutral solution to ensure dissolution of
all of the dye into the solution. After removal of the clean glass substrate
from the solution, the solution was centrifuged to precipitate silica
particles. The clean supernatant solution was decanted into one of the
matched pair of quartz cells for a spectroscopic measurement. To
increase accuracy in the quantitative analysis, we independently
determined the molar extinction coefficient of each HC-n at a
concentration similar to that of the extracted solution. For this, a mixed
solvent with a composition similar to that of the extracted solution was
prepared according to the following procedure. Freshly calcined silicalite
(20 µg), 10 pieces of clean glass plates (18 × 25 mm2), and hydrofluoric
acid (10 mL consisting of 2 mL of 49% HF and 8 mL of distilled
(27) Baerlocher, Ch., Meier, W. M., Olson, D. H., Eds. Atlas of zeolite framework
types, 5th revised ed.; Elsevier: New York, 2001; p 184.
(28) Quartz crystals generate SH along the y-axis. Therefore, a Y-cut quartz
crystal is often used as a transmission SHG reference for Maker’s fringe
experiment. See: (a) Boyd, R. W. Nonlinear Optics, 2nd ed.; Academic:
London, 2003; p 48. (b) Shen, Y. R. The Principles of Nonlinear Optics;
Wiley: New York, 1988; p 101.
9
676 J. AM. CHEM. SOC. VOL. 126, NO. 2, 2004

Inclusion of Dyes into Silica Zeolite Films
A R T I C L E S
Comparison of the Affinities of Silicalite-1 Channels to n-
Octadecane and HC-3. A methanol solution of n-octadecane (molec-
ular weight ) 254.5) was prepared by dissolving 12.73 mg of
n-octadecane in 25 mL of methanol (2 mM). Freshly calcined silicalite-1
powder (50 mg) was introduced into 5 mL of the above solution. After
the heterogeneous solution was allowed to equilibrate for 3 h at room
temperature, the solution was filtered through a cotton-packed pipet.
A known amount of n-heptadecane was added into the clear solution
of n-octadecane as the internal standard, and the remaining amount of
n-octadecane in the solution was determined by gas chromatography.
The quantitative analysis of the included amount of HC-3 into
silicalite-1 powder was similarly carried out by equilibrating 50 mg of
freshly calcined silicalite-1 powder in a methanol solution of HC-3 (2
mM, 5 mL) for 3 h followed by measuring the absorbance of the
supernatant solution on a UV-vis spectrophotometer.
Instrumentation. SEM images of zeolites and the zeolite-coated
glass plates were obtained from a FE-SEM (Hitachi S-4300) at an
acceleration voltage of 20 kV. A platinum/palladium alloy (in the ratio
of 8 to 2) was deposited with a thickness of about 15 nm on top of the
samples. The EDX analysis of the samples was carried out on a Horiba
EX-220 Energy Dispersive X-ray Micro Analyzer (model: 6853-H)
attached to the above FE-SEM. The X-ray diffraction patterns for the
identification of the zeolite films were obtained from a Rigaku
diffractometer (D/MAX-1C) with the monochromatic beam of Cu KR.
The UV-vis spectra of the samples were recorded on a Shimadzu UV-
1
3101PC. FT-IR spectra were recorded on a Jasco FT/IR 620. H and
13C NMR spectra of the HC-n dyes were obtained from a Varian Jemini
500 NMR spectrometer. Elemental analyses of the HC-n dyes were
performed on a Carlo-Erba EA1108. Mass spectral data were obtained
from an Applied Biosystem MALDI TOF mass spectrometer (Pro-
teomics Solution I Voyager-DE STR). Quantitative analysis of the
included amount of n-octadecane into silicalite-1 was performed on a
Hewlett-Packard 6890 series gas chromatograph equipped with a flame
ionization detector. The fundamental laser pulses (1064 nm, 40-ps pulse
width, and 10-Hz repetition rate) were generated from a Continuum
PY61 mode-locked Nd:YAG laser.
Figure 2. A typical X-ray diffraction pattern (A) and the SEM images (B,
cross section; C, top view) of SL/G showing the alignment of the b-axis of
the silicalite-1 film perpendicular to the glass plane. The inset in panel A
shows the X-ray powder diffraction pattern of the silicalite-1 powder, which
was obtained by finely grinding the film that was collected by scrapping
the SL/G.
the total area of the glass plate) decreased with an increase in
the Al content; that is, they were 93%, 81%, and 57% for the
films with Si/Al ratios of ∞, 50, and 25, respectively. Under
the experimental conditions described in the Experimental
Section, the film thicknesses of the ZSM-5 films were also ∼400
nm.
Results and Discussion
Characteristics of SL/Gs and ZSM-5 Films. Silicalite-1
films readily grew on both sides of glass plates with their b-axes
(straight channels) orienting perpendicular to the glass planes.
Thus, as shown in Figure 2A, the diffraction lines appeared at
a regular interval (2θ ) 8.85, 17.8, 26.85, 34.85, and 45.55°)
corresponding to the (0 2 0), (0 4 0), (0 6 0), (0 8 0), and (0 10
0) planes of the crystals. To make sure that the above regularly
spaced X-ray diffraction pattern arises not due to the formation
of an unknown zeolite film but due to the orientation of the
silicalite-1 film with the b-axis perpendicular to the glass plane,
the films were scraped off the glass plates and ground into fine
powders to analyze the nature of the film. The X-ray diffraction
pattern of the powders revealed that the film is indeed comprised
of pure silicalite-1, as shown in the inset of Figure 2A. The
SEM images shown in Figure 2B,C further confirmed that the
silicalite-1 films are oriented with b-axes perpendicular to the
substrate planes. The film thickness significantly varied depend-
ing on the batch. Although slightly, the thickness also varied
within a film depending on the spot. The typical thicknesses
were 400-500 nm, and the film thickness of the SL/G used in
this report was ∼400 nm.
λ
max and E of the HC-n Visible Band. The HC-n dyes were
readily soluble in methanol, and the methanol solution showed
a strong visible absorption band at ∼479 nm as shown in Figure
3A for HC-18. The determined ꢀ values at ∼479 nm ranged
from 39 515 to 45 160 (M-1 cm-1) as listed in Table 1. As
noticed, the tail length had almost no effect on the λmax and ꢀ,
with a slight exception of HC-3, whose corresponding values
are slightly smaller than those of the longer-tail homologues. It
is therefore concluded that the hemicyanine head, 4-[4-(dimeth-
ylamino)styryl]pyridinium, is the part that determines the λmax
and ꢀ values of the visible transition.
â
HC-n for n ) 6, 12, and 22. The plots of the ratio between
the SH intensity (Iout) and the square of the incident beam
intensity (Iin2), that is, Iout/Iin2, with respect to the concentration
was made for PNA and HC-n (n ) 6, 12, and 22), respectively
(Figure 4). In such a situation where a NLO dye is dissolved in
2
a solvent, it is well established that Iout/Iin can be expressed
with the following:
The quality of the film became inferior with an increase in
the Al content in the gel (see Figure SI-3 in the Supporting
Information). Consequently, the degree of the area of the glass
plate covered by the b-oriented MFI-type film (with respect to
Iout/Iin2 ) g[(âsolvent)2Csolvent + (âdye)2Cdye] ×
exp[-(2σinlin + σoutlout)Cdye] (1)
9
J. AM. CHEM. SOC. VOL. 126, NO. 2, 2004 677

A R T I C L E S
Kim et al.
nm,4 although the reported values in fact vary from 300 × 10-30
to 2300 × 10-30 esu at 1064 nm.29-31
Length of HC-n and Included Numbers of HC-n per
Channel (NC) of Silicalite-1 and ZSM-5 Films. The estimated
maximum lengths of the HC-n dyes listed in Table 1 were
estimated using the Chem Draw program. The increment was
about 0.12 nm per methylene (CH2) unit. The included number
of each HC-n molecule per channel (NC) after dipping a SL/G
in a 1 mM solution of each HC-n for 1 d is also listed in Table
1. As noticed, NC increased upon increasing n from 3 to 6 but
rapidly decreased upon further increasing n. When the film
thickness (∼400 nm) was taken into account, even the oc-
cupancy of silicalite-1 channels by HC-6 with the largest NC
was only ∼12% by length, which corresponds to 2 wt %. By
the same analogy, the channel occupancies by HC-18, HC-22,
and HC-24 were as small as 3%, 1%, and 1%, respectively, by
length, which corresponds to 0.3%, 0.1%, and 0.1%, respec-
tively, by weight. This indicates that only very small portions
of silicalite-1 channels are occupied by the included HC-n
molecules.
Interestingly, NC of HC-18 into Na-ZSM-5 films increased
with an increase in the Al content. For instance, whereas the
NC of HC-18 into an SL/G was 3.5 (Table 1), the corresponding
numbers into Na-ZSM-5 films (Si/Al ) 50 and 25) were 3.9
and 4.4, respectively. We attribute the above phenomenon to
the ion exchange-aided incorporation of the positively charged
HC-18 dye into the negatively charged channels by ion exchange
of Na+.
Figure 3. The UV-vis spectra of 0.02 mM HC-18 in methanol (A), HC-
18-SL/G coated with DMSO as the refractive index matching fluid (B),
and the aqueous methanol solution of HC-18 extracted from HC-18-SL/G
(C).
Irreversible Inclusion of HC-n into Silicalite-1 Channels.
The colorless SL/Gs slowly picked up a pink hue upon
immersion into methanol solutions of HC-n. The ready inclusion
of HC-n into SL/Gs was confirmed by the UV-vis spectra of
the DMSO-coated HC-n-SL/Gs, which were identical to the
spectrum of HC-n in methanol, as typically shown in Figure
3B for HC-18. In the case of HC-6 with the largest NC, we
monitored the amount of inclusion of the dye into a SL/G with
time (Figure 5). The result showed that the rate of inclusion
was high before 4 h but decreased significantly after 4 h despite
the fact that there are still enough rooms (>88%) in the channels.
Interestingly, if the dyes had once entered the silicalite
channels, they never came out of the channels into solution.
For instance, the intensities of the visible bands of HC-3-SL/G
and HC-18-SL/G at 479 nm did not decrease even after keeping
them in fresh methanol for 3 days, and the UV-vis spectra of
the supernatant solutions also did not show any indication of
leaching of HC-3 and HC-18 from silicalite-1 to methanol. The
above result shows that the silicalite-1 channel has a very strong
affinity toward HC-n dyes regardless of the chain length,
indicating that inclusion of HC-n into the silicalite-1 channel is
a nonequilibrium, irreversible process in methanol.
where g denotes the geometric factor, Csolvent and Cdye denote
the concentration of the solvent and the NLO dye, respectively,
âsolvent and âdye denote the â of the solvent and the NLO dye,
respectively, σin and σout denote the absorption cross sections
for incident beam and the SH, respectively, and lin and lout denote
the path lengths of the incident beam and the generated SH
within the solution, respectively. Based on eq 1, and the fact
that the solvent does not absorb the incident beam, each
relationship between Iout/Iin2 and concentration shown in Figure
4 can be expressed with the following general formula:
y ) (A + Bx) exp(-Cx)
(2)
2
where y and x denote Iout/Iw and Cdye, respectively, and A, B,
and C represent the parameters to be determined by curve fitting.
The obtained parameters for each curve are listed in the inset
of each panel of Figure 4. As noticed from the parameter values,
the obtained values for A are essentially zero, consistent with
the fact that the solvent used in this study (methanol) has a
negligible SHG activity (âsolvent ) (0.69 ( 0.07) × 10-30 esu).29
Therefore, by assigning the parameter A as zero, we determined
that the âHC-n/âPNA ratios were expressed in terms of the
Although silicalite-1 does not have an ion-exchange capacity,
it can be suspected that the HC-n dyes were included into the
silicalite-1 channels as a result of ion exchange of the cations
that might be present in the channels due to the presence of
inadvertently introduced aluminum sites. To test such a pos-
sibility, we dispersed HC-6-SL/Gs into a 1 M aqueous solution
of NaCl as a means to back exchange HC-6 with Na+, because
parameter B as (BHC-n/BPNA)
1/2, which were 22.3, 22.2, and 22.1
for n ) 6, 12, and 22, respectively. The fact that the obtained
ratios are very similar to each other indicates that the âHC-n
values are essentially identical regardless of n. By applying the
reported value of âPNA ((34.5 ( 4) × 10-30 esu),23,24 we
determined the âHC-n values to be (765 ( 89) × 10-30 esu at
1064 nm. The above âHC-n values are similar to the one reported
by Ashwell et al., which is (836 ( 358) × 10-30 esu at 1064
(30) Marowsky, G.; Chi, L. F.; Mo¨bius, D.; Steinhoff, R.; Shen, Y. R.; Dorsch,
D.; Rieger, B. Chem. Phys. Lett. 1988, 147, 420-424.
(31) Bubeck, C.; Laschewsky, A.; Lupo, D.; Neher, D.; Ottenbreit, P.; Paulus,
W.; Prass, W.; Ringsdorf, H.; Wegner, G. AdV. Mater. 1991, 3, 54-59.
(29) Duan, X.-M.; Okada, S.; Oikawa, H.; Matsuda, H.; Nakanishi, H. Mol.
Cryst. Liq. Cryst. 1995, 267, 89-94.
9
678 J. AM. CHEM. SOC. VOL. 126, NO. 2, 2004

Inclusion of Dyes into Silica Zeolite Films
A R T I C L E S
Table 1. Effect of Alkyl Chain Length (n) on the Optical Properties of HC-n and the Amount of Inclusion into Silicalite-1 Film of 400-nm
Thickness
a
g
h
h
i
j
n
λmax
ꢀb
âc
lengthd
N (×1014)e
NCf
Ipp
Isp (×10-4 g
)
d33
d31
d33/d31
d33/NC
s
DUA
3
6
9
12
15
18
22
24
475.5
479.0
479.0
480.0
479.5
479.5
479.0
479.0
39 515
43 355
43 775
42 460
45 160
44 340
43 180
42 320
1.7
2.1
2.4
2.8
3.1
3.5
4.0
4.2
4.64
15.56
10.37
5.70
4.30
2.59
6.4
23.1
15.4
8.2
5.7
3.5
0.3
0.1
1.6
3.8
7.0
7.9
0.9
1.9
1.12
0.50
2.25
3.59
4.99
5.30
1.71
2.57
0.18
0.02
0.15
0.63
0.88
1.51
1.90
2.14
0.08k
0.01k
0.09
0.20
0.38
0.66
0.91
0.95
769 ((89)
766 ((89)
0.5
2.2
1.7
3.0
0.4
0.9
0.02
0.04
0.04
0.05
0.02
0.03
113
90
125
106
86
0.97
0.97
0.98
0.98
0.97
0.97
761 ((88)
0.61
0.88
0.9
1.2
86
c
a In nm. b Molar extinction coefficient in MeOH, M-1 cm-1
.
×30-30 esu at 1064 nm in methanol. d In nm. e The total number of HC-n in a SL/G.
f Number of HC-n in each 400-nm long channel of silicalite-1 film. g In % with respect to that of a Y-cut 3-mm-thick quartz crystal. h In pm V-1 i The order
.
parameter. j The degree of uniform alignment of HC-n, defined by the experimental-to-theoretical ratio of d33 values, d33(E)/d33(T). k The average orientational
angle (θ ) 7.7°) was taken to derive the theoretical d33 values.
Figure 4. The plots of the intensities of the incoherent SH radiations (Iout) of PNA (A) and HC-6 (B), HC-12 (C), and HC-22 (D) with respect to the
concentrations of the dye.
the supernatant solutions showed that Na+ ion does not induce
extraction of HC-n (n ) 6, 18, and 22) from the silicalite matrix
to the solution.
In contrast to the case of HC-n-SL/Gs, 29% of the included
HC-18 was extracted from the ZSM-5 film (Si/Al ) 50) to the
solution upon immersion into 1 M NaClO4 in methanol solution
for 24 h. The result shows that most of the HC-n cations are
included into ZSM-5 channels via ion exchange.
I2ω/I2ω (quartz), d33, and d31 of HC-n-SL/G. The relative
SH intensities (expressed in % with respect to a Y-cut 3-mm-
thick quartz) are more specifically expressed as Ipp and Isp,
depending on the polarization direction of the incident beam,
where the former and the latter denote the intensities of p-
polarized SH beams generated from the p- and s-polarized
fundamental laser beam, respectively. The typical plot of Ipp
and Isp of HC-18-SL/G versus the angle of the incident beam
(with respect to the surface normal) is shown in Figure 6A. All
of the other HC-n-SL/Gs showed nearly identical profiles. The
completely destructive interferences in the profiles indicate that
the two silicalite-1 films on the opposite sides of glass plates
are nearly identical. The average angles at which the maximum
values of Ipp and Isp appeared were 67.4° and 55.9°, respectively.
Figure 5. Change of the absorbance of HC-6-SL/G at 479 nm with time
upon immersion of a SL/G into the methanol solution of HC-6 (1 mM).
the cationic dye has some solubility in water. However, the
UV-vis analysis of the supernatant solution confirmed that
HC-6 never came out of the silicalite film into the solution even
after stirring the solution for 24 h at room temperature. Similarly,
the HC-n-SL/Gs (n ) 6, 18, and 22) were also immersed in a
1 M methanol solution of NaClO4 for 24 h with an intention to
back exchange HC-n with Na+ in methanol in which the dyes
are quite soluble. In this case again, the UV-vis analyses of
9
J. AM. CHEM. SOC. VOL. 126, NO. 2, 2004 679

A R T I C L E S
Kim et al.
Figure 6. The plots of Ipp and Isp (inset) of HC-18-SL/G (in %, with respect to the corresponding value of Y-cut 3-mm quartz) versus the angle of the
incident beam (A) and the plots of d33 (B), d33/NC (C), and d33(E)/d33(T) (D) versus n.
The Ipp values ranged from 0.1% to 7.9%, while those of Isp
ranged from 0.4 × 10-4% to 3.0 × 10-4% (Table 1). The fact
that the observed Ipp values are ∼104 times larger than the Isp
values indicates that most of the dyes are positioned vertically
with the long axis perpendicular to the glass plane, consistent
with the aforementioned fact that most of the straight channels
are oriented vertically with respect to the glass plane (Figure
2).
In general, the Ipp values of HC-n-SL/Gs were relatively
smaller when n e 6 than when n g 9. In close relation to our
work, Laeri and the co-workers reported that they observed SH
signals from HC-2-loaded ALPO4-5 crystals during their study
of dye-incorporating zeolites as microlasers.32 Although they
did not report the quantitative SHG intensity, we expect that
the SH intensity of the HC-2-loaded ALPO4-5 crystals will be
similar to the value of HC-3-SL/G. Although the difference was
small, Ipp tended to decrease from 0.3 to 0.1 upon an increase
in n from 3 to 6 despite the fact that Nc increased by ∼4-fold.
This result suggests that HC-6 enters silicalite-1 channels with
a more random orientation than HC-3. However, Ipp progres-
sively increased from 0.1% to 7.9% with an increase in n from
6 to 18, despite the progressive decrease in NC. Interestingly,
consistent with the sharp decrease of NC, Ipp sharply decreased
from 7.9% to 0.9% and 1.9% upon a further increase in n from
18 to 22 and to 24. Although small, the Isp value also showed
a trend similar to that of Ipp with increasing n.
fact that DMSO can serve as a good index matching fluid, whose
refractive index is 1.48 at 20 °C. The values are listed in Table
1 and graphically displayed in Figure 6B. Like Ipp, the d33 value
initially decreased with an increase in n from 3 to 6, progres-
sively increased with an increase in n from 6 to 18, but sharply
decreased upon a further increase in n. However, d33/NC
progressively increased with an increase in n (Table 1 and Figure
6C). The above results are highly reproducible even with other
batches of SL/Gs. We propose that the above phenomenon arises
as a result of the increase in the tendency of the HC-n molecule
to enter the hydrophobic silicalite-1 channels with the hydro-
phobic tail first as the tail length increases.
The Ipp values observed from HC-18-loaded ZSM-5 films
were 1.7% and 0.5%, respectively. Obviously, the Ipp values
are much smaller than that of HC-18-SL/G (7.9%) despite the
fact that NC values are higher (3.9 and 4.4, respectively) in the
ZSM-5 films than in SL/G (3.5) (vide supra). One obvious factor
that caused the Ipp values to be lower than that of HC-18-SL/G
is poorer coverage of glass plates with b-oriented ZSM-5 films.
However, even after the lower coverage of ZSM-5 films (Si/Al
) 50) on glass plates (81%) is taken into consideration, the
resultant Ipp value is much lower than the expected value. This
phenomenon is attributed to the increase in the degree of
randomly oriented incorporation of HC-18 into the channels of
ZSM-5 due to the increase in the hydrophilicity of the channels
as a result of introduction of anionic centers in the framework
and the charge-balancing cations. On the basis of the above
observation, we rather concentrated on the aligned incorporation
of HC-n dyes into SL/Gs.
The two tensor components of quadratic nonlinear suscep-
tibility, d33 and d31 of HC-n-SL/Gs, were determined with
Maker’s fringe method25 by comparing the experimentally
observed relative Ipp and Isp values with the intensity of a 3-mm-
thick Y-cut quartz whose d11 is 0.3 pm V-1. For this, 1.48 was
used as the refractive index of silicalite-1 film based on the
Remarkably High Order Parameters of HC-n Dyes in SL/
Gs. The calculated d33/d31 ratios were always higher than 85,
and the average was 109, which is ∼2-5 times higher than
those of Langmuir-Blodget (LB) films of nonlinear optical
(NLO) dyes, which are ∼30-50 times higher than those of
poled polymers imbedded or grafted with NLO dyes, which are
(32) (a) Vietze, U.; Krauss, O.; Laeri, F.; Ihlein, G.; Schu¨th, F.; Limburg, B.;
Abraham, M. Phys. ReV. Lett. 1998, 81, 4628-4631. (b) Braun, I.; Ihlein,
G.; Laeri, F.; No¨ckel, J. U.; Schulz-Ekloff, G.; Schu¨th, F.; Vietze, U.; Weiss,
O¨ .; Wo¨hrle, D. Appl. Phys. B 2000, 70, 335-343.
9
680 J. AM. CHEM. SOC. VOL. 126, NO. 2, 2004

Inclusion of Dyes into Silica Zeolite Films
A R T I C L E S
usually less than ∼3.1 The estimated average tilted angle of
HC-n (θ) in the silicalite-1 channel based on the above very
high average d33/d31 ratio was 7.7°,33 indicating that the long
axes of the hemicyanine heads are tilted by the angle from the
direction of the straight channels of silicalite-1 film (surface
normal). The value is smaller than that of PNA in the straight
channels of Sb-SL (11°),20 and that of HC-18 in an LB film
(18-40°).4 The reason the θ value of the hemicyanine head is
smaller than that of PNA in silicalite-1 channels is conceivable
from the fact that the length of the former is much larger than
the length of the latter.
hydrophobic silicalite-1 channels with the hydrophobic tail first
as the tail length increases. The same proposal was made to
account for the progressive increase of d33/NC (vide supra). In
support of our proposal, it has been well established that
silicalite-1 channels are hydrophobic and hence prefer alkanes
to the relatively polar molecules such as alcohols.34
To be more certain, we also compared the included amounts
of HC-3 and n-octadecane, respectively, into silicalite-1 powders
after stirring for 3 h in each methanol solution under the
consideration that HC-3 and n-octadecane represent the hemi-
cyanine head and hydrophobic alkyl tail, respectively. Analyses
revealed that the included amounts of HC-3 and n-octadecane
were 0.01 mg (0.3 µmol) and 3.64 mg (143 µmol) per gram of
silicalite-1, respectively, showing that the included number of
n-octadecane was as much as 476 times larger than that of HC-3
within the period of 3 h, despite the fact that the estimated length
of n-octadecane (2.4 nm) is longer than that of HC-3 (1.7 nm).
In close relation to the above, we also found that the initial
rate of inclusion of HC-n into silicalite-1 powders from each
dilute methanol solution of HC-n progressively increased with
increasing n although the number of HC-n eventually included
into SL/G after 15 h progressively decreased with increasing n
as mentioned earlier. For example, when 20 mg of silicalite-1
powder (instead of SL/G) was immersed into each 10 mL aliquot
of a dilute methanol solution of HC-n (0.02 mM) for 1 h at
room temperature, the amounts of each HC-n included into
silicalite-1 powders were 52, 45, 75, 98, 105, 143, 162, and
165 nmol per gram of silicalite-1 for n ) 3, 6, 9, 12, 15, 18,
22, and 24, respectively.
The reversal of the order of included numbers of HC-n with
time suggests that the intrachannel migration of HC-n becomes
difficult as the alkyl chain length increases, unlike n-octadecane,
presumably due to the increase in the effective width of HC-n
as n increases. Coupled with the fact that the NC values of HC-
22 and 24 are close to 1, the increase in the difficulty of
intrachannel diffusion of HC-n with increasing n made us
suspect the possibility of HC-22 and HC-24 existing on the
surface of the silicalite-1 films as monolayers by sticking their
hydrophobic chains into the hydrophobic channels.
The comparison of the smallest and the largest DUA values,
that is, 0.01 for n ) 6 and 0.95 for n ) 24, revealed the
remarkable fact that the increase in n leads to as much as 95-
fold increases in DUA. As noticed, the estimated DUA value
for HC-24 came out close to 1. The value may be reduced to
∼60% if the largest reported average âHC-n value (2300 × 10-30
esu)29-31 is used for the estimation of d33(T). Despite the above
uncertainty in obtaining the exact value of DUA, we believe
that the d33(T) value of HC-24 matches very closely with the
experimentally observed value, considering the extreme dif-
ficulties in obtaining the absolute â values of NLO dyes and
the exact number densities of the HC-n dyes in silicalite-1 films.
Although the uncertainty has yet to be resolved, the fact that
the increment in DUA becomes very small upon increasing n
from 22 to 24 strongly supports that the DUA of HC-24 has
reached close to the maximum, that is, 1.
The degree of orientational order of nonlinear chromophores
can be quantified by the order parameter s which is expressed
by eq 3.
s ) [3 cos2 θ - 1]/2
(3)
From the derived average θ of 7.7°,33 the average order
parameter (s) of HC-n in SL/Gs becomes 0.97, which is
markedly higher than those of NLO dyes in poled polymers
whose typical values are ∼0.2.1,2 We believe that s of HC-n in
SL/Gs is the highest ever found among the SHG materials
comprised of organic NLO dyes.
Marked Increase in the Degree of Uniform Alignment
(DUA) of HC-n with Increasing n. We calculated the
maximum theoretical d33 values of HC-n-SL/Gs based on eq 3
by taking N of HC-n in each HC-n-SL/G, the average value of
â
HC-n (765 × 10-30 esu, or 3216 × 10-40 m4 V-1), the average
tilted angle of HC-n dyes with respect to the surface normal (θ
) 7.7°), and the microscopic local field correction l ) 1.40
into account, and assuming that all of the included HC-n dyes
were aligned with the alkyl tail part pointing to the glass
substrate.
The calculated theoretical maximum d33 values [d33(T)] were
14.40 (3), 70.71 (6), 24.48 (9), 18.13 (12), 13.31 (15), 8.03 (18),
1.88 (22), and 2.71 pm V-1 (24), respectively, for each n shown
in the parentheses. The experimental-to-theoretical ratio of d33
[d33(E)/d33(T)] is defined as DUA. The DUA values are listed
in Table 1 (last entry), and they are plotted against n in Figure
6D. As noticed, DUA generally increased with increasing n.
We propose that such a phenomenon arises as a result of the
increase in the tendency of the HC-n molecule to enter the
(33) To obtain the average tilted angle of HC-n dyes in the silicalite-1 channels,
we assumed that the long axes of the hemicyanine heads are tilted with an
average specific angle θ from the channel direction after inclusion into the
straight channels running perpendicular to the glass plane. d31 and d33 can
then be expressed in terms of âHC-n as eqs 4 and 5, respectively, by
assuming that âHC-n has only one dominant component â ) âêêê along the
long-axis direction ê (the direction of linear dipole moment), and the
microscopic local field correction l is independent of direction and frequency
(sample is isotropic and dispersionless),
d33 ) NâHC-n cos3 θ l3/2
(4)
d31 ) NâHC-n[ cos θ - cos3 θ ]l3/4
(5)
where N is the number density of HC-n (total number of HC-n within a
unit volume, which is 1 cm × 1 cm × 400 nm in our case, see Table 1)
and l is given by (n2 + 2)/3, where n is the refractive index of zeolite
which was estimated to be 1.48 (see text). The angle bracket indicates the
orientational average of the molecules. Because HC-n dyes should have a
very narrow distribution f(θ) of orientation, like f(θ) ) sδ(θ - θo), as a
result of confinement within the narrow straight channels of silicalite-1,
we have the following relationship (eq 6).
In an independent series of experiment, we also found that
Ipp of HC-18-SL/G increases as the loaded amount of HC-18
(34) (a) Piccione, P. M.; Laberty, C.; Yang, S.; Camblor, M. A.; Navrotsky,
A.; Davis, M. E. J. Phys. Chem. B 2000, 104, 10001-10011. (b) Eroshenko,
V.; Regis, R.-C.; Soulard, M.; Patarin, J. J. Am. Chem. Soc. 2001, 123,
8129-8230.
d33/d31 ) 2/tan2 θ ) 109
(6)
From the above equation, θ is derived to be 7.7.
9
J. AM. CHEM. SOC. VOL. 126, NO. 2, 2004 681

A R T I C L E S
Kim et al.
incorporating much larger numbers of HC-22 and HC-24 into
the channels, perhaps by employing thicker silica zeolites with
wider channels and uniform orientation. Furthermore, we found
that HC-n-SL/Gs have long-term and high thermal and me-
chanical stabilities, which most of the LB and poled polymer
NLO films are lacking.1,2 For instance, we found that HC-n-
SL/Gs retained their initial SHG activities even after keeping
them in the atmosphere for 1 year or in an oven at 120 °C for
24 h, and even after rubbing the surface with fingers many times.
Unlike AlPO4-5, ZSM-5, and Sb-SL powders loaded with
PNA, the SL/Gs loaded with PNA by vapor phase inclusion
(8%) did not exhibit SH activity, indicating the silicalite-1
channels have nearly the same affinities toward the nitro and
the amino groups as was the case to pure silica ZSM-12.21 This
result therefore represents the first example that shows the
symmetry breaking of a pure silica zeolite (free of Al or Sb) by
aligned inclusion of a dipolar NLO dye.
Figure 7. The parabolic relationship between Nc and the SHG intensity
(Ipp) of HC-18-SL/G.
increases. Thus, the observed Ipp values were 0.7%, 1.6%, 2.6%,
4.7%, and 7.5% for NC values of 1.2, 1.7, 2.2, 3.0, and 3.8,35
respectively. The result is also graphically shown in Figure 7.
As noticed, Ipp increases parabolically as NC increases. Knowing
the fact that SHG intensity is proportional to the square of the
number density of the NLO dye, the above result indicates that
HC-18 molecules enter the silicalite-1 channels with the same
DUA regardless of the degree of loading.
It is also important to note that, in the case of the dye-loaded
AlPO4-5 crystals, the loading of the dyes into AlPO4-5 should
exceed certain thresholds such as 3% for PNA and 13% for
MNA (by weight) for the composites to be SHG active. The
requirement of concentration thresholds indicates that the NLO
dyes have to undergo reorganization at high concentrations for
them to produce a net bulk dipole moment.14,15 In this respect,
the fact that HC-n-SL/Gs (n ) 22 and 24) exhibit SHG activities
even with 0.1% loading (by weight) reveals the fact that the
alignment of the dyes in the silcalite-1 channels is not a
secondary phenomenon but is determined right from the
beginning at the time of inclusion.
The d33 value of HC-18-SL/G (5.3 pm V-1) is more than 10
times larger than the d11 value of quartz (0.3 pm V-1), or the
d33 value of KDP (∼3 pm V-1), despite the fact that there are
only 3.5 molecules in each channel. However, the observed d33
values are much smaller than those observed from LB films,
which are in the range of 35 and 750 pm V-1 1,2,4-8
The reason
.
for the d33 values of HC-n-SL/Gs being smaller than those of
self-assembled molecular films is rather obvious considering
the fact that the density of HC-n dyes in 400-nm thick silicalite-1
films is very small, in particular, for those of long chain HC-n
dyes with high DUA values. For instance, in the case of HC-
24-SL/G, NC is essentially one. The comparison of the ratio of
the volumes occupied by HC-24 in HC-24-SL/G and HC-18 in
a LB film, that is, (1.35 nm2 × 400 nm)/(0.42 nm2 × 3.5 nm)
) 367, with the ratio of the corresponding d33 values, 2.6/750
) 1/300, reveals that the lower d33 value of HC-24-SL/G
originates from very low N. However, because the volume of
each silicalite-1 channel does not change even after inclusion
of larger numbers of HC-24 within each channel, we believe
that our methodology bears a great potential to be developed
into methods of preparing inorganic-organic composite NLO
films with very high d33 values by developing methods of
Overall, this report describes a method for the incorporation
of a dipolar NLO dye (hemicyanine) with a high â value into
a pure silica zeolite (silicalite-1) in high DUA by modification
of the NLO dye. The successful demonstration of the use of
zeolite films (rather than powders) as the inorganic hosts for
the aligned inclusion of hemicyanine dyes verifies the great
potential of the NLO-dye-including zeolites to be developed
into practically viable SHG materials.
Acknowledgment. We thank the Ministry of Science and
Technology (MOST) for supporting this work through the
Creative Research Initiatives (CRI) program.
Supporting Information Available: Experimental tables and
figures (PDF). This material is available free of charge via the
Internet at http://pubs.acs.org.
(35) The fact that this Nc value is higher than that shown in Table 1 is because
SL/Gs were produced from a different batch.
JA037772Q
9
682 J. AM. CHEM. SOC. VOL. 126, NO. 2, 2004