Supramolecular Organization of Dye Molecules in Zeolite L Channels: Synthesis, Properties, and Composite Materials

Article abstract of DOI:10.1002/chem.201504404

Sequential insertion of different dyes into the 1D channels of zeolite L (ZL) leads to supramolecular sandwich structures and allows the formation of sophisticated antenna composites for light harvesting, transport, and trapping. The synthesis and properties of dye molecules, host materials, composites, and composites embedded in polymer matrices, including two- and three-color antenna systems, are described. Perylene diimide (PDI) dyes are an important class of chromophores and are of great interest for the synthesis of artificial antenna systems. They are especially well suited to advancing our understanding of the structure-transport relationship in ZL because their core fits tightly through the 12-ring channel opening. The substituents at both ends of the PDIs can be varied to a large extent without influencing their electronic absorption and fluorescence spectra. The intercalation/insertion of 17 PDIs, 2 terrylenes, and 1 quaterrylene into ZL are compared and their interactions with the inner surface of the ZL nanochannels discussed. ZL crystals of about 500 nm in size have been used because they meet the criteria that must be respected for the preparation of antenna composites for light harvesting, transport, and trapping. The photostability of dyes is considerably improved by inserting them into the ZL channels because the guests are protected by being confined. Plugging the channel entrances, so that the guests cannot escape into the environment is a prerequisite for achieving long-term stability of composites embedded in an organic matrix. Successful methods to achieve this goal are described. Finally, the embedding of dye-ZL composites in polymer matrices, while maintaining optical transparency, is reported. These results facilitate the rational design of advanced dye-zeolite composite materials and provide powerful tools for further developing and understanding artificial antenna systems, which are among the most fascinating subjects of current photochemistry and photophysics. Rylenes in nanochannels: The synthesis and properties of organic dyes incorporated into zeolite L (ZL) crystals are described. The intercalation/insertion of 17 perylene diimides, two terrylene diimides, and one quaterrylene diimide into ZL is compared and their interactions with the inner surface of the ZL nanochannels are discussed (see figure).

Full text of DOI:10.1002/chem.201504404

DOI: 10.1002/chem.201504404
Full Paper
&
Host–Guest Systems
Supramolecular Organization of Dye Molecules in Zeolite L
Channels: Synthesis, Properties, and Composite Materials
Pengpeng Cao,[a] Oleg Khorev,[b] AndrØ Devaux,[a] Lucie Sägesser,[c] Andreas Kunzmann,[d]
Achim Ecker,[c] Robert Häner,[b] Dominik Brühwiler,[c] Gion Calzaferri,*[b] and Peter Belser*[a]
Abstract: Sequential insertion of different dyes into the 1D
channels of zeolite L (ZL) leads to supramolecular sandwich
structures and allows the formation of sophisticated antenna
composites for light harvesting, transport, and trapping. The
synthesis and properties of dye molecules, host materials,
composites, and composites embedded in polymer matrices,
including two- and three-color antenna systems, are de-
scribed. Perylene diimide (PDI) dyes are an important class
of chromophores and are of great interest for the synthesis
of artificial antenna systems. They are especially well suited
to advancing our understanding of the structure–transport
relationship in ZL because their core fits tightly through the
12-ring channel opening. The substituents at both ends of
the PDIs can be varied to a large extent without influencing
their electronic absorption and fluorescence spectra. The in-
tercalation/insertion of 17 PDIs, 2 terrylenes, and 1 quaterry-
lene into ZL are compared and their interactions with the
inner surface of the ZL nanochannels discussed. ZL crystals
of about 500 nm in size have been used because they meet
the criteria that must be respected for the preparation of an-
tenna composites for light harvesting, transport, and trap-
ping. The photostability of dyes is considerably improved by
inserting them into the ZL channels because the guests are
protected by being confined. Plugging the channel entran-
ces, so that the guests cannot escape into the environment
is a prerequisite for achieving long-term stability of compo-
sites embedded in an organic matrix. Successful methods to
achieve this goal are described. Finally, the embedding of
dye–ZL composites in polymer matrices, while maintaining
optical transparency, is reported. These results facilitate the
rational design of advanced dye–zeolite composite materials
and provide powerful tools for further developing and un-
derstanding artificial antenna systems, which are among the
most fascinating subjects of current photochemistry and
photophysics.
Introduction
guests, and hence, for influencing the properties of the result-
ing molecule–ZL composites. Experience, however, tells a differ-
ent story. The diversity of guest–ZL composites exhibiting
a large range of properties is impressive.[1–14] It emerges from
the fact that the variety of molecules that fit into these narrow
channels is large, and that confinement favors the formation of
sandwich-type structures (Scheme 1), including the possibility
for rational stopcock modification.[1,5,11,15–17] The size and mor-
phology of the ZL host crystals can be tuned from about
30 nm to several micrometers, and from elongated to disk-
shaped.[18–23] The composites can be organized into extended
ordered structures; thus transferring microscopic qualities to
the macroscopic scale.[24–29] The luminescence properties of
nanosized rare-earth ZL composites can be enormously influ-
enced by plugging the channel entrances with an imidazoli-
um-based stopcock.[17] Substitution of some of the charge-
compensating potassium cations by imidazolium cations
(IMZ+) can strongly influence the spectral characteristics of
pH-sensitive dyes.[30] A simple covalent modification pattern of
the inserted dyes allows not only control over the contact dis-
tance between the dyes, so that exciton interaction can be
switched on and off,[31–34] but also influences the environment
of neighboring dyes in a sandwich structure. The interaction of
a functional group of the guest with the ZL channel is not well
Molecules slipping through the 0.71 nm wide channel open-
ings of zeolite L (ZL), which widen to 1.26 nm at their largest
extension before closing again in a periodic manner, lose
much of their freedom of movement. Movement is essentially
reduced to 1D back and forth stumbling, if the length of the
molecules approaches or exceeds that of two unit cells (u.c.s).
The first conclusion would be that there are no options left for
fine-tuning of the characteristic properties of the inserted
[a] Dr. P. Cao, Dr. A. Devaux, Prof. P. Belser
Department of Chemistry, University of Fribourg
Ch. du MusØe 9, 1700 Fribourg (Switzerland)
E-mail: peter.belser@unifr.ch
[b] Dr. O. Khorev, Prof. R. Häner, Prof. G. Calzaferri
Department of Chemistry and Biochemistry
University of Bern, Freiestrasse 3, 3012 Bern (Switzerland)
[c] L. Sägesser, Prof. A. Ecker, Dr. D. Brühwiler
Institute of Chemistry and Biological Chemistry
Zurich University of Applied Sciences (ZHAW) 8820 Wädenswil (Switzerland)
[d] Dr. A. Kunzmann
Optical Additives GmbH, Flurweg 9
3073 Gümligen (Switzerland)
Supporting information and ORCID(s) from the author(s) for this article are
available on the WWW under http://dx.doi.org/10.1002/chem.201504404.
Chem. Eur. J. 2016, 22, 4046 – 4060
4046
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 1. Synthesis of dye–ZL composites. A) Sequential loading strategy,
leading to antenna systems with acceptor dyes in the middle of the chan-
nels. B) Schematic view of the hexagonal arrangement of the 1D ZL chan-
nels. C) Schematic illustration of different types of dye–ZL composites
(D=donor, A=acceptor): a) D–ZL; b) D,A–ZL sandwich, with the acceptor lo-
cated in the middle of the host; b’) A,D–ZL sandwich; b’’) random (D,A)–ZL;
and c) D1,D2,A-ZL sandwich. The colors indicate the locations of different
dyes.
understood. A major exception is the carbonyl group, which
binds to the zeolite extra-framework potassium cations, and is
thus responsible for dye stabilization in the channels.[35] The ZL
channels can accommodate more than one type of dye. They
can either be loaded one after another, forming a sandwich
dye–ZL composite, or simultaneously, leading to a composite
containing a random dye mixture. This study focuses on the
sandwich-type structure in which different dyes form distinct
domains, since the molecules are too large to pass each other
inside the channels. This sandwich dye–ZL composite structure
has proven to be invaluable for the realization of materials
showing efficient Fçrster resonance energy transfer (FRET)[36]
antenna characteristics.[30,37] We show in Scheme 1C typical
patterns, a, b, b’, b’’, and c, that can be realized by inserting
dyes into the channels of hexagonal ZL crystals. Dyes that
absorb light at a shorter wavelength are referred to as donors
(D), whereas those that absorb light at a longer wavelength as
acceptors (A). A system marked as dye–ZL contains only one
kind of dye. A composite containing two dyes is referred to as
dye2,dye1–ZL, in which the sequential insertion order is indi-
cated by the ordering of the dye names from right to left. In
this case, dye1 would have been inserted before dye2. Crystals
of about 500 nm in length and diameter are convenient for
many applications and have been chosen because their size
meets the criteria that must be respected for the preparation
of antenna composites for light harvesting, transport, and trap-
ping. They consist of about 67000 strictly parallel channels,
each of which is formed by 666 u.c.s of 0.75 nm in length.
There is space in each channel for a maximum of 333, 222, or
200 molecules of HR, DMP, or m-DXP, respectively (Scheme 2).
High dye loadings have been realized in many cases. The maxi-
mum occupation numbers are, however, rarely reached, mainly
for thermodynamic and kinetic reasons. We distinguish be-
tween one-, two-, and three-dye composites, but we do not
discuss more complex patterns. Acceptors are distributed ran-
domly inside the crystals in composite b“ (Scheme 1). Sequen-
tial insertion for synthesizing sandwich-structure composites is
simple, in principle. However, finding the right loading condi-
tions can be more difficult.[37] Perylene-3,4:9,10-tetracarboxylic
Scheme 2. Top: Comparison of the length of three dyes, HR (1.5 nm), DMP
(2 nm), and m-DXP (2.4 nm), with the u.c. length of ZL. The two yellow verti-
cal lines mark the 1.5 nm length of two u.c.s. Bottom: Illustration of the van
der Waals surfaces of tb-DXP (2.7 nm; upper) and ADE-XP (3.1 nm; lower)
with the 12-ring channel opening of ZL. Numbers in parentheses are the
van der Waals lengths of the molecules (see Tables 1 and 2 for structures).
diimides (PDIs; Table 1) represent classical structure types of
colorants that have attracted tremendous interest in funda-
mental and industrial research.[38]
Assembling these dyes in a precise multichromophore
supramolecular scaffold with DNA has been investigated,[39]
and has recently led to the discovery of two homochromo-
phoric H-aggregates.[40] Insertion of PDIs into the channels of
ZL in different laboratories resulted in a variety of composites
with remarkable properties.[1,6,8,23,30,31] Most of our PDI–ZL ex-
periments have been performed with DXP and tb-DXP. These
dyes cannot be inserted at room temperature. Their largest
van der Waals diameter is about 0.76 nm, and therefore, ex-
ceeds the 0.71 nm van der Waals diameter of the ZL channel
opening.[41] Therefore, insertion is performed under vacuum
conditions at about 2008C. At this temperature, the breathing
vibrational modes of the 12-ring channel entrance of ZL are
fully activated[42] and the PDIs can slip through. In zeolite no-
menclature, a 12-ring consists of 12T atoms (T=Si, Al) plus 12
bridging oxygen atoms.[41,42] Conditions and structural details
that enhance or hinder insertion and transport in ZL channels
are, however, not well understood. We therefore report a com-
parison between 17 differently substituted PDIs, two terrylene
diimides (TDIs), and a quaterrylene diimide (QDI; Tables 1 and
2). PDIs are especially well suited for studying the structure–
transport relationship of dyes in ZL because their core fits
tightly through the 12-ring channel opening and the substitu-
ents at both ends of the dyes can be easily varied without sig-
nificantly influencing their electronic absorption and fluores-
cence spectra. Sequential insertion of different dyes leads to
sandwich structures, and hence, allows the formation of so-
phisticated antenna composites for light harvesting, transport,
and trapping. Herein, we describe the synthesis of a rationally
Chem. Eur. J. 2016, 22, 4046 – 4060
www.chemeurj.org
4047
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 1. Insertion of PDIs into the channels of ZL. The insertion temperature was 2608C and the zeolite host, ZL, was of barrel type, unless stated other-
wise. The absorption spectra of the PDIs are very similar, with a molar extinction coefficient of about 88200mÀ1 cmÀ1 at 525 nm in CH2Cl2.
Structure
Name/loading[a] [%]
Structure
Name/loading[a] [%]
tdc-XP/29 (40)
dmpa-XP/43 (67)
(1808C)
(2208C)
DXP/48 (67)
ZLTS
dm-DXP/57 (67)
m-DXP/41 (67)
ip-DXP/51 (67)
tb-DXP/48 (67)
ZLTS
b-DXP/17 (67)
(dmp)-DXP/31 (67)
bone-DXP/40 (80)
o-bone-DXP/40 (80)
(2508C)
(2508C)
DEXP/19 (67)
ADE-XP/<2 (53)
dm-XP/19 (67)
(3008C)
DIXP/<3 (73)
tb-XP/ZLTS
b-XP/ZLTS
decomp (67)
decomp (67)
[a] The effective and target loadings are given as percentages, the latter in parenthesis, with respect to the theoretical maximum loading.
Table 2. Dyes used for the preparation of different dye–ZL composites. The solvent was CH2Cl2, unless stated otherwise. The ZL host was of barrel type.
Structure
Name/loading[a] [%]
Absorption l [nm] (e [dm3 molÀ1 cmÀ1])
Emission l [nm]
dmpa-XT/0 (2808C)
653 (106000) 600 (55796) 555 (7053)
672 (exc. 620)
tb-DXT/4.5 ( 7.5)
(2708C)
653 (106000) 600 (55796) 555 (7053)
764 (155000) 708 (72272)
527 (20000)
673 (exc. 620)
828 (exc. 690)
566 (exc. 490)
tb-DXQ/5 (17)
(2708C)
HR/60 (2008C)
(0.5IMZ+/u.c.)
DMP/60 (2008C)
363 (41420)[b]
430 (exc. 340)
[a] Loadings, and target loadings in parentheses, are given in terms of 4.5 u.c./site for tb-DXT, 5 u.c./site for tb-DXQ, and 3 u.c./site for HR and DMP. [b] The
solvent was 1-butanol.
designed library of substituted PDIs and TDIs, as well as a QDI,
and correlate the substituent type with insertion behavior and
properties of the resultant composites, including two- and
three-color antenna systems.
Chem. Eur. J. 2016, 22, 4046 – 4060
www.chemeurj.org
4048
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
It is convenient to describe the loading of a dye–ZL compo-
site in terms of the occupation probability or loading, p, which
is defined as the ratio between the numbers of occupied sites
divided by the total number of sites available. The value of p
ranges from zero for an empty ZL to one for a fully loaded ZL.
A site expresses the number of u.c.s occupied by a guest. The
dye concentration in terms of molar volume as a function of p
is expressed by Equation (1), in which nS is the number of u.c.s
that can be occupied by a guest:[1]
Scheme 3. General synthesis of PDI dyes. Reagents and conditions:
a) Zn(OAc)2, imidazole/quinoline/diethanolamine, 190–2508C, 4–48 h, 7–95%
yield.
1408C, 24 h, 8%), tb-XP (diethanolamine, 1808C, 6 h, 78%),
m-DXP (Zn(OAc)2, quinoline, 2308C, 4 h, 73%), and b-XP
(Zn(OAc)2, quinoline, 2308C, 4 h, 95%).
ꢀ
ꢁ
p
nS
mol
L
The following PDI dyes were prepared with amines de-
scribed in literature. In parentheses are given the condition for
the last step of the synthesis (dye formation). tdc-XP (Zn(OAc)2,
imidazole, 1908C, 4 h, 73%),[44] tb-DXP (Zn(OAc)2, quinoline,
2308C, 4 h, 66%),[45,46] and dmpa-XP (Zn(OAc)2, imidazole,
1908C, 24 h, 8%).[47] The following series of reaction schemes
show the preparation of PDI dyes with amines that are insuffi-
ciently or not described in the literature.
cðpÞ ¼ 0:752
ð1Þ
The numbers in the composite names, such as dye–ZL.31, indi-
cates the dye loading p. In a sandwich composite of two dyes,
we write dye1,dye2–ZL.05,.12, in which .05 refers to dye2 and
.12 to dye1. The default 3.6 charge-compensating cations per
u.c. of ZL, which can be replaced by means of cation exchange,
are K+. We often replace some of them by the organic cation
IMZ+.[43] The name of such composites is, for example,
dye1,dye2–ZL.05,.12(0.5IMZ+), which means that 0.5K+ per
u.c. have been replaced by IMZ+.
Compound 2 (Scheme 4A) was reduced with zinc powder in
combination with CaCl2 in EtOH to give 3 in 67% yield. Com-
pound dm-DXP was then prepared in 32% yield through the
condensation of 3 with 1, by using Zn(OAc)2 in quinoline.[48] In-
terestingly, this was the only PDI derivative of the series that
readily formed crystals (CH2Cl2, concentrated solution, RT).
Results and Discussion
Synthesis of dyes
The preparation of the investigated dyes can be divided into
the following classes (Tables 1 and 2): 1) commercially available
dyes, 2) PDIs, 3) TDIs, and 4) QDIs.
Commercially available dyes
The following commercially available dyes were investigated
herein: 2,9-bis(2,6-dimethylphenyl)anthra[2,1,9-def:6,5,10-d’e’f’]-
diisoquinoline-1,3,8,10(2H,9H)tetraone (DXP), 2,9-bis(3,5-dime-
thylphenyl)anthra[2,1,9-def:6,5,10-d’e’f’]diisoquinoline-1,3,8,10-
(2H,-9H)-tetraone (Pigment Red 149, for which we have used
the abbreviation dm-XP), 14H-anthra[2,1,9-mna]thioxanthen-
14-one (Hostasol Red, for which we have used the abbreviation
HR), and 1,4-bis(4-methyl-5-phenyloxazol-2-yl)benzene (DMP).
PDI dyes
Most PDI dyes were prepared by applying a simple condensa-
tion reaction in which perylene-3,4,9,10-tetracarboxylic dianhy-
dride (1) was treated with appropriate amines in quinoline or
imidazole as solvent with zinc acetate as a catalyst at 180–
2408C for several hours under an inert atmosphere, and the
yields were in the range of 7 to 95% (Scheme 3). Therefore,
the main challenge was the preparation of the corresponding
amines.
Scheme 4. A) Synthesis of dm-DXP. Reagents and conditions: a) Zn, CaCl2,
EtOH, 758C, 16 h, 67%; b) 1, Zn(OAc)2, quinoline, 2408C, 48 h, 32%. B) Syn-
thesis of (dmp)-DXP. Reagents and conditions: a) Boc2O (Boc=tert-butyloxy-
carbonyl), EtOH, RT, 48 h, quant.; b) Pd(OAc)2, Ph3P, K2CO3, THF/H2O (2:1),
808C, 24 h, 91%; c) 4m HCl/dioxane, H2O, 908C, 1 h, 32%; d) 1, Zn(OAc)2,
quinoline, 2408C, 48 h, 13%. C) Synthesis of b-DXP. Reagents and conditions:
a) Butyric acid, polyphosphoric acid (PPA), 1758C, 3.5 h, 33%; b) Et3SiH,
CF3CO2H, RT, 2 h, 85%; c) 1, Zn(OAc)2, quinoline, 2408C, 48 h, 19%.
The PDI dyes prepared from commercially available amines
were as follows (reaction conditions are given in parentheses):
DXP (Zn(OAc)2, quinoline, 2308C, 4 h, 73%), DEXP (Zn(OAc)2,
quinoline, 2208C, 24 h, 47%), DIXP (Zn(OAc)2, imidazole,
Chem. Eur. J. 2016, 22, 4046 – 4060
www.chemeurj.org
4049
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
The amino group of 4-bromo-2,6-dimethylaniline (4) was
protected with Boc (Scheme 4B) by using Boc2O in EtOH to
give 5 quantitatively. The latter was used in a Suzuki coupling
protocol with 2,6-dimethylphenylboronic acid (6) to produce
biphenyl derivative 7 in 91% yield. Standard Boc removal con-
ditions (trifluoroacetic acid (TFA) in CH2Cl2) proved inefficient,
so compound 7 was deprotected by using 4m HCl in dioxane
at 908C to give 8 in 32% yield. Compound (dmp)-DXP was
then prepared in 13% yield through the condensation of 8
with 1 by using Zn(OAc)2 in quinoline. Compound 9 was acy-
lated with butyric acid in PPA (Scheme 4C), which produced
10 in 33% yield. The latter was reduced with triethylsilane in
TFA to give 11 in 85% yield. Compound b-DXP was then pre-
pared in 19% yield through the condensation of 11 with 1, by
using Zn(OAc)2 in quinoline.
Compound 12 (Scheme 5A) was nitrated by using a combi-
nation of fuming nitric acid/acetic anhydride in acetic acid to
yield a mixture of regioisomers 13 and 14 in 71% yield. The Rf
values of the regioisomers were too similar for separation by
flash chromatography; therefore, the mixture was reduced
with zinc powder in ethanol. The resultant aniline derivatives
15 and 16 give, after purification, compound 15 in 12% yield.
Compound ip-DXP was then prepared in 7% yield through the
condensation of 15 with 1, by using Zn(OAc)2 in quinoline.
Compound ADE-XP (Scheme 5B) was prepared by the conden-
sation of 17 with 18 at 2008C for 4 h to give 19 in 48% yield.
Subsequent condensation of 19 with 1 by using zinc acetate
as a catalyst in quinoline at 2208C gave ADE-XP in 13% yield.
The amino group of 4 was protected with Cbz (Scheme 5C)
by using CbzCl and NaHCO3 in THF to give 20. The coupling of
20 and 3-butyn-2-ol with [Pd(PPh3)2Cl2] and CuI in THF/Et3N
gave 21 in 15% yield. The Cbz protecting group of 21 was
cleaved and the triple bond was fully reduced simultaneously
with H2 and Pd/C in MeOH to give 22 in 76% yield. Compound
22 was then condensed with 1 by using Zn(OAc)2 in quinoline
to give 23 in 12% yield. The diol was then oxidized to give di-
ketone bone-DXP in 25% yield by using PCC in CH2Cl2. Com-
pound o-bone-DXP (Scheme 5D) was prepared through the
condensation of 4-amino-3,5-dimethylphenol (24) with 1 by
using Zn(OAc)2 in quinoline to give 25 in 60% yield. Subse-
quent alkylation with chloroacetone by using K2CO3 and KI in
DMF gave o-bone-DXP in 57% yield. The synthetic procedures
for tdc-XP, dmpa-XP, b-XP, tb-XP, m-DXP, tb-DXP, DIXP, and
DEXP are described in the Supporting Information.
Scheme 5. A) Synthesis of ip-DXP. Reagents and conditions: a) fuming HNO3,
Ac2O, acetic acid, RT, 2 h, 71%; b) Zn, CaCl2, EtOH, 758C, 16 h, 12%; c) 1,
Zn(OAc)2, quinoline, 2408C, 48 h, 7%. B) Synthesis of ADE-XP. Reagents and
conditions: a) 2,6-diethylaniline (17; commercially available), 1-bromoada-
mantane (18; commercially available), 2008C, 4 h; b) 1, Zn(OAc)2, quinoline,
2208C, 24 h, 13%. C) Synthesis of bone-DXP. Reagents and conditions:
a) benzyl chloroformate (CbzCl), NaHCO3, THF, RT, 8 h, 78%; b) 3-butyn-2-ol,
[PdCl2(PPh3)2], CuI, THF/Et3N, 558C, 2 d, 15%; c) H2, Pd/C, MeOH, RT, 2 d,
76%; d) 1, Zn(OAc)2, imidazole, 2408C, 48 h, 12%; e) pyridinium chlorochro-
mate (PCC), CH2Cl2, RT, 5 h, 25%. D) Synthesis of o-bone-DXP. Reagents and
conditions: a) Zn(OAc)2, quinoline, 2508C, 48 h, 60%; b) chloroacetone,
K2CO3, KI, DMF, 808C, 12 h, 57%.
TDI dyes
The TDI dyes are more challenging to synthesize.[49] An appro-
priate dianhydride precursor, such as 1 for the preparation of
PDI dyes, is missing. Therefore, a step-by-step procedure to
construct the TDI core was the strategy used to obtain the de-
sired dyes, tb-DXT and dmpa-XT.
The aniline derivative (26; Scheme 6A)[45,46] was condensed
with 1 by using Zn(OAc)2 in quinoline at 2308C for 4 h to give
tb-DXP in 66% yield. For the synthesis of the TDI derivative tb-
DXT, first, 1,8-naphthalic anhydride (32) was condensed with
26 in acetic acid at reflux to give 27 in 56% yield. Then com-
pound 26 was reacted with 1 in the presence of water, Zn(O-
Ac)2·2H2O, and imidazole, leading to partial decarboxylation–
condensation to yield the monoimide derivative 28. The latter
was then fused with 27 by using tBuONa/DBN in diglyme[49] to
give tb-DXT in 20% yield. Compounds 31 and 32 were cou-
pled at 1508C to give 33 in 80% yield. In a similar reaction,
under catalytic conditions in an imidazole/water mixture at
1908C for 24 h, compounds 1 and 31 coupled to give 34 in
52% yield. Final coupling between 33 and 34 in toluene with
DBN/tBuONa at 1308C for 3 h gave dmpa-XT in 8% yield.
Chem. Eur. J. 2016, 22, 4046 – 4060
www.chemeurj.org
4050
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
The 17 dyes we have investigated fall into 4 categories. In the
first, we have two examples (tdc-XP and dmpa-XP; Table 1)
with aliphatic groups attached to nitrogen. Both can be insert-
ed at 180 and 2208C, but decompose at higher temperatures.
We distinguish between the target loading, which refers to the
composition of the reaction mixture, and the effective loading,
which refers to the final product. The fact that tdc-XP enters
the zeolite better than one might expect can be explained by
assuming that the aliphatic C7 tails coil in a manner that
allows them to slip through the narrow channel opening. The
most important class consists of 9 dyes with methyl groups at
the 2- and 6-positions of the imide phenyl groups, which pre-
vent strong p-stacking. All 9 of them can be inserted at 2608C
with an effective to target loading ratio higher than 0.6, with
the exception of b-DXP and (dmp)-DXP, which enter more re-
luctantly. The 2 dyes with carbonyl groups at both ends (bone-
DXP and o-bone-DXP) enter more easily and are therefore in-
serted at 2508C. The use of ethyl at the 2- and 6-positions in-
stead of methyl still allows the DEXP to enter, but it seems
that the molecule becomes too stiff if adamantyl is added, so
that ADE-XP cannot be inserted in a significant amount. It is
less surprising that DIXP, with isopropyl at the 2- and 6-posi-
tions, cannot be inserted. The final category consists of PDIs
without substituents at these positions, namely, tb-XP, b-XP,
and dm-XP. These dyes lack the 2- and 6 methyl groups, and
therefore, form stable aggregates as a result of strong p-stack-
ing. Only dm-XP can be inserted to a significant amount. How-
ever, a higher temperature is needed, which it appears to toler-
ate under the applied vacuum conditions. The other two dyes
decompose at 2608C either at the surface or inside of the
host. We investigated the ratio of effective loading to target
loading for values of 2, 40, 60, and 80% for o-bone-DXP at
2508C. The result was always 0.5. This means that we do not
reach the saturation region and a higher loading is possible if
Langmuir behavior is assumed, which is reasonable. This be-
havior has been observed for the insertion of cationic dyes
through cation exchange.[1] If we correlate the ratio of effective
loading against target loading (each 50%) in the series dm-
DXP, ip-DXP, tb-DXP, m-DXP, and b-DXP, we find values of 0.86,
0.76, 0.72, 0.62, and 0.26. This leaves us with the impression
that the aliphatic character of the substituted phenyl ring is
crucial for the movement of dyes on the external ZL surface
and in the channels. Butyl-substituted b-DXP does not seem to
follow this line for reasons we do not yet understand. To fur-
ther investigate the structure to loading behavior, we tested
two TDIs and one QDI. We found that tb-DXT could be inserted
at 2708C with an effective loading to target loading ratio of
0.6. The value for tb-DXQ was 0.3, whereas dmpa-XT could not
be inserted (Table 2). This means that 1) the tert-butyl (tb)-tail
supports transport within the channels, 2) aliphatic groups di-
rectly attached to nitrogen are less favorable, and 3) the length
of the molecules is less critical than one might expect. To
better understand the loadings of tb-DXT and tb-DXQ, it is
useful to realize that they correspond to 6 or 7 molecules per
channel and approximately 440000 dye molecules per ZL crys-
tal. Synthesizing ZL may sometimes be an obstacle for using
this versatile material as a host. We have therefore not only
Scheme 6. A) Synthesis of tb-DXP, tb-DXT, and tb-DXQ. Reagents and condi-
tions: a) Zn(OAc)2, quinoline, 2408C, 48 h, 56%; b) 32, CH3CO2H, reflux, 24 h,
56%; c) 1, Zn(OAc)2·2H2O, H2O, imidazole, 2458C, 4 d, 88%; d) tBuONa, 1,5-
diazabicyclo[4.3.0]non-5-ene (DBN), diglyme, 1558C, 24 h, 20%; e) Br2, chlor-
obenzene, 508C, 4 h, quant.; f) [Ni(cod)2], 2,2’-bipyridine (bipy), 1,5-cyclooc-
tadiene (cod), DMF, 708C, 48 h, 53%; g) KOH, d-glucose, EtOH, 1208C, 3 h,
85%. B) Synthesis of dmpa-XT. Reagents and conditions: a) 1, 31, Zn(OAc)2,
imidazole/H2O, 1908C, 24 h, 52%; b) 31, 32, Zn(OAc)2, imidazole, 1508C,
4.5 h, 80% for 33; c) 33, 34 in dry toluene, tBuONa, DBN, 1308C, 3 h, 8%.
QDI dye
The QDI derivative (tb-DXQ) was prepared by using a method
developed by Müllen et al. (Scheme 6A).[50,51] Bromination of
28 gave compound 29 and subsequent Yamamoto homocou-
pling[52] with [Ni(cod)2] and bipy gave biaryl 30, which was
then cyclized to give tb-DXQ with KOH[53] in 85% yield.
Insertion of dyes into the ZL nanochannels
Neutral dyes are inserted into the channels of ZL by dissolving
them in a volatile solvent, usually CH2Cl2, and then soaking the
crystals in this solution. The sample is dried under vacuum to
remove the solvent and water. The dyes adsorb on the ZL crys-
tal surfaces during this process. The evacuated glass container
is sealed and kept for the required time (hours to days) at
a temperature at which the dyes become sufficiently mobile.
The temperature regime applied to the PDIs ranges from 180
to 3008C, depending on their mobility and thermal stability.
Chem. Eur. J. 2016, 22, 4046 – 4060
www.chemeurj.org
4051
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
tested ZL, but also the commercially available product ZLTS.
The latter was found to be useful for basic test experiments. If
purity, morphology, size, and size distribution are important,
then it is advantageous to use ZL crystals prepared according
to one of several published methods.[18–23] Apart from a relative-
ly small bathochromic shift, it is a general observation that the
shape of the absorption and fluorescence spectra of molecules
inside the ZL channels changes little compared with the shape
in solution. In many cases, the vibrational structure seen in so-
lution is well preserved in the dye–ZL composites, whereas in
other examples some broadening is seen.[1,54] Larger spectral
changes are often caused by saturation effects, self-absorption
of highly loaded samples,[23] and J-aggregate coupling.[31–34]
The spectra of DMP, tb-DXP, and HR shown in Figure 1 fit well
into this general pattern. Although some broadening and red-
shifting is seen in the spectra of DMP–ZL and HR–ZL, the shift
and broadening of the tb-DXP–ZL spectrum is small and the vi-
brational structure is preserved.
Figure 2. Optical fluorescence microscopy images of 5 mm long ZL crystals
partially loaded with dyes: top: DMP–ZL, middle: m-DXP–ZL, bottom: HR–
ZL. The images on the left were observed without a polarizer; the middle
row when applying a horizontally oriented polarizer, as indicated by the
double arrow; and in the right row the polarizer was oriented perpendicular-
ly.
These three dyes were chosen for the construction of
a three-dye antenna cascade because they showed perfect
spectral overlap (Table 3), which is a requirement for a good
antenna. Their length and shape is such that they align inside
the ZL channel with their S0–S1 electronic transition moment
parallel to the channel axis, which is the best possible situation
for efficient FRET.[36,55]
The same result as that shown for m-DXP–ZL was observed for
DXP–ZL, dm-XP–ZL, and tdc-XP–ZL. Thus, we can safely
assume that it applies to all PDI–ZL composites reported in
Table 1. The crystals were only partially loaded, so that the
dyes did not have time to diffuse into the middle part of the
long crystals. We therefore see high intensity at both ends of
the crystals and a dark part in the middle.
This was tested by inserting the dyes into ZL crystals of
about 5 mm in length, which allowed the optical dichroism to
be measured by means of fluorescence microscopy. The results
of such measurements (Figure 2) demonstrate the anisotropy.
The absorption and fluorescence spectra of all PDIs in de-
gassed CH2Cl2 and those of the PDI–ZL composites change
very little, and the measured extinction coefficients are very
similar. Moreover, all show high fluorescence quantum yields in
solution and as dye–ZL composites, with the exception of o-
bone-DXP, which has a quantum yield about 10 times lower in
solution and in ZL. This matches observations by Langhals
et al.,[56] who reported that a similar para-methoxy-substituted
dye showed a much lower quantum yield than that of DXP.[23]
In Figure 3, the absorption spectra of tb-DXP–ZL, tb-DXT–ZL,
and tb-DXQ–ZL are compared. The spectra in wavelengths
(Figure 3, top) give the impression that spectral changes are
larger for tb-DXT and tb-DXQ than those for tb-DXP. This
Table 3. Spectral overlap, Jꢀn, and Fçrster radius, R0, data for the dyes dis-
cussed herein.
Donor/Acceptor
Jꢀn [cm3 mÀ1
]
R0 [nm]
(k2 =4)
R0 [nm]
(k2 =1)
DMP (homo)
DMP/tb-DXP
DMP/HR
tb-DXP (homo)
tb-DXP/HR
HR (homo)
5.4”10À15
4.1
6.8
5.8
7.4
7.0
5.6
3.2
5.4
4.6
5.9
5.6
4.4
1.15”10À13
4.53”10À14
2.03”10À13
1.37”10À13
3.82”10À14
Figure 1. Absorption (c) and fluorescence (d) spectra of dyes in solution and those of dye–ZL composites. Left: DMP (1-butanol) and DMP–ZL.2. Middle:
tb-DXP (CH2Cl2) and tb-DXP–ZL.0.002 (dark grey). Right: HR (CH2Cl2) and HR–ZL.1. The spectra of the dye–ZL composites were measured as oil–glass sandwich
(OGS)[23] layers and are indicated as ZL.
Chem. Eur. J. 2016, 22, 4046 – 4060
www.chemeurj.org
4052
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
channel entrances to customize its properties or just for seal-
ing, preventing guests from leaving the channels.[1,5,60,61]
Sealing the channels for molecules such as DXP, rare-earth
complexes, or HR was achieved by adsorbing branched poly-
ethyleneimine (PEI) on the ZL surface.[30,62,63] This procedure is,
however, less specific than modification with trialkoxysilyl and
fine-tuning is not supported. It also has the effect of increasing
the size of the particles substantially because it cannot be ap-
plied selectively on the base and also because multilayer coat-
ing is necessary. It was found to hinder dispersibility in organic
matrices.
We have observed that the grafting of trialkoxysilanes onto
ZLs, either in a stopcock manner on the base or for surface
modification on the coat, leads to superior properties and
makes them more versatile for handling as functional compo-
site particles. Modification of ZL surfaces with trialkoxysilyl
agents has been used to synthesize oriented monolayers of
zeolite crystals on different substrates.[24–28] We have therefore
tested the molecules collected in Table 4 for surface modifica-
tion. The stopcock modification reaction is explained for APTES
in Scheme 7. APTES adsorbs specifically at the basal surface of
the ZL crystals at the acidic[1,30] channel entrances, which en-
courages protonation of the amine, and hence, its entering
more deeply into the channel. It then binds covalently through
condensation with an OH group at the surface. It remains diffi-
cult to experimentally determine details and the number of co-
valent siloxane bonds that fix the stopcock head to the chan-
nel entrance.
Figure 3. Absorption spectra of tb-DXP (blue), tb-DXT (green), and tb-DXQ
(red) as 5”10À6 m solutions (d) in CH2Cl2 and as dye–ZL composites
(c), given in wavelengths (top) and wavenumbers (bottom).
wrong impression is corrected when plotting the spectra in
wavenumber (Figure 3, bottom). The characteristic vibronic
structure of PDI dyes, also present in TDI and QDI dyes, has
been used to analyze their interaction with the inner surface of
the ZL nanochannels by comparing the positions of the 0–0’,
0–1’, 0–2’, and 0–3’ transitions, which can be identified in
dilute solutions of these dyes. This allows us to draw the con-
clusion that the energy difference, DE[n’À(n’+1)], between
the 0–n’ and 0–(n’+1) transitions is the same in solution and
in ZL. These observations can be interpreted in terms of
a weak interaction of the chromophoric core of these dyes
with the inner surface of the ZL nanochannels (for details, see
the Supporting Information). We have observed, however, that
the cocations exert a stronger influence on the fluorescence
spectra of tb-DXT and tb-DXQ than that seen for the PDIs; this
demands a dedicated investigation.
Based on extensive theoretical studies, we know that AlÀOH
groups at the entrances are the preferred modification sites
and that bipodally modified entrances are possible, but may
suffer from strain, whereas tripodal modifications can be ruled
out.[64]
Table 4. Molecules for stopcock and surface modification.
(3-aminopropyl)triethoxysilane (APTES)
(3-cyanopropyl)triethoxysilane (CPTES)
Stopcock and surface modification
The surface hydroxyl groups of zeolites can be used for the
modification and fine-tuning of the particle properties. The
grafting of silane coupling agents with trialkoxysilyl or trichlor-
osilyl groups to zeolites and mesoporous materials has been
investigated extensively.[13,25,57–60] As a consequence of the hex-
agonal structure of ZL crystals, which can be approximated by
cylinders, the coat and base surfaces have different reactivity. It
is therefore possible to selectively modify them. Selective
modification has been used to organize zeolite microcrystals
into 2D functional entities, such as monolayers, multilayers,
and patterned monolayers, on various substrates.[24–28] It has
also been used to selectively modify the base that bears the
(3-methacryloxypropyl)trimethoxysilane (MPTMS)
(n-octadecyl)triethoxysilane (C18TES)
1-methyl-3-[2-(triethoxysilyl)ethyl]-1H-imidazol-3-ium (MeImiSil)
1-butyl-3-[2-(triethoxysilyl)ethyl]-1H-imidazol-3-ium (BuImiSil)
Chem. Eur. J. 2016, 22, 4046 – 4060
www.chemeurj.org
4053
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
one month, whereas absolutely no leakage was observed for
the stopcock-plugged samples. We know, however, that APTES
permits the entry of small molecules, such as water, acetoni-
trile and methanol. Therefore, more complete sealing is also
achieved for small molecules by using the imidazolium-based
stopcock molecules MeImiSil and BuImiSil (Table 4).
Scheme 7. Insertion of APTES and the formation of covalent bonds with sila-
nol groups at the ZL channel entrances.
Embedding dye–ZL composites into matrices and photo-
chemical stability
The results of leaking tests performed with ZL crystals
loaded with HR are shown in Figure 4. The samples were
stored in transparent glass bottles, at RT, on the laboratory
bench. HR was chosen for these experiments because it leaked
out much faster than tb-DXP in uncoated samples, so test re-
sults obtained with HR–ZL samples were more conclusive. The
procedure used for plugging the channels had no influence on
the absorption and fluorescence spectra of the composites.
The unsealed HR–ZL sample showed significant leakage after
The embedding of ZL particles into a polymer matrix to obtain
transparent layers can be difficult, mainly because of their ten-
dency to agglomerate; therefore, specific embedding methods
have been reported.[65]
Dye–ZL composites sealed with APTES are much easier to
disperse in PMMA films than unsealed ones. Composites with-
out APTES modification show a greater tendency to form ag-
glomerates under similar conditions. Upon examining optical
microscopy images of such films, we have observed that a solu-
tion of PMMA in toluene or acetonitrile would often contain
a rather large amount of undissolved material (Figure 4A).
These PMMA particles are problematic because they act as ag-
glomeration nuclei for the composites (Figure 4D). However,
once the solution of PMMA is filtered, the cast film is free of
such polymer blocks, and the dye–ZL crystals agglomerate
much less than in the previous case (see Figure 4B and C, re-
spectively). Both films have a dye–ZL concentration of 1 mg
per 100 mg of PMMA. Good dispersion of the primary particles
is of great importance for optical applications, since larger ag-
glomerates will lead to increased light scattering of the dye–
ZL/PMMA suspension. The APTES-modified dye–ZL composites,
which were further functionalized by grafting MPTMS, CPTES,
and C18TES (Table 4), showed good dispersibility in nonpolar
organic solvents and in PMMA. C18TES-modified samples
showed the best behavior in these experiments.
The stability of the transparent dye–ZL PMMA matrices
under irradiation is important. We report a comparison of
100 mm thick films on a glass substrate containing dissolved
tb-DXP and dispersed tb-DXP,HR–ZLTS exposed in an accelerat-
ed light stability simulator at 358C by using a 380 nm cutoff
filter. The data in Figure 5 illustrate remarkable stabilization of
the composite. The initial optical densities of the tb-DXP and
tb-DXP,HR–ZLTS PMMA layers were 0.6 and 0.5, respectively. No
significant changes to the shape of the absorption spectra
were seen in either case. Composites modified with APTES
show the same behavior as the unmodified samples. The
decay is always faster at the beginning and then levels off. A
similar observation was made for HR–ZLTS composites and
other samples. This indicates that reactive species are present
at the beginning, but are gradually used up. The decay rate
slows as a consequence and perhaps becomes controlled by
transport kinetics. This is supported by the observation that
layers of larger optical density are photochemically more
stable. It is therefore necessary to compare samples of similar
optical density.
Figure 4. Top: Results from leaking tests. UV/Vis absorption spectra of the fil-
tered washing solution of open and stopcock-plugged HR–ZL samples
tested after one month of storage: HR–ZLTS.20(0.75IMZ+) (black), HR–
ZLTS.20(0.75IMZ+)@BuImiSil (red), HR–ZLTS.20(0.75IMZ+)@MeImiSil (blue), and
HR–ZLTS.20(0.75IMZ+)@APTES (green). The blue and red spectra have been
displaced to better show their shape. Bottom: Microscopy images of thin
poly(methyl methacrylate) (PMMA) films with and without dye–ZL. All films
were cast from solutions of PMMA in toluene at a concentration of
100 mgmLÀ1 PMMA. A) Pure PMMA film cast without preliminary filtration.
B) Cast from a PMMA solution that was filtered through a 0.22 mm syringe
filter. C) HR–ZLTS.16(0.5IMZ+)@APTES/PMMA film cast from a filtered solution
of PMMA. D) Film cast from the same material as that used in C), but with-
out prior filtration.
The decomposition mechanism is not well understood de-
spite efforts to understand the photochemical degradation of
Chem. Eur. J. 2016, 22, 4046 – 4060
www.chemeurj.org
4054
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Different organizational patterns of dye–ZL composites
The combinations b, b’, and c shown in Scheme 1 were synthe-
sized by using sequential procedure A. This can, in principle,
lead to antenna composites for light harvesting and trans-
port[1,23,30] that differ in the photophysical properties. System b“
can only be synthesized if a temperature regime exists for
which both dyes can be inserted simultaneously, so that
a random distribution of the small number of acceptors results.
This has so far only been realized for cationic dyes of equal
size and shape that can be inserted by ion exchange from
aqueous solution.[34] The syntheses of b and b’ can also differ
because the dyes discussed herein have to be inserted at ele-
vated temperatures and it is important that the dye inserted
first survives the temperature regime needed for the insertion
of the next one.
Figure 5. Photochemical stability upon exposure in an accelerated light sta-
bility simulator at 358C. Left: Optical density measured at l=536 nm, scaled
to 1 at 0 h irradiation time. Orange and green rectangles: two sets of experi-
mental data for tb-DXP,HR–ZLTS.04,.27(0.5IMZ+) embedded in PMMA; blue
circles: tb-DXP dissolved in PMMA; d: calculated data. Right: Absorption
spectra measured after 0 (red solid line), 8, 24, 48, 79, and 103 h of exposure
of tb-DXP,HR-ZLTS.04,.27(0.5IMZ+) embedded in PMMA.
Furthermore, dye A experiences a different environment in
the composite A,D–ZL than that in D,A–ZL because in the
latter it is embedded in the environment of D. Not all dyes are
sensitive to this, but HR is and has therefore been used to
demonstrate this difference. The results for the three combina-
tions (b, b’, and b“) with tb-DXP as the donor and HR as the
acceptor are shown in Figure 7. The absorption spectra have
the same shape, with the exception of the tail at lꢀ580 nm.
remarkably stable Lumogen perylene dyes embedded in
PMMA.[66,67] Foreign molecules present in PMMA, including
oxygen and traces of solvents used to prepare the layers, may
play a role. As mentioned before, APTES modification prevents
leakage of dyes embedded in ZL, but does not prevent small
molecules from entering the channels. Modification with stop-
cocks containing a larger tail is needed to fully seal the chan-
nels.
In Figure 6 a photograph of a 100 mm layer of tb-DXP,HR–ZL
embedded in PMMA on a glass substrate is shown under illu-
mination from the top. The PMMA/ZL layer was deposited by
using a film applicator. A 1:1 solution of 1-butanol/acetonitrile
Figure 6. Photograph of a thin PMMA film (100 mm) doped with tb-DXP,HR–
ZL.04,.25(0.5IMZ+). on a glass substrate under UV illumination.
Figure 7. Comparison of the absorption (c) and fluorescence (d, excit-
ed at l=490 nm) spectra for composites of type b, b’, and b“ (Scheme 1)
with tb-DXP as the donor and HR as the acceptor: tb-DXP,HR–
ZL.01,.23(0.5IMZ+) (green), HR,tb-DXP–ZL.15,.04(0.5IMZ+) (red), and (tb-
DXP,HR)ZL(.14,.05)(0.5IMZ+) (blue).
containing 20 wt% PMMA and 1 wt% of the composite was
used to apply the film. The image in Figure 6 illustrates the
wave-guiding properties of the device: the edges can be clear-
ly seen to glow orange. Such effects of the layer are of great
interest for applications as, for example, luminescent solar con-
centrators (LSCs). An LSC is a wave-guiding plate containing lu-
minescent chromophores. Light that falls on the face of the
plate is absorbed and subsequently re-emitted at longer wave-
length. The emitted light is trapped by total internal reflection
and guided to the edges of the plate, where it can be, for ex-
ample, collected and converted into electricity by a photovolta-
ic cell.[68,69]
The intensity of the tail depends on the relative amount of HR
with respect to tb-DXP, which is similar for the (tb-DXP,HR)–
ZL(.14,.05)(0.5IMZ+) and HR,tb-DXP–ZL.15,.04(0.5IMZ+) sam-
ples, but lower for tb-DXP,HR–ZL.01,.23(0.5IMZ+). The tb-DXP
emission is only seen as a short-wavelength tail, whereas the
intense band stems from the acceptor HR. The main difference
is in the fluorescence maxima. The maximum of tb-DXP,HR–
ZL.01,.23(0.5IMZ+) in which HR is embedded between tb-DXP
molecules, and therefore, essentially feels the environment of
the tb tails appears at the same position as observed in
CH2Cl2.[30] The largest redshift is seen in the HR,tb-DXP–
ZL.15,.04(0.5IMZ+) sample, in which HR is located at both ends
of the ZL channels, and is therefore not protected by the tb
tails. The fluorescence quantum yield of tb-DXP,HR–
ZL.01,.23(0.5IMZ+) is very good, in the order of 90%, and the
It has been well understood for many years that a major
energy loss in such a device is caused by the overlap between
the absorption and emission spectra of the chromophores.[69]
One way to minimize this loss is to use antenna materials,
such as those described herein. Absorption and emission spec-
tra are separated by employing a large amount of the absorb-
ing dye and a very small amount of an emitting dye.
Chem. Eur. J. 2016, 22, 4046 – 4060
www.chemeurj.org
4055
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
FRET efficiency of this sample has been determined to be 82%
(for details, see the Supporting Information). The fluorescence
quantum yield of the other two samples is lower: in the order
of 50%. The fluorescence quantum yield depends on the prep-
aration technique and on the properties of the ZL host. We
assume that optimization is also possible to obtain samples
with large quantum yields in these two cases. At present, we
focus, however, on composites with HR in the middle. Compo-
sites of type b’ with very high energy transfer and fluorescence
quantum yield have been prepared with the cationic dyes oxo-
nine and oxazine 1 as acceptor dyes (see Figure 9 of ref. [23]),
for which it is more difficult to realize arrangement b (PDIs as
donors) because the dyes may decompose at the temperature
needed to insert the PDIs. In general, it seems to be advanta-
geous to use neutral dyes if robust materials of high thermal
and photochemical stabilities are envisaged.
of FRET was set to equal 1 to show the maximum R0 possible
for the system. The value of R0 is only moderately influenced
by small changes in fD*. Using Equation (3), we find, for exam-
ple, that R0 decreases by about 10% if fD* drops by a factor
of 2.
ꢀ
ꢁ
1=6
k2
ð2Þ
ð3Þ
R0 ¼ TF ꢀD Jꢀn
n4
*
1=6
R0ðꢀD Þ ¼ R0ð1Þ Â ꢀD
*
*
The large spectral overlap of the DMP/tb-DXP, tb-DXP/HR,
and tb-DXP (homo) systems, in combination with a large value
for k2, leads to the most relevant Fçrster radii in the order of
7 nm for transfer along a channel. The value is 5.5 nm for verti-
cal transfer to a chromophore in one of the six surrounding
nearest neighbor channels. This means that all Fçrster radii of
relevance lie between the values of R0(k2 =4) and R0(k2 =1), as
reported in Table 3.[61] These are very favorable values for the
construction of efficient three-dye antenna composites of type
c (Scheme 1). The average distance between the centers of ad-
jacent tb-DXP molecules in a ZL channel at a loading, p, of 0.5
is 6 nm, and the distance between the centers of two channels
is 1.84 nm thus very efficient FRET can occur. To test this exper-
imentally, we prepared and analyzed the following composites:
DMP,HR–ZL, DMP,tb-DXP–ZL, tb-DXP,HR–ZL, and DMP,tb-
DXP,HR–ZL. Spectra of these samples are shown in Figures 8
and 9. The excitation spectrum of the DMP,HR–ZL sample in
Figure 8A shows that the sensitivity below l=450 nm is con-
Light-harvesting antenna system
The parallel orientation of the electronic transition dipole mo-
ments of DMP, tb-DXP, and HR (Figure 2), with respect to the
channels of ZL, consequently has the largest possible Fçrster
radii (R0) for FRET because the orientation parameter (k2) in
Equation 2 has a maximum value of four. The spectral overlaps
(Jꢀn) and Fçrster radii (R0) are given in Table 3. Equation 2 has
been used to calculate R0.[36,55] In this case, TF is equal to
8.785”10À25 mol, the refractive index (n) equals 1.43, and the
fluorescence quantum yield (fD*) of the donor in the absence
Figure 8. Absorption (c), fluorescence (d), and excitation (g) spectra of dye–ZL composites. The fluorescence spectra were measured upon excitation
at l=490 nm and the excitation spectra were observed at l=690 nm. A) Absorption and excitation spectrum of HR–ZL.04(0.5IMZ+) (red), and excitation
spectrum of DMP,HR–ZL.04,.15(0.5IMZ+) (blue). B) Absorption and fluorescence spectra of HR–ZL.04(0.5IMZ+) (red) and DMP,HR–ZL.04,.15(0.5IMZ+) (blue).
C) Comparison of the absorption (c) and fluorescence (d) spectra of tb-DXP–ZL.02 (green) and tb-DXP,HR–ZL.04,.27(0.5IMZ+). D) Spectra from C) have
been scaled to the same value at their respective maxima.
Chem. Eur. J. 2016, 22, 4046 – 4060
www.chemeurj.org
4056
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 8. HR with substituents at the extremities of the long axis.
in the middle of the channels, which means that only organiza-
tional patterns b and c lead to strongly luminescent antenna
composites, does not apply for cationic dyes such as oxonine
and oxazine 1. However, these have so far, however, resulted in
thermally less stable composites.[23] It appears to be worth-
while to test whether modified HR dyes (R,R’-HR; Scheme 8)
substituted with appropriate R groups would allow inversion
of the organization.
Figure 9. Three-color antenna system DMP,tb-DXP,HR–
ZLTS.04,.27,.15(0.5IMZ+). Blue c, absorption; blue d, excitation, ob-
served at l=680 nm; red c, fluorescence excited at l=360 nm.
siderably enhanced by the addition of DMP, and Figure 8B in-
dicates that the shapes of the HR absorption and the fluores-
cence spectra are not affected. The results in Figure 8A and B
demonstrate that energy transfer from DMP to HR is significant
without the addition of tb-DXP. The spectra in Figure 8C and D
demonstrate excellent energy-transfer efficiency from tb-DXP
to HR in the tb-DXP,HR–ZL composite. The fluorescence quan-
tum yields of optimized samples of this type are often larger
than 80%.[30]
Conclusion
Sequential insertion of different dyes into the 1D channels of
ZL leads to sandwich structures, and thus allows the formation
of sophisticated antenna composites for light harvesting, trans-
port, and trapping. The basic concepts of dye alignment in 1D
channel systems can, in principle, be applied to other zeolite
framework types, such as AlPO4-5, and is also applicable to
mesoporous materials.[71] Armbruster et al. were able to resolve
the crystal structure of such a dye–Mordenite composite,[72]
and Mintova et al. stressed the importance of progress in zeo-
lite synthesis to promote advanced applications.[73] More gen-
eral aspects of host–guest systems based on nanoporous crys-
tals, including inorganic guests, have also been previously re-
ported.[74] An important prerequisite for the integration of dye-
loaded zeolites into devices, for example, for light-harvesting,
is the availability of strategies for the synthesis of crystals with
various well-defined sizes and aspect ratios. There has been
substantial success in the field of tuning the size and aspect
ratio of ZL crystals.[17–22] Furthermore, concepts for closing the
channel entrances rely on the surface chemistry of silicates,
thereby, so far, excluding aluminophosphate zeolites, such as
AlPO4-5. This means that ZL remains the only host for which
a synthetic strategy (Scheme 1) has been successful. Stopcock
chemistry has recently, however, also been applied to metal–
organic frameworks.[75]
These two observations form the basis for understanding
the remarkable antenna properties we have observed in
DMP,tb-DXP,HR–ZL composites illustrated in Figure 9. The
shape of the fluorescence band, observed under excitation at
l=360 nm corresponds to that of HR, with a small feature at
shorter wavelength due to the tb-DXP. The band looks identi-
cal under excitation at l=490 nm. The excitation spectrum,
observed at l=680 nm, follows the absorption spectrum. In-
terestingly, the intensity of the short-wavelength contribution,
mainly stemming from the absorption of DMP, is larger than
expected. This is probably caused by the scattering contribu-
tion at short wavelength, which is much lower in the range of
emission of the HR acceptor. The essential conditions for ob-
taining highly luminescent composites with HR as the acceptor
are that 1) 0.5 of the 3.6 exchangeable potassium cations per
u.c. of ZL are substituted by IMZ+, and 2) HR is located in the
middle of the channels, and thus, embedded between tb-DXP
molecules, where it essentially feels the environment of the ali-
phatic tb tails.
Substituting one potassium cation per u.c. with IMZ+ leads
to the same result, whereas adding more IMZ+ may cause
problems because it starts to occupy too much of the available
space in the channels.
Our synthetic procedures—from molecules and hosts to
composite materials and composite-polymer matrices, includ-
ing two- and three-color antennas—demonstrate the impor-
tant flexibility of the strategy for creating a very promising
class of composites with tunable spectroscopic and photo-
physical properties. PDIs are important chromophores for the
synthesis of artificial antenna systems and have therefore been
considered in detail. The final step in the dye synthesis was, in
general, a condensation reaction of a perylene precursor with
the appropriate amines under high temperature and catalytic
conditions. Therefore, the main challenge was the preparation
of the corresponding amines. Of the 17 PDIs tested, 11 could
be easily inserted into the ZL channels at elevated tempera-
tures, 2 entered more reluctantly, and 4 could not be inserted
Although most PDIs, and also HR and DMP, support insertion
temperatures of 2608C under vacuum conditions, IMZ+ de-
composes above about 2008C, and thus damages the dye–ZL
composites. The necessarily lower insertion temperature pro-
longs the synthetic procedure. PDIs that seem to travel more
easily inside the channels (e.g., bone-DXP) are therefore of spe-
cial interest. Organic cations other than IMZ+, such as those
used by Li et al. to modify Laponite,[70] could be tested for fine-
tuning the synthetic conditions and properties of the dye–ZL
composites. The observation that the acceptor HR should be
Chem. Eur. J. 2016, 22, 4046 – 4060
www.chemeurj.org
4057
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
at all. The pattern we have observed improves phenomenolog-
ical understanding considerably, so that results become more
predictable, which enhances the possibilities for designing de-
sired functionality. Controlling the substitution pattern of in-
serted dyes not only allows monitoring of the contact distance
between nearest neighbors (so that, e.g., exciton interaction
can be switched on and off), but also fine-tuning of transport
properties, as well as altering the nanochannel environment.
The fact that composites bearing reasonably good characteris-
tics can be prepared with commercial zeolite underlines the
usefulness of ZL as a host material. However, if purity, mor-
phology, particle size, and particle size distribution matter,
then it is advantageous to use well-described and reproducible
published synthetic procedures. Sealing the channel entrances,
so that the dye molecules cannot exit, is a prerequisite for ach-
ieving long-term stability of dye–ZL composites. We have
shown that a variety of procedures can be used successfully.
Plugging the channels, for example, with APTES is sufficient for
sealing the channels against dye leakage. However, it does not
prevent small molecules (such as acetonitrile, ethanol, water
and oxygen) from entering and exiting the channels. As
a result, the photochemical stability of the material, in particu-
lar, may be severely affected. Long-term photostability is an
important issue for LSC devices. Sealing against smaller mole-
cules is possible, for example, with the two imidazolium stop-
cocks we have reported as examples. Thus, we conclude that
this route can be extended by designing specialized stopcocks
for fine-tuning to achieve the desired properties. Embedding
the ZL particles in a polymer matrix to obtain transparent
layers can also be difficult because of their tendency to ag-
glomerate. Composites sealed with APTES are much easier to
disperse in PMMA films than unsealed ones. Additional modifi-
cation of the coat with C18TES further improves this quality.
Our results deepen our understanding of these systems exten-
sively and facilitate the rational design of advanced dye–zeolite
composite materials, which are of great interest as smart com-
posites for LSCs in optoelectronics, sensing, and diagnostics.
We provide powerful tools for further developing and under-
standing artificial antenna systems, which are among the most
fascinating subjects of current photochemistry and photophy-
sics.
Synthesis and postsynthetic treatment of ZL (LTL type)
The ZL crystals used in most of the experiments performed in this
study were of barrel-type with a length and diameter of about
500 nm. Such crystals had a length that could accommodate a max-
imum of about 166 dye molecules per channel if each dye occu-
pied 4 u.c.s. This made it possible to prepare dye–ZL composites
with high light absorption capabilities, while the channel length
was still short enough to keep the time required for diffusion pro-
cesses short. The ZL crystals were synthesized by following a re-
ported procedure.[20] However, we also tested to what extent com-
mercial ZL materials, such as HSZ-500 from TOSOH Corporation
(type HSZ-500 KOA, length ꢀ400 nm, width ꢀ500 nm) could be
used. Commercial materials do not feature well-defined morpholo-
gies and may contain larger amounts of amorphous particles and
cations other than K+.
SEM images of self-synthesized ZL and of commercial ZL from
TOSOH Corporation used herein are shown in Figure 10. In both
cases, it was necessary to perform a postsynthesis cation exchange
with KNO3 to ensure that the composition of the charge-compen-
sating ions of the ZL was well defined (see the Supporting Infor-
mation for more details).
Figure 10. SEM images of ZL used in this study. Left: Self-synthesized (ZL).
Right: Commercial HSZ-500KOA,TOSOH Corporation (ZLTS).
Synthesis of dye–ZL composites
Depending on their nature, organic dyes were inserted into the
channels of ZL either by ion exchange or through an adsorption
process. Cationic species were incorporated into ZL by first dispers-
ing the host material in an appropriate solvent, such as deionized
water. The required amount of dye was then added to the ZL sus-
pension as a stock solution, usually dissolved in the same solvent.
Ion exchange was carried out either at room temperature or under
gentle heating to increase the reaction rate. However, one had to
be careful when heating the dye-containing ZL dispersion because
the organic molecules could decompose if the exchange condi-
tions were too harsh. The critical temperature regime depended
very much on the kind of dyes used and on the conditions. An ex-
ample of this behavior was the cationic dye oxazine 1 (Ox1+). If
the ion-exchange process was carried out at a temperature above
708C, most of the dye decomposed, and only very small loadings
could be achieved. Performing the same procedure at room tem-
perature or 408C led to higher Ox1+ loadings because the dye did
not decompose under these conditions.[23] Neutral dyes were in-
serted into the ZL channels through adsorption in an evacuated
and heated ampoule. In this process, the appropriate amount of
dye was first adsorbed on the outer surface of the ZL by dispersing
the host material in a solution of dye (see the Supporting Informa-
tion).
Experimental Section
General
The synthesis of the dye–ZL composites consists of several steps
outlined in this section. An extension of the experimental details is
given in the Supporting Information. Some samples have been
prepared by using commercially available ZL. This is indicated by
adding TS as an index [e.g., dye1,dye2–ZLTS.05,.12(0.5IMZ+)].
Synthesis of dyes
Details of the synthesis of the dyes can be found in the Supporting
Information.
Chem. Eur. J. 2016, 22, 4046 – 4060
www.chemeurj.org
4058
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
tor (Elcometer K3505). In the case of film casting, the solution of
PMMA in toluene had a concentration of 100 mgmLÀ1 PMMA, and
the amount of dye–ZL composite used was 1 mg per 100 mg of
PMMA (1 wt%). A more concentrated solution of PMMA was
needed when using the film applicator. A 30 wt% solution in a 1:1
mixture of 1-butanol and acetonitrile, along with 1 wt% concentra-
tion of ZL, provided the best results when depositing films on
glass substrates. The film applicator was set to deposit layers with
a wet thickness of 250 mm, which corresponded to films with
a thickness of 100 mm after drying and shrinking.
Plugging of channel entrances and modification of the
surface
Plugging of channel entrances: Condensation of alkoxysilanes
with the ZL silanol groups was used to modify the ZL coat and
channel entrances. In such a condensation reaction, which could
be carried out in a large variety of solvents, the alkoxysilane would
bind to a surface SiÀOÀH group of the ZL under elimination of an
alcohol, RÀOH. A tertiary amine was used to catalyze the reaction.
The ZL channel entrances were sealed by binding APTES to the
composite. In
a typical experiment, the dye–ZL composite
(400 mg) was dispersed in toluene (70 mL) by means of a rotor/
stator dispersion tool (Ultra Turrax T18 basic, IKA) for 20 min at
16000 rpm. Once the ZL was well dispersed, APTES (51 mL) was
added to the suspension. The condensation reaction was then car-
ried out at room temperature for 3 h under constant treatment
with the dispersion tool. The reacted material was then centrifuged
(15 min at 2100 rpm) and washed twice with toluene. The material
was then dried in a vacuum oven at 608C overnight (16 h). The
molecular structures of the imidazole-containing stoppers (MeImiSil
and BuImiSil) are given in Table 4. Both compounds were prepared
according to procedures reported in ref. [76]. The experimental
protocol to introduce the imidazolium stopcocks was similar to
that of the sealing experiment with APTES. Sealing was tested in
a leaking experiment, in which the treated and untreated HR–ZL
composites were dispersed in a solution of 1-butanol/acetonitrile
(1:1) and stirred at room temperature for 8 h. To ensure a good dis-
persion, as well as to test the effectiveness of the stopcock, the
mixture was sonicated for 1 h. The suspension was then centri-
fuged for 15 min at 2100 rpm and the supernatants were filtered
over a 0.22 mm PTFE syringe filter to remove any residual small ZL
particles. The amount of leaked dye was then determined by meas-
uring the UV/Vis spectrum of the filtered supernatant. The samples
were stored in transparent glass bottles at room temperature and
normal light exposure.
Acknowledgements
Financial support by the Swiss Commission for Technology and
Innovation (KTI/CTI, Project 12902.1 PFNM-NM) and the
Schweizerische Bundesamt für Energiewirtschaft BFE (no. SI/
500995-01) is gratefully acknowledged.
Keywords: dyes/pigments · FRET · host–guest systems ·
supramolecular chemistry · zeolites
[1] G. Calzaferri, S. Huber, H. Maas, C. Minkowski, Angew. Chem. Int. Ed.
2003, 42, 3732–3558; Angew. Chem. 2003, 115, 3860–3888.
[2] A. Bertucci, H. Lülf, D. Septiadi, A. Manicardi, R. Corradini, L. De Cola,
Adv. Healthcare Mater. 2014, 3, 1812–1817.
[3] a) N. S. Kehr, S. Atay, B. Ergün, Macromol. Biosci. 2015, 15, 445–463;
b) S. Hashimoto, J. Phys. Chem. Lett. 2011, 2, 509–519; c) D. Brühwiler,
G. Calzaferri, T. Torres, J. H. Ramm, N. Gartmann, L.-Q. Dieu, I. López-
Duarte, M. V. Martínez-Díaz, J. Mater. Chem. 2009, 19, 8040–8067; d) M.
Pauchard, S. Huber, R. MØallet-Renault, H. Maas, R. Pansu, G. Calzaferri,
Angew. Chem. Int. Ed. 2001, 40, 2839–2842; Angew. Chem. 2001, 113,
2921–2924.
[4] N. VilaÅa¸ R. Amorim, A. F. Machado, P. Parpot, M. F. R. Pereira, M. Sardo,
J. Rocha, A. M. Fonseca, I. C. Neves, F. Baltazar, Colloids Surf. B 2013, 112,
237–244.
Surface modifications: The surface of some APTES-modified dye–
ZL composites was further functionalized by grafting additional
types of alkoxysilanes. The sealing procedure described above left
some unreacted ZL silanol and alkoxysilane groups. These could be
used for additional condensation reactions with molecules such as
MPTMS, CPTES, and C18TES (Table 4). For example, C18TES was
successfully bound to an APTES-modified dye–ZL composite by
first dispersing the material (170 mg) in toluene (70 mL) for 20 min
at 16000 rpm by using the dispersion tool. Then C18TES (200 mL)
was added to the suspension, before stirring the mixture for 3 h at
room temperature with the dispersion tool at 16000 rpm. The
product was collected through centrifugation (15 min at 2100 rpm)
and washed twice with toluene before being dried in a vacuum
oven at 608C for 16 h.
[5] M. Tsotsalas, K. Kopka, G. Luppi, S. Wagner, M. P. Law, M. Schäfers, L.
De Cola, ACS Nano 2010, 4, 342–348.
[6] F. Cucinotta, A. Guenet, C. Bizzarri, W. Mróz, Ch. Botta, B. Miliµn-Medina,
J. Gierschner, L. De Cola, ChemPlusChem 2014, 79, 45–57.
[7] a) A. Mech, A. Monguzzi, F. Meinardi, J. Mezyk, G. Macchi, R. Tubino, J.
Am. Chem. Soc. 2010, 132, 4574–4576; b) P. Cao, Y. Wang, H. Li, X. Yu, J.
Mater. Chem. 2011, 21, 2709–2714; c) T. Wen, W. Zhang, X. Hu, L. He, H.
Li, ChemPlusChem 2013, 78, 438–442.
[8] R. N. Mahato, H. Lülf, M. H. Siekman, S. P. Kersten, P. A. Bobbert, M.
de Jong, L. De Cola, W. G. van der Wiel, Science 2013, 341, 257–260.
[9] a) L. Gigli, R. Arletti, G. Tabacchi, E. Fois, J. Vitillo, G. Martra, G. Agostini,
S. Quartieri, G. Vezzalini, J. Phys. Chem. C 2014, 118, 15732–15743.
[10] B. Schulte, M. Tsotsalas, M. Becker, A. Studer, L. De Cola, Angew. Chem.
Int. Ed. 2010, 49, 6881–6884; Angew. Chem. 2010, 122, 7033–7036.
[11] J. M. Beierle, R. Roswanda, P. M. Erne, A. C. Coleman, W. R. Browne, B. L.
Feringa, Part. Part. Syst. Charact. 2013, 30, 273–279.
[12] a) S. Hashimoto, M. Hagiri, N. Matsubara, S. Tobita, Phys. Chem. Chem.
Phys. 2001, 3, 5043–5051; b) B. Bussemer, I. Dreiling, U. W. Grummt,
G. J. Mohr, J. Photochem. Photobiol. A 2009, 204, 90–95.
[13] G. Schulz-Ekloff, D. Wçhrle, B. van Duffel, R. A. Schoonheydt, Micropo-
rous Mesoporous Mater. 2002, 51, 91–138.
[14] Ch. Sprung, B. Weckhuysen, J. Am. Chem. Soc. 2015, 137, 1916–1928.
[15] O. Bossart, L. De Cola, S. Welter, G. Calzaferri, Chem. Eur. J. 2004, 10,
5771–5775.
[16] I. López-Duarte, L.-Q. Dieu, I. Dolamic, M. V. Martínez-Díaz, T. Torres, G.
Calzaferri, D. Brühwiler, Chem. Eur. J. 2011, 17, 1855–1862.
[17] P. Li, Y. Wang, H. Li, G. Calzaferri, Angew. Chem. Int. Ed. 2014, 53, 2904–
2909; Angew. Chem. 2014, 126, 2948–2953.
Embedding into PMMA and preparation of layers
An alternative to using OGS[23] to record absorption or emission
spectra of dye–ZL composites were thin ZL-doped PMMA films.
The refractive index of PMMA was very close to that of the dye–ZL
composites, so that light scattering from non-agglomerated ZL
particles could be minimized. Such thin films were prepared by
first dispersing the sealed dye–ZL composite in an appropriate sol-
vent, such as toluene or 1-butanol. This suspension was then
added to a solution of PMMA in toluene or acetonitrile and mixed
well. The best mixing results were obtained by using a rotor/stator
dispersion tool. The films were then prepared by either casting in
a Petri dish or depositing on a glass substrate with a film applica-
[18] T. Ohsuna, B. Slater, F. Gao, J. Yu, Y. Sakamoto, G. Zhu, O. Terasaki, D.
Vaughan, S. Qiu, C. Catlow, Chem. Eur. J. 2004, 10, 5031–5040.
[19] O. Larlus, V. P. Valtchev, Chem. Mater. 2004, 16, 3381–3389.
Chem. Eur. J. 2016, 22, 4046 – 4060
www.chemeurj.org
4059
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[20] A. Zabala Ruiz, D. Brühwiler, T. Ban, G. Calzaferri, Monatsh. Chem. 2005,
136, 77–89.
[21] A. I. Lupulescu, M. Kumar, J. D. Rimer, J. Am. Chem. Soc. 2013, 135,
6608–6617.
[22] T. Ban, M. Takamura, M. Morikawa, Y. Ohya, Mater. Chem. Phys. 2013,
137, 1067–1072.
[53] W. Bradley, F. W. Pexton, J. Chem. Soc. 1954, 4432–4435.
[54] E. Fois, G. Tabacchi, A. Devaux, P. Belser, D. Brühwiler, G. Calzaferri, Lang-
muir 2013, 29, 9188–9198.
[55] G. Calzaferri, A. Devaux, in Supramolecular Photochemistry: Controlling
Photochemical Processes (Eds.: V. Ramamurthy, Y. Inoue), Wiley, Hobo-
ken, 2011, Chapter 9, pp. 285–387.
[23] A. Devaux, G. Calzaferri, I. Miletto, P. Cao, P. Belser, D. Brühwiler, O.
Khorev, R. Häner, A. Kunzmann, J. Phys. Chem. C 2013, 117, 23034–
23047.
[24] A. Zabala Ruiz, H. Li, G. Calzaferri, Angew. Chem. Int. Ed. 2006, 45, 5282–
5287; Angew. Chem. 2006, 118, 5408–5413.
[25] K. B. Yoon, Acc. Chem. Res. 2007, 40, 29–40.
[26] F. Cucinotta, Z. Popovic’, E. A. Weiss, G. M. Whitesides, L. De Cola, Adv.
Mater. 2009, 21, 1142–1145.
[56] A. Rademacher, S. Märkle, H. Langhals, Chem. Ber. 1982, 115, 2927–
2934.
[57] a) J. S. Park, Y.-J. Lee, K. B. Yoon, J. Am. Chem. Soc. 2004, 126, 1934–
1935; b) A. Kulak, Y.-J. Lee, Y. S. Park, K. B. Yoon, Angew. Chem. Int. Ed.
2000, 39, 950–953; Angew. Chem. 2000, 112, 980–983.
[58] a) N. R. E. N. Impens, P. van der Voort, E. F. Vansant, Microporous Mesopo-
rous Mater. 1999, 28, 217–232; b) T. Kawai, K. Tsutsumi, Colloid Polym.
Sci. 1998, 276, 992–996.
[27] S. Hashimoto, K. Samata, T. Shoji, N. Taira, T. Tomita, S. Matsuo, Micropo-
rous Mesoporous Mater. 2009, 117, 220–227.
[28] Y. Wang, H. Li, Y. Feng, H. Zhang, G. Calzaferri, T. Ren, Angew. Chem. Int.
Ed. 2010, 49, 1434–1438; Angew. Chem. 2010, 122, 1476–1480.
[29] S. Fibikar, G. Luppi, V. Martínez-Junza, M. Clemente-LØon, L. De Cola,
ChemPlusChem 2015, 80, 62–67.
[59] a) W. Xu, Q. Luo, H. Wang, L. C. Francesconi, R. E. Stark, D. L. Akins, J.
Phys. Chem. B 2003, 107, 497–501; b) N. K. Mal, M. Fujiwara, Y. Tanaka,
Nature 2003, 421, 350–353; c) J. Liu, X. Feng, G. E. Fryxell, L.-Q. Wang,
A. Y. Kim, M. Gong, Adv. Mater. 1998, 10, 161–165.
[60] T. Ban, D. Brühwiler, D. Calzaferri, J. Phys. Chem. B 2004, 108, 16348–
16352.
[30] A. Devaux, G. Calzaferri, P. Belser, P. Cao, D. Brühwiler, A. Kunzmann,
Chem. Mater. 2014, 26, 6878–6885.
[31] M. Busby, A. Devaux, C. Blum, V. Subramaniam, G. Calzaferri, L. De Cola,
J. Phys. Chem. C 2011, 115, 5974–5988.
[61] G. Calzaferri, Langmuir 2012, 28, 6216–6231.
[62] A. Guerrero-Martínez, S. Fibikar, I. Pastoriza-Santos, L. M. Liz-Marzàn, L.
De Cola, Angew. Chem. Int. Ed. 2009, 48, 1266–1270; Angew. Chem.
2009, 121, 1292–1296.
[32] G. Calzaferri, D. Brühwiler, T. Meng, L.-Q. Dieu, V. L. Malinovskii, R. Häner,
Chem. Eur. J. 2010, 16, 11289–11299.
[63] Y. Wang, Y. Yue, H. Li, Q. Zhao, Y. Fang, P. Cao, Photochem. Photobiol. Sci.
2011, 10, 128–132.
[33] M. Busby, C. Blum, M. Tibben, S. Fibikar, G. Calzaferri, V. Subramaniam, L.
De Cola, J. Am. Chem. Soc. 2008, 130, 10970–10976.
[34] G. Calzaferri, K. Lutkouskaya, Photochem. Photobiol. Sci. 2008, 7, 879–
910.
[35] E. Fois, G. Tabacchi, G. Calzaferri, J. Phys. Chem. C 2010, 114, 10572–
10579.
[64] G. Tabacchi, E. Fois, G. Calzaferri, Angew. Chem. Int. Ed. 2015, 54, 11112–
11116, Angew. Chem. 2015, 127, 11264–11268.
[65] a) S. Suµrez, A. Devaux, J. BaÇnuelos, O. Bossart, A. Kunzmann, G. Calza-
ferri, Adv. Func. Mater. 2007, 17, 2298–2306; b) J. Schneider, D. Fanter,
M. Bauer, C. Schomburg, D. Wçhrle, G. Schulz-Ekloff, Microporous Meso-
porous Mater. 2000, 39, 257–263.
[36] a) Th. Fçrster, Naturwissenschaften 1946, 33, 166–175; b) Th. Fçrster,
Ann. Phys. 1948, 437, 55–75.
[66] G. Griffini, L. Brambilla, M. Levi, M. Del Zoppo, S. Turri, Sol. Energy Mater.
Sol. Cells 2013, 111, 41–48.
[37] M. Pauchard, A. Devaux, G. Calzaferri, Chem. Eur. J. 2000, 6, 3456–3470.
[38] a) T. Weil, T. Vosch, J. Hofkens, K. Peneva, K. Müllen, Angew. Chem. Int.
Ed. 2010, 49, 9068–9072; Angew. Chem. 2010, 122, 9252–9278; b) D.
Jänsch, C. Li, L. Chen, M. Wagner, K. Müllen, Angew. Chem. Int. Ed. 2015,
54, 2285–2289; Angew. Chem. 2015, 127, 2314–2319.
[39] H. Bittermann, D. Siegemund, V. L. Malinovskii, R. Häner, J. Am. Chem.
Soc. 2008, 130, 15285–15287.
[40] C. B. Winiger, S. M. Langenegger, G. Calzaferri, R. Häner, Angew. Chem.
Int. Ed. 2015, 54, 3643–3647.
[41] a) International Zeolite Association, http://www.iza-structure.org; b) Ch.
Baerlocher, L. B. McCusker, D. H. Olson, Atlas of Zeolite Framework Types,
6th ed., Elsevier, Amsterdam, 2007.
[67] N. Tanaka, N. Barashkov, J. Heath, W. N. Sisk, Appl. Optics 2006, 45,
3846–3851.
[68] L. Danos, T. J. Meyer, P. A. Kittidachan, L. Fang, T. S. Parel, N. Soleimani, T.
Markvart, in Materials Challenges: Inorganic Photovoltaic Solar Energy
(Ed. S. J. C. Irvine), RCS, Cambridge, 2015, Chapter 9: Photon Frequency
Management Materials for Efficient Solar Energy Collection, p. 297–331.
[69] a) M. G. Debije, P. P. Verbunt, Adv. Energy Mater. 2012, 2, 12–35; b) T.
Dienel, C. Bauer, I. Dolamic, D. Brühwiler, Sol. Energy 2010, 84, 1366–
1369; c) J. C. Goldschmidt, M. Peters, A. Bçsch, H. Helmers, F. Dimroth,
S. W. Glunz, G. Willeke, Sol. Energy Mater. Sol. Cells 2009, 93, 176–182.
[70] D. Yang, Y. Wang, Y. Wang, Z. Li, H. Li, ACS Appl. Mater. Interfaces 2015,
7, 2097–2103.
[42] C. Marcolli, G. Calzaferri, Appl. Organomet. Chem. 1999, 13, 213–226.
[43] IMZ+: 1-ethyl-3-methyl imidazolium chloride counterion.
[44] M. Myahkostupov, V. Prusakova, D. G. Oblinsky, G. D. Scholes, F. N. Cas-
tellano, J. Org. Chem. 2013, 78, 8634–8638.
[45] J. Liu, Y. Li, Y. Li, N. Hu, J. Appl. Polym. Sci. 2008, 109, 700–707.
[46] B. L. Small, R. Rios, E. R. Fernandez, D. L. Gerlach, J. A. Halfen, M. J.
Carney, Organometallics 2010, 29, 6723–6731.
[71] a) G. Ihlein, F. Schüth, O. Krauss, U. Vietze, F. Laeri, Adv. Mater. 1998, 10,
1117; b) T. H. Noh, H. Lee, J. Jang, O.-S. Jung, Angew. Chem. Int. Ed.
2015, 54, 9284–9288; Angew. Chem. 2015, 127, 9416–9420.
[72] P. Simoncic, T. Armbruster, P. Pattison, J. Phys. Chem. B 2004, 108,
17352–17360.
[73] M. Zaarour, B. Dong, I. Naydenova, R. Retoux, S. Mintova, Microporous
Mesoporous Mater. 2014, 189, 11–21.
[47] R. Mishra, J. M. Lim, M. Son, P. Panini, D. Kim, J. Sankar, Chem. Eur. J.
2014, 20, 5776–5786.
[74] F. Laeri, F. Schüth, U. Simon, M. Wark, Host–Guest Systems Based on
Nanoporous Crystals, Wiley-VCH, Weinheim, 2003.
[48] H. Langhals, Chem. Ber. 1985, 118, 4641–4645.
[49] F. Nolde, J. Qu, C. Kohl, N. G. Pschirer, E. Reuther, K. Müllen, Chem. Eur. J.
2005, 11, 3959–3967.
[75] H. Wang, J. Xu, D.-S. Zhang, Q. Chen, R.-W. Wen, Z. Chang, X-H. Bu,
Angew. Chem. Int. Ed. 2015, 54, 5966–5971; Angew. Chem. 2015, 127,
6064–6068.
[50] Y. Geerts, H. Quante, H. Platz, R. Mahrt, M. Hopmeier, A. Bohm, K.
Müllen, J. Mater. Chem. 1998, 8, 2357–2369.
[76] Y. Lu, S. S. Moganty, J. L. Schaefer, L. A. Archer, J. Mater. Chem. 2012, 22,
4066–4072.
[51] H. Quante, K. Müllen, Angew. Chem. Int. Ed. Engl. 1995, 34, 1323–1325;
Angew. Chem. 1995, 107, 1487–1489.
[52] T. Yamamoto, A. Morita, Y. Miyazaki, T. Maruyama, H. Wakayama, Z. H.
Zhou, Y. Nakamura, T. Kanbara, S. Sasaki, K. Kubota, Macromolecules
1992, 25, 1214–1223.
Received: November 2, 2015
Published online on February 11, 2016
Chem. Eur. J. 2016, 22, 4046 – 4060
www.chemeurj.org
4060
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim