Solid-state gas sensors developed from functional difluoroboradiazaindacene dyes

Article abstract of DOI:10.1002/chem.200801911

This article describes the synthesis and characterization of several new difluoroboradiazaindacene (BODIPY) dyes functionalized at the central 8-position by a phenyliodo, phenylheptynoate or phenylheptynoic fragment and at the 3- or 3/5-position(s) by 4-d

Full text of DOI:10.1002/chem.200801911

FULL PAPER
DOI: 10.1002/chem.200801911
Solid-State Gas Sensors Developed from Functional
Difluoroboradiazaindacene Dyes
AHCTUNGTRENNUNG
Raymond Ziessel,*[a] Gilles Ulrich,[a] Anthony Harriman,[b] Mohammed A. H. Alamiry,[b]
Beverly Stewart,[b] and Pascal Retailleau[c]
Abstract: This article describes the syn-
thesis and characterization of several
tained on successive addition of acid
and base. The difunctionalized deriva-
tive is especially interesting in this re-
spect and shows two well-resolved pKa
values of 5.10 and 3.04 in acetonitrile.
Addition of the first proton causes only
small spectral changes and deactivates
the molecule towards addition of the
second proton. It is this latter step that
accommodates the large change in ab-
sorption and emission properties, due
to the reversible extinction of the intra-
molecular charge-transfer character in-
herent to this type of dye. The main
focus of the work is the covalent an-
choring of the dyes to inert, porous
polyacrylate beads so as to form a
solid-state sensor suitable for analysis
of gases or flowing liquids. The final
material is highly stable—its perfor-
mance is undiminished after more than
one year—and fully reversible over
many cycles. The sensitivity is such that
reactions can be followed by the naked
eye and the detection limit is about
600 ppb for HCl and about 80 ppb for
ammonia. Trace amounts of diphos-
gene can be detected, as can alkylating
agents. The sensing action is indiscrimi-
nate and also operates when the beads
are dispersed in aqueous media.
new
difluoroboradiazaindacene
(BODIPYꢀ) dyes functionalized at
the central 8-position by a phenyliodo,
phenylheptynoate or phenylheptynoic
fragment and at the 3- or 3/5-posi-
tion(s) by 4-dimethylaminophenylstyryl
residue(s). Single-crystal structural de-
terminations confirm the planarity of
the dyes, while the absorption and fluo-
rescence spectroscopic properties are
highly sensitive to the state of protona-
tion (or alkylation) of the terminal ani-
lino donor group(s). Reversible color
tuning from green to blue for absorp-
tion and from colorless (i.e., near-IR
region) to red for fluorescence is ob-
Keywords: analytical methods
dyes/pigments fluorescence
functional beads · sensors
·
·
·
Introduction
A wide variety of chemical sensors capable of monitoring
trace amounts of atmosphere-borne pollutants and various
analytes at ambient temperature have been described and
tested under rigorous conditions.[1,2] The more advanced sys-
tems tend to use polymer films loaded with dyes[3] that un-
dergo well-defined color changes in the presence of the
target substrate and make use of common dyes such as
metal-free porphyrins,[4–6] Phenol Red,[7] oxazines,[8] Reich-
ardt dyes,[9] sulfophthalein,[10] or Nile Red.[11] In certain
cases, polymeric materials such as poly(2-methoxyaniline),[12]
[a] Dr. R. Ziessel, Dr. G. Ulrich
Laboratoire de Chimie Molꢁculaire
Ecole Europꢁenne de Chimie, Polymꢂres et Matꢁriaux
CNRS, 25 rue Becquerel, 67087 Strasbourg Cedex 02 (France)
Fax : (+33)3-90-24-26-89
E-mail: ziessel@chimie.u-strasbg.fr
[b] Prof. A. Harriman, Dr. M. A. H. Alamiry, B. Stewart
Molecular Photonics Laboratory
polyACTHNUGRTNENUG ACHUTNGTRENNUGN
(aniline),[13] or poly(pyrrole)[14] play active roles in the
School of Natural Sciences, Bedson Building
University of Newcastle
Newcastle upon Tyne, NE1 7RU (United Kingdom)
sensing process, rather than simply providing an inert sup-
port. The vast majority of these systems employ optical ab-
sorption spectral changes to provide the monitoring signal,
although related systems are known in which the sensing
action is triggered by a change in fluorescence intensity and/
or spectral distribution.[3,8] A persistent problem encoun-
tered in many hybrid materials, such as dyes entrapped in
[c] Dr. P. Retailleau
Laboratoire de Cristallochimie
ICSN—CNRS, Bꢃt 27 1 avenue de la Terrasse
91198 Gif-sur-Yvette, Cedex (France)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200801911.
Chem. Eur. J. 2009, 15, 1359 – 1369
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1359

sol–gels,[7] copolymers,[6] or composite films,[5] relates to the
dye leaching from the host material[15] and this situation
serves to limit the shelf life of the device. It also poses
severe problems for the measurement of low concentrations
of substrate, especially when used in the fluorescence mode.
The only logical solution to this particular problem is to
attach the dye to the host by way of robust covalent linkag-
es.
In searching for suitable chromophores that could be
adapted for the simultaneous measurement of absorption
and fluorescence changes, such behavior is taken as being
representative of the simplest form of orthogonal detection,
imposed by the presence of acids, bases, or toxins such as
phosgene, our attention has been drawn to the difluorobor-
aindacenes (BODIPY) class of dyes.[16,17] These materials,
which are stable, easily functionalized and highly fluores-
cent, have been developed as laser dyes,[18] selective fluores-
cent tags,[19] molecular sensors,[20] donor–acceptor dyads,[21]
solar concentrators,[22] and components for electrolumines-
cent devices.[23] Of particular interest is the realization that
their optical properties can be fine-tuned by chemical modi-
fication of the organic core,[24] or at the boron center.[25] For
instance, the attachment of electron-rich residues at the 8-
position, or at the boron atom, has little effect on the ab-
sorption spectral profile. In contrast, grafting unsaturated
groups to the 2,6-positions serves to red-shift the absorption
maximum by about 50 nm for each attachment.[26] A spec-
tacular hypsochromic shift accompanies substitution at the
3,5-positions by vinylic[27,28] or acetylenic fragments.[29] More-
over, BODIPY-based dyes enter readily into charge-transfer
processes such that their absorption and emission properties
can be further adjusted by controlling the electron affinity
of appended donor and/or acceptor groups.[30,31] This last
AHCTUNGTREGrNNNU ealization forms the basis of the operating mechanism for
the sensors reported here.
We are particularly interested in the concept of monitor-
ing the concentration of species present as contaminants in
a flowing effluent stream, either gaseous or liquid. This is
best achieved using a porous support functionalized with a
predetermined loading of the sensor. The sensor must be
firmly attached to the support and undergo distinctive and
reversible color changes that themselves lead to marked dif-
ferences in the fluorescence properties. We are aware that a
tremendous wealth of information has accumulated over the
past few decades regarding the design of chemical sensors
for the in situ monitoring of analytes, although most refer to
static liquids, and protocols have been established for testing
such materials under various conditions.[1–3] In developing
the present system, we have given considerable attention to
the sensing mechanism and have set out to ensure orthogo-
nal detection. In particular, the color change has been care-
fully optimized for maximum visibility, while the fluores-
cence changes are equally dramatic. The work should be
considered as being “proof-of-concept” for which future ra-
tional modifications would allow for the detection of specific
substrates by building the chromophore into a suitable
Abstract in French: Ce manuscrit dꢀcrit la synthꢁse et la ca-
ractꢀrisation de nouvelles sondes fluorescentes communꢀment
appelꢀes BODIPY qui ont ꢀtꢀ ciselꢀes en position centrale
par un groupe phꢀnyl-iodo, phꢀnylheptynoate ou phꢀnylhep-
tanoic acide et en positions 5 ou 3/5 avec des fragments 4-di-
mꢀthylaminostyryle. Ces derniers fragments favorisent une
excellente conjugaison qui est ꢂ lꢃorigine de la couleur verte
des sondes. Des structures molꢀculaires obtenues par diffrac-
tion aux rayon X sur monocristal confirment la planaritꢀ des
ces objets, tandis que les propriꢀtꢀs dꢃabsorption et de fluores-
cence sont sensibles ꢂ la protonation ou lꢃalkylation des grou-
pements aminꢀs terminaux. Un changement de couleur rꢀver-
sible du vert au bleu par absorption et de lꢃinvisible au rouge
en fluorescence est obtenu en prꢀsence dꢃun flux dꢃair pimen-
tꢀ avec des traces dꢃacides. La rꢀversibilitꢀ est observꢀe en
prꢀsence de traces de base. Le dꢀrivꢀ di-fonctionnalisꢀ est
particuliꢁrement intꢀressant dans la mesure ou deux pKa sont
bien rꢀsolus ꢂ 5,10 et 3,04 unitꢀs dans lꢃacꢀtonitrile. Lꢃaddi-
tion du premier proton induits des changements spectroscopi-
ques modestes et diminue la basicitꢀ du deuxiꢁme groupe
amino. Cette derniꢁre protonation induit des changements
majeurs en absorption et ꢀmission et inhibe le transfert de
charge intramolꢀculaire. LꢃintꢀrÞt majeur de ces sondes rꢀside
dans leur potentiel dꢃÞtre liꢀ de faÅon covalente ꢂ des billes de
polymꢁres et dꢃutiliser ces billes pour la dꢀtection dꢃꢀlꢀments
toxiques dans des effluents gazeux. En particulier il a ꢀtꢀ dꢀ-
montrꢀ que la dꢀtection visuelle de trace de HCl (environ 600
ppb) ou dꢃammoniaque (environ 80 ppb) est possible sans
lꢃutilisation de systꢁmes de dꢀtection particuliers. De faÅon si-
milaire des traces de diphosgꢁne et dꢃautre agent vꢀsicant et
alkylant est possible. Ces systꢁmes sont trꢁs stables (durꢀe su-
pꢀrieure ꢂ un an) et fonctionnent ꢀgalement en phase aqueu-
se. Aucune sꢀlectivitꢀ de dꢀtection nꢃa ꢀtꢀ obtenue mais con-
ceptuellement les systꢁmes sont particuliꢁrement sensibles et
peuvent Þtre modifiꢀs ꢂ faÅon.
AHCTUNGTRENNGrUN eceptor site using well-known strategies.
Results and Discussion
Synthesis and X-ray characterization: Taking account of the
arguments raised above, and making due allowance for re-
lated demands needed for the design of a viable orthogonal
sensor, dyes 1 and 2 were designed. Dye 1 is regarded as a
reference compound with which to judge the significance of
increased conjugation, dye 2 is intended as the prototype
with which to examine our general strategy, while the iodo-
benzene unit attached at the meso-position will provide the
linkage to the porous polymer beads. These dyes were syn-
thesized from the 8-iodophenyl-1,3,5,7-tetramethyl-4-bora-
3a,4a-diaza-s-indacene derivative by condensation with 4-di-
methylaminobenzaldehyde by using piperidine as catalyst.[27]
Both the monosubstituted dye, 1, which is blue, and the di-
AHCTUNGTRENNUNG
1360
www.chemeurj.org
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1359 – 1369

Solid-state Gas Sensors
FULL PAPER
ACHTUNGTRENNUNGamino groups present in 2 is kinetically controlled and the
use of CH3I in acetonitrile gives rise to both the mono-alky-
lated compound 3a (green, 20%) and the dialkylated com-
pound 3b (blue, 55%). The alkylated dyes are stable for
several weeks, but show some degradation after long-term
storage in air at room temperature. Dyes 4 and 5 are
equipped with an alternative anchor. We draw attention to
the similarity between 1 and BODIPY-dyes reported by
Boens et al.[28] The main points of interest in our work con-
cern attaching the dye to an inert support and examining
the advantages of equipping the dye with two identical vinyl
arms.
Compound 1 crystallizes in the monoclinic space group,
P21, whereas crystals of 2 fall within the centrosymmetric
triclinic space group. The two pyrrole rings that make up
the BODIPY core are quasi planar, for both compounds,
with their mean deviations from the least-squares molecular
plane being 0.021 and 0.067 ꢅ, respectively. The structure of
2 is slightly distorted, because the anilino substituents act as
levers, with the maximum deviation from planarity of 0.12
and À0.11 ꢅ being observed for C5 and C7, respectively
(from one outermost pyrrole ring). The iodobenzene unit
lies orthogonal to the BODIPY nucleus due to the presence
of methyl groups on either side, with the dihedral angles
being 86.4 and 83.38 for 1 and 2, respectively (Figure 1).
This orthogonal arrangement is a common feature of struc-
tures computed at various levels for all of the compounds 1–
5; the average dihedral angle is calculated at 858.
Figure 1. Top: ORTEP view of compound 1. Bottom: ORTEP view of
compound 2. Displacement ellipsoids are drawn at the 50% probability
level and H atoms are shown as small spheres of arbitrary radii.
AHCTUNGTREGUNdNN ihedral angles between these mean planes and the
BODIPY nucleus are slightly twisted (10.38 for 1; 12.5 and
17.38 for 2). In the monosubstituted derivative 1, the anilino
subunit is almost aligned with the BODIPY nucleus. How-
ever, in the V-shaped molecule 2 these subunits are directed
above and below the mean plane of the BODIPY platform,
making approximate angles of À10 and 158, respectively.
The molecular arrangement in the crystal structure deter-
mined for 1 is set primarily by directional iodine···methyl in-
In both 1 and 2, the bond lengths and angles around the
boron atom are as might be expected on the basis of other
BODIPY structures.[32,33] The effect of the vinyl group can
À
be seen clearly by comparing the bond lengths for C4 N1
¯
À
À
À
and C5 N2 and for C1 N1 and C8 N2 (Table S1 in the
termolecular interactions of 3.62 ꢅ along the 201 direction,
À
Supporting Information). Note that the I1 C4A bond
linking one molecule at general position x, y, z to two mole-
cules at positions (1Àx, yÀ1/2, Àz) and (1Àx, y+1/2, Àz).
Cohesion is reinforced through short-range interactions be-
tween F2 and hydrogen atoms located at C6A and C9B with
respective positions #i (xÀ1, y, z) and #ii (À1Àx, y+1/2,
1Àz). Additional secondary interactions include C···H p-ring
edge-to-face interactions (e.g., between C5A···H5A and the
lengths are in compliance with the mean value 2.099
(0.026) ꢅ found from a survey of 1075 structures contained
in the CCDC database. The eleven atoms of the anilino sub-
units are essentially co-planar in each structure, with mean
deviations from the respective least-squares molecular plane
of 0.035 ꢅ in 1 and 0.04 and 0.08 ꢅ in 2. The respective
Chem. Eur. J. 2009, 15, 1359 – 1369
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1361

R. Ziessel et al.
Table 1. Spectroscopic data for the compounds measured at room
central six-membered ring at position 1+x, y, z (distance
3.12 ꢅ and angle 162.38), and between C7B···H7B and the
aniline ring at position #ii (distance 2.98 ꢅ and angle
163.38)). The crystal packing can be viewed as chains of tet-
rameric units (see Supporting Information, Figure S1a), with
the iodobenzene–BODIPY moieties stacking along the b
screw axis, appearing head-to-head around position z=0,
and the anilino subunits being edge-to-face linked at z=1/2
(see Supporting Information, Figure S1b).
AHCTUNGTRENNUNG
temperature.[a]
Property
1
2[b]
3a
3b
4
5
labs [nmÀ1
]
596
706
675
617
60800
630
8.0
0.62
8
703
81800
762
3.6
0.21
6
702
82000
762
3.6
0.22
6
[c]
emax [mÀ1 cmÀ1
]
78000 83000 70000
lflu [nmÀ1
]
639
3.2
0.85
27
767
3.0
0.18
6
756
1.2
0.07
6
tF [nsÀ1
]
fF
kRAD [107 sÀ1
]
[d]
kNR [107 sÀ1
]
5
27
77
5
22
22
[d]
For compound 2, the V-shaped molecules are intertwined
across an inversion center (the angle of ca. 508 between the
two arms is sufficient for insertion of a symmetry-related
molecule) and lie parallel to the plane (120), with all iodo-
benzene groups aligned along the same direction. The mole-
cules are arranged with an inclination of 31.18 and without
p–p overlap. This gives the appearance of a staircase-like
pattern, with the aniline subunits as the steps and the iodo-
benzene units as banisters. Adjacent staircases propagating
along the c direction display zigzag sheets (see Supporting
Information Figure S2). The most noticeable stabilization of
the molecular packing can be attributed to a C···H interac-
tion, with a contact distance of 2.61 ꢅ, between the methyl
C9B of the aniline molecule at position Àx, Ày, 1Àz and the
central BODIPY ring. The aniline rings stack over the vinyl
groups, thereby augmenting the overall stabilization of the
structure. We note in passing that single crystals of 1 and 2
are nonfluorescent at ambient temperature. This is in
marked contrast to the ready detection of emission from sol-
utions of these dyes and indicates that the crystalline materi-
al is subjected to strong fluorescence quenching by short-
range interactions.
Molecular modeling studies made for the various com-
pounds clearly indicate a trans arrangement of the vinyl
double bond, as observed by both X-ray diffraction and
1H NMR spectroscopy; calculations made with the STO
3G* basis set conclude that the trans form of 1 is more
stable than the cis form by 22.6 kJmolÀ1. As expected, a
high barrier (Ea =215 kJmolÀ1) is calculated for rotation
around the vinyl double bond (see Supporting Information
Figure S3). The modeling work indicates that the styryl resi-
due in the trans form prefers to align with the dipyrrin nu-
cleus, there being two such geometric arrangements, each of
comparable energy, separated by a barrier of 6Æ2 kJmolÀ1.
These two stable structures have closely comparable excita-
tion energies and are inter-converted by rotation around the
[a] Determined in CH2Cl2, except for 3b where measurements were
made in CH3CN. [b] Same e and l were measured after addition of a
drop of triethylamine or acetic acid. [c] A second peak lies at 562 nm
(emax =36,700mÀ1 cmÀ1). [d] Calculated using the following equations:
k
RAD =fF/tF, kNR =(1ÀfF)/tF.
sorption centered at around 700 nm as found for 2, since
this is the key to understanding much of the ensuing analyti-
cal chemistry. This transition contains an important contri-
bution from intramolecular charge-transfer effects arising
from interaction between the amino donor and the
BODIPY-based acceptor. Such an effect favors establishing
quinoidal resonance structures for charge delocalization and
has been studied in detail for a compound closely related to
1.[28] Alkylation of both amino groups induces a strong hyp-
sochromic shift and the charge-transfer absorption band dis-
appears, since the N lone pair is no longer available to act
as a donor. Consequently, the absorption maximum for 3b is
found at 617 nm. Interestingly, the mono-alkylated com-
pound 3a exhibits a blue shift of 30 nm for absorption and
10 nm for emission spectra relative to 2 (Table 1). To the
best of our knowledge, these are the first ammonium-type
BODIPY derivatives to be well characterized. The absorp-
tion maxima recorded for 4 and 5, which differ only in re-
spect of the nature of the meso substituent, are in the far-
red region and, as expected, are strongly reminiscent of 2
(Table 1). This latter finding can be used to raise the idea
that attaching the dye to an inert support is unlikely to per-
turb the photophysical properties.
The substituents affect the molecular polarizability, as in-
dicated in Figure 2. Relative to the phenyl analogue, both
the N,N-dimethylanilino group and the corresponding
N,N,N-trimethylammonium derivative possess considerably
higher polarizability, with the former being the highest. It is
crucial for the design of an intelligent sensor that these
groups work in a cooperative fashion so as to amplify the
spectral changes that accompany protonation or alkylation
À
C C bond linking the vinyl group to the pyrrole ring (see
Supporting Information Figure S4). The excitation energy
computed for the cis form is only 10 kJmolÀ1 higher than
that of the trans form.
Spectral properties: Each of the compounds displays several
strong p–p* absorption transitions in the 300–750 nm
region; the lowest energy absorption maxima (labs) and
molar absorption coefficients (emax) are collected in Table 1.
Dye 1 has similar spectroscopic features to those exhibited
by the analogous compound lacking the iodine substitu-
ent.[30] Particular attention should be paid to the strong ab-
Figure 2. Contour plots showing the computed molecular polarizability
for 2 (left-hand side), 3a (center) and 3b (right-hand side). Light=
AHCTUNGERTGPNNUN ositive: Dark=Negative.
1362
www.chemeurj.org
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1359 – 1369

Solid-state Gas Sensors
FULL PAPER
of the donor. Thus, the vinyl residue is intended to increase
the conjugation length and thereby push the absorption and
fluorescence maxima to longer wavelength. The anilino
donor is intended to shift these maxima, especially the fluo-
rescence band, to lower energy by way of charge-transfer in-
teractions and to provide the reactive site. The ammonium
cation is intended to serve the dual purpose of removing the
charge-transfer interactions and shortening the conjugation
length by an inductive effect. It is only by combining these
various phenomena that the full impact of the sensor will
become apparent. The polarizability plots indicate that this
level of mutual co-operation is viable. Related displays of
the molecular HOMOs and LUMOs are given in the Sup-
porting Information (Figure S5).
Each new derivative fluoresces in fluid solution at ambi-
ent temperature and the emission maxima (lflu) recorded in
CH2Cl2 are collected in Table 1. For 1, the Stokesꢆ shift is
1130 cmÀ1, which is indicative of the excited state showing
increased charge-transfer character, while the fluorescence
quantum yield (fF) is rather high. The singlet-excited state
decays by mono-exponential kinetics with a lifetime (tF) of
number, decreases markedly. It is this decrease that leads to
an enhancement of kNR, in agreement with the energy-gap
law,[34] and the expected reduction in kRAD. The reduced ex-
citation energy is clearly a consequence of the longer effec-
tive conjugation length and is not to be confused with an in-
creased change in dipole moment.
The absorption and emission maxima recorded for 3a in
CH2Cl2 fall between those determined for 1 and 2. Com-
pound 3a has two vinyl substituents, but only one terminal
donor group. The effect of removing one donor group is to
push lmax from 706 to 675 nm and to lower lflu from 767 to
756 nm. This is due to the inductive effect of the ammonium
ion and serves to lower the effective conjugation length. The
Stokesꢆ shift is 1590 cmÀ1 and is the highest of all the com-
pounds described here. The dipole moment is raised relative
to 1, because the BODIPY-based acceptor is made more
electron affinic by the second vinyl substituent. In 2, the
dipole moment is kept modest by having two donors com-
peting for the same acceptor. The derived kRAD and kNR
values now represent a compromise between the increased
dipole moment (lowering kRAD but raising kNR) and the in-
creased excitation energy (raising kRAD but lowering kNR).
On removing the remaining terminal donor, as for 3b, the
absorption and emission spectra move to higher energy and
the Stokesꢆ shift falls to 335 cmÀ1. This value indicates the
absence of a significant geometry change on excitation and
is consistent with the lack of intramolecular charge-transfer
character. Both fF and tF increase to values typical of con-
ventional BODIPY-based dyes, allowing for the relatively
low excitation energy. It is important to note that whereas
fluorescence from 2 is outside of the visible region, that
from 3b is clearly apparent to the naked eye. It should also
be emphasized that the vinyl groups inherent to 2 do not
provide extra decay channels, such as isomerization.
3.2 ns. For this compound, the radiative rate constant (kRAD
)
has a value of 2.7ꢇ108 sÀ1; this value is in reasonable agree-
ment with other BODIPY derivatives bearing a single vinyl
group at the 2-position. Indeed, the photophysical properties
recorded for 1 are in good accord with those reported for a
closely-related BODIPY-based dye, taking due account of
the pronounced solvent dependence reported previously.[28]
The rate constant for nonradiative decay (kNR) is relatively
unimportant for this compound. Increasing the conjugation
length, as for 2, pushes the emission wavelength to lower
energy, but has no effect on the Stokesꢆ shift. There is a
marked decrease in the fluorescence quantum yield relative
to 1, a sharp drop in kRAD and a corresponding increase in
kNR. Compounds 4 and 5 exhibit fF and tF values that
remain comparable to those observed for 2, indicating that
the iodine atom is too remote to influence the photophysical
properties. The Stokesꢆ shifts measured for 4 and 5 are also
similar to that of 2, as might be expected if the nature of the
fluorophore remains constant (Table 1).
pH Titration: During the pH titration of 1 in acetonitrile,
with HCl as the proton source, the absorption maximum
AHCTUNGTREGuNNUN ndergoes a hypsochromic shift from 597 to 553 nm, the
fluorescence maximum moves from 717 to 561 nm, and the
fluorescence quantum yield increases from 0.09 to 1.0 upon
protonation of the terminal donor N atom. Under the same
conditions, kRAD increases from 7.5ꢇ107 to 2.4ꢇ108 sÀ1. On
the assumption that HCl is fully dissociated in acetonitrile,
the titration data correspond to a pKa of 2.25 for the donor
N atom. We have been concerned with extending this pH ti-
tration to 2, knowing that N-alkylation occurs in two well-
resolved steps. Control experiments confirmed that addition
of acid (i.e., HCl or CF3COOH) had no effect on the ab-
sorption and fluorescence spectra recorded for 3b in aceto-
nitrile or CH2Cl2.
Addition of excess HCl to 2 in CH3CN results in a hypso-
chromic shift from 700 to 620 nm, this absorption maximum
being reminiscent of that found for 3b. The fluorescence
maximum shows a similar shift, moving from 780 to 630 nm,
and a substantial increase in both fF (from 0.18 to 0.68) and
tF (from 3.0 to 5.6 ns). There is a sharpening and intensifica-
tion of the absorption and fluorescence bands on formation
It is informative at this point to make a critical compari-
son of the photophysical properties of 1 and 2 and subse-
quently to broaden the discussion to include 3a. On the
basis of the detailed studies made for somewhat similar de-
rivatives, it can be argued that the excited-singlet state of 1
has considerably more polar character than the correspond-
ing ground state. The measured kRAD and kNR values can be
taken as referring to decay of a charge-transfer state, with
the actual values being related to the magnitude of the
change in dipole moment that occurs on excitation. An in-
creased dipole moment, as realized by moving to a solvent
with higher dielectric constant, will increase kNR and de-
crease kRAD. This is also what happens when the second elec-
tron-donating arm is added to 1. Here, the dipole moment
remains unaffected, but the excitation energy, defined as the
intersection point between normalized absorption and fluo-
rescence spectra after conversion from wavelength to wave-
Chem. Eur. J. 2009, 15, 1359 – 1369
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1363

R. Ziessel et al.
of the dicationic species (Figure 3) and a corresponding in-
crease in kRAD from 6.0ꢇ107 to 1.2ꢇ108 sÀ1. However, the ti-
tration clearly proceeds by way of an intermediate species
that bears the hallmarks of the monoprotonated species
(Figure 4).
Figure 3. Absorption and fluorescence spectra recorded for 2 at different
stages of protonation: neutral species (dark grey curve), monocationic
species (light grey curve) and dicationic species (black curve) in CH3CN.
Thus, stepwise addition of HCl to 2 in CH3CN leads to
the progressive replacement of the strong absorption band
centered at 700 nm with a new band centered at 670 nm
(Figure 4). This latter band is relatively broad and retains a
significant amount of charge-transfer character. There are
several clear isosbestic points. Further addition of HCl re-
sults in the loss of the 670 nm band and the concomitant
evolution of an absorption band centered at 620 nm. No fur-
ther changes occur on addition of excess acid, but there are
at least three evident isosbestic points. The 620 nm band is
notably narrower and more intense than the original 700 nm
band and is taken to be indicative of the loss of charge-
transfer character. There are accompanying changes in the
near-UV region. The absorption spectra derived for the
mono- and diprotonated species are comparable to those
observed for the corresponding N-methylated analogues
(Table 1). Similar absorption spectral changes were seen
when the proton source was CF3COOH and in CH2Cl2.
The titration was also followed by fluorescence spectros-
copy (Figure 5). For 2 in CH3CN, the fluorescence maximum
lies at 785 nm and corresponds to a Stokesꢆ shift of
1630 cmÀ1. Addition of small amounts of HCl causes the
fluorescence intensity to decrease markedly and, although
difficult to see by eye, there is a shift in the emission maxi-
mum to 772 nm. On further addition of HCl, the emission
maximum moves to 630 nm and the intensity grows steadily.
The 630 nm band is clearly due to the diprotonated species,
for which the Stokesꢆ shift is only 280 cmÀ1. As noted for 3b,
this species possesses little if any charge-transfer character.
The intermediate species can be attributed to the monopro-
tonated material, for which the Stokesꢆ shift is increased to
1970 cmÀ1, and that retains a relatively broad spectral profile
Figure 4. Titration of 2 in CH3CN with HCl. The lower panel shows the
initial part of the titration, while the upper panel focuses on the second
protonation step. The curve labeled with an asterix corresponds to the
starting solution. Sufficient HCl is then added to convert the neutral spe-
cies to the monocation so that the second protonation can be followed.
Figure 5. Fluorescence spectra recorded for 2 after progressive addition
of HCl in CH3CN (excitation was made at the isosbestic point at
548 nm).
consistent with considerable charge-transfer character. Thus,
a major consequence of diprotonation is that the dipole
1364
www.chemeurj.org
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1359 – 1369

Solid-state Gas Sensors
FULL PAPER
moment (i.e., the Stokesꢆ shift) vanishes. Protonation of the
first N atom removes that particular donor site and should
thereby increase the dipole moment. This, in turn, will in-
crease the state of hybridization of the second N atom,
which will become more difficult to protonate. It is recog-
nized, however, that the ammonium ion exerts an inductive
effect that will keep the dipole moment to a modest level. It
is this concerted behavior that allows separation of the suc-
cessive protonation steps. Although there will be an associ-
ated electrostatic effect, the two N atoms are too far apart
for this to be the primary cause of the disparate reactivity of
these donor sites.
Analysis of the absorption spectral changes caused by ad-
dition of HCl in CH3CN in terms of two successive protona-
tion steps gives pKa values of 3.04 and 5.10, respectively
(see Supporting Information Figures S6 and S7). These are
averaged values obtained by fitting the data to conventional
pKa transitions by using SPECFIT and also by using the ra-
tiometric method,[36] taking advantage of the well-preserved
isosbestic points. In the former case, the entire spectral
window was used, but in the latter case data analysis was re-
stricted to ten individual wavelengths. The fluorescence ti-
tration data were analyzed using SPECFIT to give pKa
values of 3.0 and 5.2. Unfortunately, the first pKa is difficult
to assess by fluorescence spectroscopy, because it is on the
limit of our detection window, but the second protonation
step can be followed easily by the appearance of emission
around 630 nm. Overall, the derived pKa values for the two
steps are consistent among the various measurements.
washing, the beads were isolated by sieving. Linkage of the
dye to the beads was inspired by peptide synthesis,[37] and
was made feasible by using a solution of 5 (1.3 mmol), N-(3-
dimethylaminopropyl)-N’-ethyl-carboimide
hydrochloride
(10 mol%) and 5-dimethylaminopyridine (15 mol%) in a
mixture of THF (5 mL) and CH3OH (1 mL). After shaking
overnight, analysis of the solution by absorption spectropho-
tometry allowed us to conclude that 42 mol% of 5 had been
grafted onto 40 mg of beads. This corresponds to an approx-
imate loading of 13 nmol of dye per bead. After prolonged
washing, the resultant beads were green. In contrast, blank
experiments carried out with dye and Amberzyme Oxirane
beads, but without prior treatment with 1,3-diaminopropane,
revealed no surface loading and no color change. The specif-
ic surface area, as determined by the chemisorption of N2,
of the microspheres is about 220 m2gÀ1 and the cumulative
surface area of the pores is about 190 m2gÀ1, clearly corre-
sponding to highly porous material. Scanning electron mi-
croscopy (SEM) showed the mean diameter to be about
200 mm. The size distribution of the pores lies in the range
of 50 to 100 nm, indicating that the dye can be bound to
both outer and inner surfaces. In some instances, SEM
shows the presence of strings of larger pores (Figure 6). This
level of loading is reproducible and the bound dye is stable
with respect to leaching from the surface after prolonged ex-
posure in air to light or solvent.
ACHTUNGTRENNUNG
a
switched-on at quite low concentrations of acid. In contrast,
diprotonation induces substantial spectral shifts, but de-
mands moderately high acid concentrations. It is notable
that the first pKa value is higher than that determined for a
derivative of 1, despite the comparable experimental condi-
tions.[28] This behavior might have been expected in view of
the significantly lower dipole moment exhibited by 2, which
will favor a higher pKa value. Even so, it is a fairly large
change in reactivity that serves to demonstrate the sensitivi-
ty of the protonation process. Likewise, the difference in
pKa values noted for addition of one and two protons indi-
cates that the ammonium cation exerts a strong effect on
the electronic properties of the molecule.
Figure 6. Examples of SEM photographs of the functionalized bead sur-
face (left) and a zoom of the surface showing the pore distribution
(right).
Covalent attachment to porous beads: To fix the dye to a
macroscopic support it proved necessary to master the
chemistry of the iodo function without harming the fluores-
cence properties. Thus, cross-coupling of 2 with ethyl hept-6-
ynoate by using Pd0 as catalyst under mild conditions afford-
ed 4 in excellent yield. Hydrolysis of the ester was straight-
forward and gave 5 in quantitative yield. In an effort to link
this dye to Amberzyme OxiraneTM beads, which have a low
dispersity with average diameter of 200 mm, we first treated
the oxirane-functionalized beads with excess 1,3-diamino-
propane in anhydrous THF in the presence of LiClO4.[38]
After vigorous shaking, followed by subsequent drying and
Examination of the capability of the loaded beads to
detect trace quantities of acid present in a gas stream is il-
lustrated in Figure 7. The beads are packed into an open-
ended microcapillary and exposed to the gas flow. On illu-
mination with a 365 nm UV lamp, the fluorescence of the
dye adhered to the surface of the beads lies at 770 nm, but
could not observed by the naked eye. Instead, the beads
appear violet-blue due to very weak background fluores-
cence from the treated surface (Figure 7b); note untreated
beads do not fluoresce. When the gas flow was spiced with a
Chem. Eur. J. 2009, 15, 1359 – 1369
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1365

R. Ziessel et al.
notably a laser diode as excitation source and a photocell as
detector, even without signal amplification. Systematic expo-
sure of the beads to aqueous solutions allowed the pKa for
the acid/base transition to be identified as lying in the 2.4 to
2.6 pH range, as determined by the “naked-eye”. Clearly,
this refers to addition of the second proton and is reasona-
bly consistent with the results of the pH titrations carried
out for 2 in acetonitrile.
The same concept can be applied to the detection of trace
amounts of highly toxic chemicals and to vesicant warfare
gases, provided they contain acidic or alkylating groups.
Thus, loaded beads packed into a glass capillary and illumi-
nated at 365 nm display pronounced fluorescence at around
800 nm that can be monitored with a photocell, but is out-
side of the range detectable by the human eye (Figure 8).
Figure 7. Open-ended capillary tube packed with macroscopic beads
functionalized with 13 nmol of 5 as observed with an optical microscope
at 4ꢇ magnification. From the top: a) Exposure of the beads to a current
of air loaded with a trace of HCl gas. b) The same capillary tube but ob-
served by fluorescence with excitation at 365 nm. c) Exposure of the pro-
tonated beads to a current of air loaded with a trace amount of NH3. d)
The same capillary tube but observed by fluorescence with excitation
with a non-filtered 365 nm bench UV lamp. Note, the blue color ob-
served in b) and d) is due to weak inherent fluorescence from surface-
treated beads. Fluorescence from the green dye occurs around 800 nm
and cannot be observed by the naked eye.
Figure 8. Steady-state fluorescence spectra recorded for beads packed
into a quartz capillary tube; normal beads in air (grey curve) and beads
exposed briefly to a trace amount of diphosgene vapor (black curve).
The inserts show a single bead as observed by an optical fluorescence
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
trace of HCl, the beads instantaneously turned blue and
there was concomitant appearance of an intense red fluores-
cence (Figure 7c). The lower limit for detection of HCl in
air, as determined by the naked eye for illumination with a
low intensity UV source, lies in the 400 to 800 ppb range.
The uncertainty is due to the very low concentration of acid.
Each experiment was repeated five times and the quoted
range represents the lower and upper limits of detection
that were encountered. Furthermore, when the gas flow con-
tains trace amounts of ammonia, the blue beads revert to
their original color and the red fluorescence disappears
(Figure 7).
This on–off color switching is stable for at least one year
and there was no decrease in efficacy after more than ten
complete cycles carried out in rapid succession. The detec-
tion limit for ammonia is between 60 and 120 ppb by using
the same “naked eye” approach. The resultant NH4Cl de-
posits on the capillary walls or is removed by the gas flow.
This detection limit was improved by two orders of magni-
tude by using routine optical spectroscopic equipment, most
However, when trace amounts of diphosgene were allowed
to enter the capillary tube there was an immediate color
change from green to blue.[39] This transition was accompa-
nied by the appearance of intense red fluorescence; this
emission is centered at 650 nm and can be seen easily by the
human eye (Figure 8). It was not possible to estimate the
lower detection level for phosgene, because of the difficulty
in handling this material under quantitative conditions in
our laboratory. The lower level, however, is at least as good
as that found for HCl. This protocol can be adapted for
facile detection of other highly toxic materials such as nitro-
gen and sulfur mustard gases, phosphine, arsines, and HCN.
The visual detection process (red fluorescence) is unaf-
fected by humidity or even when the gas stream is saturated
with water. Hence, when the beads are dispersed in aqueous
solution the on–off detection process continues to operate
on addition of trace amounts of HCl or HCOOH. The use
of triethylamine or piperidine vapors in place of ammonia is
also effective in extinguishing the red fluorescence. We em-
1366
www.chemeurj.org
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1359 – 1369

Solid-state Gas Sensors
FULL PAPER
phasize, however, that there is no selectivity for any particu-
lar acid or base in either solution or gaseous states.
shaped geometry of 2, in which the initial electronic system
is best represented as push–pull–push. After monoalkyla-
tion, or monoprotonation, the electronic system becomes
push–pull–pull with a concomitant enhancement of the
dipole moment.
Conclusion
The current set-up has been designed to operate as a simple
saturation-type optical detector for monitoring the presence
of trace amounts of acidic or electrophilic pollutants. The
dye-coated beads are highly stable and exhibit a green-to-
blue color transformation on exposure to acid; the reverse
process is available for the detection of bases, such as am-
monia or organic amines, simply by pre-exposure of the
beads to acid. There is a concomitant change in the fluores-
cence maximum from 800 to 650 nm; this has the effect of
moving the emission from the far-red to the visible region.
By using a capillary tube, the device is easily saturated and
this makes for facile on–off sensing that can be followed by
the naked eye. The green to blue color change is nicely com-
plemented by the emergence of an intense red emission,
thereby fulfilling the basic requirement for orthogonal sens-
ing. Under spectroscopic conditions, the most important de-
tection mode involves the parallel monitoring of the absorb-
ance change at 720 nm and the accompanying modulation of
the fluorescence intensity at 650 nm. An advantage of the
capillary tube is that the dye is quickly saturated, but many
facile modifications are possible. Thus, the glass capillary
can be replaced with a porous tube and used to monitor aer-
osols or circulating liquids. At the present stage, the sensor
is not specific towards the nature of the acid or base, but it
should be emphasized that this prototype could be chemical-
ly modified at the aromatic amine site so as to produce
shape-specific molecular pockets[40] able to recognize certain
species.
The main motivation for synthesizing a BODIPY-based
dye equipped with two vinylic arms relates to the position-
ing of the relevant optical absorption and fluorescence max-
imum. This represents a deliberate attempt to generate the
most visual changes upon contacting the beads with the sub-
strate. An unexpected feature of this strategy relates to the
relative inactivation of the second nitrogen atom caused by
alkylation or protonation of the first anilino N atom. This
leads to a marked disparity in the respective pKa values. Al-
though reaction at the first N atom does not cause dramatic
spectral changes it is notable that the pKa is sufficiently high
for sensitive detection of substrates at low concentration.
The second pKa value is such that high concentrations of
substrate are required. In terms of sensor technology, this
situation could be exploited to develop dual-purpose detec-
tors. Here, reaction at the first N atom would send a warn-
ing signal to the operator to the effect that low concentra-
tions of the substrate had entered the system. Reaction at
the second N atom would be used to signal that the process
must be shut down. In a chemical sense, the disparate pKa
values relate to a changeover from the strongly donating
nature of the amino group to the inductive effect inherent
to the ammonium group. The effect is amplified by the V-
Experimental Section
Synthesis and characterization of compounds 1 and 2: Prepared accord-
ing to previously published procedures for isolation of monosubstituted
compounds,[26] from 8-(4-phenyliodo)-tetramethyldifluoroboradipyrrome-
thene (500 mg, 1.11 mmol) and 4-dimethylaminobenzaldehyde (362 mg,
2.44 mmol) in a mixture of toluene (20 mL) and piperidine (0.5 mL) con-
taining a single crystal of p-TsOH at 1208C for 1 d. Chromatography on
silica gel, eluting with a gradient of dichloromethane-petroleum ether
(v/v 30:70) to dichloromethane, gave 1 (168 mg, 26%) as a deep-blue
solid and 2 (459 mg, 58%) as a deep-green solid after recrystallization
from a dichloromethane–cyclohexane mixture.
Data for 1: 1H NMR ([D6]DMSO, 400 MHz): d=7.76 (d, 3J=17.2 Hz,
1H), 7.62 (d, 3J=8.8 Hz, 2H), 7.44 (d, 3J=17.2 Hz, 1H), 7.38 (d, 3J=
8.8 Hz, 2H), 7.12 (d, 3J=8.7 Hz, 2H), 6.82 (d, 3J=8.7 Hz, 2H), 6.75 (s,
1H), 6.04 (s, 1H), 3.14 (s, 6H), 2.62 (s, 3H), 1.51 (s, 3H), 1.48 ppm (s,
3H); ESI-MS in CH3OH + 1% TFA: m/z (%): 582.2 (100) [M+H]+; el-
emental analysis calcd (%) for C28H27N3IBF2: C 57.86, H 4.68, N 7.23;
found: C 57.57, H 4.44, N 7.00.
Data for 2: 1H NMR ([D6]DMSO, 400 MHz): d=7.93 (d, 3J=8.3 Hz,
2H), 7.47 (d, 3J=8.8 Hz, 4H), 7.43 (d, 3J=17.1 Hz, 2H), 7.30 (d, 3J=
17.1 Hz, 2H), 7.24 (d, 3J=8.3 Hz, 2H), 6.89 (s, 2H), 6.79 (d, 3J=8.8 Hz,
4H), 3.00 (s, 12H), 1.43 ppm (s, 6H); ESI-MS in CH3OH + 1% TFA:
m/z (%): 713.2 (100) [M+H]+; elemental analysis calcd (%) for
C37H36N4IBF2: C 62.38, H 5.09, N 7.86; found: C 62.12, H 4.83, N 7.57.
Synthesis and characterization of compounds 3a and 3b: The green dye
2 (100 mg, 0.14 mmol) dissolved in acetonitrile (10 mL) was allowed to
react with CH3I (3 mL) for 48 h at RT. The course of reaction was fol-
lowed by TLC on silica gel, using a mixture of acetonitrile/water as
eluant (85:15). After disappearance of the starting material, the deep-
blue solution was evaporated to dryness and dissolved in a mixture of
water/methanol. A tenfold excess of KPF6 dissolved in water was added
dropwise, resulting in precipitation of the salt. The organic phase was
evaporated to dryness and the precipitate was washed with water and di-
ethyl ether. Purification was ensured by column chromatography on alu-
mina using acetonitrile/water as solvent 85:15 v/v. The first compound to
be eluted was 3a, which, after recrystallization in acetone, was isolated as
a greenish-blue solid (26 mg, 20%). The second compound to be eluted
from the column was recrystallized by slow evaporation of acetone from
a
mixture of acetone/cyclohexane to give 3b as deep-blue crystals
(80 mg, 55%).
Data for 3a: 1H NMR ([D6]DMSO, 400 MHz): d=7.72 (d, 3J=17.1 Hz,
2H), 7.60 (d, 3J=8.8 Hz, 4H), 7.39 (d, 3J=17.1 Hz, 2H), 7.41 (d, 3J=
8.8 Hz, 4H), 7.15 (d, 3J=8.6 Hz, 2H), 6.89 (d, 3J=8.6 Hz, 2H), 6.84 (s,
1H), 6.17 (s, 1H), 4.38 (s, 9H), 3.10 (s, 6H), 1.54 (s, 3H), 1.49 ppm (s,
3H); ESI-MS in CH3CN: m/z (%): 727.3 (100) [MÀPF6]+; elemental
analysis calcd (%) for C38H39N4IBF2PF6: C 52.32, H 4.51, N 6.42; found:
C 52.52, H 4.66, N 6.57.
Data for 3b: 1H NMR ([D6]DMSO, 400 MHz): d=7.97 (d, 3J=8.4 Hz,
2H), 7.49 (d, 3J=8.7 Hz, 4H), 7.45 (d, 3J=17.0 Hz, 2H), 7.26 (d, 3J=
17.0 Hz, 2H), 7.17 (d, 3J=8.4 Hz, 2H), 6.99 (s, 2H), 6.83 (d, 3J=8.7 Hz,
4H), 4.23 (s, 18H), 1.54 ppm (s, 6H); ESI-MS in CH3CN: m/z (%): 887.2
(80) [MÀPF6]+, 371.1 (100) [MÀ2PF6]2+; elemental analysis calcd (%)
for C39H42N4IBF2P2F12: C 45.37, H 4.10, N 5.43; found: C 45.22, H 3.89, N
5.29.
Synthesis and characterization of compound 4: Compound 4 Prepared
from compound 2 (100 mg, 0.140 mmol) and HC=CC4H8COOEt (32 mg,
0.210 mmol) in a mixture of THF (5 mL), diisopropylamine (2 mL), [Pd-
AHCTUNGTRENN(GUN PPh3)2Cl2] (6 mg) and CuI (6 mg) at RT for one night. After that time,
Chem. Eur. J. 2009, 15, 1359 – 1369
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1367

R. Ziessel et al.
the mixture was evaporated to dryness and treated with water (5 mL),
before being extracted with dichloromethane (3ꢇ50 mL). The organic
phase was dried over MgSO4 and evaporated to dryness. The pure com-
pound was obtained by chromatography on silica gel, eluting with di-
chloromethane-petroleum ether (v/v 60:40) to give 4 as a deep-green
solid (100 mg, 97%) after recrystallization from a dichloromethane–
cyclohexane mixture. 1H NMR ([D6]DMSO, 400 MHz): d=7.89 (d, 3J=
8.2 Hz, 2H), 7.43 (d, 3J=8.7 Hz, 4H), 7.37 (d, 3J=16.9 Hz, 2H), 7.28 (d,
3J=16.9 Hz, 2H), 7.18 (d, 3J=8.2 Hz, 2H), 6.82 (m, 6H), 4.17 (q, 3J=
7.1 Hz, 2H), 2.94 (s, 12H), 2.38–2.25 (m, 4H), 1.89–1.65 (m, 4H), 1.42 (s,
6H), 1.28 ppm (t, 3J=7.1 Hz, 3H); ESI-MS: m/z (%): 739.2 (100)
[M+H]+, 719.3 (20) [MÀF]+; elemental analysis calcd (%) for
C46H49N4O2BF2: C 74.79, H 6.69, N 7.58; found: C 74.52; H 6.39; N 7.27.
beads I (1 g, co-polymer containing ethyleneglycoldimethacrylate 70%
and glycidylmethacrylate 30%) were suspended in a solution of anhy-
drous THF (5 mL), diaminopropane (5 mL), and LiClO4 (200 mg) and
shaken overnight at RT. The deposited white beads were washed with
THF, dichloromethane, methanol, and diethyl ether. The resultant sur-
face-modified beads II were dried under high vacuum overnight.
Dye 5 (2 mg) was dissolved in a solution containing THF (5 mL), metha-
nol (1 mL), N-(3-dimethylaminopropyl)-N’-ethyl-carboimide hydrochlo-
ride (0.3 mg, 10 mol%) and 4-dimethylaminopyridine (0.2 mg, 15
mol%). Beads II (40 mg) were added to this fresh mother liquor (3 mL),
and the resultant mixture was shaken for 8 h. After decantation, the
beads were washed copiously with THF (2ꢇ15 mL), dichloromethane
(2ꢇ25 mL), methanol (15 mL), and diethyl ether (15 mL). The resultant
beads were green and the absorbance of the residual solution was used
to determine the level of dye loaded onto the surface. On average, a
single bead contains 13 nmol of the dye (about 2 mg per bead of 200 mm
size).
Synthesis and characterization of compound 5: Saponification of
4
(70 mg, 0.09 mmol) in a mixture of THF (5 mL) and methanol (3 mL)
was carried out using NaOH at 608C for 16 h. The course of reaction was
followed by TLC, clearly showing the formation of a green polar com-
pound at the expense of the nonpolar starting material. On completion
of the reaction, the pH was adjusted to 7.0 by using a solution of dilute
HCl. The target acid was extracted with dichloromethane and the organic
phase dried over MgSO4. The analytically pure sample was obtained
Blank experiment: Beads I (40 mg) were added to the fresh mother
liquor (3 mL), and the resulting mixture was shaken for 8 h and treated
as described above. The recovered beads remain white and the solution
has the same concentration of dye as the mother liquor, according to ab-
sorption spectroscopy.
after recrystallization from
a mixture of dichloromethane, methanol
(trace) and cyclohexane providing of 5 as deep-green crystals (63 mg,
99%). 1H NMR ([D6]DMSO, 400 MHz): d=7.94 (d, 3J=8.3 Hz, 2H),
7.47 (d, 3J=8.7 Hz, 4H), 7.39 (d, 3J=17.2 Hz, 2H), 7.34 (d, 3J=17.2 Hz,
2H), 7.27 (d, 3J=8.3 Hz, 2H), 6.89 (m, 6H), 3.03 (s, 12H), 2.47–2.18 (m,
4H), 1.92–1.67 (m, 4H), 1.50 ppm (s, 6H); ESI-MS: m/z (%): 711.2 (20)
[M+H]+, 691.3 (100) [MÀF]+; elemental analysis calcd (%) for
C44H45N4O2BF2: C 74.36, H 6.38, N 7.88; found: C 74.19, H 6.17, N 7.62.
General conditions: 400 MHz 1H NMR spectra were recorded at room
temperature by using the residual proton resonances in deuterated sol-
vents as internal references. Chromatographic purification was conducted
using Silica gel Si-60 (40–63 mm). Thin-layer chromatography (TLC) was
performed on silica gel plates coated with fluorescent indicator. All mix-
tures of solvents are given as v/v ratios. Synthetic details, including com-
pound characterization, are given in the Supporting Information.
Gas detection: To determine the “naked eye” detection limit for gaseous
HCl (99.9% purity) and ammonia (99% purity), a Schlenk style tech-
nique was developed. Pure gases were stored in a calibrated flask, de-
gassed under high vacuum before being filled at atmospheric pressure
and room temperature. The flask was sealed with a syringe valve. Known
volumes of pure gas were withdrawn by using gas-tight syringes and di-
luted with known quantities of dry air. Each experiment was repeated
five times and the quoted detection range corresponds to the lower and
upper limits of the abilities of different observers to detect the color
changes using both absorption and fluorescence approaches.
Spectroscopic titrations: Absorption and fluorescence spectral titrations
were made with solutions of 2 in CH3CN following addition of standard
solutions of HCl in CH3CN. A known volume of the dye solution was
stirred at 208C and aliquots of HCl solution added by means of a cali-
brated syringe. The solution was equilibrated before recording the spec-
trum. Several different concentrations of acid were used as the titration
progressed. The results were corrected for dilution and analyzed with
SPECFIT. Each titration was repeated four times. Corresponding titra-
tions were carried out with trifluoroacetic acid in both CH3CN and
CH2Cl2.
Spectroscopic measurements: Electronic absorption and emission spectra
were measured under ambient conditions using commercial instruments.
Fluorescence spectra were recorded with a Yvon–Jobin Fluorolog tau-3
equipped with a near-IR photodiode detector and with gratings blazed at
700 nm. Spectral imperfections were corrected by reference to a standard
lamp. Solvents for spectroscopy were spectroscopic grade and were used
as received after checking for impurities. A wide variety of excitation
wavelengths were used, according to the species under investigation.
Spectral titrations were carried out with excitation at isosbestic points.
Fluorescence quantum yields were measured relative to Cresyl Violet for
2[41] and tetramethoxy-bis-isoindolodipyrromethene-difluoroborate for 2,
4, and 5.[42] Luminescence lifetimes were measured by using the time-cor-
related, single-photon counting mode coupled to a Stroboscopic system.
The excitation source was a thyratron-gated flash lamp filled with nitro-
gen gas. No filter was used for the excitation. The instrumental response
function was determined with a scattering solution.
Acknowledgement
We thank the CNRS, ANR and the University Louis Pasteur of Stras-
bourg for financial support of this work. Dr. J. Hawecker (Rohm and
Haas, France) is thanked for help with the beads. Drs. T. Dintzler and D.
Begin (LMSPC, Strasbourg) are thanked for the SEM and BET
X-ray studies: Crystallographic measurements were carried out at
293(2) K on a Nonius kappa-CCD diffractometer by using monochromat-
ic graphite MoKa radiation (l=0.71073 ꢅ). Accurate unit cell parameters
and orientation matrices were obtained from least-squares refinement.
Reflections were processed with Denzo[42] and scaled in Scalepack[43]
after post-refinement of the unit cell parameters. The structures were
solved by direct methods (SHELXS 97) and refined by full-matrix, least-
squares techniques on F2 using SHELXL 97.[44] Differences in diffraction
power were observed between thin-plate crystals for compound 1 and 2,
AHCTUNGTREGUNmNN easurements.
[1] a) A. P. De Silva, H. Q. N. Gunaratne, T. Gunnlaugsson, A. J. M.
Huxley, C. P. McCoy, J. T. Rademacher, T. E. Rice, Chem. Rev. 1997,
97, 1515–1566; b) Chemosensors of Ion and Molecular Recognition
(Eds.: J.-P. Desvergne, A. W. Czarnik) Kluwer, Dordrecht, 1997;
c) Chemical Sensors and Biosensors for Medical and Biological Ap-
plications (Ed.: U. E. Spichiger-Keller), Wiley-VCH, Weinheim,
1998; d) P. D. Beer, P. A. Gale, Angew. Chem. 2001, 113, 502–532;
Angew. Chem. Int. Ed. 2001, 40, 486–516; e) H. Tsukube, S. Shino-
da, Chem. Rev. 2002, 102, 2389–2404; f) Molecular Fluorescence:
Principles and Applications (Ed.: B. Valeur), Wiley-VCH, Wein-
heim, 2002.
À
resulting in C C bond length precision about three times better for the
molecular structure of compound 1 (0.006 vs 0.018 ꢅ). Details of the
data collection and structure refinements are given in the Supporting In-
formation. Quantum Chemical calculations were performed with SPAR-
TAN version 06 (Wavefunction Inc).
Functionalization of the polyacrylate beads (AMBERZYMEꢀ
ACHTUNGTRENNUNGOXIRANE from Rohm and Haas, France): The Amberzyme oxirane
[2] See, for example: a) Coord. Chem. Rev. 2000, 205, 1–232 (whole
issue); b) A. B. Ellis, D. R. Walt, Chem. Rev. 2000, 100, 2477–2738;
1368
www.chemeurj.org
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1359 – 1369

Solid-state Gas Sensors
FULL PAPER
c) C. Bargossi, M. C. Fiorini, M. Montalti, L. Prodi, N. Zaccheroni,
Coord. Chem. Rev. 2000, 208, 17–32; d) L. Prodi, F. Bolletta, M.
Montalti, N. Zaccheroni, Coord. Chem. Rev. 2000, 205, 59–83;
e) A. P. De Silva, G. D. Mclean, T. S. Moody, S. M. Weir in Hand-
book of Photochemistry and Photobiology Vol. 3 (Ed.: H. S. Nalwa),
American Institute of Physics, Stevenson Ranch 2003, pp. 217–238;
f) P. Jiang, Z. Guo, Coord. Chem. Rev. 2004, 248, 205–229; g) L.
Prodi, New J. Chem. 2005, 29, 20–31.
[24] W. Qin, M. Baruah, M. van De Auweraer, F. C. Schryver, N. Boens,
J. Phys. Chem. A 2005, 109, 7371–7384.
[25] G. Ulrich, C. Goze, M. Guardigli, A. Roda, R. Ziessel, Angew.
Chem. 2005, 117, 3760–3764; Angew. Chem. Int. Ed. 2005, 44, 3694–
3698.
[26] C.-W. Wan, A. Burghart, J. Chen, F. Bergstrçm, L. B.-A. Johansson,
M. F. Wolford, T. G. Kim, M. R. Topp, R. M. Hochstrasser, K. Bur-
gess, Chem. Eur. J. 2003, 9, 4430–4441.
[3] a) O. S. Wolfbeis in Fiber Optic Chemical Sensors and Biosensors II,
CRC, Boca Raton, 1991; b) P. Simon, F. Kvanik in Optical Sensors
for Industrial, Environmental and Clinical Applications (Eds.: R.
Narayanaswamy, O. S. Wolfbeis), Springer, Heidelberg, 2003,
pp. 441–465.
[27] a) R. P. Haughland, H. C. Kang, U.S. Patent 4.74.339, 1988; b) K.
Rurack, M. Kollmansberger, J. Daub, New J. Chem. 2001, 25, 289–
292; c) Z. Dost, S. Atilgan, E. U. Akkaya, Tetrahedron 2006, 62,
8484–8488; d) S. Atilgan, Z. Ekmekci, A. L. Dogan, D. Guc, E. U.
Akkaya, Chem. Commun. 2006, 4398–4400.
[4] M. G. Baron, R. Narayanaswamy, S. C. Thorpe, Sens. Actuators B
1996, 34, 511–515.
[5] K. Nakagawa, T. Kitagawa, Y. Sadaoka, Sens. Actuators B 1998, 52,
10–14.
[6] H. Supriyatno, M. Yamashita, K. Nakagawa, Y. Sadaoka, Sens. Ac-
tuators B 2002, 85, 197–204.
[7] E. Wang, K.-F. Chow, W. Wang, C. Wong, C. Yee, A. Persad, J.
Mann, A. Bocarsly, Anal. Chim. Acta 2005, 534, 301–306.
[8] R. Gvishi, R. Reisfeld, Chem. Phys. Lett. 1989, 156, 181–186.
[9] Y. Sadaoka, Y. Sakai, Y. Murata, Talanta 1992, 39, 1675–1679.
[10] Y. Sadaoka, Y. Sakai, M. Yamada, J. Mater. Chem. 1993, 3, 877–881.
[11] T. A. Dickinson, J. White, J. S. Kauer, D. R. Walt, Nature 1996, 382,
697–700.
[28] M. Baruah, W. Qin, C. Flors, J. Hofkens, R. A. L. Vallꢁe, D. Bel-
jonne, W. M. Van de Auweraer, M. De Borggraeve, N. Boens, J.
Phys. Chem. A 2006, 110, 5998–6009.
[29] T. Rohand, W. Qin, N. Boens, W. Dehaen, Eur. J. Org. Chem. 2006,
4658–4663.
[30] K. Rurack, M. Kollmannsberger, J. Daub, Angew. Chem. 2001, 113,
396–399; Angew. Chem. Int. Ed. 2001, 40, 385–387.
[31] A. Coskun, E. Deniz, E. U. Akkaya, Org. Lett. 2005, 7, 5187–5189.
[32] Z. Shen, H. Rçhr, K. Rurack, H. Uno, M. Spieles, B. Schulz, G.
Reck, N. Ono, Chem. Eur. J. 2004, 10, 4853–4871.
[33] Y.-H. Yu, A. B. Descalzo, Z. Shen, H. Rçhr, Q. Liu, Y.-W. Wang, M.
Spieles, Y.-Z. Li, K. Rurack, X.-Z. You, Chem. Asian J. 2006, 1,
176–187.
[12] E. Scorsone, S. Christie, K. C. Persaud, F. Kvasnik, Sens. Actuators
B 2004, 97, 174–181.
[13] E. Pringsheim, E. Terpetsching, O. S. Wolfbeis, Anal. Chim. Acta
1997, 357, 247–252.
[14] a) M. Mazur, P. Krysinski, K. Jackowska, Thin Solid Films 1998, 330,
167–172; b) L. Bansal, M. El-Sheriff, J. Yuan, Sensors 2007, 7,
3100–3118.
[15] A. T. Canada, L. R. Allain, D. B. Beach, Z. Xue, Anal. Chem. 2002,
74, 2535–2540.
[16] A. Loudet, K. Burgess, Chem. Rev. 2007, 107, 4891–4932.
[17] a) G. Ulrich, R. Ziessel, A. Harriman, Angew. Chem. 2008, 120,
1202–1219; Angew. Chem. Int. Ed. 2008, 47, 1184–1201; b) R. Zies-
sel, G. Ulrich, A. Harriman, New J. Chem. 2007, 31, 496–501.
[18] F. Lꢈpez-Arbeloa, J. Banuelos Prieto, V. Martinez Martinez, T.
[34] R. Englman, J. Jortner, Mol. Phys. 1970, 18, 145–181.
[35] We are well aware that the use of terminology such as pKa is inap-
propriate for organic solutions and we do so only to relate our find-
ings to prior literature work. The proton concentrations were deter-
mined by dilution of concentrated HCl on the basis of complete
AHCTUNGTREGUNdNN issociation.
[36] a) “Astronomical Data Analysis Software and Systems III”: G. A.
Kriss, Astron. Soc. Pac. Conf. Ser. 1994, 61, 437–441; b) W. Qin, M.
Baruah, W. M. De Borggraeve, N. Boens, J. Photochem. Photobiol.
A 2006, 183, 190–197.
[37] M. Chini, P. Crotti, F. Macchia, Tetrahedron Lett. 1990, 31, 4661–
4664.
[38] J. Rebek, Jr., D. Feitler, J. Am. Chem. Soc. 1974, 96, 1606–1607.
[39] Carried out under anhydrous conditions to exclude any traces of
acid.
ACHTUNGTRENNUNG rbeloa Lopez, I. Lopez Arbeloa, ChemPhysChem 2004, 5, 1762–
1771.
[40] A. Coskun, E. Deniz, E. U. Akkaya, Tetrahedron Lett. 2007, 48,
5359–5361.
[41] J. Olmsted III, J. Phys. Chem. 1979, 83, 2581–2584.
[42] G. Ulrich, S. Goeb, A. De Nicola, P. Retailleau, R. Ziessel, Synlett
2007, 1517–1520.
[19] The Handbook—A Guide to Fluorescent Probes and Labelling Tech-
nologies (Ed.: R. P. Haughland), Invitrogen/Molecular Probes, 2005.
[20] J. L. Bricks, A. Kovalchuk, C. Trieflinger, M. Nofz, M. Bꢉschel, A. I.
Tolmachev, J. Daub, K. Rurack, J. Am. Chem. Soc. 2005, 127,
13522–13529.
[43] Z. Otwinowski, W. Minor, Methods Enzymol. 1997, 276 307–326.
[44] G. M. Sheldrick, Acta Crystallogr. Sect. A 2008, 64, 112–122
[21] A. Coskun, E. U. Akkaya, J. Am. Chem. Soc. 2006, 128, 14474–
14475.
[22] A. Harriman, G. Izzet, R. Ziessel, J. Am. Chem. Soc. 2006, 128,
10868–10875.
Received: September 25, 2008
[23] R. Y. Lai, A. J. Bard, J. Phys. Chem. B 2003, 107, 5036–5042.
Published online: December 29, 2008
Chem. Eur. J. 2009, 15, 1359 – 1369
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1369