Highly selective chemical vapor sensing by molecular recognition: Specific detection of C1-C4 alcohols with a fluorescent phosphonate cavitand

Article abstract of DOI:10.1002/anie.201100738

Fiat lux on alcohols: Molecular-level resolution was achieved for the detection of short-chain alcohols in the vapor phase using a fluorescent cavitand sensor. The transduction mechanism, activated exclusively by the complexation mode, is provided by the change of the electronic density on the fluorophore caused by the formation of an intracavity hydrogen bond between the cavitand P-O and the alcohol OH group.

Full text of DOI:10.1002/anie.201100738

Communications
DOI: 10.1002/anie.201100738
Supramolecular Sensors
Highly Selective Chemical Vapor Sensing by Molecular Recognition:
Specific Detection of C1–C4 Alcohols with a Fluorescent Phosphonate
Cavitand**
Francesca Maffei, Paolo Betti, Damiano Genovese, Marco Montalti, Luca Prodi,* Rita De Zorzi,
Silvano Geremia, and Enrico Dalcanale*
In the last few years there has been a huge demand to monitor
different chemical species in the vapor phase, such as
environmental pollutants, hazardous chemicals, food
aromas, explosives, and volatiles in breath for disease
diagnosis.[1] Chemical vapor sensors are among the most
promising devices to be exploited for these applications,
because they have the great advantage of allowing an online
measure suitable for remote control.[2] In this context, the
need to develop sensors specific for different classes of
analytes is well-recognized and confirmed by the considerable
research efforts spent for the preparation of more and more
efficient devices.[3] The crucial parameter to define the success
of a given sensor is therefore selectivity,[4] and for this reason
the strategy to prepare the sensing material following the
principle of supramolecular chemistry has quickly gained
increasing importance.[5] However, the realization of selective
chemical vapor sensors requires particular attention since
they operate at the gas–solid interface. Any given analyte,
upon moving from the vapor to the solid phase, experiences a
dramatic increase in non-specific dispersion interactions,
which tend to override any specific complexation event
responsible for the selective responses.[6] As a result, the
sensor selectivity drops and false positive responses soar. A
possible solution to this general problem relies on trans-
duction modes activated exclusively by the molecular recog-
nition event.
Following this approach, we present herein a new solid-
state fluorescent sensor based on phosphonate cavitand
Mi[C2H5, H, fluorophore][7] (Scheme 1) for detecting short-
chain alcohols in the gas phase. Phosphonate cavitands are
Scheme 1. Structure of fluorescent model compound 1 and the fluo-
rescent phosphonate cavitands Mi and Mo.
molecular receptors that present one or more PV moieties as
bridging units.[7] In previous studies, we have shown that there
are two key factors affecting the sensing performances of
mono-,[8] di-,[9] and tetraphosphonate[10] cavitands toward
alcohols: 1) the simultaneous presence of hydrogen bonding
[*] F. Maffei, Dr. P. Betti, Prof. E. Dalcanale
Dipartimento di Chimica Organica ed Industriale and INSTM
UdR Parma, Universitꢀ di Parma
=
with one of the P O groups and CH–p interactions with the
p-basic cavity, which require an inward orientation of the
Viale Usberti 17 A, 43124 Parma (Italy)
Fax: (+39)0521-905-472
E-mail: enrico.dalcanale@unipr.it
=
P O bridges; and 2) a cavity that provides a permanent free
=
volume for the analyte around the inward P O groups, which
is pivotal for effective hydrogen bonding.[5d,10] Increasing the
Homepage: http://www.ssmg.unipr.it/
=
number of inward-facing P O groups enhances the sensor
D. Genovese, Dr. M. Montalti, Prof. L. Prodi
Dipartimento di Chimica “G. Ciamician”, Latemar Unit
Universitꢀ di Bologna
Via Selmi 2, 40126 Bologna (Italy)
Fax: (+39)051-209-9456
responses through the entropic stabilization of host–guest
complexes, but does not change the observed selectivity
trend.
At first, a systematic study was undertaken to assess the
complexation properties of phosphonate cavitands towards
alcohols in the solid state. The compact tetraphosphonate
cavitand Tiiii[H, CH3, CH3] was chosen as host for its
tendency to crystallize. Crystallization trials of this cavitand
were performed by the vapor diffusion method with sitting
drops in Linbro multiwell plates containing trifluoroethanol
(TFE) as solvent. The addition of a short-chain alcohol in the
reservoir solution, through vapor diffusion in the drop,
allowed the easy and fast growth of monoclinic crystals of
E-mail: luca.prodi@unibo.it
Dr. R. De Zorzi, Prof. S. Geremia
Centro di Eccellenza in Biocristallografia
Dipartimento di Scienze Chimiche, Universitꢀ di Trieste
Viale Giorgeri 1, 34127 Trieste (Italy)
[**] This work was supported by the EC through the ITN Project
FINELUMEN (PITN-GA-2008-215399), by MIUR (PRIN projects),
Fondazione Cassa di Risparmio in Bologna, and INSTM.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201100738.
4654
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 4654 –4657

the C1–C4 alcohol series. Taken alone,
none of them is sufficient to bind
alcohols in the solid state. This result
can be extended to the case of mono
and diphosphonate cavitands, as their
interaction mode with alcohols in the
solid state is the same.[5d]
The next step was the selection of a
transduction mechanism activated
exclusively by this specific complex-
ation mode to suppress the contribu-
tion of dispersion interactions to the
overall response registered by QCM
(quartz crystal microbalance) transdu-
cers.[6b] It is well known that, among
the different sensors, those based on
luminescence present many advan-
tages, such as high sensitivity, low
cost, ease of operation, and versatil-
ity.[3,11,12] The introduction of a fluo-
rescent moiety on the receptor was
necessary because the absorbance and
fluorescence of the phosphonate cav-
itand family is too far in the UV region
to be used for practical applications.[13]
Figure 1. Top and side views of the crystal structures of the six Tiiii[H, CH3, CH3]·alcohol
complexes.
the corresponding host–guest complex. A whole series of
isomorphic crystal structures of the TFE disolvate of C1–C4
alcohol–cavitand complexes were obtained (Figure 1). All six
host–guest complexes present the same interaction pattern:
an H-bond between the alcoholic OH group and one of the
We chose to introduce at the phosphonate site a fluorophore,
similar to the commercial 2-anilinonaphthalene-6-sulfonic
acid (2,6-ANS), because of the charge-transfer character of its
excited state. The rationale of this design was based on the
belief that the formation of the hydrogen bond between the
=
=
P O units, and CH–p contacts between the p-rich cavity and
P O and the alcohol OH group could decrease the electronic
one methyl group of the alkyl chain. The relevant geometric
parameters describing the host–guest interaction are reported
in Table 1. In the case of 1-butanol, the contemporary onset of
density on the phosphorus atom to a such an extent to modify
the energy of excited state of the fluorophore. As in the
excited state of 2,6-ANS a charge transfer from the aniline to
the naphthalene moiety occurs, and for this reason the
formation of the hydrogen bond was expected to make the
charge transfer easier, leading to a red-shift of the emission
band. This design was also conceived to offer a high
specificity, as only the formation of a hydrogen bond could
Table 1: Geometric parameters describing the host–guest interaction in
C1–C4 alcohol–cavitand complexes.[a]
Alcohol
H-bond [ꢁ]
Oalc oop [ꢁ]
CH3alc oop [ꢁ]
=
MeOH
EtOH
2-PrOH
1-PrOH
2-BuOH
1-BuOH
2.704 (4)
2.765 (9)
2.816 (9)
2.730 (3)
2.75 (1)
0.039 (3)
1.150 (3)
0.926 (3)
1.409 (3)
1.357 (8)
1.926 (7)
À1.264 (4)
À1.028 (4)
À0.964 (4)
À1.341 (4)
À1.43 (1)
cause such a spectral shift. A single P O unit was introduced
on the cavitand to funnel the H-bond perturbation on a single
site, to maximize the desired red-shift.[14] Monophosphonate
cavitands offer the additional advantage of being water
=
insensitive, as at least two P O units are necessary to bind
2.85 (1)
À1.521 (7)
water efficiently.[10,15]
=
[a] Hydrogen-bond distances (Oalcohol···O P) and distances of the
hydroxy and methyl groups from the mean plane of oxygen atoms of
The target cavitand Mi[C2H5, H, fluorophore], from now
onward referred to as Mi (Scheme 1), which presents at the
upper rim the inward oriented P O and three methylene
=
P O groups (out-of-plane, oop) are given. Atoms inside the cavity have a
=
negative sign for the oop distances.
bridges, was synthesized by introducing the fluorescent
phosphonate moiety on a trimethylene-bridged resorcinar-
ene[16] (see the Supporting Information for details of the
preparation). The out isomer, Mo, obtained as a byproduct in
the last step, was used as control system to exclude unspecific
responses, and to demonstrate the need of the inward
both interactions with the cavitand requires the deep
insertion of the methyl group of the guest into the cavity
and the weakening of the H-bond. The presence of a single
additional methylene unit is sufficient to completely suppress
the crystallization process of the complex, as observed in the
case of 1-pentanol (Supporting Information, Figure S8).
The solid-state study clearly indicates that the two
interactions responsible for the high selectivity of phospho-
nate cavitands towards alcohols occur simultaneously only in
=
orientation of the P O unit to synergistically activate both
H-bonding and CH–p interactions.
The identification of the two isomers was achieved by 1H
and 31P NMR spectroscopy. The 1H NMR spectrum shows an
upfield shift of the resonances belonging to the naphtalenic
Angew. Chem. Int. Ed. 2011, 50, 4654 –4657
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4655

Communications
moiety oriented inside the cavity in Mo, which is the result of
fluorescence intensity changes of the Mi film exposed to
different alcohols at various concentrations in N2 are shown in
Figure 3. The intensity changes were comparable for the
whole C1–C4 alcohol series with the exception of 1-butanol
(Figure 3), which caused lower responses. This result implies
the shielding effect exerted by the aromatic rings of the
resorcinarene skeleton. For the same reason, the 31P NMR
spectrum of Mo also presents a significant upfield shift
compared to Mi, as already reported by Dutasta et al.[17]
The fluorescent model compound 1, synthesized in two
steps starting from aniline and 6-bromo-2-naphthol (see the
Supporting Information), was used as a model system to
demonstrate the importance of the cavity surrounding the
=
that its H-bond with the P O of Mi is less effective, as inferred
by the crystal structure of Tiiii[H, CH3, CH3]·n-butanol.
=
P O unit in the recognition event.
PVC thin films containing 0.2% w/w of cavitands Mi, Mo,
and model compound 1 were deposited on glass substrates by
spin coating. Dioctyl sebacate was added to these matrixes as
plasticizer before deposition to enhance the layer permeabil-
ity. To measure the surface fluorescence of a solid substrate
subject to a continuous gas flow, we designed a cell with some
special features (Supporting Information, Figure S2,S3). The
orientation of the cell was optimized to minimize the
reflection of the excitation light and to maximize the
fluorescence signal. An attenuator filter was used to diminish
the number of incoming photon to decrease problems related
to photochemical stability, thus allowing long measurements
in continuous flow. The single-wavelength emission was
monitored at 460 nm, where the highest intensity excursion
is observed upon complexation.
Upon excitation at 350 nm, the Mi film showed an intense
broad and unstructured band emission with a maximum at
414 nm. Upon exposure to different alcohols, the maximum of
this band was red-shifted by 5 nm, with a more pronounced
difference in the tail of the spectrum (Supporting Informa-
tion, Figure S4). Although relatively small, this difference was
sufficient to monitor the concentration of the alcohols in the
gas phase. Figure 2 shows the profile of the fluorescence
intensity of a PVC film containing Mi exposed to a flux of
pure N2 alternated with a flux of ethanol in N2 with an
increasing concentration from 40 ppm to 630 ppm. The
changes in fluorescence intensity were fast and fully rever-
sible. Under the same conditions, films of Mo and of 1 showed
negligible changes in the emission maximum. The relative
Figure 3. Relative fluorescence intensity changes (lexc =350 nm,
lexc =460 nm) of a PVC film containing the receptor Mi exposed to
different alcohols in N2: black trace: methanol; red: ethanol; green:
1-propanol; yellow: 2-butanol; blue: 1-butanol.
The histograms of Figure 4 are even more revealing of the
agreement between solid-state structures and sensor
responses. The Mo film showed negligible responses to all
C1–C4 alcohols, as expected owing to the disconnection of the
two interaction modes. The drop in fluorescence intensity
experienced by the Mi film exposed to 500 ppm of 1-butanol
Figure 2. Fluorescence intensity (lexc =350 nm, lexc =460 nm) of a PVC
film containing the receptor Mi subject to a pure N2 flux alternated
with a ethanol flux in N2 with increasing concentration (from 40 to
630 ppm; see also Figure 3).
Figure 4. Relative fluorescence intensity changes (lexc =350 nm,
lexc =460 nm) of PVC films containing the receptor Mi (dark gray) or
Mo (light gray) exposed to different alcohols in N2: methanol, ethanol,
1-propanol, 2-butanol, 1-butanol, 1-pentanol (500 ppm each).
4656 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 4654 –4657

and 1-pentanol is indicative of their limited complexation in
the layer. The inversion of response between Mi and Mo films,
observed in the case of 1-pentanol, supports the nonspecific
origin of this fluorescence change, which can be attributed to
extra-cavity H-bonding. The high sensor selectivity was
demonstrated by comparing the responses of the Mi and
Mo films to high concentrations of ethanol, acetone, and
hydrocarbons: very low responses were obtained for both
films in the case of acetone, n-pentane, and n-heptane, and for
the Mo film in the case of ethanol (Supporting Information,
Figure S5,S6). Competition experiments between ethanol and
water showed that the low responses of the Mi layer to water
are totally suppressed in the presence of ethanol vapors
(Supporting Information, Figure S7).
In conclusion, this work demonstrates that it is possible to
achieve high selectivity in chemical vapor sensing by harness-
ing the binding specificity of a cavitand receptor. The key
requirement for transferring the molecular recognition prop-
erties from the solid-state to the gas–solid interface is the
selection of the transduction mechanism, which must be
turned on exclusively by the desired complexation mode with
the analyte. In our case, the H-bonding of the alcohol to the
[1] a) F. Rꢀck, N. Barsan, U. Weimar, Chem. Rev. 2008, 108, 705 –
725; b) J. W. Grate, Chem. Rev. 2008, 108, 726 – 745.
[2] Sensors, A Comprehensive Survey, Vol. 2 (Eds.: W. Gꢀpel, J.
Hesse, J. N. Zemel), VCH, Weinheim, 1991.
[3] L. Prodi, New J. Chem. 2005, 29, 20 – 31.
[4] H. Hierlemann, R. Gutierrez-Osuna, Chem. Rev. 2008, 108, 563 –
613, and references therein.
[5] a) F. C. J. M. Van Veggel in Comprehensive Supramolecular
Chemistry, Vol. 10 (Eds.: J. L. Atwood, J. E. D. Davies, D. D.
MacNicol, F. Vꢀgtle, D. N. Reinhoudt), Pergamon, Oxford, 1996,
pp. 171 – 185; b) F. L. Dickert, U. P. A. Bꢁumler, H. Stathopulos,
Anal. Chem. 1997, 69, 1000 – 1005; c) J. J. Lavigne, E. V. Anslyn,
Angew. Chem. 2001, 113, 3212 – 3225; Angew. Chem. Int. Ed.
2001, 40, 3118 – 3130; d) L. Pirondini, E. Dalcanale, Chem. Soc.
Rev. 2007, 36, 695 – 706; e) E. V. Anslyn, J. Org. Chem. 2007, 72,
687 – 699.
[6] a) W. Grate, G. C. Frye in Sensors Update, Vol. 2 (Eds.: H.
Baltes, W. Gꢀpel, J. Hesse), Wiley-VCH, Weinheim, 1996,
pp. 37 – 83; b) M. Tonezzer, M. Melegari, G. Maggioni, R. Milan,
G. Della Mea, E. Dalcanale, Chem. Mater. 2008, 20, 6535 – 6542.
[7] For the nomenclature adopted for phosphonate cavitands see: R.
Pinalli, M. Suman, E. Dalcanale, Eur. J. Org. Chem. 2004, 451 –
462.
[8] R. Pinalli, F. F. Nachtigall, F. Ugozzoli, E. Dalcanale, Angew.
Chem. 1999, 111, 2530 – 2533; Angew. Chem. Int. Ed. 1999, 38,
2377 – 2380.
=
P O induces a detectable red-shift of the fluorescence
emission of the 2,6-ANS fluorophore directly linked to the
phosphonate acceptor. The source of selectivity can be
dissected into three components. First, the ubiquitous non-
specific layer adsorption, being luminescence-silent, does not
contribute to the overall response, as it did in QCM devices.[6b]
Second, in the layer the intracavity H-bonding in Mi is highly
favored over the extra-cavity form in Mo or 1, owing to the
cavity free-volume effect. Third, the synergy between CH–p
interactions and H-bonding in Mi leads to a strong bias
toward C1–C4 alcohol detection. The molecular level reso-
lution of this last contribution is outstanding, as it allows the
discrimination of alcohols on the basis of a single methylene
unit difference. In this way, the responses owing to nonspecific
interactions of the analytes and competitive binding by
interferents have been almost completely removed. Although
it is still necessary to improve the characteristics of the
fluorescence moiety to increase sensitivity and the signal-to-
noise ratio, we think that Mi is an important step forward to
the design of more efficient chemical vapor sensors.
As most organic and polymer-based sensors detect
analytes mainly on the basis of polarity,[6a] the approach
described herein is a viable solution to the general problem of
discriminating analytes by chemical class, rather than by
polarity, in vapor sensing. This approach can be extended to
many different classes of organic receptors, thus opening the
way for the rational design of sensor materials as function of
the analytes to be detected.
[9] M. Suman, M. Freddi, C. Massera, F. Ugozzoli, E. Dalcanale,
J. Am. Chem. Soc. 2003, 125, 12068 – 12069.
[10] M. Melegari, M. Suman, L. Pirondini, D. Moiani, C. Massera, F.
Ugozzoli, E. Kalenius, P. Vainiotalo, J.-C. Mulatier, J.-P. Dutasta,
E. Dalcanale, Chem. Eur. J. 2008, 14, 5772 – 5779.
[11] 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) L. Prodi, F. Bolletta, M. Montalti, N.
Zaccheroni, Coord. Chem. Rev. 2000, 205, 59 – 83; c) S. W.
Thomas III, G. D. Joly, T. M. Swager, Chem. Rev. 2007, 107,
1339 – 1386.
[12] For luminescent organometallic vapochromic sensors, see: C. E.
Strasser, V. J. Catalano, J. Am. Chem. Soc. 2010, 132, 10009 –
10011, and references therein.
[13] E. Biavardi, G. Battistini, M. Montalti, R. M. Yebeutchou, L.
Prodi, E. Dalcanale, Chem. Commun. 2008, 1638 – 1640.
=
[14] A Tiiii cavitand with a single P O group functionalized with the
fluorophore and three standard phenyl groups would have lead
to an undesired dilution of the H-bonding perturbation on four
sites, with three of them luminescence silent. On the other hand,
=
a Tiiii cavitand bearing the fluorophore on each P O unit would
have two counter-indications: a likely self-quenching in the solid
state, and a lower DI/I and thus a lower s/n ratio, as only 1/4 of
the signal would change upon complexation.
[15] a) M. Melegari, C. Massera, F. Ugozzoli, E. Dalcanale, Cryst-
EngComm 2010, 12, 2057 – 2059; b) C. Massera, M. Melegari, E.
Kalenius, F. Ugozzoli, E. Dalcanale, Chem. Eur. J. 2011, 17,
3064 – 3068.
[16] a) P. Timmerman, H. Boerrigter, W. Verboom, J. G. Van Hum-
mel, S. Harkema, D. N. Reinhoudt, J. Inclusion Phenom. Mol.
Recognit. Chem. 1994, 19, 167 – 191; b) E. Menozzi, M. Busi, C.
Massera, F. Ugozzoli, D. Zuccaccia, A. Macchioni, E. Dalcanale,
J. Org. Chem. 2006, 71, 2617 – 2624.
Received: January 28, 2011
Published online: April 14, 2011
[17] P. Delangle, J.-C. Mulatier, B. Tinant, J.-P. Dutasta, Eur. J. Org.
Chem. 2001, 3695 – 3704.
Keywords: alcohols · cavitands · chemical vapor sensors ·
.
fluorescence sensors · molecular recognition
Angew. Chem. Int. Ed. 2011, 50, 4654 –4657
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4657