affected by additional parameters, such as the sample pH, the
choice of polymer matrix, the addition of a catalyst, and the
use of perm-selective coatings. Thus, sensors for aqueous
amines have to operate at high pH because the reactands do
not respond to alkylammonium ions. Furthermore, sensors
for amines do not require any catalyst and since they are not
very reactive with water and alcohols, they are less sensitive
to these analytes (Table 2). Whereas the sensor membranes
used for ethanol measurements respond to ethanol in the
molar range, the response of alcohol sensor layers to amines
is in the mM range. Furthermore, the sensor membranes for
ethanol only show sufficient response in the presence of a
catalyst.13–15 Nevertheless, sensor membranes composed of
tridodecylmethylammonium chloride, ETHT 4001 and ETHT
4004 are useful in alcohol sensing because, in beverages and
in bioreactor broth (the target samples for alcohol sensors),
amines are found in the mM range only.13,14 Furthermore, the
pH in beverages and in bioreactor broth is neutral or acidic.
Therefore, any interfering amine would be present in the
ammonium form, which does not interact with the reactands.
Both the sensors for amines and alcohols respond to
humidity, albeit with relatively small signal changes. By chang-
ing the polymer matrix to a more hydrophilic one, the response
to humidity can be increased while the response to lipophilic
alcohols remains relatively small. As with all these types of
sensor membranes, ionic analytes do not interfere. In the case
of the amine sensor no ionic additives are present so that ionic
analytes can not be extracted into the sensor layer. In the case
of the ethanol sensor, where cationic additives improve the
response behaviour, the sensor layers are coated with ion-
impermeable microporous PTFE coatings. Thus, the presented
sensor layers respond to neither hydronium ions nor carbonate
which are well-known to interact with trifluoroacetophenone
derivatives.22
this approach is that, unlike with labelling, selectivity problems
due to chemical modification (via labelling) of the analyte do
not arise. However, a colour change due simply to the close
proximity of the analyte and the dye cannot be guaranteed in
all cases. The limitations of both approaches could be
overcome by preparing molecular imprints with copolymerised
reactands. Thus, the number of possible analytes capable of
interacting with the reactand can be influenced by changing
the shape of the molecular imprint, while optical recognition
depends on the chemical reaction of the dye with the analyte.
The presented approach based on reactands is not limited
to the chemical reactivity of trifluoroacetophenone derivatives.
Conceivably, other chemical reactions can also be used as the
basis for reversible optical sensors, such as the reaction of
non-fluorescent hydrazine derivatives with aldehydes to form
fluorescent hydrazones, the bisulfite addition reaction, Schiff
base formation, or the Michael addition. We are currently
exploring these reactions as the basis for novel optical sensors
to detect neutral analytes.
References
1
O. S. Wolfbeis (ed.), Fiber Optic Chemical Sensors and Biosensors,
CRC Press, Boca Raton, FL, 1991.
2
U. E. Spichiger-Keller, Chemical Sensors and Biosensors for
Medical and Biological Applications, Wiley VCH, Weinheim,
Germany, 1998.
3
4
5
6
C. Gojon, B. Dureault, N. Hovnanian and C. Guizard, Sens.
Actuators, B, 1997, 38–39, 154.
E. Pringsheim, E. Terpetschnig and O. S. Wolfbeis, Proc.
Europt(r)ode IV, (Munster) 1998, 235.
¨
E. Pringsheim, E. Terpetschnig and O. S. Wolfbeis, Chimia, 1998,
7–8, 398.
O. S. Wolfbeis, I. Klimant, T. Werner, C. Huber, U. Kosch, C.
Krause, G. Neurauter and A. Durkop, Sens. Actuators, B, 1998,
¨
51, 17.
7
M. Kuratli, M. Badertscher, B. Rusterholz and W. Simon, Anal.
Chem., 1993, 65, 3473.
Outlook: what other applications are possible?
8
9
M. Kuratli and E. Pretsch, Anal. Chem., 1994, 66, 85.
K. Wang, K. Seiler, J. P. Haug, B. Lehmann, S. West, K. Hartman
and W. Simon, Anal. Chem., 1991, 63, 970.
The new reactands are interesting synthons for non-linear
optical (NLO) structures that exhibit a chemical functionality.
Dyes based on the azobenzene or stilbene structure comparable
to the dyes presented here have been extensively studied since
the early 90s.23–25 In order to show a non-vanishing first-order
hyperpolarisability and a second-order non-linear susceptibil-
ity (x2), the material must exhibit an asymmetric electronic
10 K. Seiler, K. Wang, M. Kuratli and W. Simon, Anal. Chim. Acta,
1991, 244, 151.
11 U. E. Spichiger, M. Kuratli and W. Simon, Biosens. Bioelectron.,
1992, 7, 715.
12 R. Wild, D. Citterio, J. Spichiger and U. E. Spichiger, J. Biotechn.,
1996, 50, 37.
13 G. J. Mohr, D. Citterio and U. E. Spichiger, Sens. Actuators, B,
1998, 49 226.
distribution at
a molecular and a macroscopic level.
Consequently, isotropic solids, such as, e.g., amorphous NLO-
polymers, require an orientation process (poling) to avoid the
overall centrosymmetry of the dipoles in the material. This
means that the covalent link of the dye to the polymer, e.g.
by methacrylate groups, is very important in obtaining a stable
orientation of the dye in the matrix. Both requirements, namely
having a high dipole moment and allowing polymerization,
are met by the dye monomers, ETHT 4012 and 4014. They
could be especially interesting for use in NLO since their
optical properties can be ‘switched on and off’ by specific
chemical reactions.
For molecularly imprinted polymers (MIPs), the new
reactands could provide a means of combining the recognition
of the analyte by the imprint matrix with the optical transduc-
tion by the indicator dye. So far, optical signal changes have
usually been obtained by using labelled analytes or non-specific
indicator dyes. In the first case, an imprint matrix for a specific
analyte is saturated with a dye-labelled analyte. Upon exposure
to the sample solution containing the analyte, the labelled
analyte is competitively exchanged with the analyte and
detected by its absorbance or fluorescence.26 This approach
cannot be used in all cases because labelled analytes do not
necessarily fit into the imprint.
14 G. J. Mohr and U. E. Spichiger, Anal. Chim. Acta, 1997, 344, 215.
15 G. J. Mohr and U. E. Spichiger, Anal. Chim. Acta, 1997, 351, 189.
16 G. J. Mohr, C. Demuth and U. E. Spichiger, Anal. Chem., 1998,
70, 3868.
17 G. J. Mohr and U. E. Spichiger, Mikrochim. Acta, 1998, 130, 29.
18 G. J. Mohr, N. Tirelli C. Lohse and U. E. Spichiger, Adv. Mater.,
1998, 10, 1353.
19 G. J. Mohr, N. Tirelli and U. E. Spichiger, Anal. Chem., 1999,
71, 1534.
20 M. L. M. Schilling, H. D. Roth and W. C. Herndon, J. Am. Chem.
Soc., 1980, 102, 4271.
21 J. Sangster, Octanol–Water Partition Coefficients: Fundamentals
and Physical Chemistry, Wiley, West Sussex, England, 1997.
22 C. Behringer, B. Lehmann, J. P. Haug, K. Seiler, W. E. Morf,
K. Hartman and W. Simon, Anal. Chim. Acta, 1990, 233, 41.
23 D. Brown, A. Natansohn and P. Rochon, Macromolecules, 1995,
28, 6116.
24 M. Zhao, M. Bautista and W. T. Ford, Macromolecules, 1991,
24, 844.
25 G. S’Heeren, A. Persoons, P. Rondou, M. Van Beylen and
C. Samyn, Eur. Polym. J., 1993, 29, 975.
26 R. Levi, S. McNiven, S. A. Piletsky, S.-H. Cheong, K. Yano and
I. Karube, Anal. Chem., 1997, 69, 2017.
27 P. Turkewitsch, B. Wandelt, G. D. Darling and W. S. Powell,
Anal. Chem., 1998, 70, 2025.
Paper 9/01961H
In the second case, the analyte interacts with a dye that is
copolymerised with the imprint matrix.27 The advantage of
2264
J. Mater. Chem., 1999, 9, 2259–2264