Plasticizer-free optode membranes for dissolved amines based on copolymers from alkyl methacrylates and the fluoro reactand ETH(T) 4014

Article abstract of DOI:10.1021/ac9812983

A new class of fluorogenic copolymers has been synthesized that reversibly interacts with dissolved aliphatic amines resulting in a change in fluorescence. The copolymers are composed of different methacrylate monomeric units, such as hexyl methacrylate a

Full text of DOI:10.1021/ac9812983

Anal. Chem. 1999, 71, 1534-1539
Plasticizer-Free Optode Membranes for Dissolved
Amines Based on Copolymers from Alkyl
Methacrylates and the Fluoro Reactand ETHT 4014
Gerhard J. Mohr,*,† Nicola Tirelli,‡ and Ursula E. Spichiger-Keller†
Centre for Chemical Sensors, ETHZ Technopark, Technopark Strasse 1, CH-8005 Zurich, Switzerland, and Institute for
Polymers, ETH Zu¨rich, Universita¨tstrasse 6, CH-8092 Zurich Switzerland
tion) prompted us to develop a new approach to optical sensor
layers in which no plastification at all is needed.
A new class of fluorogenic copolymers has been synthe-
sized that reversibly interacts with dissolved aliphatic
amines resulting in a change in fluorescence. The copoly-
mers are composed of different methacrylate monomeric
units, such as hexyl methacrylate and the dye monomer
4-[N,N-bis(11-methacroyloxyundecyl)amino]-4 ′-trifluoro-
acetylstilbene (ETHT 4 0 1 4 ). Upon exposure to aqueous
amine solutions, thin layers of the copolymers show a
decrease in fluorescence at a wavelength around 5 2 0 nm
and an increase in fluorescence at around 4 2 0 nm. The
change in fluorescence is based on the nucleophilic
addition of amines to the trifluoroacetyl group of the
reactand, thus causing a change in the degree of electron
delocalization. Since the sensor layers are composed of
copolymerized components, they exhibit superior opera-
tional stability to sensors based on plasticized P VC, and
their lack of plasticizers enhances the shelf lifetime.
In the past decade, there have been several attempts to use
copolymerized compounds in optical sensing. Munkholm et al.
made use of acryloylfluorescein copolymerized with acrylamide-
methylenebis(acrylamide) for optical sensing of pH.7 Hisamoto
et al. introduced a method to prepare poly(hydroxyethyl meth-
acrylates) (polyHEMA) with copolymerized solvatochromic pH
indicator dyes for sensing of pH and the polarity of the sample
solution.8 They copolymerized the components directly on the
glass support by exposure to UV light. Ambrose and Meyerhoff
developed an optical polyion probe by using photo-cross-linked
decyl methacrylate.9 However, so far there have been no attempts
to use copolymers for selective sensing of neutral species such
as alcohols or amines.
Chromogenic and fluorogenic molecules with structures com-
parable to the compounds presented in this paper have been used
in recent years due to their nonlinear optical (NLO) properties.
Mostly, they were produced by copolymerization of methacrylic
dyes with methacrylates.10-15 The results obtained using NLO
copolymers have motivated us to develop optical sensors based
on structurally related copolymers. Methacrylates were chosen
as the basic monomer materials since they are commercially
available and easily synthesized and polymerized. Furthermore,
we synthesized a methacrylate derivative of a reactand that is
sensitive to nucleophilic species such as amines. However, in
contrast to the NLO-based polymers, which require high Tg values
for a stable orientation of the dyes, polymers for optical sensors
require high elasticity and, therefore, low Tg values. Thus, we have
prepared copolymers composed of butyl, hexyl, and dodecyl
We have recently introduced a new type of optical sensor based
on chromogenic and fluorogenic reactands for the determination
of analytes, such as primary amines, alcohols, and humidity.1-6
Unlike the “ligands” used in ion-selective sensors, which interact
by complexing the analyte reversibly, the interaction between the
analyte and the reactive compound includes a reversible chemical
reaction. Since the reactive host compound is incorporated into a
polymer film, the determination of the analyte is feasible in a
reagent-free system. Therefore, the reactive host compound is
named “reactand” in analogy to “ligands” for complexing agents.
Optical sensors have traditionally been prepared with thin layers
of plasticized PVC as the polymeric matrix in which the lipophilic
reactands were dissolved. However, the inherent disadvantages
of plasticizers (such as evaporation, leaching, and phase separa-
(7) Munkholm, C.; Walt, D.; Milanovich, F. P.; Klainer, S. M. Anal. Chem. 1986,
58, 1427.
(8) Hisamoto, H.; Manabe, Y.; Yanai, H.; Tohma, H.; Yamada, T.; Suzuki, K.
Anal. Chem. 1 9 9 8 , 70, 1255.
(9) Ambrose, T. M.; Meyerhoff, M. E. Anal. Chem. 1 9 9 7 , 69, 4092.
(10) S’Heeren, G.; Persoons, A.; Rondou, P.; VanBeylen, M.; Samyn, C. Eur.
Polym. J. 1 9 9 3 , 29, 975.
* Corresponding author: (e-mail) gerhard@chemsens.pharma.ethz.ch, http:/ /
www.chemsens.ethz.ch.
† Centre for Chemical Sensors.
‡ Institute for Polymers.
(1) Mohr, G. J.; Demuth, C.; Spichiger, U. E. Anal. Chem. 1 9 9 8 , 70, 3868.
(2) Mohr, G. J.; Lehmann, F.; Grummt, U.-W.; Spichiger, U. E. Anal. Chim.
Acta 1 9 9 7 , 344, 215.
(3) Mohr, G. J.; Spichiger, U. E. Anal. Chim. Acta 1 9 9 7 , 351, 189.
(4) Mohr, G. J.; Citterio, D.; Spichiger, U. E. Sens. Actuators, B 1 9 9 8 , 49, 226.
(5) Mohr, G. J.; Spichiger, U. E. Mikrochim. Acta 1 9 9 8 , 130, 29.
(6) Spichiger-Keller, U. E. Chemical Sensors and Biosensors for Medical and
Biological Applications; Wiley VCH: Weinheim, Germany, 1998.
(11) Altomare, A.; Ciardelli, F.; Ghiloni, M. S.; Solaro, R.; Tirelli, N. Macromol.
Chem. Phys. 1 9 9 7 , 198, 1739.
(12) Dworczak, R.; Fabian, W. M. F.; Kieslinger, D.; Junek, H. Dyes Pigments
1 9 9 7 , 34, 13.
(13) Brown, D.; Natansohn, A.; Rochon, P. Macromolecules 1 9 9 5 , 28, 6116.
(14) Zhao, M.; Bautista, M. O.; Ford, W. T. Macromolecules 1 9 9 1 , 24, 844.
(15) Tirelli, N.; Suter, U. W.; Altomare, A.; Solaro, R.; Ciardelli, F.; Follonier, S.;
Bosshard, Ch.; Gu¨ nter, P. Macromolecules 1 9 9 8 , 31, 2152.
1534 Analytical Chemistry, Vol. 71, No. 8, April 15, 1999
10.1021/ac9812983 CCC: $18.00 © 1999 American Chemical Society
Published on Web 03/16/1999

oil was purified by flash chromatography using n-hexane/ ethyl
acetate (2:1 v/ v) as the eluent, yielding 9.0 g of a pale yellow
liquid: 1H NMR (CDCl3) δ (ppm) 7.27 (m, 2 H, dCH), 6.63 (d, 1
H, dCH), 6.57 (m, 2 H, dCH), 5.53 (d, 1 H, dCH), 4.96 (d, 1 H,
dCH), 3.63 (t, 4 H, CH2), 3.24 (t, 4 H, CH2), 1.58 (m, 8 H, CH2),
1.32 (m, 28 H, CH2). Calcd for C30H53NO2 (459.75): C, 78.37; H,
11.62; N, 3.05. Found: C, 78.20; H, 11.42; N, 3.03.
4-[N,N-Bis(1 1-hydroxyundecyl)amino]-4 ′-trifluoroacetyl-
stilbene (2 ). A mixture of 2.11 g (4.6 mmol) of 1 , 1.17 g (4.6
mmol) of 4-bromo-R,R,R-trifluoroacetophenone, 10 mg of palladium
diacetate, 25 mg of tri-o-tolylphosphine, and 4 mL of triethylamine
was heated at 115 °C for 22 h in a capped heavy-wall Pyrex tube
that was flushed with dry nitrogen. Water and dichloromethane
were added to the cooled solution, and the aqueous layer was
extracted twice with 100 mL of dichloromethane. The combined
dichloromethane solutions were washed three times with distilled
water, dried over magnesium sulfate, and evaporated to dryness.
After purification by flash chromatography (n-hexane/ ethyl ac-
etate, 2:1 v/ v), a pure fraction of 2.1 g of 2 as an amorphous red
solid, mp 53-57 °C was obtained: 1H NMR (CDCl3) δ (ppm) 8.03
(d, 2 H, dCH), 7.58 (d, 2 H, dCH), 7.44 (d, 2 H, dCH), 7.25 (d,
1 H, dCH), 6.88 (d, 1 H, dCH), 6.63 (d, 2 H, dCH), 3.64 (m, 4 H,
-CH2), 3.31 (m, 4 H, -CH2), 2.17 (s, 2 H, OH), 1.58 (m, 8 H,
CH2), 1.32 (m, 28 H, CH2). Calcd for C38H56NO3F3 (631.86): C,
72.23; H, 8.93; N, 2.22. Found: C, 72.38; H, 9.07; N, 2.31.
Figure 1. Synthesis of the fluorogenic monomer 4-[N,N-bis(11-
methacroyloxyundecyl)amino]-4′-trifluoroacetylstilbene (ETHT 4014).
methacrylates, rather than methyl methacrylate. The optical
sensors obtained by copolymerization were investigated for their
sensitivity to dissolved amines, as well as for their selectivity,
stability, and response. Amines were the target analytes because
a wide range of amines are pollutants in the pharmaceutical and
chemical industry. They are used in the preparation of fertilizers,
pharmaceuticals, surfactants, and colorants and can be found in
agricultural areas. The presence of amines may also be indicative
of food quality since they are produced during the degradation
of fish products.
4-[N,N-Bis(11-methacryloxyundecyl)amino]-4′-(trifluoro-
acetyl)stilbene (3 ; ETHT 4 0 1 4 ). A solution of 0.15 g of
methacryloyl chloride in 0.4 mL of tetrahydrofuran (THF) was
added drop by drop to a mixture of 0.32 g (0.5 mmol) of 2 , 0.28
g of triethylamine, 2.5 mg of hydroquinone, and 6 mL of dry THF
(cooled to 0 °C) and stirred under nitrogen. The reaction mixture
was heated to 45 °C for 4 h. Then, a solution of 0.15 g of
methacryoyl chloride in 0.4 mL of THF was added at 0 °C and
the reaction mixture was kept at 45 °C for 12 h. After evaporation
of the tetrahydrofuran, the residue was dissolved in dichloro-
methane and washed once with saturated aqueous sodium
hydrogen carbonate solution and three times with water. The
resulting solution was dried over sodium sulfate, and the solvent
was removed under reduced pressure. The remaining liquid was
purified by flash chromatography (n-hexane/ ethyl acetate, 95:5
v/ v), yielding 0.24 g of the pure viscous compound: 1H NMR
(CDCl3) δ (ppm) 8.05 (d, 2 H, dCH), 7.58 (d, 2 H, dCH), 7.44
(d, 2 H, dCH), 7.24 (d, 1 H, dCH), 6.88 (d, 1 H, dCH), 6.63 (d,
2 H, dCH), 6.10 (s, 2 H, dCH), 5.55 (d, 2 H, dCH), 4.17 (t, 4 H,
CH2), 3.28 (m, 4 H, CH2), 1.93 (s, 6 H, CH3), 1.58 (m, 8 H, CH2),
1.30 (m, 28 H, CH2). Calcd for C46H64NO5F3 (768.01): C, 71.94;
H, 8.40; N, 1.82. Found: C, 72.06; H, 8.52; N, 1.82.
EXPERIMENTAL SECTION
Reagents. All reagents were of synthetic reagent grade. Flash
chromatography was performed with silica gel 60 (63-200 µm)
from Fluka (Buchs, Switzerland). n-Butyl methacrylate and n-
dodecyl methacrylate were obtained from Fluka. n-Hexyl meth-
acrylate came from Polysciences, Inc. (Warrington, PA).
Tyramine [4-(2-aminoethyl)phenol], spermidine [N-(3-amino-
propyl)-1,4-diaminobutane], cadaverine (1,5-diaminopentane), his-
tamine [2-(4-imidazolyl)ethylamine], and phenethylamine (2-
phenylethylamine) were of analytical grade and were obtained
from Fluka. Amine solutions were prepared by dissolving the
appropriate amount of each amine in 0.1 M sodium hydroxide
solution. Due to the high pH of 13.0, the amines were mainly
present in the electrically neutral form and not in the ammonium
form. The correct amine concentrations were calculated by using
the Henderson-Hasselbach equation and the pKa value of each
amine.17 The selectivity of the sensor layers to the different amines
was calculated according to ref 1.
Synthesis of the Fluororeactand (see Figure 1 ). N,N-Bis-
(1 1 -hydroxyundecyl)-4 -vinylaniline (1 ). A mixture of 4.0 g (30
mmol) of vinylaniline (90% purity), 15.2 g (60 mmol) of 11-bromo-
1-undecanol, 15.0 g (116 mmol) of N-ethyldiisopropylamine, and
20 mL of dimethylformamide was stirred at 110 °C for 24 h under
argon atmosphere. After cooling to room temperature, the product
was poured onto 100 mL of distilled water, leading to a separation
of the product from the aqueous phase. The product was put into
100 mL of dichloromethane, washed twice with distilled water,
and dried over magnesium sulfate. After evaporation, the resulting
Copolymer Synthesis. CP 1 , CP 2 , and CP 3 . The copoly-
merizations were carried out in degassed acetone or benzene
solutions in an argon atmosphere at 60 °C in the presence of 2,2′-
azobisisobutyronitrile (AIBN). For CP 1 , 30 mmol of freshly
distilled butyl methacrylate, 30 mg of AIBN, and 9.2 mg of ETHT
4014 were dissolved in 30 mL of acetone, and the resultant mixture
was refluxed for 23 h. In the case of CP 2 , 30 mmol of freshly
distilled hexyl methacrylate, together with 11 mg of ETHT 4014
in acetone, were used. For CP 3 , 16.4 mg of ETHT 4014, and 30
mmol of dodecyl methacrylate were reacted in benzene. The
resulting copolymer solutions were cooled to room temperature
(16) Spichiger, U. E.; Demuth, C. Anal. Chim. Acta 1 9 9 7 , 355, 259.
(17) Lide, D. R. Handbook of Chemistry and Physics, 74th ed.; CRC Press: Boca
Raton, 1993-1994; Vol. 8, pp 43-45.
Analytical Chemistry, Vol. 71, No. 8, April 15, 1999 1535

Table 1. Compositions of the Chromogenic Copolymers CP1-CP3
copolymer
monomer
ETHT 4014 (%)
Tg (°C)
Mn
Mw/ Mn
CP 1
CP 2
CP 3
butyl methacrylate
hexyl methacrylate
dodecyl methacrylate
0.21
0.18
0.13
41
-4
< -20
42 000
53 000
120 000
2.1
2.7
3.7
and poured into 150 mL of n-hexane (CP 1 ), acetonitrile (CP 2 ),
or acetone (CP 3 ) to precipitate the copolymers. The solvent was
decanted, and the copolymers were redissolved and reprecipitated
twice in the respective solvents. The dye content was calculated
by dissolving a defined amount of the copolymer in di-n-butyl ether
and correlating the observed absorbance with the molar absor-
bance of ETHT 4014 in di-n-butyl ether. The amounts of ETHT
4014 in the copolymers, CP 1 -CP 3 , were then found to be 0.21,
0.18, and 0.13 wt %, respectively (Table 1).
Membrane P reparation. Sensor membranes M1 -M3 were
obtained by dissolving 100 mg of the corresponding copolymers
CP 1 and CP 2 in 1.0 mL of ethyl acetate, and CP 3 in 1.0 mL of
cyclohexane. A dust-free glass plate was placed in a spin-coating
device18 with a solvent-saturated atmosphere. Then, 0.3 mL of the
solution was transferred onto the rotating glass support. The
resulting membranes were placed in ambient air for drying.
A peeling-off of the copolymer layers was observed after several
hours of measurement in the flow-through cell, especially in the
case of M1 . Therefore, glass plates cleaned in a mixture of
concentrated sulfuric acid and 30% hydrogen peroxide (3:1 v/ v)
were silanized by exposing them to hexamethyldisilazane vapor
for 15 h in a desiccator at room temperature and at ambient
pressure. After the silanization process, the lipophilic copolymer
layers adhered well to the lipophilized glass plates and no peeling-
off upon exposure to aqueous solutions was observed. The
thickness of the layers was found to be 0.65, 0.30, and 0.15 µm,
respectively, for M1 -M3 . The partial diol formation of the
copolymer layers upon exposure to water requires conditioning
of the sensor layer in aqueous solution before measuring dissolved
aqueous amines.
Figure 2. Schematic representation of the chemical reactions of
trifluoroacetophenone derivatives with water and amines.
angle light scattering detector (Chromatix KMX-6). The thickness
of the sensor layers on the glass plates was measured by using
an Alpha-Step 200 surface profiler from Tencor Inst. (Mountain
View, CA) with an accuracy of 0.05 µm.
Apparatus. Melting points were measured on a Bu¨ chi melting
point apparatus (Flawil, Switzerland) and are uncorrected. 1H
NMR spectra were recorded on a Varian XL 200 Gemini spec-
trometer with chemical shifts given in δ (ppm) relative to TMS.
The fluorescence measurements were performed on a Perkin-
Elmer LS 50 B spectrofluorometer by using a specially designed
flow-through cell16 and pumping the sample solutions at a flow
rate of 1.5 mL min-1 using a Perpex peristaltic pump (Jubile,
Switzerland). The differential scanning calorimetry (DSC) mea-
surements were performed on a Perkin-Elmer DSC 7; all the
samples were first heated to 140 °C and then the inflection points
in a second heating scan at 10°/ min were reported as glass
transition temperature values with an accuracy of (0.5 °C. The
molecular weights and molecular weight distributions were
determined using polystyrene standards in THF polymer solutions
by GPC on a Knauer HPLC System with a PLGel mixed C 5-µm
column, a viscosimetry detector (Viscotek Differential model 502),
a differential refractometry detector (Knauer DRI), and a small-
RESULTS AND DISCUSSION
Optical P roperties of the Fluorogenic Reactand ETHT
4 0 1 4 in Solution and in the Copolymer. The synthesis of the
reactand was performed via alkylated 4-vinylaniline, which was
coupled to 4-bromo-R,R,R-trifluoroacetophenone by a palladium-
catalyzed Heck reaction (Figure 1). This synthetic procedure
provided the methacrylate derivative of a reactand within a simple
three-step reaction. This finding contradicts the claim that such
a synthesis would be impossible because 4-vinylaniline derivatives
polymerize too readily during palladium-catalyzed coupling.19 The
reactand resulting from the synthetic procedure consisted of a
stilbene chromophore and terminal donor-acceptor substituents.
The alkylamino group served as an electron donor, and the long
alkyl chains attached to the nitrogen atom were used for the
introduction of the polymerizable methacrylate groups. The
acceptor part of the fluorescent reactand was the trifluoroacetyl
group, where the C-nucleus is attacked selectively by nucleophilic
reagents, such as alcohols and amines (see Figure 2). The change
in fluorescence intensity of ETHT 4014 was based on the nucleo-
(18) Seiler, K. Ionenselektive Optodenmembranen; Fluka Chemie AG: CH-9470
Buchs, Switzerland, 1991.
(19) Hassner, A.; Birnbaum, D.; Loew, L. M. J. Org. Chem. 1 9 8 4 , 49, 2546.
1536 Analytical Chemistry, Vol. 71, No. 8, April 15, 1999

Table 2. Optical Properties of ETHT 4014 in Solution
ET(30)
(molar transition energy)
20
solvent
λ
max (abs) (nm)
ꢀ (L mol-1 cm-1
)
λ
max (em) (nm)
(kcal mol-1
)
cyclohexane
di-n-butyl ether
toluene
diethyl ether
ethyl acetate
430
436
441
437
444
27 000
34 300
23 900
35 600
34 800
487
523
530
545
597
30.9
33.0
33.9
34.5
38.1
Table 3. Optical Properties and Response Characteristics of Copolymer Membrane Layers, M1-M3
λ
max (nm)
λ
max (nm)
max sensitivity
to 1-butylamine (mM)
LOD
(mM)
forward
reverse
polymer
(HAM)a
(TFA)b
responsec (min)
responsed (min)
M1
M2
M3
422
421
420
528
522
514
2.5-250
3.1-310
5.0-500
0.8
1.0
1.6
4-8
6-10
8-12
8-15
10-15
15-20
a
Emission maximum of the hemiaminal (HAM) form of the fluoro reactand. b Emission maximum of the trifluoroacetyl (TFA) form of the
fluoro reactand. c Time for 95% of the signal change to occur in changing from lower to higher amine concentrations (∆c typically 50 mM). d Time
for 95% of the signal change to occur in changing from higher to lower amine concentrations (∆c typically 50 mM).
philic addition of amines to the trifluoroacetyl group of the
reactand, thereby forming a hemiaminal or zwitterion and bringing
about a change in the degree of electron delocalization.
The structurally related reactands ETHT 4003 and 4004, which
have already been used in plasticized PVC layers, were shown to
exhibit a pronounced solvatochromism. Similar behavior was
found for the methacrylate derivative ETHT 4014 (Table 2). This
behavior allowed us to investigate the decreasing polarity of the
copolymers by comparing the fluorescence emission maximums
of the copolymers with the reactand in solution. The emission
maximums decreased from 528 to 522 and 514 nm, respectively,
in going from butyl and hexyl methacrylate to dodecyl methacryl-
Figure 3. Fluorescence excitation and emission spectra of M2 after
(a, b) and before (c, d) contact with 500 mM 1-butylamine at pH 13.0.
When changing from 0.1 M NaOH to aqueous 1- butylamine, the
trifluoroacetyl form (c, d) is converted into the hemiaminal (a, b).
ate. These values were at shorter wavelengths than the maximum
of ETHT 4014 in PVC plasticized by bis(2-ethylhexyl)sebacate (561
nm). The fluorescence maximums of the copolymers were
comparable to the emission maximums of the reactand in toluene
(530 nm) and di-n-butyl ether (523 nm) and, consequently, gave
an indication of the low polarity inside the copolymer layer (Tables
2 and 3). The well-known ET(30) scale, based on the molar
transition energies of the pyridinium-N-phenoxide betaine dye,
which was developed by Dimroth, Reichardt, and co-workers, is
helpful in discussing solvent polarity in the vicinity/ microenvi-
ronment of a solvatochromic dye.20 It constitutes a more compre-
hensive measure of polymer polarity than the dielectric constant
of single components since it reflects more reliably the complete
picture of all intermolecular forces acting between solute and
solvent. When the observed fluorescence maximums of ETHT 4014
in different solvents (see Table 2) were linearly correlated with
the ET(30) values of the respective solvents, then the ET(30) values
for the membrane bulk using ETHT 4014 as an indicator were
calculated to be 33.5 kcal mol-1 in M1 , 33.1 kcal mol-1 in M2 ,
and 32.6 kcal mol-1 in M3 .
in the copolymers M1 -M3 were all located at around 453 and
468 nm, whereas the emission maximums were at 528 nm in the
case of M1 , and at 522 nm for M2 and 514 nm for M3 (Table 3).
Exposure to aqueous amines resulted in a decrease in fluores-
cence intensity of the reactand due to the conversion of the
trifluoroacetyl group of the stilbene into a hemiaminal (Figure
2). At the same time, the conversion of the reactand resulted in
an increase in the fluorescence of the hemiaminal form of the
stilbene with an excitation maximum at ∼390 nm and an emission
maximum at ∼420 nm (Table 3, Figure 3). Since the fluorescence
intensity of the hemiaminal was at a shorter wavelength, the
fluorescence of the trifluoroacetyl form was chosen for further
investigations.
The fluorescence emission of all three copolymer membranes
(M1 -M3 ) decreased significantly at ∼530 nm on exposure to
solutions containing 1-butylamine (when excited at 450 nm) and
was fully reversible. Due to the favorable mechanical stability and
low Tg value, M2 was characterized in more detail. The relative
decrease in the fluorescence of M2 in changing from 0 to 0.5 M
1-butylamine was 90%. The forward response time, t95, was in the
P erformance of Amine-Sensitive Membranes M1 -M3 .
The fluorescence excitation maximums of the fluorogenic reactand
(20) Reichardt, C. Solvents and solvent effects in organic chemistry, 2nd ed.; Wiley
VCH: Weinheim, Germany, 1990.
Analytical Chemistry, Vol. 71, No. 8, April 15, 1999 1537

Table 4. Selectivity (log Kopt) of M1-M3 for Amines,
Ethanol, and Carbonate
analyte
M1
0.0
M2
0.0
M3
0.0
1-hexylamine
1-butylamine
1-propylamine
ethylamine
methylamine
ammonia
diethylamine
triethylamine
aniline
ethanol
carbonate
spermidine
histamine
cadaverine
tyramine
-1.2
-1.8
-2.3
-2.6
-3.5
-2.7
-2.1
-1.7
-4.2
<-4.5
-1.2
-1.9
-2.4
-2.7
-3.5
-2.9
-2.1
-1.8
-4.2
<-4.5
-3.5
-3.8
-2.1
-2.5
-0.9
-0.8
-1.2
-1.9
-2.5
-2.8
-3.7
-2.9
-2.0
-1.8
-4.2
<-4.5
Figure 4. Short-time repeatability of the response of M2 on
exposure to 1-butylamine: a, 0.1 M sodium hydroxide solution; b, 10
mM 1-butylamine; c, 50 mM 1-butylamine; d, 100 mM 1-butylamine
(at pH 13.0), showing the reproducibility and the response time
measured at an excitation and emission wavelength of 450 and 520
nm, respectively.
benzylamine
phenethylamine
dark at room temperature. The membrane exhibited remarkable
operational stability, and no leaching of the copolymerized ETHT
4014 was observed during 1 week of continuous measurement.
The sensitivity to the amines was mainly governed by the
balance between lipophilicity and sterical hindrance at the reactive
trifluoroacetyl site. The first factor determined the distribution of
the amines between the polymer and the aqueous phase, and the
second factor influenced the ease of reaction. Consequently,
lipophilic primary amines easily interacted with the dye, whereas
the signal changes with bulky secondary and tertiary amines were
much lower. The sensitivity of M1 -M3 to primary amines
decreased with decreasing polarity of the copolymer. In going
from M1 to M3 , the points of inflection of the calibration plots to
1-butylamine (not shown) shifted from 25 to 31 and 50 mM,
respectively (compare the figures for sensitivity in Table 3). The
copolymer membranes also showed large signal changes upon
exposure to aromatic amines. In the case of aniline, it was assumed
that fluorescence quenching accounted for the enhanced sensitiv-
ity because not only was the fluorescence of the trifluoroacetyl
form decreased but the fluorescence of the hemiaminal form was
quenched and was much smaller than in the case of the aliphatic
amines. This behavior has already been observed with amine
sensor membranes based on trifluoroacetyl dyes in plasticized
PVC.1
The sensitivity of M2 to biogenic amines was very low in the
case of histamine, spermidine, tyramine, and cadaverine, although
spermidine exhibited three amino groups. We assume that the
high polarity of the amines prevented their extraction from the
aqueous phase into the highly unpolar copolymer membranes.
However, M2 was extremely sensitive to the more lipophilic and
unpolar benzylamine and phenethylamine. In the case of tyramine,
the high pH of the sample solution deprotonated the phenolic
hydroxy group, and the resulting anionic molecule was not
extracted into the lipophilic polymer matrix. This lack of sensitivity
to the anion was due to the lack of lipophilic cationic sites within
the membrane (e.g., lipophilic ammonium compounds), which
would compensate for the extracted anion and enable the anion
exchange.
Figure 5. Response function of M2 on exposure to aqueous amines
at pH 13.0 measured at an excitation wavelength of 450 nm and an
emission of 520 nm: (A) ammonia; (MA) methylamine; (EA) ethy-
lamine; (PA) 1-propylamine; (BA) 1-butylamine; (HA) 1-hexylamine.
The solid lines represent the response functions calculated according
to ref 1, which gives the symbols for the experimental results.
range of 6-10 min, whereas the time for the reverse response
was in the range of 10-15 min. The short-time reproducibility of
M2 is shown in Figure 4. The relative standard deviations for 10,
50, and 100 mmol 1-butylamine were found to be 1.9%, 1.3%, and
1.4% (n ) 20), respectively, during 48 h of continuous measure-
ment. The limit of detection (LOD) of M2 for 1-butylamine was
1.0 mM. The selectivity of M2 for primary amines is shown in
Figure 5 and the log Kopt values (logarithm of the relative
selectivities between the interfering amines obtained by calculating
the horizontal distance between the amine response curves at the
point of inflection; see also Figure 5)1 in Table 4. As has been
shown for amine sensor layers based on ETHT 4001 and 4004,
exposure to water results in the partial formation of a diol which
brings about small changes in both absorbance and fluorescence.1
However, due to the significantly higher nucleophilicity of amines,
exposure to aqueous amines results in the formation of a
hemiaminal or zwitterion. This causes a large decrease in
absorbance and fluorescence of the trifluoroacetyl form of the
reactand as well as an increase in the hemiaminal form.
M2 exhibited neither bleaching of the dye, chemical decom-
position, nor peeling-off of the copolymer layer from the glass
support for a period of longer than 6 months when stored in the
Cross-Sensitivity of M1 -M3 to Carbonate and Ethanol.
All three copolymer membranes were investigated for their
1538 Analytical Chemistry, Vol. 71, No. 8, April 15, 1999

selectivity to potential interferences. Trifluoroacetophenone de-
rivatives are well-known for their selective interaction with
carbonate in ion-selective electrodes21 and they respond to
alcohols.2-4 Consequently, the response of M1 -M3 to a 0.1 M
solution of carbonate at pH 13.0 was investigated. No signal
change was observed. All three membranes exhibited a significant
response to ethanol, albeit with more preference for amines than
ethanol due to the higher nucleophilicity of amines (Table 4). The
sensitivity of membranes to alcohols is low compared with the
sensitivity to amines. Whereas the sensor membranes used for
ethanol measurement respond to ethanol in the molar range,3,4
the response of M1 -M3 to amines is in the millimolar range.
Furthermore, the sensor membranes for ethanol only show
sufficient response in the presence of a catalyst albeit at neutral
pH. In the present investigation, no such catalyst was used.
There is a significant effect of pH in that the concentration of
amines in aqueous solution is pH-dependent. The pKa value of
aliphatic amines is generally ∼10.5.17 At higher pH, virtually all
amines are present in the electrically neutral form and large
changes in the signals of the sensor membranes on exposure to
amines are observed. At lower pH values than 8.0, there is no
response to amines because they are present in the “unreactive”
protonated ammonium form. Nevertheless, the presence of basic
amines in water can increase the pH up to ∼12 unless the analyte
solutions are buffered, and thus, they can be investigated by the
presented sensor membranes.
Comparison between Copolymerized Reactands and Re-
actands in P lasticized P VC. Generally, copolymers with good
flexibility and low Tg values are obtained by using methyl
methacrylate and adding different amounts of alkyl acrylates. The
latter lowers the Tg of the resulting copolymers and increases the
flexibility of the material. The disadvantage of such an approach
is that the polymerization reactivity of methacrylates and acrylates
is different, producing copolymers whose composition may be
different from the composition of the monomer feed.22 Thus, the
products of the copolymerizations may exhibit unpredictable
properties. As a consequence, we chose to prepare copolymers
from single methacrylates whose Tg values are already low, i.e.,
butyl, hexyl, and dodecyl methacrylate. These monomers were
then copolymerized with small amounts of ETHT 4014 (typically
0.1-0.2%), resulting in copolymers whose physical properties were
practically undisturbed by the presence of the reactand. This
meant that the properties of the copolymers (such as the Tg value)
were predictable.
layers from plasticized PVC. The forward response of M1 -M3
was in the range of 4-12 min and the reverse response within
8-20 min. The response of membranes composed of ETHT 4004
and PVC plasticized with bis(2-ethylhexyl)sebacate was 5-7 min
for the forward and 5-10 min for the reverse response. At the
same time, the LODs decreased: the LOD of PVC-based mem-
branes for 1-butylamine was 3 mM, whereas the LOD of M2 for
the same analyte was 1 mM. The selectivity pattern of M1 -M3
was similar to membranes based on plasticized PVC, with a
preference for primary amines over secondary and tertiary amines.
However, the copolymer-based sensor layers exhibited several
advantages over PVC-based membranes, namely: (a) all compo-
nents were copolymerized and thus no leaching was observed;
(b) due to the covalent linkage, no crystallization, migration, or
aggregation of the dye was found, a behavior often observed with
optical sensor membranes based on plasticized PVC; (c) the
copolymers did not require plasticizers, thus enhancing the
operational and shelf lifetimes, and (d) the present approach is
not limited to sensing neutral analytes because copolymerizable
additives, such as methacrylate derivatives of quaternary am-
monium ions and borates, are available.23,24 These cationic and
anionic sites are necessary for the proper functioning of ion-
selective optodes based on the mechanisms of ion exchange and
coextraction. Consequently, selective ion sensors with high
operational and shelf lifetimes are conceivable.6
The mechanical stability of membranes based on copolymers
and of those based on plasticized PVC is comparable. The only
significant difference is the lower adherence of the copolymers
on glass plates, due to the copolymers’ high hydrophobicity. The
problem can be overcome by silanization of the glass supports,
and the use of supports such as polyester or transparent poly-
methacrylates is also conceivable.
The fluorescent reactands ETHT 4003 and 4004 have already
been shown to respond to amines in organic solvents,1 but since
the copolymer layers M1 -M3 are soluble in almost any organic
solvent except water and alcohols, they cannot be applied for
monitoring amines in organic solvents. An application of copoly-
mers for amine-sensing in organic solvents will, however, be
feasible by preparing highly cross-linked copolymer layers which
are insoluble in organic solvents.8,23,24
ACKNOWLEDGMENT
We thank Bosch Telecom GmbH, Germany, for their financial
support for part of the work. We also thank Profs. Elizabeth A.
H. Hall (University of Cambridge, Institute of Biotechnology) and
Renate Dworczak (Karl-Franzens Universit¨at, Graz, O¨ sterreich)
for stimulating discussions.
In summary, the response behavior of membranes incorporat-
ing the reactands linked to the polymer (M1 -M3 ) is comparable
to the structurally related reactands (ETHT 4001 and 4004)
dissolved in PVC/ bis(2-ethylhexyl) sebacate. However, immobiliz-
ing the reactands brings about response times that are by a factor
of 1.5-2 slower although M1 -M3 are ∼10 times thinner than
Received for review November 24, 1998. Accepted
February 2, 1999.
AC9812983
(21) Behringer, C.; Lehmann, B.; Haug, J. P.; Seiler, K.; Morf, W. E.; Hartman,
K.; Simon, W. Anal. Chim. Acta 1 9 9 0 , 233, 41.
(22) Mohr, G. J.; Tirelli, N.; Lohse, C.; Spichiger, U. E. Adv. Mater. 1 9 9 8 , 10,
1353.
(23) Antonisse, M. M. G.; Lugtenberg, R. J. W.; Egberink, R. J. M.; Engbersen,
J. F. J.; Reinhoudt, D. N. Anal. Chim. Acta 1 9 9 6 , 332, 123.
(24) Brunink, J. A. J.; Lugtenberg, R. J. W.; Brzozka, Z.; Engbersen, J. F. J.;
Reinhoudt, D. N. J. Electroanal. Chem. 1 9 9 4 , 378, 185.
Analytical Chemistry, Vol. 71, No. 8, April 15, 1999 1539