Preparation and self-assembly of chiral porphyrin diads on the gold electrodes of quartz crystal microbalances: A novel potential approach to the development of enantioselective chemical sensors

Article abstract of DOI:10.1002/1521-3765(20020603)8:11<2476::AID-CHEM2476>3.0.CO;2-E

Porphyrin diad 1 was synthesized by reaction of the acyl chloride of porphyrin 2 and trans-1,2-dithiane-4,5-diol. The Co complex of this diad was studied as a potential enantioselective receptor for chiral recognition in solution and in the solid state. In solution both enantiomers of limonene induce significant changes in the visible and circular dichroism (CD) spectra of [Co₂(1)], while a different behavior is observed in the case of the enantiomeric pair of trans-1,2-diaminocyclohexane. A different efficiency of [CO₂(1)] chiral recognition is obtained for these compounds, with a remarkable degree of enantiodiscrimination observed in the case of limonene. Self-assembled monolayers of [Co₂(1)] were deposited onto the gold electrodes of quartz crystal microbalances to be used as sensing materials of nanogravimetric sensors operating in the gas phase. The enantio-discrimination properties of these sensors towards the enantiomeric pairs of chiral analytes have been studied. While in the case of analytes bearing donor ligand atoms we did not observe a remarkable enantioselectivity, a significant degree of chiral discrimination was observed in the case of limonene; this result is particularly encouraging for the potential development of enantioselective chemical sensors for use in an array configuration.

Full text of DOI:10.1002/1521-3765(20020603)8:11<2476::AID-CHEM2476>3.0.CO;2-E

FULL PAPER
Preparation and Self-assembly of Chiral Porphyrin Diads on the Gold
Electrodes of Quartz Crystal Microbalances: A Novel Potential Approach to
the Development of Enantioselective Chemical Sensors
Roberto Paolesse,*[a] Donato Monti,[a] Laura La Monica,[a] Mariano Venanzi,[a]
Antonella Froiio,[a] Sara Nardis,[a] Corrado Di Natale,[b] Eugenio Martinelli,[b] and
Arnaldo D×Amico[b]
Abstract: Porphyrin diad 1 was synthe-
sized by reaction of the acyl chloride of
porphyrin 2 and trans-1,2-dithiane-4,5-
diol. The Co complex of this diad was
studied as a potential enantioselective
receptor for chiral recognition in solu-
tion and in the solid state. In solution
both enantiomers of limonene induce
significant changes in the visible and
circular dichroism (CD) spectra of
[Co2(1)], while a different behavior is
observed in the case of the enantiomeric
pair of trans-1,2-diaminocyclohexane. A
different efficiency of [Co2(1)] chiral
recognition is obtained for these com-
pounds, with a remarkable degree of
enantiodiscrimination observed in the
case of limonene. Self-assembled mono-
layers of [Co2(1)] were deposited onto
the gold electrodes of quartz crystal
microbalances to be used as sensing
materials of nanogravimetric sensors
operating in the gas phase. The enantio-
discrimination properties of these sen-
sors towards the enantiomeric pairs of
chiral analytes have been studied. While
in the case of analytes bearing donor
ligand atoms we did not observe a
remarkable enantioselectivity, a signifi-
cant degree of chiral discrimination was
observed in the case of limonene; this
result is particularly encouraging for the
potential development of enantioselec-
tive chemical sensors for use in an array
configuration.
Keywords: enantiomeric
recogni-
tion ¥ monolayers ¥ porphyrinoids
¥ self-assembly ¥ sensors
an enhancement of the selectivity properties and an optimi-
zation of the transduction mechanism. Of course some
important limitations should be taken into account, such as
the low saturation limit of the resulting sensor. This aspect,
however, is not necessarily a drawback, but rather represents
an advantage in the presence of low concentrations of
analytes.
Among the different molecules proposed as sensing mate-
rials we have been particularly focused on porphyrins.[4] In the
last few years we have developed sensor arrays composed of
quartz crystal microbalances (QCMs) coated with metal
complexes of porphyrins for the analysis of volatile organic
compounds (VOCs).[5] Metalloporphyrins have been partic-
ularly useful in these applications, because of their stability
and their versatility as binding agents.[6] Various examples of
the self-assembly of porphyrins onto inorganic surfaces have
been reported in the literature.[7] We have been able to
functionalize the QCM by depositing self-assembled mono-
layers of thiol-substituted porphyrins onto the gold electrodes
of the QCM disks; the resulting sensors showed satisfying
sensing properties for the detection of VOCs.[8]
Introduction
The modification of inorganic surfaces by chemisorption of
organic compounds is an interesting approach to the fabrica-
tion of hybrid organic/inorganic materials, the properties of
which are almost unexplored.[1] When adequately developed,
these materials could have a tremendous impact in the
development of technologies in which the versatility of
organic compounds is a requirement.[2]
One of the fields in which this goal seems to be close to
realization is the development of chemical sensors.[3] In these
devices interactions at the sensor surface are clearly of critical
importance: the possibility of a direct bond between the
sensing material and the transducer surface can greatly
improve sensor performance. For example, it can allow both
[a] Dr. R. Paolesse, Dr. D. Monti, L. La Monica, Prof. M. Venanzi,
A. Froiio, S. Nardis
Dipartimento di Scienze e Tecnologie Chimiche
Universita di Roma Tor Vergata
¡
Via della Ricerca Scientifica, 00133 Rome (Italy)
Fax : (‡39)6-72594328
E-mail: roberto.paolesse@uniroma2.it
These promising results led us to explore a further step in
the exploitation of self-assembled, functionalized QCMs. In
the previously developed QCMs, single porphyrins were
[b] Prof. C. Di Natale, E. Martinelli, Prof. A. D×Amico
Dipartimento di Ingegneria Elettronica
¡
Universita di Roma ™Tor Vergata∫, 00133 Roma (Italy)
2476
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0811-2476 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 11

2476 2483
chemisorbed onto the electrode surface; our present approach
involves the deposition of a chiral porphyrin diad in order to
create a more complex receptor for the recognition of VOCs.
Our goal is to test such a device for enantiomeric recognition.
This aspect, though particularly challenging, is remarkably
important when applied to the analysis of real matrices, as
enantiomers are chemically and physically identical in an
achiral environment, differing only by the spatial arrange-
ment of their elements. In biological systems enantiomeric
discrimination is performed by interaction of chiral molecules
with enantiomerically pure receptors;[9] the progress of
supramolecular chemistry has given a great impulse to the
development of a similar host guest concept with simple
molecular systems, and several elegant systems have been
onto the gold pad of QCM sensors. The binding behavior of
such a diad was first tested in solution, while its sensing and
chiral discrimination properties toward some enantiomeric
pairs of chiral compounds were measured by QCM sensors.
Results and Discussion
Synthesis and binding studies in solution: The porphyrin diad
1 was prepared following the synthetic pathway depicted in
Scheme 1. Porphyrin 2 reacted with oxalyl chloride in
dichloromethane solution to give the corresponding porphy-
rin acyl chloride; this intermediate was then treated with
dithiane to afford 1 in satisfactory yield. In the same way, 3
proposed to enhance chiral discrimination in solution.[10]
A
similar host guest approach has been used in analytical
chemistry to obtain discrimination and separation of mixtures
of enantiomers by GC and HPLC chromatographic meth-
ods.[11] This has generally been achieved by using stationary
phases functionalized with chiral receptor sites. On the other
hand, while a huge number of chemical sensors have been
developed, only a few examples have been reported in which
chiral discrimination is performed by chemical sensors.[12] In
these studies the researchers used sensing materials composed
of an inert matrix into which chiral receptors were dispersed;
these materials have been derived from those exploited as
stationary phases in the GC technique.
CO2H
N
NH
N
+
+
N
HN
H
CHO
CHO
CO2H
2
However in the case of some real application, such as the
aroma of foods, in which chiral compounds often represent a
fundamental component of the blend, the use of conventional
analytical methods is usually hampered by the overwhelming
complexity of their chemical composition.[13] In this case,
mimicking the human nose, which is able to discriminate
between the enantiomers of limonene simply on the basis of
their different smell, constitutes a desirable goal.
Since the seminal paper of Persaud and Dodd,[14] artificial
olfaction has been achieved by using an array of different
chemical sensors, each with different selectivity properties
towards the various chemical species present in the gaseous
phase.[15] In this way the whole sensor array gives information
related to the bulk environment, and the subsequent data
analysis interprets this information on the basis of the
chemical composition.[16] An advantageous feature of such a
system is that the absolute selectivity of each individual sensor
is not required.[17]
oxalyl chloride
N
NH
N
OH
SH
HN
N
NH
N
C
O
O
HN
SH
CO2Cl
3
HO OH
S S
Using this approach, our long-term goal is to develop
chemical sensors with enantiomeric recognition properties
towards some classes of chiral analytes. These devices will be
designed for sensor array exploitation, in order to develop an
electronic nose devoted to the analysis of real matrices, in
which chiral analytes can play an important role.
N
HN
N
N
H
H
N
H
N
N
N
We decided to first explore the use of QCMs functionalized
by monolayers of a chiral receptor directly linked to the
sensor surface. This method is particularly appealing, because
it could avoid the occurrence of spurious interactions as a
result of the bulk of the sensing materials; this would greatly
improve their recognition performance.
O
O
C
C
O
O
S
S
1
In this work we present the synthesis of a chiral porphyrin
diad suitable for deposition as a self-assembed monolayer
Scheme 1. Synthetic pathway to 1.
Chem. Eur. J. 2002, 8, No. 11
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0811-2477 $ 20.00+.50/0
2477

R. Paolesse et al.
FULL PAPER
was prepared by reaction of the intermediate porphyrin acyl
chloride with thiophenol. Co complexes of 1 and 3 were
obtained by following standard metalation procedures.[18] We
chose Co as the coordinated metal, because Co porphyrins
have been successfully used as sensing materials in electronic
nose applications.[4]
The UV-visible spectrum of the [Co2(1)] diporphyrin
receptor features two bands in the Soret region at 435 and
412 nm, in a 1:4 intensity ratio. The invariance of the spectrum
profile over a wide concentration range (1 Â 10À7 5 Â 10À5 m)
safely rules out the occurrence of an aggregation effect on the
observed phenomenon. On the basis of the spectral profile the
occurrence of an equilibrium between two different confor-
mations, one open and one closed, can be hypothesized. In the
open conformation the porphyrin macrocycles are probably
held apart, resulting in a noninteracting, energetically mono-
mer-like structure, which generates the Soret band at 412 nm.
In the closed structure, the two macrocycles are probably held
in a J-type conformation, resulting in a red-shifted, somewhat
broadened absorption band at 435 nm.[19]
Figure 2. Titration plot (l ˆ 437 nm) of [Co2(1)] (1.8  10À5 m), at various
concentrations of S-(À)-limonene in CHCl3. The points are experimental
and the curve is calculated for 1:1 complex formation (see text).
Analogous experiments carried out with cyclohexene, a
related olefinic structure bearing only one double bond,
showed a much weaker interaction (K < 10mÀ1); this supports
the above-given hypothesis. Evidently, the two unsaturated
bonds of limonene act cooperatively, resulting in a stronger
interaction.
Parallel circular dichroism (CD) spectroscopic experiments
gave further compelling evidence for the inclusion of the
chiral analytes within the receptor walls. The [Co2(1)]
diporphyrin system features an intense, negative dichroic
The addition of S-(À)- or R-(‡)-limonene to a solution of
[Co2(1)] causes dramatic changes in the UV-Visible spectra of
the receptor. Figure 1 shows the peculiar evolution of the
band ([W] ˆ 1.5  104 degcm2 dmolÀ1) at 440 nm, (2  10À5
m
chloroform solution, Figure 3), along with a weaker ([W] ˆ
1.2 Â 103 degcm2 dmolÀ1) negative band centered at 415 nm.
The red-shifted absorption probably corresponds to the
closed, J-like conformation, whereas the less intense band
arises from the open structure.
Figure 1. UV-visible spectral changes (Soret band selected region) of
[Co2(1)] (1.8 Â 10À5 m), at various concentrations of S-(À)-limonene in
CHCl3. The arrows indicate the changes upon increasing limonene
concentration.
spectral pattern (i.e. the two Soret bands at 412 and 435 nm)
into a new, red-shifted band at 437 and 438 nm for S-(À)- and
R-(‡)-limonene, respectively. A significant hyperchromic
effect is also present. A red shift of the two Q visible bands
is also evident upon limonene addition. These results can be
explained in terms of inclusion of the unsaturated species
within the receptor walls, as a consequence of p p interaction
between the limonene double bonds and the aromatic system
of the porphyrin platforms.[20] The overall association constant
values K, for the binding of the limonene enantiomers to the
diporphyrin receptor, can be calculated from the spectral
changes (Figure 2). The experimental points can be fitted by a
1:1 binding curve.[21] The preference for the S-(À) over the R-
(‡) enantiomer, demonstrated by their relative K values 6200
and 1600mÀ1, respectively, is in close agreement with that
observed in the solid-state experiments (vide infra).
Figure 3. a) CD spectra of [Co2(1)] (1.8 Â 10À5 m, CHCl3, 298 K); b) in the
presence of R-(À)-limonene 5.0  10À3 m; c) in the presence of S-(‡)-
limonene 5.0 Â 10À3 m.
The addition of an excess of S-(À)-limonene promotes
a
decrease
of
the
CD
amplitude
([W] ˆ 9.8 Â
103 degcm2 dmolÀ1). Furthermore, the addition of R-(‡)-
limonene results in an even more pronounced decrease of
the CD band intensity, up to 7.8 Â 103 degcm2 molÀ1. This
finding can be explained by the formation of supramolecular
complexes that are characterized by a lower degree of
ellipticity. Remarkably, the CD-coupled band at 415 nm,
2478
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0811-2478 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 11

Chemical Sensors
2476 2483
disappears upon the formation of a limonene [Co2(1)], in a
close analogy to that observed in the UV-visible spectroscopic
studies. This finding corroborates the hypothesis that the
guest is sandwiched within the porphyrin platforms of the
receptor.
The larger CD hypochromic effect observed in the case of
the R-(‡) enantiomer would indicate that the formation of
the inclusion complex occurs with higher structural reorgan-
ization (i.e. unfavorable cage complementarity). This lack of
pre-organization of the structure would constitute an insur-
mountable drawback in the solid-state interaction, dictating
the preference for the S-(À) enantiomer.
Figure 4. UV-visible spectral changes of [Co2(1)] (1.8 Â 10À5 m, in CHCl3 ;
continuous line) upon addition of S-(À)-limonene (3.6 Â 10À3 m; dashed
line), and trans-1,2-DACy (1.2 Â 10À4 m; dotted line).
UV-visible spectroscopic studies were also carried out in
the case of trans-R,R-(À)- and trans-S,S-(‡)-1,2-diaminocy-
clohexane (‡/ À DACy). Spectral changes are exhibited by
the diporphyrin receptor upon nitrogen-to-metal coordina-
tion, in close analogy to those observed in the case of the
interaction with the limonene enantiomers. A striking differ-
ence resides in the Soret bands of the resulting DACy
[Co2(1)] supramolecular complexes, which feature blue-shift-
ed maxima at 431 nm (Figure 4). This finding implies the
formation of a complex in which the porphyrin platforms are
held in a geometry that would probably be tighter than that
achieved by the complexes with limonenes.
The [Co2(1)] receptor features a rather small degree of
enantioselectivity toward the binding of trans-R,R-(À)- and
trans-S,S-(‡)-1,2-diaminocyclohexane stereoisomers, K being
3.3 and 2.8 Â 104 mÀ1, respectively, which corresponds to a ratio
as small as 1.2.[22]
of [Co2(1)] or [Co(3)], or by recording the visible spectrum in
the case of glass substrates. The experimental set up for QCM
measurements is reported in Scheme 2.
The amount of the sensing material deposited on the QCMs
can be determined from the change in the resonating
frequency of the quartz. In the absence of viscoelastic effects,
the relationship between quartz frequency shift and adsorbed
mass is, in a first approximation, linear, as described by the
Sauerbrey Equation [Eq. (1)].[23] The constant Kq is given by
Equation (2), in which fo is the fundamental frequency of the
QCM, vs is the sound velocity on the quartz, A is the QCM
area, and 1c is the density of the deposited molecular film.
Df ˆ ÀKqDm
2f 2o
(1)
(2)
Thin-film depositions: Porphyrin monolayers were deposited
onto gold surfaces by following two different procedures. In
the first approach the diad was first synthesized and then
deposited onto the gold surface by dipping the substrate into a
solution of 1 in chloroform. In the second approach the gold
surface was first modified by chemisorp-
tion of dithiane; this functionalized sur-
face was then dipped into a solution of
the porphyrin acyl chloride in CH2Cl2 to
obtain the diad in situ.
1
Kq ˆ
vs 1cA
Equation (1) holds only if the viscoelastic effect can be
neglected. This condition depends on the thickness of the
We took this latter approach to
deposit monolayers of the monomeric
4-Co-5'-(4''-carboxyphenyl)-10',15',20'-
triphenylporphyrinyl]-1-mercaptoben-
zene ([Co(3)]). In this case the surface
was modified by adsorption of thiophe-
nol and then the surface was treated with
porphyrin acyl chloride to obtain the
relative porphyrin monolayer. We used
[Co(3)] to compare its properties with
those of [Co2(1)] in the subsequent
sensor measurements.
We modified the gold electrodes pres-
ent on both sides of the QCMs, then used
for sensing measurements, and transpar-
ent gold layers deposited on glass sub-
strates. The formation of monolayers was
monitored by the frequency change of
Scheme 2. Experimental set up for QCM measurements.
the QCMs, caused by the chemisorption
Chem. Eur. J. 2002, 8, No. 11
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0811-2479 $ 20.00+.50/0
2479

R. Paolesse et al.
FULL PAPER
adsorbed mass and it may be assumed valid in case of a
molecular monolayer.
The Kq constant can be either calculated from Equation (2)
or determined experimentally for each QCM by deposition of
a known amount of mass onto the quartz surface. We followed
experiment was carried out with a bare quartz and we
obtained the same result, confirming that the cleaning
procedure removes all the porphyrin diads not chemisorbed
on the gold surface.
These results seem to indicate complete coverage of the
surface and a close-packing of the molecules, with the
porphyrin rings oriented in an almost perpendicular fashion
to the gold surface.
We also have to consider the possibility of the occurrence of
aggregation phenomena among the macrocycles linked to the
surface; this phenomenon is generally not advantageous for
sensing applications, as the presence of porphyrin stacking
does not allow an optimal interaction of the volatile analytes
with the macrocycles.
this latter procedure and obtained
1.00(Æ0.02) HzngÀ1.
a
value of
After deposition of [Co2(1)], we obtained a frequency shift
of about 500 Hz in the case of the two-step approach, while
the single-step method afforded smaller changes, with a
frequency shift around 350 Hz. All these depositions were
carried out in triplicate, by dipping three different QCMs into
the same organic solution of the diad, and the frequency
changes obtained were reproducible within 5%.
Longer deposition times did not increase the frequency
changes, while shorter times gave smaller values, indicating
incomplete coverage of the gold surface.
In the case of porphyrins, optical spectroscopy is a powerful
method for the investigation of aggregation and coordination
phenomena.[25, 26] For this reason we deposited monolayers of
[Co2(1)] onto transparent gold surfaces, and the optical
spectrum obtained is shown in Figure 5.
According to Equation (1), and considering the measured
quartz constant, the amount of [Co2(1)] self-assembled onto
the gold electrode of the QCM may be estimated to be in the
500(Æ10) ng range. From this value we can calculate the
moles and then the number of the molecules deposited onto
the quartz surfaces. For a total gold surface of 1.25 cm2
(considering both sides of the QCM), an area of 66 ä2 per
molecule can be estimated. Considering that a single porphy-
rin molecule has an area of 225 ä2, the value we obtained is
even lower than that expected for a porphyrin ring oriented
perpendicular to the surface. However, previous STM meas-
urements showed that the electrode surface is very rough,
with the characteristic grain structure of evaporated gold.[24]
This should negatively influence the accuracy of the determi-
nation of the real surface area, and can explain the small value
estimated. In fact, if we consider the gold crystallites as
hemispheres, we can estimate the real surface from simple
Figure 5. Visible spectra of [Co2(1)] deposited on gold transparent layer
(dashed line), upon addition of triethylamine vapors (dotted line), upon
addition of limonene vapors (continuous line).
geometrical considerations by the relation Sreal ˆ 2pr2Sgeometrical
/
pr2, in which 2pr2 is the surface of a hemisphere and
Sgeometrical/pr2 is the total number of hemispheres. From this
relation we obtain that Sreal ˆ 2Sgeometrical, without considering
the interstitial space between the hemispheres. As a conse-
quence the area per molecule assumes the more reasonable
value of more than 130 ä2. On the other hand we have to
exclude the possibility that the high mass value could be
caused by the presence of further porphyrin diads aggregated
onto the layer directly bound to the surface. It is well known,
in fact, that porphyrins show a strong tendency to aggregation
by p p stacking of their aromatic rings.[21a, 25] However these
molecules are linked by weaker interaction and, if present,
they are effectively removed by the washing protocol adopted
after the deposition. To control the efficiency of the procedure
we measured the resonating frequency of the QCMs before
and after each washing step and we detected no variation after
the first cleaning treatment. To further check this point we
deposited additional layers of the porphyrin diad by a spray-
coating technique onto one of the [Co2(1)]-functionalized
QCMs.[4f] After drying the surface with an N2 stream, we
treated this quartz by the usual washing protocol; after this
treatment we obtained the starting resonating frequency
value for the [Co2(1)] QCM, indicating that the further layers
were completely removed by the cleaning protocol. The same
In the solid state the Soret band is centered at 432 nm and
the splitting observed in solution is not present; however the
band is somewhat broadened and, as a consequence, the less
intense, red-shifted band may not be resolved. Exposure of
the functionalized glass substrates to different volatile com-
pounds demonstrated the interactions of the porphyrin layers
with these analytes. In the case of triethylamine (Figure 5)
there is a small red shift of the Soret band, indicating the
coordination of the amine to the metal center. It is interesting
to note that this red shift is analogous to that observed in
solution for Co porphyrin complexes.[26] To study the influence
of the metal coordination on the sensing properties we
performed the same experiment with VOCs without donor
atoms, such as limonene. In this case a significant sharpening
of the Soret band was observed, without significant shift of the
absorption band. This feature is similar to that observed in
solution, supporting a similar binding mechanism in the
recognition phenomenon (Figure 5).
QCM studies. We studied the sensing performance of the
functionalized QCMs described in the previous section
towards the detection of VOCs (for the experimental set up
2480
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0811-2480 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 11

Chemical Sensors
2476 2483
Table 1. Average sensor responses of QCM modified with different self-
assembled monolayers of porphyrins upon exposure to enantiomeric pairs
of chiral VOCs.
see Scheme 2). For these studies we used eight different
QCMs housed in a 10 mL measurement chamber: two of them
were modified with 3, and four of them with [Co2(1)],
deposited by the two different approaches. We used each
sensing material in duplicate in order to test the reproduci-
bility of the performances. The other two positions in the
array were occupied by a bare QCM, and by a [Co(tpp)]
(tpp ˆ tetraphenylporphyrin) spray-coated QCM. In this
way all of the sensors were exposed to the analytes in the
same conditions, allowing for a direct comparison of the
performances of the self-assembled QCM with both a bare
QCM, and with a sensor currently used in the electronic
nose.[4]
In the first phase, we carried out studies on the sensitivity of
the QCMs towards simple molecules, such as triethylamine. In
this case we observed a good sensitivity of the sensors toward
this analyte, but we obtained saturation of the self-assembled
QCMs in all cases, even in the presence of very low
concentrations of triethylamine. This low saturation limit is
not unexpected in the case of a sensing material monolayer; it
is important to note that the saturation value is in good
agreement with the theoretical value obtained from Equa-
tion (1), considering a 1:1 binding stoichiometry between each
Co porphyrin and triethylamine. At higher concentrations of
the analyte we observe an increase of the sensor responses,
due in this case to the deposition of further layers of the
analyte on the QCM surface, as expected for a BET-like
(BETˆ Brunauer Emmet Teller) behavior.[27]
Sensing material
analyte
DF [Hz]
DF(À)/DF(‡)
3
R,S,R-(À)-menthol
S,R,S-(‡)-menthol
S-(À)-limonene
48 (Æ3)
42 (Æ3)
60 (Æ5)
55 (Æ4)
51 (Æ4)
45 (Æ3)
20 (Æ1)
22 (Æ3)
53 (Æ4)
62 (Æ6)
233 (Æ9)
65 (Æ3)
69 (Æ7)
59 (Æ5)
61 (Æ4)
44 (Æ7)
72 (Æ6)
84 (Æ3)
108 (Æ7)
51 (Æ5)
58 (Æ3)
51 (Æ4)
58 (Æ6)
48 (Æ2)
1.1
1.1
1.1
1.1
0.8
3.6
1.2
1.4
0.9
2.1
1.1
1.2
R-(‡)-limonene
R,R-(À)-1,2-DACy
S,S-(‡)-1,2-DACy
R,S-(À)-ephedrine
S,R-(‡)-ephedrine
R,S,R-(À)-menthol
S,R,S-(‡)-menthol
S-(À)-limonene
1 (two step)
R-(‡)-limonene
R,R-(À)-1,2-DACy
S,S-(‡)-1,2-DACy
R,S-(À)-ephedrine
S,R-(‡)-ephedrine
R,S,R-(À)-menthol
S,R,S-(‡)-menthol
S-(À)-limonene
1 (one step)
R-(‡)-limonene
R,R-(À)-1,2-DACy
S,S-(‡)-1,2-DACy
R,S-(À)-ephedrine
S,R-(‡)-ephedrine
ing that QCMs modified with 3 did not show significant
differences between racemate and enantiomers, as expected
in the absence of chiral discrimination.
Because limonene is the more volatile among the different
compounds tested, we exposed the sensor array to cyclo-
hexene-saturated vapors to exclude the different volatility as
the reason for the higher sensor responses; in this case the
frequency shifts were significantly low, excluding saturation
effects in the chiral discrimination of limonene.
We then studied sensor responses towards some enantio-
meric pairs of chiral compounds, namely limonene, menthol,
1,2-DACy, and ephedrine; we chose compounds with similar
molecular structure and with or without donor atoms, in order
to determine the importance of metal coordination in the
enantiosensing recognition process.
The sensor array was exposed to saturated vapors of each
compound. The measurement of each isomer was performed
five times and the results obtained are reported in Table 1.
Surprisingly, with the menthol, 1,2-DACy, and ephedrine
enantiomeric pairs, we did not obtain significant differences in
the sensor responses; the frequency change ratios being close
to one. This finding can be explained by considering the
saturation limit of each sensor, calculated as reported in the
previous section. We obtained, in fact, a [Co2(1)]:analyte
binding stoichiometry of 1:2 for all of these compounds,
probably indicating that the two Co porphyrins behave
independently in the coordination of the analytes to the metal
centers.
Remarkably, a significant chiral discrimination is observed
in the case of limonene, a structure closely related to menthol,
the most evident difference being the lack of an oxygen donor
site. The DF(À)/DF(‡) value for the enantiomeric pair was as
large as 3.6, and this result nicely confirms what we observed
in solution. It is interesting to note that we obtained a binding
stoichiometry of about 1:3 for S-limonene, indicating that a
complex binding mechanism involving p p interaction is
probably in operation, as is the case in solution.
Conclusion
Diad 1 was straightforwardly synthesized starting from
dithiane and 2. This diad was used to modify the gold surface
present on quartz crystal microbalances (QCMs) by following
two different approaches. The formation of the organic layer
was monitored by the frequency change of the QCM and by
UV-visible spectroscopy. The results obtained showed a good
coverage of the metal surface.
These QCMs showed good sensitivities to different volatile
organic compounds (VOCs), with a low saturation limit.
Enantiodiscrimination studies were carried out in the solid
state by exposure of the functionalized QCMs to vapors of
some enantiomeric pairs. While we did not observe chiral
discrimination in the case of menthol, ephedrine, and 1,2-
DACy, we obtained a remarkable degree of selectivity in the
case of limonene. This result is particularly encouraging for
the development of sensor arrays to be exploited in the
analysis of real matrices in which chiral VOCs play an
important role.
When we exposed the QCMs to a racemic mixture of
limonene, we obtained sensor responses of intermediate value
between the two pure enantiomers. It is also worth mention-
Chem. Eur. J. 2002, 8, No. 11
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0811-2481 $ 20.00+.50/0
2481

R. Paolesse et al.
FULL PAPER
6 mmol) was added to the reaction mixture under stirring. After 2 h the
solvent was evaporated in vacuo to give a dark brown residue, which was
then dissolved in dry CH2Cl2 (20 mL). A solution of trans-1,2-dithiane-4,5-
diol (5.8 mg, 0.38 mmol) in dry CH2Cl2 (4 mL) was rapidly added along
with two drops of anhydrous pyridine. The mixture was left stirring
overnight. The solvent was then stripped off in vacuo, and the residue
dissolved in CH2Cl2 (20 mL), washed with brine, and dried (Na2SO4). The
solvent was reduced to a small volume, and the residue was purified by
chromatography (SiO2), eluting with CH2Cl2 in order to give the title
diporphyrin 1 (0.03 mmol, 0.043 g, 40% yield). The unreacted 2 was
Experimental Section
Instrumentation: 1H NMR spectra were recorded with Bruker AC300P
(300 MHz) or Bruker AM400 (400 MHz) spectrometers. Chemical shifts
are given in ppm relative to tetramethylsilane (TMS). Routine UV-visible
spectra were measured on a Varian Cary 50 spectrophotometer, whereas
more delicate measurements were performed on a Perkin Elmer l spec-
trophotometer equipped with a temperature-controlled cell holder. CD
spectra were recorded on a JASCO J600 spectropolarimeter; all spectra,
averaged over eight consecutive scans, were reported as molar ellipticity
(degrees cm2 dmolÀ1). Mass spectra (FAB) were recorded on a VG Quattro
spectrometer with m-nitrobenzyl alcohol (NBA, Aldrich) as a matrix in the
positive ion mode.
recovered by elution with
(300 MHz, CDCl3, 258C): d ˆ À2.71 (brs, 4H), 3.24 (brs, 4H), 3.95 (brm,
2H), 7.60 (m, 20H), 8.06 (m, 18H), 8.71 (m, 16H); UV/Vis (CH2Cl2): lmax
a
CH2Cl2/MeOH 2% mixture. 1H NMR
ˆ
419, 513, 551, 589, 649 nm; MS (FAB): m/z: 1435; elemental analysis calcd
Materials: Silica gel 60 (70 230 mesh) was used for column chromatog-
raphy. Reagents and solvents (Aldrich, Merck, or Fluka) were of the
highest grade available and were used without further purification.
Solvents used for the spectrophotometric measurements, acetonitrile and
chloroform Uvasol (Merck), were stored over activated molecular sieves.
(%) for C94H64N8O4S2: C 78.75, H 4.50, N 7.82; found: C 78.99, H 4.39, N 7.74.
Metalation reactions: The general procedure for the synthesis of the Co
derivatives of the above porphyrins was as follows: a CHCl3 solution of the
substrate was placed in a round-bottomed flask equipped with a reflux
condenser and a nitrogen inlet tube. The mixture was brought to reflux
temperature and then an excess of a saturated methanolic solution of
Co(OAc)2 (2 mL) was added under stirring. The progress of the reaction
was monitored by UV-visible spectroscopy. At the end of the reaction the
mixture was allowed to cool to room temperature, and the solvent was
removed in vacuo. The residue was dissolved in CHCl3 (50 mL), washed
with brine (3 Â 50 mL), and dried (Na2SO4). The solvent was removed in
vacuo, and the residues, that is, the metalated porphyrins, were subse-
quently purified by crystallization from CH2Cl2/hexane mixture.
UV-visible spectroscopy studies in solution: UV-visible spectroscopic
studies of the interactions of chiral analytes with [Co2(1)] in solution were
carried out in chloroform by following the absorbance changes in the 350
450 nm range (porphyrin derivative Soret band) upon guest complexation
(Tˆ 258C).
UV-visible spectroscopic titration: A typical procedure is as follows:
aliquots of titrating solution of the guest (1 Â 10À2 1 Â 10À3 m), prepared by
dissolving the required amount of olefin in a freshly prepared solution of
[Co2(1)] (1 5 Â 10À6 m) in chloroform, were added portionwise with a
micro syringe to the diad solution (2.5 mL, 1 cm quartz cell, 25.08C). The
absorbance variation (Soret band) was monitored at different concentra-
tions of added guest. The stability constants (K) were calculated by using a
standard equation for a 1:1 complexation:[28] DA ˆ De[[Co2(1)]]K[G]/
(1 ‡ K[G]), where DA ˆ A0 À Ai; De ˆ e0 À e1; [G] is the concentration of
the added guest. The computer aided nonlinear least-square fitting analyses
[Co(2)]: Starting from 2 (0.8 g, 1.5 mmol), we obtained [Co(2)] (1.04 g,
1.4 mmol, 93% yield) as a purple crystalline solid. UV/Vis (CH2Cl2): lmax
412, 529, 554 nm; MS (FAB): m/z (%): 716; elemental analysis calcd (%)
ˆ
for C45H28N4O2Co: C 75.52, H 3.94, N 7.83; found: C 75.40, H 3.89, N 7.77.
[Co2(1)]: Starting from 1 (50 mg, 0.07 mmol) we obtained [Co2(1)] (46 mg,
0.03 mmol, 45% yield). UV/Vis (CH2Cl2): lmax (log e) ˆ 412 (4.95), 435
(5.1), 548 (4.11), 590 (3.78) nm; MS (FAB): m/z (%): 1548; elemental
analysis calcd (%) for C94H60N8O4S2Co2: C 72.96, H 3.91, N 7.24; found C
72.88, H 3.93, N 7.19.
¾
were performed with the program Kaleidagraph to give K and De. The
experiments were run in duplicate and were reproducible within the range of
2
5%. Analyses carried out at different wavelengths (i.e., 412 and 435 nm)
Monolayer of [Co(3)]: A quartz microbalance (QCM) disk was dipped into
a solution of 4-mercaptophenol (12.6 mg) in dichloromethane (100 mL)
and left standing for 12 h. The disk was then rinsed with CH2Cl2 and dipped
gave results in good agreement within 15%.
5-(4'-Carboxyphenyl)-10,15,20-triphenylporphyrin (2): A solution of pyr-
role (14.5 g, 0.21 mol) in acetic acid (100 mL) was added dropwise to a
solution of 4-carboxybenzaldehyde (8.1 g, 0.054 mol) and benzaldehyde
(20 g, 0.19 mol) in refluxing acetic acid (250 mL), over a period of
30 minutes. The reaction mixture was stirred at reflux temperature for an
additional 3 h, then cooled to room temperature and left standing
overnight. The crystalline residue, separated from the bulk solution, was
collected by filtration, washed with methanol, and then purified by
chromatography on SiO2. The first elution with CHCl3 gave the
5,10,15,20-tetraphenylporphyrin (H2TPP) co-product as a purple band,
while the desired porphyrin was subsequently eluted with a 2% methanol/
chloroform mixture, to give 2 (0.5 g, 0.76 mmol, 1.4% yield) as a purple
crystalline solid. Another crop of product was obtained from the mother
liquor as follows: the solvent was evaporated under reduced pressure and
the black tarry residue dissolved in 200 mL of CHCl3. A saturated solution
of aqueous NaHCO3 was added to the mixture that was left stirring until
the evolution of gas ceased. The separated organic layer was washed with a
10% aqueous Na2CO3 solution (3 Â 200 mL) and then with brine until
neutrality. To remove polymeric byproducts, silica gel (200 g) was added to
the solution, and the resulting mixture stirred for 30 minutes. The solid
mass was washed with CHCl3 (1% MeOH) until a UV/vis check run
showed the disappearance of the typical porphyrin absorption band (l ˆ
419 nm) of the product 2 in the eluate. The washings were collected,
in a 10À3
m (0.07 mmol, 0.05 g in 70 mL of CH2Cl2) solution of the
corresponding acyl chloride of 2 in CH2Cl2. After standing for 1 day, the
quartz disk was washed with CH2Cl2, sonicated in CH3OH, washed again
with CH2Cl2, and dried under N2.
Monolayer of [Co2(1)] (Method 1 or one-step method): The QCM disk was
dipped into a 10À3 m CH2Cl2 solution of [Co2(1)] (0.046 g, 0.03 mmol,
30 mL) for 12 h and then washed with CH2Cl2, sonicated in CH3OH,
washed again with CH2Cl2, and dried under a stream of N2.
Monolayer of [Co2(1)] (Method 2 or two-step method): The quartz disk
was dipped into a 10À3 m CH2Cl2 solution of trans-1,2-dithiane-4,5-diol
(7 mg, 0.05 mmol, 50 mL) for 12 h before being washed with CH2Cl2 and
again dipped into a 10À3 m solution of the corresponding acyl chloride of 2
for 12 h. The QCM disk was finally cleaned as reported above.
Sensor measurements: QCM sensors were AT-cut quartz crystals resonat-
ing at a fundamental frequency of 10 MHz. Frequency readout and
electronic conditioning were performed with an electronic oscillating
circuit, whose frequency was measured by
a frequency counter
(HP53131A). The frequency counter was connected, through a GP-IB
link, with a PC that was set to supervising the measurement and collecting
the sensor responses.
Eight sensors were housed in a 10 mL measurement chamber. In the case of
triethylamine, saturated vapours of the analyte were obtained by bubbling
N2 into liquid amine; these vapours were diluted to different concentrations
by a mass flow controller, with dry synthetic air as the carrier gas, and
allowed to flow into the measurement chamber. The measurements were
carried at a constant temperature (308 K), controlled by a Julabo F40
thermostat. Vapour-phase concentrations of the analyte were calculated by
the Antoine equation.[30]
reduced to
a small volume by rotary evaporation, and purified by
chromatography on SiO2, first by eluting with CHCl3 to recover H2TPP
and then with a 2% methanol/chloroform mixture to obtain the title
porphyrin 2 (3 g, 4.5 mmol, 9% total yield). Spectroscopic characterization
was in perfect agreement with data reported in the literature.[29]
4,5-Bis[5-(4-carboxyphenyl)-10,15,20-triphenylporphyrinyl]-trans-1,2-di-
thiane (1): All manipulations were carried out under an inert atmosphere
by using standard Schlenk techniques. Porphyrin 2 (50 mg, 0.076 mmol) was
dissolved in dry and distilled CH2Cl2 (20 mL) in a 100 mL two-necked
round-bottomed flask. An excess of a 2m solution of oxalyl chloride (3 mL,
In the case of chiral compounds a different experimental protocol had to be
followed, owing to both the lower amount and the low volatility of these
analytes. In this case a sealed vial containing the analyte was opened into a
2482
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0811-2482 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 11

Chemical Sensors
2476 2483
2 L flask held at the constant temperature of 308 K. No significant
variations of the headspace composition were observed after 1 h, as
checked by GC analysis. Vapour-phase concentrations, calculated by the
Antoine equation, were in good agreement with those measured by GC
(internal standard dodecane). The headspace was then pumped into the
measurement chamber by a micro pump at a constant flow rate of
0.2 LminÀ1. The equilibrium response time for the sensors was in the order
of 1 minute; after each measurement the chamber was purged by a flow of
dry synthetic air. Typical sensor response is shown in Figure 6.
Sakata, Langmuir 1998, 14, 5335; g) T. A. Postlehwaite, J. E. Hutch-
ison, K. W. Hatchock, R. W. Murray, Langmuir 1995, 11, 4109; h) J. E.
Hutchison, T. A. Postlehwaite, C.-H. Chen, K. W. Hatchock, R. S.
Ingram, W. Ou, R. W. Linton, R. W. Murray, Langmuir 1997, 13, 2143;
i) D. L. Pilloud, C. C. Moser, K. S. Reddy, P. L. Dutton, Langmuir
1998, 14, 4809; l) D. A. Offord, S. B. Sachs, M. S. Ennis, T. A.
Eberspacher, J. H. Griffin, C. E. D. Chidsey, J. P. Collman, J. Am.
Chem. Soc. 1998, 120, 4478; m) D. T. Gryko, C. Clausen, J. S. Lindsey,
J. Org. Chem. 1999, 64, 8635; n) Z. Zhang, R. Hu, Z. Liu, Langmuir
2000, 16, 1158; o) Z. Zhang, S. Hou, Z. Zhu, Z. Liu, Langmuir 2000,
16, 537; p) N. Kanayama, T. Kanbara, H. Kitano, J. Phys. Chem. B
2000, 104, 271.
[8] R. Paolesse, C. Di Natale, A. Macagnano, F. Davide, T. Boschi, A.
D×Amico, Sens. Actuators B 1998, 47, 70.
[9] J. D. Rawn, Biochemistry, Patterson, Burlington, 1989.
[10] a) J.-M. Lehn, Supramolecular Chemistry, Wiley-VCH, Weinheim,
1995; b) Comprehensive Supramolecular Chemistry Vol. 2 (Eds.: J.-M.
Lehn, J. L. Atwood, J. E. D. Davies, D. D. MacNicol, F. Vˆgtle),
Pergamon, Oxford, 1996.
[11] a) P. Schreier, A. Bernreuther, M. Huffer, Analysis of Chiral Organic
Molecules: Methodology and Applications, de Gruyter, Berlin, 1995,
p. 371; b) J. Snopek, E. Smolkova-Keulemansova, T. Cserhati, K.
Gahm, A. Stalcup in Comprehensive Supramolecular Chemistry, Vol. 3
(Eds.: J.-M. Lehn, J. L. Atwood, J. E. D. Davies, D. D. MacNicol, F.
Vˆgtle, T. Szejtli, T. Osa), Pergamon, New York, 1996, p. 515.
[12] a) J. Ide, T. Nakamoto, T. Moriizumi, Sens. Actuators A 1995, 49, 73;
b) K. Bodenhˆfer, A. Hierlmann, J. Seemann, G. Gauglitz, B.
Koppenhoefer, W. Gˆpel, Nature 1997, 387, 577; c) K. Bodenhˆfer,
A. Hierlmann, J. Seemann, G. Gauglitz, W. Gˆpel, Anal. Chem. 1997,
69, 3058; d) K. Bodenhˆfer, A. Hierlmann, M. Juza, V. Schurig, W.
Gˆpel, Anal. Chem. 1997, 69, 4017; e) I. P. May, M. P. Byfield, M.
Lindstrˆm, L. F. W¸nsche, Chirality 1997, 9, 225.
[13] C. Di Natale, A. Macagnano, F. Davide, A. D×Amico, R. Paolesse, T.
Boschi, M. Faccio, G. Ferri, Sens. Actuators B 1997, 44, 521.
[14] K. C. Persaud, G. Dodd, Nature 1982, 299, 352.
[15] Sensors andSensory Systems for an Electronic Nose (Eds.: J. W.
Gardner, P. N. Bartlett), Kluwer Academic, Dordrecht (NL), 1992.
[16] W. Gˆpel, Sens. Actuators B 1998, 52, 125.
Figure 6. Typical frequency changes of a [Co2(1)]-modified QCM sensor
upon exposure to VOCs.
Acknowledgement
This work was supported by CNR MADESS II project 00.00781PF48.
[17] A. D×Amico, C. Di Natale, R. Paolesse, Sens. Actuators B 2000, 68, 324.
[18] J. W. Buchler in The Porphyrins (Ed.: D. Dolphin), Academic Press,
New York, 1978, p. 389, and references therein.
[1] a) A. Ulman, An Introduction to Ultrathin Organic Films: From
Langmuir Blodgett to Self-Assembly, Academic Press, San Diego,
1991; b) A. Ulman, Chem. Rev. 1996, 96, 1533.
¬
[19] For some recent reports on this topic see: a) J. M. Ribo, J. M. Bofill, J.
[2] a) J. H. Fendler, Chem. Mater. 1996, 8, 1616; b) A. Kumar, G. M.
Whitesides, Science 1994, 263, 60; c) Nanoparticles andNanostruc-
turedFilms (Ed.: J. H. Fendler), Wiley-VCH, Weinheim, 1998.
[3] W. Gopel, K. D. Schierbaum in Sensors, Vol. 2 (Eds.: W. Gopel, T. A.
Jones, M. Kleitz, J. Lundstrom, T. Seiyama), Wiley-VCH, Weinheim,
1991, p. 1.
Crusats, R. Rubires, Chem. Eur. J. 2001, 7, 2733; b) R. Rubires, J.-A.
¬
Ferrera, J. M. Ribo, Chem. Eur. J. 2001, 7, 436.
[20] a) C. A. Hunter, J. K. M. Sanders, J. Am. Chem. Soc. 1990, 112, 5525;
b) H.-J. Schneider, M. Wang, J. Org. Chem. 1994, 59, 7464.
[21] Although an excellent adherence of the experimental points to a 1:1
binding curve is evident, the relative values could probably be the
result of a series of consecutive conformational and binding equilibria.
Detailed studies, which are beyond the immediate scope of this work,
are in progress, and the results will be reported elsewhere.
[22] For a recent overview on the chiral recognition of porphyrin derivatives
see: a) H. Ogoshi and T. Mizutani, Acc. Chem. Res. 1998, 31, 81; b) X.
Huang, K. Nakanishi, N. Berova, Chirality 2000, 12, 237.
[23] F. Sauerbrey, Zeit. Physik. 1959, 155, 206.
[24] A. Sgarlata, A. Angelaccio, N. Motta, R. Paolesse, C. Di Natale, A.
D×Amico, Surf. Sci. 2000, 466, 167.
[25] R. F. Pasternack, P. R. Huber, P. Boyd, G. Engasser, L. Francesconi, P.
Fasella, G. Cerio Venturo, L. deC. Hinds, J. Am. Chem. Soc. 1972, 94,
4511.
[26] M. Nappa, J. S. Valentine, J. Am. Chem. Soc. 1978, 100, 5075.
[27] S. Brunauer, P. H. Emmet, E. Teller, J. Am. Chem. Soc. 1938, 60, 309.
[28] K. A. Connors, Binding Constants–The Measurements of Molecular
Complex Stability, Wiley, New York, 1987.
[4] a) J. A. J. Brunink, C. Di Natale, F. Bungaro, F. A. M. Davide, A.
D×Amico, R. Paolesse, T. Boschi, M. Faccio, G. Ferri, Anal. Chim. Acta
1996, 325, 53; b) C. Di Natale, A. Macagnano, G. Repole, A. D×Amico,
R. Paolesse, T. Boschi, Mater. Sci. Eng. C 1998, 5, 209; c) C. Di Natale,
R. Paolesse, A. Macagnano, V. I. Troitsky, T. S. Berzina, A. D×Amico,
Anal. Chim. Acta 1999, 384, 249; d) C. Di Natale, R. Paolesse, A.
Macagnano, A. Mantini, C. Goletti, A. D×Amico, Sens. Actuators B
1998, 52, 162; e) A. D×Amico, C. Di Natale, R. Paolesse, A.
Macagnano, A. Mantini, Sens. Actuators B 2000, 65, 209; f) C.
Di Natale, R. Paolesse, A. Macagnano, A. Mantini, P. Mari, A.
D×Amico, Sens. Actuators B 2000, 68, 319; g) M. Andersson, M.
Holmberg, I. Lundstrˆm, A. Lloyd-Spetz, P. Martensson, R. Paolesse,
C. Falconi, C. Di Natale, A. D×Amico, Sens. Actuators B, in press.
[5] A. D×Amico, C. Di Natale, A. Macagnano, F. Davide, A. Mantini, E.
Tarizzo, R. Paolesse, T. Boschi, Biosens. Bioelectron. 1998, 13, 711.
[6] The Porphyrin Handbook (Eds.: K. M. Kadish, K. M. Smith, R.
Guilard), Academic, 2000.
[7] a) J. E. Hutchison, T. A. Postlehwaite, R. W. Murray, Langmuir 1993,
9, 3277; b) J. Zak, H. Yuan, M. Ho, L. K. Woo, M. D. Porter, Langmuir
1993, 9, 2772; c) T. Akiyama, H. Imahori, Y. Sakata, Chem. Lett. 1994,
1447; d) L.-H. Guo, G. Mclendon, H. Razafitrimo, Y. L. Gao, J. Mater.
Chem. 1996, 6, 369; e) W. Han, S. Li, S. M. Lindsay, D. Gust, T. A.
Moore, A. L. Moore, Langmuir 1996, 12, 5742; f) H. Imahori, H.
Norieda, S. Ozawa, K. Ushida, H. Yamada, T. Azuma, K. Tamaki, Y.
[29] a) S. Matile, N. Berova, K. Nakanishi, S. Novkova, I. Philipova, B.
Blagoev, J. Am. Chem. Soc. 1995, 117, 7021; b) C. E. Kibbey, M. E.
Meyerhoff, Anal. Chem. 1993, 65, 2189.
[30] J. Riddick, A. Bunger in Techniques of Chemistry, Vol. II (Ed.: A.
Weissberger), Wiley Interscience, New York, 1970.
Received: June 15, 2001
Revised: January 14, 2002 [F3339]
Chem. Eur. J. 2002, 8, No. 11
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0811-2483 $ 20.00+.50/0
2483