A novel approach to the molecular imprinting of polychlorinated aromatic compounds

Article abstract of DOI:10.1021/ja9818295

The aim of this investigation was to determine whether relatively weak interactions, such as hydrogen bonds to aromatic chlorine atoms and interactions involving aromatic π electrons could be exploited within artificial receptors, constructed using the te

Full text of DOI:10.1021/ja9818295

13342
J. Am. Chem. Soc. 1998, 120, 13342-13348
A Novel Approach to the Molecular Imprinting of Polychlorinated
Aromatic Compounds
Markus Lu1bke, Michael J. Whitcombe,* and Evgeny N. Vulfson
Contribution from the FMS Department, Institute of Food Research, Earley Gate, Whiteknights Road,
Reading, Berkshire RG6 6BZ, UK
ReceiVed May 26, 1998
Abstract: The aim of this investigation was to determine whether relatively weak interactions, such as hydrogen
bonds to aromatic chlorine atoms and interactions involving aromatic π electrons could be exploited within
artificial receptors, constructed using the technique of molecular imprinting. For the purposes of this investigation
we chose 2,3,7,8-tetrachlorodibenzodioxin (TCDD) as the model target. Imprinted polymers have been prepared
with two new templates designed to create recognition sites for TCDD. The first of these, the bis-N-(4-
vinylphenyl)urea derivative of 2,8-dichloro-3,7-diaminodibenzodioxin, employed a carbonyl spacer to introduce
aromatic amines into the polymer after reductive cleavage of the template. The second, N-(2-(3,7,8-
trichlorodibenzodioxinyl))-2-methacryloyloxybenzamide, incorporated a salicylic acid spacer and introduced
a methacrylic acid residue into the polymer following hydrolysis. Both amine and acid groups were positioned
in such a way as to interact with TCDD through the formation of weak hydrogen bonds to aromatic chlorine
atoms. A second recognition element was introduced into the binding sites by the inclusion of a polymerizable,
electron-rich, aromatic ether capable of forming π-π interactions with the electron-deficient dioxin molecule.
Polymers imprinted with either template showed significantly higher uptake of TCDD than the corresponding
nonimprinted controls, even at concentrations as low as 2 nM.
Introduction
tion,5 removal of undesired components from complex mix-
tures,6 and separation or concentration of proteins7 and whole
cells8 have also been described.
Molecular imprinting has emerged as a powerful technique
for the creation of recognition elements in highly cross-linked
polymeric matrices.1 Generally this methodology is based on
the simple but elegant principle of using the functionality of a
target molecule (template) to assemble its own recognition site
by forming interactions with “complementary” functional groups
of appropriate polymerizable monomers. These interactions are
then “frozen in” by polymerization, carried out in the presence
of a high concentration of cross-linker. Subsequent removal of
the template creates the polymer’s binding site with the precise
spatial arrangement of functional groups to ensure highly
selective recognition of the target molecule. Most of the recent
work in this area has focused on the preparation of imprinted
polymers for chiral resolutions,2 but numerous other applications
in the design of catalysts3 and sensors,4 in solid-phase extrac-
The majority of polymers for these applications have been
prepared using the noncovalent imprinting methodology where
the template-monomer complex is formed in situ by association
of the target molecule with relatively simple functional mono-
mers such as methacrylic acid, acrylamide,9 vinylpyridines,10
and others11 present in the polymerization mixture. However,
to achieve efficient complexation before and during polymer-
ization, the compound of interest must contain suitable func-
tionality to provide sufficiently strong interactions with the
constituent monomers.12 As a result highly functionalized
molecules, typically containing multiple carboxy, amino, amido,
and keto groups, remain the most popular targets for imprinting.
(5) (a) Sellergren, B. Trends Anal. Chem. 1997, 16, 310-320. (b) Matsui,
J.; Okada, M.; Tsuruoka, M.; Takeuchi, T. Anal. Commun. 1997, 34, 85-
87. (c) Muldoon, M. T.; Stanker, L. H. ACS Symp. Ser. 1997, 657, 314-
330. (d) Andersson, L. I.; Paprica, A.; Arvidsson, T. Chromatographia 1997,
46, 57-62.
(6) Whitcombe, M. J.; Alexander, C.; Vulfson, E. N. Trends Food Sci.
Technol. 1997, 8, 140-145.
(7) (a) Mallik, S.; Plunkett, S. D.; Dhal, P. K.; Johnson, R. D.; Pack, D.;
Shnek, D.; Arnold, F. H. New J. Chem. 1994, 18, 299-304. (b) Hjerten,
S.; Liao, J. L.; Nakazato, K.; Wang, Y.; Zamaratskaia, G.; Zhang, H. X.
Chromatographia 1997, 44, 227-234.
(8) (a) Aherne, A.; Alexander, C.; Payne, M. J.; Perez, N.; Vulfson, E.
N. J. Am. Chem. Soc. 1996, 118, 8771-8772. (b) Alexander, C.; Vulfson,
E. N. AdV. Mater. 1997, 9, 751-755.
(9) (a) Yu, C.; Mosbach, K. J. Org. Chem. 1997, 62, 4057-4064. (b)
Yu, C.; Ramstro¨m, O.; Mosbach, K. Anal. Lett. 1997, 30, 2123-2140.
(10) Kempe, M.; Fischer, L.; Mosbach, K. J. Mol. Recog. 1993, 6, 25-
29.
(11) (a) Tanabe, K.; Takeuchi, T.; Matsui, J.; Ikebukuro, K.; Yano, K.;
Karube, I. J. Chem. Soc., Chem. Commun. 1995, 2303-2304. (b) Matsui,
J.; Doblhoff-Dier, O.; Takeuchi, T. Anal. Chim. Acta 1997, 343, 1-4. (c)
Yano, K.; Nakagiri, T.; Takeuchi, T.; Matsui, J.; Ikebukuro, K.; Karube, I.
Anal. Chim. Acta 1997, 357, 91-98.
* Author to whom correspondence should be sent. E-mail:
michael.whitcombe@bbsrc.ac.uk.
(1) (a) Wulff, G. Angew. Chem., Int. Ed. Engl. 1995, 34, 1812-1832.
(b) Shea, K. J. Trends Polym. Sci. 1994, 2, 166-173. (c) Mayes, A. G.;
Mosbach, K. Trends Anal. Chem. 1997, 16, 321-332. (d) Vulfson, E. N.;
Alexander, C.; Whitcombe, M. J. Chem. Br. 1997, 33, 23-26.
(2) (a) Kempe, M.; Mosbach, K. J. Chromatogr. A 1995, 694, 3-13.
(b) Kempe, M.; Mosbach, K. J. Chromatogr. A 1995, 691, 317-323. (c)
Nicholls, I. A.; Andersson, L. I.; Mosbach, K.; Ekberg, B. Trends Biotechnol.
1995, 13, 47-51.
(3) (a) Mu¨ller, R.; Andersson, L. I.; Mosbach, K. Makromol. Chem. Rapid
Commun. 1993, 14, 637-641. (b) Beach, J. V.; Shea, K. J. J. Am. Chem.
Soc. 1994, 116, 379-380. (c) Gamez, P.; Dunjic, B.; Pinel, C.; Lemaire,
M. Tetrahedron Lett. 1995, 36, 8779-8782. (d) Wulff, G.; Gross, T.;
Scho¨nfeld, R. Angew. Chem., Int. Ed. Engl. 1997, 36, 1962-1964. (e) Davis,
M. E.; Katz, A.; Ahmad, W. R. Chem. Mater. 1996, 8, 1820-1839.
(4) (a) Piletsky, S. A.; Parhometz, Y. P.; Lavryk, N. V.; Panasyuk, T.
L.; El′Skaya, A. V. Sens. Actuators, B 1994, 18-19, 629-631. (b) Chen,
G. H.; Guan, Z. B.; Chen, C. T.; Fu, L. T.; Sundaresan, V.; Arnold, F. H.
Nature Biotechnol. 1997, 15, 354-357. (c) Kriz, D.; Ramstro¨m, O.;
Mosbach, K. Anal. Chem. 1997, 69, 9, A345-A349.
10.1021/ja9818295 CCC: $15.00 © 1998 American Chemical Society
Published on Web 12/30/1998

Molecular Imprinting of Polychlorinated Aromatic Compounds
J. Am. Chem. Soc., Vol. 120, No. 51, 1998 13343
Clearly the scope of this methodology would be widened
significantly if it was possible to exploit other chemical features
of the template molecule such as the ability to form very weak
hydrogen bonds, which are unlikely to be formed spontaneously
by self-association with conventional monomers, and/or to
engage in interactions involving aromatic π electrons. To
achieve this, novel functional monomers have to be designed
and an appropriate method for their introduction into the
polymer’s recognition sites developed. The main objective of
this investigation was to demonstrate the feasibility of the
preparation of imprinted polymers, utilizing recognition elements
based on such interactions.
Polychlorinated aromatic compounds would be particularly
attractive models for the purpose of this investigation on account
of two structural features, namely the presence of aromatic
chlorine atoms which are known to form exceedingly weak
hydrogen bonds in nonpolar solvents13 and an aromatic π
electron system which can interact with that of other aromatic
molecules.14 In addition, many polychlorinated aromatic com-
pounds (PCDDs, PCDFs, PCBs) are presently of great envi-
ronmental concern, and the development of a generic approach
to the preparation of polymers specific to these molecules would
be of considerable practical value. In particular, 2,3,7,8-
tetrachlorodibenzodioxin (TCDD) was chosen as a representative
target compound, because it is the most prominent member of
a large family of polychlorinated dibenzodioxins which are
currently the subject of extensive monitoring15 due to their
extraordinary toxicity toward mammals.16
Figure 1. Schematic representation of the proposed binding site of
polymers prepared with template 2. Urea-derived amine groups are
shown interacting with the aromatic chlorine atoms of TCDD. The
position of the electron-rich or electron-deficient aromatic comonomer
in the site is also shown along with the structures of pentafluorostyrene
(PFS), 1,4-bis(3/4-vinylbenzyloxy)benzene (3), and 1-methoxy-3,5-bis-
(4-vinylbenzyloxy)benzene (5).
Results and Discussion
To achieve specific recognition of TCDD both the aromatic
π electron system and the chlorine atoms must be engaged in
interactions with functional groups forming the binding site of
an imprinted polymer. This posed two immediate questions:
what functionality would be most suitable for the recognition
and how might it be introduced into the binding sites? As
hydrogen-bonded complexes between aromatic chlorine atoms
of TCDD and conventional monomers acting as hydrogen bond
donors in noncovalent imprinting (e.g., methacrylic acid,
acrylamide) were probably too weak to enable efficient com-
plexation of the template in the polymerization mixture, we
employed a modification of the “sacrificial spacer” methodology
recently developed in our laboratory.17 For this approach to work
a TCDD analogue was required, covalently linked to polymer-
izable monomer(s) such that hydrogen bond donors would be
generated in the recognition site by the chemical treatment used
to remove the template. To satisfy these requirements, 2,8-
diamino-3,7-dichlorodibenzodioxin (1) was prepared as de-
scribed in the Experimental Section and reacted with 2 mol
equiv of 4-vinylphenylisocyanate to give the diurea template
2, where the two urea bridges were designed to act as sacrificial
spacers (Scheme 1A). Hydride reduction of the template after
its incorporation into the polymer would release 1, leaving two
aromatic amines in the recognition site, positioned in such a
way as to form hydrogen bonds with the chlorine atoms of
TCDD subsequently bound to the polymer (Scheme 1B).
The aromatic π electron system of TCDD was targeted
through noncovalent interactions. To this end, several aromatic
comonomers, capable of forming π-π interactions with the
template at the stage of polymerization, were prepared and
tested. The electron-deficient 2,3,4,5,6-pentafluorostyrene (PFS)
and the electron rich 1,4-bis(3/4-vinylbenzoxy)benzene (3) were
used in the initial experiments. As a consequence of the template
structure and the expected π-π complex between the diurea
template and the monomer, the designed recognition site would
be “equipped” with two aromatic amino groups situated on both
sides of the cavity, coplanar with the bound TCDD and the
charge transfer monomer forming the “floor” of the imprint, as
illustrated schematically in Figure 1. Although the individual
interactions between these functional groups and TCDD are
weak, in combination they should lead to the formation of
specific recognition sites.
(12) This is not crucial for covalent imprinting,1a,b but the arsenal of
suitable monomers is currently limited since relatively few functional groups
are capable of forming reversible covalent bonds, e.g., boronate esters, Schiff
bases, or ketals.
(13) Smith, J. W. In The Chemistry of the Carbon-Halogen Bond; Patai,
S., Ed.; Wiley-Interscience: London, 1973; Chapter 5.
(14) (a) Pyell, U.; Garrigues, P.; Fe´lix, G.; Rayez, M. T.; Thienpont, A.;
Dentraygues, P. J. Chromatogr. 1993, 643, 169-181. (b) Kimata, K.;
Hosoya, K.; Araki, T.; Tanaka, N.; Barnhart, E. R.; Alexander, L. R.;
Sirimanne, S.; McClure, P. C.; Grainger, J.; Patterson, D. G. Anal. Chem.
1993, 65, 2502-2509.
(15) A highly sensitive aqueous antibody-based assay (ELISA) for TCDD
has recently been developed.See: (a) Sanborn, J. R.; Gee, S. J.; Gilman, S.
D.; Sugawara, Y.; Jones, A. D.; Rogers, J.; Szurdoki, F.; Stanker, L. H.;
Stoutamire, D. W.; Hammock, B. D. J. Agric. Food Chem. 1998, 46, 2407-
2416. (b) Sugawara, Y.; Gee, S. J.; Sanborn, J. R.; Gilman, S. D.; Hammock,
B. D. Anal. Chem. 1998, 70, 1092-1099.) However, TCDD and its
analogues are very poorly soluble in water, hence such an assay may have
limited applicability. There is therefore still a need to develop materials
specific to dioxins which are capable of working in organic solvents.
(16) Safe, S. H. Annu. ReV. Pharmacol. Toxicol. 1986, 26, 371-399.
(17) Whitcombe, M. J.; Rodriguez, M. E.; Villar, P.; Vulfson, E. N. J.
Am. Chem. Soc. 1995, 117, 7105-7111.
All polymers were prepared in bulk, using azobisisobuty-
ronitrile (AIBN) as radical initiator and divinylbenzene (DVB)
as cross-linking monomer (Table 1). DVB was used in prefer-
ence to other commonly employed cross-linkers such as
ethyleneglycol dimethacrylate because its polymers would be

13344 J. Am. Chem. Soc., Vol. 120, No. 51, 1998
Lu¨bke et al.
Scheme 1. (A) Synthesis of the Diurea Template 2 and (B) Schematic Representation of the Preparation of Polymers Imprinted
with Template 2 Showing the Positioning of Aromatic Amine Groups in the Recognition Site via the Urea Functionality
(incorporating a carbonyl spacer)
Table 1. Binding of TCDD (2 nM in Nonane) to Polymers Imprinted with Template 2 and Their Nonimprinted Counterpartsa
cross-linker,
DVB [mol %]
template (or functional monomer)
[mol %]
co-monomer
[mol %]
template
removal [%]
specific surface
area [m2/g]
bound TCDD
[pmol/g]
polymer
P1
P2
P3
P4
P5
P6
99.5
94.5
94.5
99.5
94.5
94.5
AS, 0.5
AS, 0.5
AS, 0.5
2, 0.5
2, 0.5
2, 0.5
none
PFS, 5
3, 5
none
PFS, 5
3, 5
149
106
242
149
143
278
5.72 ( 0.38
5.22 ( 0.48
7.50 ( 0.95
6.55 ( 0.24
6.37 ( 0.46
9.30 ( 0.82
44
26
43
a All polymers were prepared with DMSO as the porogenic solvent.
more likely to withstand the drastic conditions necessary for
template removal. DMSO was used as the porogen since it was
the only solvent among those tested to dissolve 2 sufficiently
well. Nonimprinted polymers P1-P3 were prepared in exactly
the same way as 2-imprinted polymers P4-P6, except that the
template was replaced by 4-aminostyrene (AS). The amount of
AS was chosen such that the density of amino groups incor-
porated into the polymer would be similar to that of the
imprinted polymers P4-P6 after removal of the template. Thus,
the chemical compositions of the hydrolyzed imprinted and
control polymers were essentially the same, the only difference
being the distribution of the recognition elements in the polymer
matrix (preassembled versus random, respectively). Hence an
increase in the binding of TCDD, if observed, would be due to
imprinting, as opposed to nonspecific adsorption to the polymer
surface. Both sets of polymers were subjected to the same
template removal procedure, which involved treatment with
LiAlH4 in refluxing THF.18 Nonimprinted polymers were
included to account for any effects the template removal
procedure might have on the polymer matrix itself. The template
removal in polymers imprinted with 2 was quantified by
determination of the chlorine content of the polymers before
and after the reduction and was typically about 50%.19
The binding of TCDD to polymers was assessed in batch
experiments, where 10 mg of polymer were incubated for 24 h
with 0.5 mL of anhydrous nonane containing 1 pmol (10-12
mol) of [1,6-3H2]-TCDD,20 after which time the remaining
activity in the filtrate was measured by scintillation counting.
All uptake measurements were done in 5 replicates. The amounts
of TCDD bound per gram of polymer for P1-P6 are presented
in Table 1, along with the specific surface areas of each polymer.
Entries for polymers P1-P3 show the extent of nonspecific
binding, i.e., binding that occurs “outside” an imprinted recogni-
tion site. Binding to P2 was considerably lower than that to
P3, which reflects the higher electron-donating strength of 3
over PFS. P1, prepared in the absence of comonomer, shows
similar binding to P2. These results are suggestive of the
formation of π-π interactions between TCDD and elements
of the polymer surface, with TCDD acting as an electron
(19) As the degree of cleavage was around 50% for both templates,
nonimprinted polymers were prepared with either 0.5 mol % 4-aminostyrene
for comparison with template 2 or 0.25 mol % methacrylic acid in the case
of 4 (see Tables 1 and 2).
(20) The concentration of TCDD used in uptake experiments (2 nM)
was chosen to represent that typically encountered in the analysis of extracts
from food and environmental samples.
(18) Hydride reduction has previously been used to cleave the template
from DVB-based polymers imprinted with sterol methacrylate esters. See:
Bystro¨m, S. E.; Bo¨rje, A.; Akermark, B. J. Am. Chem. Soc. 1993, 115,
2081-2083.

Molecular Imprinting of Polychlorinated Aromatic Compounds
J. Am. Chem. Soc., Vol. 120, No. 51, 1998 13345
acceptor.21 Increased binding on imprinting, and thus specific
interactions, was observed when comparing the imprinted
polymers P4-P6 with their nonimprinted counterparts P1-P3.
The ratio of binding between imprinted and nonimprinted
polymer was 1.15 in the case of the polymers P4/P1, prepared
without comonomer, 1.22 for the polymers which contained the
electron acceptor PFS (P5/P2) and 1.24 for P6/P3 (which
incorporated 3). This suggests that the effect of the hydrogen-
bonding structures in the recognition sites is similar irrespective
of the nature of the comonomer. The overall binding was highest
in the case of P6.
Scheme 2. (A) Synthesis of the Amide Template 4 and (B)
Schematic Representation of the Preparation of Polymers
Imprinted with Template 4, Incorporating a Salicylic Acid
“Smart Spacer”. Intramolecular Hydrogen Bonding Assists in
the Prearrangement of Functional Groups Which Will
Interact on Rebinding
To assess further the contribution of weak hydrogen bonding
to the recognition of TCDD, we designed and prepared another
template, amide 4, by reacting 2-amino-3,7,8-trichlorodibenzo-
dioxin22 with 2-methacryloyloxybenzoyl chloride (Scheme 2A).
In this template, salicylic acid fulfils the role of “sacrificial
spacer”, maintaining the spatial relationship between the meth-
acrylate oxygen atom and the amide template’s -NH- group
by means of highly favored intramolecular hydrogen bonding.
Evidence for this conformation comes from NMR and infrared
studies on a model compound.23 The 2-methacryloyloxyben-
zamide group is the first example of a “smart sacrificial spacer”
which, in effect, incorporates a captive noncovalent imprint.
This approach should be especially suited to the imprinting of
templates which contain primary amino groups, which will be
discussed in a later paper.24 After hydrolysis of the template,
the methacrylate ester-derived carboxylic acid in the recognition
site should be well situated to interact with a chlorine atom of
a bound TCDD molecule (Scheme 2B). The imprinted site
produced by using this template should differ in a number of
key aspects from those arising from template 2: The sites
contain one carboxyl rather than two amino groups; there will
be an additional void associated with the site due to the presence
of a larger spacer group. In addition, the trichloro-template 4,
being more electron deficient than the dichloro-derivative 2,
should form stronger π-π interactions which should increase
the likelihood of incorporating monomers such as 3 into the
recognition site.
As well as preparing polymers imprinted with template 4,
we also investigated what effect the inclusion of a new
comonomer with stronger electron-donating properties would
have on TCDD binding. As increasing the number of alkoxy
groups on the central benzene ring from two to three was
expected to yield a better electron donor than 3, 1-methoxy-
3,5-bis(4-vinylbenzyloxy)benzene (5) was synthesized and
compared to the other comonomers. The results for nonimprinted
polymers (Table 2) again showed that the incorporation of PFS
(21) It was not possible to predict the strength of association between
the dioxin templates and the charge-transfer monomers due to the lack of
published data on such systems. However, ref 14a attributes retention of
TCDD on a 1-pyreneethyl bonded silica stationary phase in HPLC to the
formation of charge-transfer complexes but no such retention occurred with
electron-deficient bonded phases. This reflects the relative binding strengths
of polymers incorporating PFS or aromatic ether comonomers and confirms
that TCDD is acting as a π-electron acceptor. In principle it could have
been possible to calculate Ka for the pyrene-TCDD couple from the
published capacity factors if some additional parameters (column void
volume and pyrene content) were provided by the authors.
gave the lowest TCDD uptake, lower even than that for
polymers prepared without comonomer. On the other hand,
copolymerization with 3, and especially 5, significantly in-
creased TCDD binding from 7.62 to 8.70 and 10.78 pmol/g,
respectively (P7-P11). These results confirmed that TCDD
binding is indeed positively correlated with the electron-donating
property of the comonomer. Consequently, 5 was used in
preference to 3 for the preparation of polymers imprinted with
template 4. Template removal was accomplished by refluxing
the polymers in methanolic KOH, and the resulting polymer-
bound potassium carboxylate functions were converted into the
protonated from by washing with dilute sulfuric acid. The extent
of template removal in the imprinted polymers P12-P15 was
again determined by chlorine microanalysis and was found to
(22) Kennel, S. J.; Jason, C.; Albro, P. W.; Mason, G.; Safe, S. H. Toxicol.
Appl. Pharmacol. 1986, 82, 256-263.
(23) Infrared data on N-(4-hexadecylphenyl)-2-methacryloyloxybenza-
mide in cyclohexane solution shows no shift for the N-H band at 3434
cm-1 in a dilution series from 10 to 0.3 mM, suggesting the absence of
intermolecular hydrogen bonding. The 1H NMR signal for the amide proton
(10 mM solution in d12-cyclohexane) shows a temperature-dependent shift
from 8.03 (20 °C) to 7.87 ppm (60 °C), resulting from increasing dissociation
of the intramolecular hydrogen bond.
(24) Klein, J.-U.; Whitcombe, M. J.; Mulholland, F.; Vulfson, E. N. In
preparation.

13346 J. Am. Chem. Soc., Vol. 120, No. 51, 1998
Lu¨bke et al.
Table 2. Binding of TCDD (2 nM in Nonane) to Polymers Imprinted with Template 4 and Their Nonimprinted Counterparts
cross-linker,
DVB [mol %]
template (or functional monomer)
[mol %]
comonomer
[mol %]
template
removal [%]
specific surface
area [m2/g]
bound TCDD
[pmol/g]
polymer
porogen
P7
P8
P9
P10
P11
P12
P13
P14
P15
THF
DMSO
THF
THF
THF
THF
DMSO
THF
99.75
99.75
94.75
94.75
94.75
99.5
99.5
94.5
94.5
MAA, 0.25
MAA, 0.25
MAA, 0.25
MAA, 0.25
MAA, 0.25
4, 0.5
4, 0.5
4, 0.5
4, 0.5
none
none
PFS, 5
3, 5
556
144
478
439
400
497
149
326
404
7.62 ( 0.63
5.23 ( 0.35
6.59 ( 0.49
8.70 ( 0.36
10.78 ( 0.57
9.17 ( 0.46
6.40 ( 0.56
7.61 ( 0.42
11.98 ( 0.32
5, 5
none
none
PFS, 5
5, 5
81
40
ND
62
THF
Table 3. Effect of Solvent on the Binding of TCDD to Polymers
Imprinted with Template 4
the decreased solubility of TCDD, and thus a shift in the
partition coefficient between solvent and polymer matrix in favor
of the latter.
bound TCDD [pmol/g]
polymer
nonane
acetonitrile
methanol
Conclusions
P7
P12
7.62 ( 0.63
9.17 ( 0.46
23.15 ( 0.31
25.81 ( 0.62
64.76 ( 0.39
68.84 ( 0.38
We have demonstrated that even very weak molecular
interactions, such as hydrogen bonds involving aromatic chlorine
atoms, can be successfully exploited in the binding of ligands
to imprinted polymers. To introduce suitable hydrogen bond
donating groups into the polymer recognition sites, two new
“sacrificial spacer” strategies (which use covalent imprinting
techniques to assemble sites which bind the target molecules
in a noncovalent fashion) have been employed in the preparation
of imprinted polymers. The first of these exploited the N-(4-
vinyl)phenyl urea group and used a carbonyl spacer (previously
used in the form of a carbonate ester to imprint cholesterol17).
The second utilized a 2-methacryloyloxybenzamide in which
salicylic acid acted as the spacer. This is the first example of a
new class of “smart spacer” in which the relative positions of
template and polymer functional groups are fixed by the
presence of intramolecular hydrogen bonding in the template
structure. Template removal of around 50% was achieved, either
by reduction in the case of the urea or hydrolysis in the case of
the amide, to produce imprinted sites bearing aromatic amine
or carboxylic acid groups, respectively. Polymers imprinted with
either template showed significantly higher uptake of TCDD
than the corresponding nonimprinted controls, even at concen-
trations as low as 2 nM.25 In addition, the inclusion of electron-
rich aromatic monomers, designed to interact via π-π inter-
actions with electron-deficient aromatic templates, was also
shown to enhance the binding of TCDD to highly cross-linked
polymers. 1-Methoxy-3,5-bis(4-vinylbenzyloxy)benzene (5) was
the most effective of the monomers investigated.
be typically around 50%. As before, nonimprinted polymers
(identical but for the inclusion of methacrylic acid (MAA) in
place of the template19) were hydrolyzed under the same
conditions as the imprinted polymers. Binding of TCDD to
polymers P7-P15 was assessed as described above.
The increased solubility of 4 over 2 allowed us to compare
the properties of polymers prepared in two polymerization
solvents, namely THF and DMSO. The nature of the porogen
has a marked influence on the morphology of the polymer. The
use of THF resulted in very high surface area polymers (326-
556 m2 g-1, Table 2), which showed increased nonspecific
binding over the same polymers prepared in DMSO (106-278
m2 g-1, Tables 1 and 2). Specific binding for this particular
template and cross-linker seems to be largely independent of
the porogen used. The ratio of binding to imprinted over
nonimprinted polymers is virtually identical for P12/P7 (pre-
pared in THF) and P13/P8 (prepared in DMSO). Similar results
were obtained for polymers prepared in the presence of
π-electron-donating comonomers.
As with template 2, polymers imprinted with 4 showed
significantly higher uptake of TCDD compared to their non-
imprinted counterparts. Binding increased by a factor of 1.20,
1.15, and 1.11 for P12/P7, P14/P9, and P15/P11, respectively.
It is interesting to note that the increase in TCDD binding
achieved with template 4 was slightly lower than that observed
with analogous polymers imprinted with 2. This is probably
due to the formation of only one weak hydrogen bond on
binding of TCDD into the recognition site as compared with
two in the latter case. However, the difference is clearly too
small to draw any definite conclusions. In addition, the two sites
were equipped with different functional groups, carboxylic acid
and aromatic amine, respectively. To gain further insight into
the contribution of hydrogen bonding to the overall binding,
the uptake of TCDD by 2-imprinted polymer P12 and the
corresponding control polymer P7 was tested in a number of
solvents. It was expected that replacement of nonane with
solvents capable of competing for hydrogen bonding with TCDD
should suppress the specific component of binding. This indeed
proved to be the case with the P12/P7 binding ratio dropping
from 1.20 in nonane to 1.06 in methanol. A significant reduction,
in relative terms, was also observed in acetonitrile. In both polar
solvents, uptake of the imprinted polymer remained higher by
a small, yet significant amount, compared to the nonimprinted
control. The overall binding of TCDD is much higher in
acetonitrile and methanol than in nonane, presumably due to
Experimental Section
Materials and Methods. 2,3,4,5,6-Pentafluorostyrene, vinylbenzyl
chloride (mixture of 3- and 4-isomers), 4-vinylbenzyl chloride, 3,5-
dihydroxyanisole, ethyl chloroformate, 2,5-dichloronitrobenzene, 2,4,5-
trichloronitrobenzene, hydrazine hydrate, methacrylic anhydride, oxalyl
chloride, 4-aminostyrene, methacrylic acid, divinylbenzene (technical
grade, mixture of meta and para isomers, 55%), and anhydrous nonane
were obtained from Aldrich. 4-Chlorocatechol was obtained from
Fluorochem Ltd. AIBN, which was recrystallized from methanol, and
anhydrous acetonitrile were purchased from Fluka. Anhydrous methanol
was prepared by drying over 3 Å molecular sieve. [1,6-3H2]-2,3,7,8-
Tetrachlorodibenzodioxin (specific activity 24.4 Ci/mmol) was obtained
as a solution of 0.52 mCi/mL in toluene/methanol (96:4) from
Greyhound Chromatography. 4-Vinylbenzoic acid was prepared from
4-bromostyrene according to a standard textbook procedure via the
Grignard reagent. All other solvents and reagents were laboratory
(25) Unfortunately, it proved to be impossible to perform Scatchard
analysis on these polymers with any reasonable degree of accuracy due to
the relatively high nonspecific adsorption of the control materials.

Molecular Imprinting of Polychlorinated Aromatic Compounds
J. Am. Chem. Soc., Vol. 120, No. 51, 1998 13347
reagent grade or better. FT-NMR spectra were obtained on a JEOL
EX-270. GC-MS was performed on a Trio 1-S, direct inlet EI-MS on
a Kratos MS890, and MALDI-TOF-MS on a Kratos Kompact Maldi
2, the sample being dispersed in 5-chloro-2-mercaptobenzothiazole. FT-
IR spectra were recorded on a Perkin-Elmer Series 1600 by diffuse
reflectance from samples dispersed in KBr, or from a liquid film
between KBr disks. Polymers were ground in a Fritsch Pulverizette
grinding mill, equipped with an agate mortar. Polymer suspensions were
filtered through Whatman syringe filters with 0.2 µm PTFE membranes
prior to scintillation counting. Scintillation counting was done on a
Packard Tri-Carb 2700TR, using Zinsser Quicksafe A scintillation
cocktail.
on completion of the reaction the mixture was cooled and filtered. The
filtrate was evaporated to yield crude 2,8-diamino-3,7-dichlorodiben-
zodioxin (1) as a white, readily oxidized solid. Under an atmosphere
of nitrogen, 0.6 mmol of 1 was dissolved in 20 mL of boiling THF.
4-Vinylphenyl isocyanate (1.22 mmol) was added dropwise and
refluxing was continued for 30 min. The mixture was allowed to cool
and filtered. The filtrate was poured into water, and the precipitate was
filtered off to yield 90 mg (26%) of an off-white solid, mp >300 °C
dec. IR (cm-1) 3308 (N-H), 1643, 1593, 1556 (3×NH-CO-NH),
1
1512 (dioxin skeleton); H NMR (d6-DMSO) δ (ppm) 5.12 (d, 2H, J
) 11 Hz, cis-CH2dCH), 5.69 (d, 2H, J ) 18 Hz, trans-CH2dCH),
6.65 (dd, 2H, J(Z) ) 11 Hz, J(E) ) 18 Hz, CH2dCH), 7.18 (s, 2H,
arom CH), 7.38, 7.43 (2×d, 2H, J ) 9 Hz, arom CH), 7.82 (s, 2H,
arom CH); 13C NMR (d6-DMSO) δ (ppm) 106.5, 109.0 (2×t), 112.1
(s), 116.4, (q), 118.1 (t), 124.9 (q), 126.7 (t), 131.2 (q), 136.2 (H2Cd
CH), 139.0, 139.9, 140.7, (3×q) 152.3 (CdO). No molecular ion could
be observed with EI- or FAB-MS; MALDI-ToF-MS showed a
molecular ion at m/z ) 572 (high-resolution mass measurement was
not possible); EI-HRMS on the principal fragments: 1, calcd for
C8H9N+ 119.0735, found 119.0740; 2, calcd for C9H9NO+ 145.0528,
found 145.0517; 3, calcd for C12H8N2O2Cl2+ 281.9963, found 281.9965.
Synthesis of Templates and Monomers. 4-Vinylphenyl isocyanate
was prepared by the Curtius rearrangement from the corresponding acyl
azide. 4-Vinylbenzoic acid (0.085 mmol) was dissolved in 500 mL of
acetone and cooled to 0 °C. Triethylamine (0.1 mol) in 40 mL of
acetone was added, followed by the dropwise addition of a solution of
0.11 mol of ethyl chloroformate in 40 mL of acetone. After the solution
was stirred for 30 min at 0 °C, 0.13 mol of sodium azide in 30 mL of
water was added. After further stirring for 1 h, the mixture was poured
into 400 mL of ice-cold water and extracted with 3 × 70 mL of cold
toluene. The toluene extracts were dried in the freezer, first over MgSO4,
then over P2O5. The dried solution was dropped slowly into a flask
suspended in a boiling water bath. After being heated for 1 h, the solvent
was removed and the product obtained by distillation under reduced
pressure as a colorless oil (52%), making sure to leave a small residue
in the distillation flask. IR (cm-1) 2270 (-NCO); 1H NMR (CDCl3) δ
1,4-Bis(3/4-vinylbenzyloxy)benzene (3). Hydroquinone (10 g), 15
g of K2CO3, and 35 g of 3/4-vinylbenzyl chloride were heated to 70
°C in DMF and stirred for 1 h. The mixture was poured into water and
extracted with diethyl ether. The organic extracts were washed with
NaOH solution and water. After drying and evaporation the residue
was recrystallized to yield 3 as a mixture of the three possible isomeric
compounds (58%), mp 91-92 °C. IR (cm-1) 3084 (dCH2), 2907 (C-
2
3
(ppm) 5.24 (dd, 1H, J ) 1 Hz, J ) 11 Hz, cis-CH2dCH), 5.70 (dd,
1H, 2J ) 1 Hz, 3J ) 18 Hz, trans-CH2dCH), 6.65 (dd, 2H, J(Z) ) 11
Hz, J(E) ) 18 Hz, cis-CH2dCH), 7.03 (d, 2H, J ) 9 Hz, arom CH),
7.34 (d, 2H, J ) 9 Hz, arom CH); 13C NMR (CDCl3) δ (ppm) 114.2
(H2CdCH), 124.8, 127.3 (2×t), 132.6, 135.2 (2×q), 135.6 (H2CdCH);
GC-MS m/z (M+) 145.
1
H), 2860 (OCH2), 1509 (aryl), 1235 (C-O), 1022 (C-O); H NMR
(CDCl3) δ (ppm) 4.99 (s, 4H, O-CH2), 5.24, 5.26 (2×d, total integr.
2H, J ) 11 Hz, cis-CH2dCH), 5.75, 5.76 (2×d, total integr. 2H, J )
18 Hz, trans-CH2dCH), 6.71, 6.72 (2×dd, total integr. 2H, J(Z) ) 11
Hz, J(E) ) 18 Hz, CH2dCH), 6.90 (d, 4H, J ) 3 Hz, arom CH), 7.4
(m, 8H, arom CH); 13C NMR (CDCl3) δ (ppm) 70.4, 70.5 (2×O-
CH2), 114.0, 114.2 (2×H2CdCH), 115.8 (CHdC-O), 136.4, 136.6
(2×H2CdCH), 125.3, 125.7, 126.4, 126.9, 127.7, 128.7 (6×t), 136.8
137.2, 137.5, 137.8 (4×q), 153.1 (CHdC-O); EI-HRMS calcd for
C24H22O3+ 342.1620, found 342.1620; isotope distribution calcd (found)
342 ) 100 (100), 343 ) 26.3 (27.3), 344 ) 3.7 (4.0).
2,8-Dichloro-3,7-dinitrodibenzodioxin was obtained according to
a modified protocol from Kennel et al.22 4-Chlorocatechol (20 mmol)
was dissolved in 20 mL of 1 N methanolic KOH, the solvent was
evaporated, and the residue dried under reduced pressure. 2,5-
Dichloronitrobenzene (20 mmol), 1 g of K2CO3, and 30 mL of HMPA
were added and the mixture was heated to 180 °C for 24 h. After the
mixture was cooled, 50 mL of water was added. The resultant precipitate
was filtered off and extracted with boiling dichloromethane. The extract
was filtered and evaporated to dryness. The residue of evaporation was
chromatographed on silica with petroleum ether. The first fraction was
collected and yielded 1.23 g (24%) of a mixture of 2,7- and
2,8-dichlorodibenzodioxin. Recrystallization from carbon tetrachloride/
petroleum ether yielded predominantly the 2,7-isomer. Recrystallization
from methanol of the evaporated filtrate from the first recrystallization
yielded 443 mg (9%) of pure 2,8-dichlorodibenzodioxin. 1H NMR
2-Methacryloyloxybenzoyl Chloride. Salicylic acid (25 g) was
dissolved in 90 mL of pyridine, the solution was cooled in ice, and
33.5 g of methacrylic anhydride was added. The mixture was stirred
overnight and then added to an excess of dilute HCl and crushed ice.
The oily product was extracted with diethyl ether and the organic
extracts were dried over MgSO4 and evaporated. Pure 2-methacryloy-
loxybenzoic acid was obtained by repeated crystallization from hexane
as colorless crystals, yield 42%; mp, IR, and 1H NMR and in accordance
with the literature.27 The acid was placed in a 50 mL round-bottomed
flask, fitted with a gas scrubber, and an excess of oxalyl chloride was
added. The reaction was initiated by the addition of one drop of DMF.
When gas production had ceased, the excess of oxalyl chloride was
removed by applying a water-pump vacuum directly to the flask. The
residue was dried under reduced pressure and the product was obtained
by vacuum distillation as a yellowish oil (65%). IR (cm-1) 1781 (Cl-
2
3
(CDCl3) δ (ppm) 6.74 (dd, 1H, J ) 1 Hz, J ) 8 Hz), 6.9 (m, 2H);
13C NMR26 (CDCl3) δ (ppm) 116.8, 117.2, 124.0 (3×t); GC-MS m/z
(M+) 252. The intermediate 2,8-dichlorodibenzodioxin (1.75 mmol),
18 mL of nitromethane, and 6 mL of trifluoroacetic anhydride were
cooled in an ice bath. Ammonium nitrate (4.3 mmol) was added slowly
while the mixture was stirred. On completion of the reaction, 6 mL of
methanol was added and the product was filtered off. Recrystallization
from chloroform yielded 443 mg (74%) of mustard yellow shiny plates,
mp 245 °C (sublimation). IR (cm-1) 3096 (C-H), 1530 (NO2), 1478
1
CdO), 1743 (O-CdO), 1601, 1451 (2×aryl); H NMR (CDCl3) δ
(ppm) 2.07 (s, 3H, CH3), 5.80 (s, 1H, cis-H3C-CdCH2), 6.37 (s, 1H,
3
4
trans-H3C-CdCH2), 7.20 (dd, 1H, J ) 8 Hz, J ) 1 Hz, C-3), 7.39
(ddd, 1H, 3J4,5 ) 8 Hz, 3J5,6 ) 8 Hz, 4J3,5 ) 1 Hz, C-5), 7.67 (ddd, 1H,
3J3,4 ) 8 Hz, 3J4,5 ) 8 Hz, 4J4,6 ) 2 Hz, C-4), 8.21 (dd, 1H, 3J ) 8 Hz,
4J ) 2 Hz, C-6); 13C NMR (CDCl3) δ (ppm) 18.2 (p), 123.7(q, C-1),
124.2 (t, C-3), 126.3 (t, C-5), 128.2 (s), 134.1 (t, C-6), 135.0 (>C)CH2),
135.9 (t, C-4), 150.4 (q, C-2), 164.3, 165.3 (2×CdO).
1
(dioxin skeleton), 1341 (NO2); H NMR (d8-THF) δ (ppm) 7.35, (s,
2H), 7.75 (s, 2H); 13C NMR (d8-THF) δ (ppm) 115.2, 120.1 (2×t),
+
123.7, 140.9, 145.0, 145.2 (4×q); EI-HRMS calcd for C12H4N2O6Cl2
341.9446, found 341.9451; isotope distribution calcd (found) 342 )
100 (100), 343 ) 14.0 (14.6), 344 ) 66.9 (64.6), 345 ) 9.3 (8.7), 346
) 11.9 (11.1), 347 ) 1.6 (1.6).
N-(2-(3,7,8-Trichlorodibenzodioxinyl))-2-methacryloyloxybenza-
mide (4). 2-Amino-3,7,8-trichlorodibenzodioxin was obtained as a white
solid by the reduction of the corresponding nitro compound (prepared
by the published procedure,22 see Scheme 2A) in a similar way as
described above for 2,8-diamino-3,7-dichlorodibenzodioxin; GC-MS22
m/z (M+) 301. The crude amine (0.45 mmol) was dissolved in 7 mL
2,8-Dichloro-3,7-bis(N′-(4-vinylphenyl)ureido)dibenzodioxin (2).
Ethanol (15 mL), 23 mL of THF, 1.5 mL of hydrazine hydrate, and
1.5 g of Raney nickel were heated to reflux and 0.61 mmol of 2,8-
dichloro-3,7-dinitrodibenzodioxin, dissolved in 17 mL of THF, was
added dropwise. Another 1.5 mL of hydrazine hydrate was added, and
(26) Grainger, J.; Liu, Z.; Sirimanne, S.; Francis, V. L.; Patterson, D.
G. Organohalogen Compd. 1993, 11, 207-210.
(27) Webr, J.; Svobodova´, J.; Hraba´k, F. Collect. Czech. Chem. Commun.
1976, 41, 738-741.

13348 J. Am. Chem. Soc., Vol. 120, No. 51, 1998
Lu¨bke et al.
of THF and 0.9 mmol of triethylamine. A solution of 0.45 mmol of
2-methacryloyloxybenzoyl chloride in 3 mL of THF/dichloromethane
(6:4) was added dropwise. After stirring overnight under nitrogen, the
mixture was filtered and the filtrate was evaporated. The evaporation
residue was chromatographed on silica with diethyl ether/hexane (1:
1). An intermediate fraction was collected, yielding 180 mg (82%) of
an off-white solid, mp 177-179 °C. IR (cm-1) 3394 (N-H), 3089
(dCH2), 2963 (C-H), 1743 (O-CdO), 1681, 1529 (2×NH-CdO),
1493 (dioxin skeleton); 1H NMR (d8-THF) δ (ppm) 2.01 (s, 3H, CH3),
5.79 (s, 1H, cis-H3C-CdCH2), 6.34 (s, 1H, trans-H3C-CdCH2), 6.82
(s, 1H, arom CH), 7.15, 7.17 (2×s, 1H, arom CH), 7.22 (dd, 1H, 3J )
8 Hz, 4J ) 1 Hz, C-3), 7.36 (ddd, 1H, 3J4,5 ) 8 Hz, 3J5,6 ) 8 Hz, 4J3,5
) 1 Hz, C-5), 7.54 (ddd, 1H, 3J3,4 ) 8 Hz, 3J4,5 ) 8 Hz, 4J4,6 ) 2 Hz,
radical initiator AIBN were mixed in a test tube and the mixture was
degassed by repeated freezing, evacuating, and thawing. The tube was
sealed under vacuum and heated in a water bath at 65 °C for 24 h. The
block of polymer was removed from the tube, washed with methanol,
dried, and ground to an average particle size of 10 µm.
Template Removal Procedures. (a) Reduction of P1-P6 with
LiAlH4. A 250 mg portion of ground polymer was suspended in 25
mL of THF and 125 mg of LiAlH4 was added. The mixture was refluxed
for 2 h. After cooling, destroying the excess LiAlH4 with water, and
acidifying with H2SO4, the mixture was filtered and the filter residue
was washed with 2% H2SO4, water, methanol, diethyl ether, hexane,
diethyl ether, and methanol. To deprotonate the cavity-borne amino
groups, the polymers were further washed with 4% ammonia in water/
methanol (1:1) and dried.
(b) Base Hydrolysis of P7-P15. A 250 mg portion of ground
polymer was refluxed in 25 mL of 1 N methanolic KOH for 6 h. After
cooling and acidifying with H2SO4, the mixture was filtered and the
filter residue was washed successively with water, methanol, diethyl
ether, hexane, diethyl ether, and methanol and dried.
Determination of the Degree of Template Removal. Samples of
imprinted polymers before and after template removal were subjected
to oxygen flask combustion and quantification of chloride by ion
chromatography.
Determination of the Specific Surface Area. Polymer surface areas
were determined from multipoint N2 adsorption isotherms and calculated
by using the BET equation. Polymers were degassed in vacuo overnight
at room temperature before measurement.
Assessment of Binding of Polymers toward TCDD. Individual
samples of polymer, 10 ( 0.1 mg, were weighed into screw cap vials,
incubated for 24 h with 0.5 mL of the appropriate solvent (either nonane,
acetonitrile, or methanol; all solvents were anhydrous) containing 1
pmol of [1,6-3H2]-TCDD, and filtered. The TCDD concentration in the
filtrate was determined by scintillation counting and gave, by subtraction
from the initial TCDD concentration, the amount of bound TCDD per
unit weight of polymer. All experiments were done in 5 replicates.
3
4
C-4), 7.83 (dd, 1H, J ) 8 Hz, J ) 2 Hz, C-6), 7.94 (s, 1H, arom
CH); 13C NMR (d8-THF) δ (ppm) 18.5 (p), 112.3 (t, C-1′), 117.4 (t,
C-4′), 118.6, 118.7 (2×t, C-5′, C-9′), 119.8 (q), 124.1 (t, C-3), 126.6
(t, C-5), 128.0 (s), 129.8 (q), 130.5 (t, C-6), 132.6 (q), 132.8 (t, C-4),
136.9 (>CdCH2), 138.5, 140.7, 141.8, 141.9, 142.5, 142.8 (6×q), 149.8
(q, C-2), 164.4, 165.7 (2×CdO); EI-HRMS calcd for C23H14NO5-
Cl3+ 488.9937, found 488.9930; isotope distribution calcd (found) 489
) 98.7 (100), 490 ) 25.3 (27.7), 491 ) 100 (99.6), 492 ) 25.1 (26.0),
493 ) 35.1 (34.9), 494 ) 8.5 (8.5), 495 ) 4.7 (4.5), 496 ) 1.0 (1.0).
1-Methoxy-3,5-bis(4-vinylbenzyloxy)benzene (5). A solution of 4.8
mmol of 3,5-dihydroxyanisole in 9.6 mL of 1 N methanolic KOH was
evaporated under a nitrogen atmosphere. The residue was dissolved in
10 mL of DMF, the solution was stirred and heated to 60 °C, and 11.6
mmol of 4-vinylbenzyl chloride was added. After 30 min, the mixture
was poured into water and extracted with diethyl ether. The organic
extract was washed with 5 N KOH, then with water, and evaporated.
The residue was chromatographed on silica with diethyl ether/hexane
(1:9). Collection of an intermediate fraction yielded 180 mg of a
colorless oil (10%). IR (cm-1) 3087 (dCH2), 2954 (C-H), 2870
1
(OCH2), 2845 (OCH3), 1602 (CdC), 1197 (C-O); H NMR (CDCl3)
δ (ppm) 3.75 (s, 3H, O-CH3), 4.99 (s, 4H, O-CH2), 5.26 (d, 2H, J )
11 Hz, cis-CH2dCH), 5.76 (d, 2H, J ) 18 Hz, trans-CH2dCH), 6.17
(d, 2H, J ) 2 Hz, arom CH), 6.23 (d, 1H, J ) 2 Hz, arom CH), 6.72
(dd, 2H, J(Z) ) 11 Hz, J(E) ) 18 Hz, CH2dCH), 7.36 (d, 4H, J ) 8
Hz, arom CH), 7.42 (d, 4H, J ) 8 Hz, arom CH); 13C NMR (CDCl3)
δ (ppm) 55.3 (O-CH3), 69.8 (O-CH2), 94.0, 94.5 (2×CHdC-O),
114.1 (H2CdCH), 126.4, 127.7 (2×t), 136.3 (q), 136.4 (H2CdCH),
Acknowledgment. The authors acknowledge financial sup-
port from the UK Ministry of Agriculture, Fisheries and Food.
We would like to thank Dr. F. Mellon and J. Eagles for the
mass spectrometric data and Dr. M. A. Claydon, Kratos
Analytical, for MALDI-ToF analyses. Furthermore, we are
indebted to Professor R. Burch and D. Gleeson of the Depart-
ment of Chemistry, University of Reading, for the surface area
measurements, and to A. Saunders, Department of Chemistry,
University of East Anglia, for the chlorine microanalyses.
+
137.3 (q), 160.6, 161.4 (2×CHdC-O); EI-HRMS calcd for C25H24O3
372.1725, found 372.1729; isotope distribution calcd (found) 372 )
100 (100), 373 ) 27.5 (28.2), 374 ) 4.2 (4.7).
Polymer Synthesis. All polymers in this study were prepared
according to the following protocol: technical grade divinylbenzene
was washed free of inhibitor with KOH solution, washed with water,
and dried. A 0.5 g sample of a mixture of the appropriate monomers,
0.5 mL of porogen, and 1 mol % (relative to double bonds) of the
JA9818295