Molecularly imprinted polymers synthesized via semi-covalent imprinting with sacrificial spacer for imprinting phenols

Article abstract of DOI:10.1016/j.polymer.2010.09.037

Semi-covalent imprinting with carbonyl group as sacrificial spacer was employed to synthesize molecularly imprinted polymer (MIP) for phenols. A series of semi-covalently imprinted polymers were prepared by varying the templates and porogens. The MIP with 4-chlorophenyl (4-vinyl)phenyl carbonate as template was proved to be the best one, with ethylene glycol dimethacrylate (EGDMA) as cross-linker, 2,2-azobisisobutyronitrile(AIBN), and chloroform as initiator and porogen, respectively. Under such conditions, the corresponding non-covalently imprinted polymer was fabricated with 4-chlorophenol (4-CP) as template and 4-vinylpyridine (4-VP) as functional monomer. The polymer prepared by semi-covalent imprinting displayed superior selectivity to the non-covalently imprinted polymer for phenols. The peak broadening and tailing had been largely reduced on the column packed with semi-covalently imprinted polymer. Meanwhile, the constant retention for these phenols and the good linearity for phenol and 4-CP augured that the semi-covalently imprinted polymer had the potential application as stationary phase for quantitative determination of phenols.

Full text of DOI:10.1016/j.polymer.2010.09.037

Polymer 51 (2010) 5417e5423
Contents lists available at ScienceDirect
Polymer
journal homepage: www.elsevier.com/locate/polymer
Molecularly imprinted polymers synthesized via semi-covalent imprinting with
sacrificial spacer for imprinting phenols
*
Peipei Qi, Jincheng Wang, Liandi Wang, Yun Li, Jing Jin, Fan Su, Yuzeng Tian, Jiping Chen
Dalian Institute of Chemical Physics, Chinese Academy of Science, Dalian 116023, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Semi-covalent imprinting with carbonyl group as sacrificial spacer was employed to synthesize molec-
ularly imprinted polymer (MIP) for phenols. A series of semi-covalently imprinted polymers were
prepared by varying the templates and porogens. The MIP with 4-chlorophenyl (4-vinyl)phenyl
carbonate as template was proved to be the best one, with ethylene glycol dimethacrylate (EGDMA) as
cross-linker, 2,2-azobisisobutyronitrile(AIBN), and chloroform as initiator and porogen, respectively.
Under such conditions, the corresponding non-covalently imprinted polymer was fabricated with
4-chlorophenol (4-CP) as template and 4-vinylpyridine (4-VP) as functional monomer. The polymer
prepared by semi-covalent imprinting displayed superior selectivity to the non-covalently imprinted
polymer for phenols. The peak broadening and tailing had been largely reduced on the column packed
with semi-covalently imprinted polymer. Meanwhile, the constant retention for these phenols and the
good linearity for phenol and 4-CP augured that the semi-covalently imprinted polymer had the
potential application as stationary phase for quantitative determination of phenols.
Received 26 May 2010
Received in revised form
12 August 2010
Accepted 15 September 2010
Available online 22 September 2010
Keywords:
Semi-covalent imprinting
Sacrificial spacer
Stationary phases
Molecularly imprinted polymer
Crown Copyright Ó 2010 Published by Elsevier Ltd. All rights reserved.
1. Introduction
Initially, (meth)acrylate ester of template is synthesized to form
a semi-covalent MIP. After removal of the template by hydrolysis,
Molecularly imprinted polymer has drawn much attention for
its superior selectivity, and it has become a potential material from
an application’s point of view. Recently, three different strategies
have been employed to prepare molecularly imprinted polymer
(MIP), that is, non-covalent, covalent and semi-covalent imprinting,
depending on the molecular interaction between the template and
functional monomer during both pre-polymerization and rebind-
ing process. Non-covalent imprinting is by far the most widespread
in the research on MIP, due to its relative simplicity of the synthesis
process and wide range of chemical functionalities. However, the
method is always compromised by the heterogeneity of binding
site distribution to some extent. The covalent imprinting signifi-
cantly lowers the non-specific interactions, but the complicated
rebinding process makes it not practical for the vast majority of
applications. An attractive option is referred to as semi-covalent
imprinting. Semi-covalent imprinting approach differs from cova-
lent imprinting in that the rebinding step is non-covalent in nature.
Synthetic strategies for the generation of MIP has been extensively
reviewed by A.G. Mayes [1].
the polymer can bind the unesterified template using non-covalent
interaction. However, template hydrolysis is often not straightfor-
ward and the steric requirements of an acid and an alcohol in
hydrogen-bonding contact are rather different from the corre-
sponding ester, which largely hamper the imprinting. Then some of
the limitations of the semi-covalent imprinting approach can be
overcome by the use of a linker group between the template and
the functional molecule, which is lost on template removal [1]. This
linker group has dual role of attaching the template to the func-
tional monomer during polymer formation and acting as spacer
between the template and polymer-bound functionality to prevent
steric crowding in the non-covalent rebinding step. The carbonyl
group of a carbonate ester was the first sacrificial spacer group to be
used in the imprinting of cholesterol [2]. Several groups have also
adopted semi-covalent imprinting with sacrificial spacer to prepare
MIP. Those MIPs have employed carbonyl spacer in template
monomers linked through urea [3], carbonate [2,4e12] and
carbamate [13e15] linkage. The dimethyl silyl group of silyl ether
and silyl esters has also been introduced to the MIP preparation as
a spacer for binding heterocycles [16]. Owing to the complicated
synthesis process prior to polymerization, this method has been
confined to imprint few compounds, such as cholesterol [2,4,6,8,9],
estrone [15], propofol [5,11], menthol [10,12], DDT [13], 2,3,7,8-
tetrachlorodibenzodioxin (TCDD) [3] and so on.
* Corresponding author. Tel./fax: þ86 411 84379562.
E-mail address: chenjp@dicp.ac.cn (J. Chen).
0032-3861/$ e see front matter Crown Copyright Ó 2010 Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.polymer.2010.09.037

5418
P. Qi et al. / Polymer 51 (2010) 5417e5423
Phenolic compounds, especially the priority pollutants
1266 (CeO) cmꢀl; 400 MHz 1H NMR (CDCl3)
d (ppm): 7.48(d, 2H,
controlled by EPA, have aroused researchers’ serious concern for
their considerable detriment to humans and environment. Up to
date, molecular imprinting technique has already been utilized for
the pre-treatment and determination of these phenols, such as
2,4-dichlorophenol (2,4-DCP) [17,18], 2,4-dimethylphenol (2,4-
DMP) [19], 2,4,6-trichlorophenol (2,4,6-TCP) [20e23], pentachlo-
rophenol [24e26], 2,4-dinitrophenol [27,28], 4-nitrophenol
[29e33] and phenol [34e36]. Among these works, most of the MIPs
were fabricated by non-covalent imprinting, except that 4-nitro-
phenol-MIP was synthesized by semi-covalent imprinting using
methacrylate of the target compounds [30]. To the best of our
knowledge, semi-covalent imprinting with carbonyl group as
sacrificial spacer has not been employed to produce MIP for
phenols. Therefore, MIP preparation via semi-covalent imprinting
may be a valuable attempt to acquire a more selective and sensitive
methodology for determination of phenolic compounds.
In the present work, 4-chlorophenyl (4-vinyl)phenyl carbonate
(4-CPC) and 4-methylphenyl (4-vinyl)phenyl carbonate (4-MPC)
were synthesized as the template for bulk polymerization to
imprint phenols. The bulk polymer presented superior recognition
selectivity with 4-CPC as the template, ethylene glycol dimetha-
crylate (EGDMA) and 2,2-azobisisobutyronitrile(AIBN) as cross-
linker agent and initiator, respectively. Meanwhile, the semi-
covalently imprinted polymer showed potential application as
stationary phase in the determination of phenols.
aryl), 7.42(d, 2H, aryl), 7.28(m, 4H, aryl), 6.70 (dd, 1H, CH]CH), 5.78
(d, eCH]CH2, cis), 5.32 (d, eCH]CH2, trans); 100 MHz 13C NMR
(CDCl3)d
(ppm): 151.88(carbonyl), 150.45 (C1), 149.55 (C10), 136.11
(C40), 135.77 (eCH, CH]CH2), 131.89 (C4), 129.78 (C3, 5), 127.47
(C30, 50), 122.42 (C2, 6), 121.02 (C20, 60), 114.71 (eCH2, CH]CH2).
2.2.3. 4-Methylphenyl (4-vinyl)phenyl carbonate (4-MPC)
The 4-MPC was prepared by the same method as the 4-CPC,
from 4-vinylphenol (2.0 g, 16.6 mmol) and 4-methylphenyl chlor-
oformate (2.83 g,16.6 mmol). The compound was obtained as white
crystal, following recrystallization from aqueous methanol, mp
84e85 ꢁC. IR (KBr) 3044 (CH), 1773 (C]O), 1596 (Ar), 1630 (C]C),
1506 (Ar), 1257 (CeO) cmꢀl; 400 MHz 1H NMR (CDCl3)
d (ppm): 7.47
(d, 2H, aryl), 7.44(d, 2H, aryl), 7.26(m, 4H, aryl), 6.75 (dd, 1H, CH]
CH), 5.76 (d, eCH]CH2, cis), 5.29 (d, eCH]CH2, trans); 100 MHz
13C NMR (CDCl3)
d (ppm): 152.29 (carbonyl), 150.62 (C1), 148.95
(C10), 136.11 (C4, 40), 135.77 (eCH, CH]CH2), 129.78 (C3, 5), 127.47
(C30, 50), 122.42 (C2, 6), 121.02 (C20, 60), 114.71 (eCH2, CH]CH2),
20.98 (eCH3).
2.2.4. Polymer synthesis
For the non-covalently imprinted polymer, the template (4-CP,
1 mmol) was dissolved in porogen (5.6 mL) in a 10-mL thick walled
glass tube. The functional monomer (4-VP) (0.425 g, 4 mmol),
cross-linking monomer (EGDMA) (3.8 mL, 20 mmol) and initiator
(AIBN) (0.04 g) were then added to the above solution. For the
semi-covalent imprinting, the template (4-CPC, 4-MPC or their
mixture, 1 mmol) was dissolved in porogen (5.6 mL) in a 10-mL
thick walled glass tube. The initiator (AIBN) (0.04 g) and the cross-
linking monomer (EGDMA) (3.8 mL, 20 mmol) were then added to
the above solution. Both the non-covalent and semi-covalent pre-
polymerization solution were sonicated and purged with oxygen-
free nitrogen for 10 min on an ice bath. The glass tubes were sealed
under nitrogen and placed in a water bath at 60 ꢁC. The reaction
was allowed to proceed for 24 h. As a reference, non-imprinted
polymers that did not contain any template were prepared simul-
taneously using the same protocol. The obtained hard polymers
were crushed, ground, and wet-sieved using acetone to obtain
2. Experimental
2.1. Chemicals
2,4-DMP, 2,4-DCP, 2,4,6-TCP, 2,4,6-trimethylphenol(2,4,6-TMP),
4-VP, EGDMA and AIBN were all from Acros Organics (Geel,
Belgium); acetonitrile was from Fisher (Loughborough, UK). The
monomer 4-VP was purified by the standard procedure to remove
stabilizers. EGDMA was extracted with 10% aqueous sodium
hydroxide and water; after drying over MgSO4, it was filtered and
distilled under reduced pressure. p-Acetoxystyrene was purchased
form TCI chemical (Tokyo, Japan). 4-Chlorophenyl chloroformate
and 4-methylphenyl chloroformate were from Chemlin Chemical
Industrial Co. (Nanjing, China). Dichloromethane and isopropanol
were of HPLC grade. Phenol, 4-chlorophenol (4-CP), 4-methyl-
phenol (4-MP), acetic acid, methanol, chloroform, toluene and
other regents were of analytical grade. Water used was purified
using a Milli-Q gradient A10 system (Millipore, Milford, MA, USA).
regularl sized particles between 45 and 63
chromatographic evaluations.
mm suitable for the
2.2.5. Template removal
For the semi-covalent polymers, the removal of template
requires hydrolysis of polymers, while the template of the non-
covalent polymers can be removed by extraction. The semi-cova-
lent polymers were hydrolyzed using the method described
previously by Whitcombe et al. [2]. The polymers were suspended
in 1 mol Lꢀ1 sodium hydroxide in methanol and heated to reflux for
6 h. The cooled suspensions were added to an excess of dilute
hydrochloric acid, and the products were filtered and washed with
water and methanol. Then both the non-covalent and pre-hydro-
lyzed semi-covalent polymers were extracted in a Soxhlet extrac-
tion apparatus with methanol/acetic acid solution (9:1, v/v)
followed by methanol. The polymers were dried in vacuum at 50 ꢁC
overnight.
2.2. Polymer preparation
2.2.1. p-Vinylphenol
This compound was prepared by the hydrolysis of p-acetox-
ystyrene with aqueous potassium hydroxide according to the
method of Corson et al. [37] and obtained as shiny colorless plates.
The product was washed thoroughly with water, dried under
vacuum and stored at ꢀ20 ꢁC until used.
2.2.2. 4-Chlorophenyl (4-vinyl)phenyl carbonate (4-CPC)
To
a cooled solution (ice bath) of 4-vinylphenol (2.0 g,
16.6 mmol) in dry THF (60 mL) and triethylamine (4 mL) containing
a trace of 2, 6-di-tert-butyl-4- methylphenol (3 mg) was added
dropwise a solution of 4-chlorophenyl chloroformate (3.17 g,
16.6 mmol) in THF (40 mL), and the mixture was stirred overnight
at room temperature. The obtained solution was filtered and the
filtrate was evaporated to yield the crude product. Recrystallization
from methanol gave the product as colorless plates, mp 111e114 ꢁC.
IR (KBr) 3099 (CH), 1764 (C]O), 1602 (Ar), 1630 (C]C), 1507 (Ar),
2.3. Characterization of the prepared polymers
2.3.1. Nitrogen sorption porosimetry measurements
Nitrogen sorption porosimetry measurements were performed
on a Nova Surface Area and Porosimetry Analyzer (Quantachrome
Instrument Corporation, USA). The specific surface area was
calculated using the standard BET method, with the specific pore
volume and average pore diameter using BJH theory.

P. Qi et al. / Polymer 51 (2010) 5417e5423
5419
2.3.2. FTIR analysis
The spectra of the polymers were measured using a Spectrum
GX spectrometer (PerkinElmer, USA) in the 4000e400 cmꢀ1 region
with a resolution of 4 cmꢀ1. The spectrum of each solution was
obtained by averaging 8 consecutive scans.
2.3.3. X-ray fluorescence (XRF) spectroscopy
Residual chlorine content of the prepared semi-covalently
imprinted polymer was determined by XRF spectroscopy on a Phi-
lips Magix XRF spectrometer. Samples were pressed as homoge-
nous tablets for XRF analysis.
2.3.4. Solid-state 13C NMR measurement
13C CP/MAS NMR spectra were recorded on Varian Infinityplus-
400 spectrometer using a 5 mm probe at a spinning speed of 6 kHz.
2000e3000 scans were accumulated with a
p/2 pulse width of
4.2 s, recycle delay of 2 s, and contact time of 7 ms.
m
2.4. Chromatographic evaluation of the polymers
To evaluate the polymers in analytical columns, ground polymer
particles were suspended in acetonitrile by sonication and then
slurry packed into 10 cm ꢂ 0.46 cm i.d. stainless steel HPLC columns
at 3000 psi using an air-driven fluid pump (Haskel) with ethanol as
the solvent. A Waters 515 ternary HPLC pump and a Waters 2487
dual
l absorbance detector were used.
Fig. 1. Schematic representation of the sacrificial semi-covalent approach used in this
research. (a) Template synthesis. (b) Polymer preparation of 4-chlorophenol- imprin-
ted polymers by semi-covalent imprinting.
2.4.1. Reversed-phase liquid chromatographic evaluation
The chromatographic evaluation of the polymers was carried
out using acetonitrile as the mobile phase at 1 mL minꢀ1. The
injection volume was 20
m
L, the detector was set at 280 nm, and the
a shoulder in the IR spectrum at 1779 cmꢀ1, and its double-
frequency absorption at 3445 cmꢀ1, which all disappeared on
hydrolysis (Fig. 2, bottom curve). The broad and strong band
appeared at 3527 cmꢀ1 (Fig. 2, top curve), indicative of the
formation of aromatic phenolic hydroxyl group on account of the
polymer hydrolysis. Other semi-covalently imprinted polymers
were also characterized by FTIR, with the same spectra obtained.
Residual chlorine content was determined to clarify the hydro-
lysis of the polymers. In the synthesized semi-covalently MIP 2,
chlorine is introduced from the reactants, 4-CPC and chloroform.
The porogen chloroform and unreacted template 4-CPC can be
released by extraction with methanol, while the removal of the
chlorine in the integrated 4-CPC required the hydrolysis of the
polymer. Thus, the chlorine contents in the polymers with different
treatments were measured. The chlorine content was 10.206 wt% in
the synthesized polymer without any treatment, and then it
decreased to 8.740 wt% after extraction with methanol, later to
2.575 wt% upon hydrolysis (Fig. S2). The decrease of the chlorine
content fully proved that the hydrolysis of the semi-covalently
imprinted polymer can remove the template, forming the func-
tional site for imprinting phenols.
analyses were performed at room temperature. Acetone was
injected as the void marker. Capacity factor, k0, was calculated by
the equation k0 ¼ (tR ꢀ t0)/t0, where tR and t0 are the retention time
of the analyte being investigated and the void marker, respectively.
The molecular imprinting factor (IF) proposed for the evaluation of
the recognition selectivity was calculated by the equation
IF ¼ k0MIP/k0NIP, where k0MIP was the capacity factor of the analyte on
the MIP and k0NIP was that on the NIP.
2.4.2. Normal liquid chromatographic evaluation
The HPLC columns packed with MIP and NIP particles were
firstly conditioned with isopropanol and dichloromethane, and
then evaluation were performed using dichloromethane as mobile
phase at the flow rate of 1.0 mL minꢀ1. The injection volume was
20 mL, the detector was set at 280 nm, and the analyses were per-
formed at room temperature. Acetone was injected as the void
marker. Capacity factor and imprinting factor were calculated by
the same equations as demonstrated in Section 2.4.1.
3. Results and discussion
Solid-state 13C NMR experiments were also carried out to
demonstrate the change of the semi-covalent MIP through hydro-
lysis. The results displayed that the carbonate carbonyl group in 4-
CPC at 151.0 ppm disappeared in the NMR spectrum of the polymer
after hydrolysis, providing abundant evidence for the hydrolysis of
the semi-covalent MIP. The detailed information was offered in
Fig. S3.
Various templates and porogens were employed to explore their
influence on selectivity of the obtained polymers and determine
the optimum combination of the MIP for imprinting phenols. Three
templates, 4-CPC, 4-MPC and 4-CPC/4-MPC (1:1, mol/mol) were
compared. As for porogen, hexane and the mixture of hexane and
toluene were used in the majority of the imprinted polymers
described in literature by semi-covalent imprinting. Hexane and
3.1. Preparation and evaluation of semi-covalently imprinted
polymers
A series of MIPs were synthesized via semi-covalent imprinting
with carbonyl group as sacrificial spacer. Fig. 1 showed the
synthesis process of the templates and the formation of those semi-
covalent MIPs, and the compositions of the polymerization
mixtures were described in Table 1. The corresponding NIPs were
prepared by the same protocol without addition of the templates.
Semi-covalent MIPs 2e7 were hydrolyzed using the standard
conditions of 6 h reflux with 1 mol Lꢀ1 NaOH in methanol [2]. The
template removal was confirmed by FTIR spectroscopy. The
aromatic carbonate carbonyl band of MIP 2 was clearly resolved as

5420
P. Qi et al. / Polymer 51 (2010) 5417e5423
Table 1
Compositions of the polymerization mixtures used for the preparation of the MIPs.
Polymer
Template
Functional monomer 4-VP
Cross-linker EGDMA
Initiator AIBN
Porogen
MIP 1
MIP 2
MIP 3
MIP 4
MIP 5
MIP 6
MIP 7
4-CP
4-CPC
4-CPC
4-MPC
4-MPC
4-MPC/4-CPC
4-MPC/4-CPC
0.425 mL
20 mmol
20 mmol
20 mmol
20 mmol
20 mmol
20 mmol
20 mmol
0.04 g
0.04 g
0.04 g
0.04 g
0.04 g
0.04 g
0.04 g
Chloroform
Chloroform
Acetonitrile
Chloroform
Acetonitrile
Chloroform
Acetonitrile
e
e
e
e
e
e
hexane/toluene (9:1, v/v) were also employed in our initial study.
Unsatisfactorily, the polymers were not rigid for further applica-
tion. Thus, chloroform and acetonitrile were extensively investi-
gated. Superficially, there was no obvious difference among these
MIPs using different templates and porogens. All were white
powders after grinding, except that the polymers prepared with
chloroform as porogen were more rigid than those prepared with
acetonitrile as porogen. The recognition selectivity of the polymers
for phenols were characterized by IF values. The results were
demonstrated in Fig. 3.
Using chloroform as porogen in the bulk polymerization, the
polymers obtained from the different templates showed the same
trends for the phenols with regard to IF value (Fig. 3). The IF values of
these phenols decreased with increasing of their structure size. This
result verifies the imprinting effect and the significant role of the
imprintingcavities.AsillustratedinFig.1b,themolecularrecognitionis
based on the hydrogen bonding interaction between the phenolic
hydroxyl group of the target analyte and the hydroxyl residue of the
polymers.Priortointeraction,thetestanalyteisrequiredtoenterinto
thevacantimprintedcavities.Thatis,toachieveefficientrecognitionin
the binding process, the compounds need to possess the phenolic
hydroxyl group and similar molecular size and structure to the
template.Thephenoliccompoundswithlargermolecularsizethanthe
template cannot be recognized efficiently, resulting in the smaller
IF values for 2,4,6-TCP and 2,4,6-TMP than other phenols. Large
molecular size prevents their entrance into the imprinted cavities,
and thus disrupts the specific interaction between the phenolic
hydroxyl group of the target analyte and the hydroxyl residue of the
polymers. However, 2,4-DCP and 2,4-DMP can give rise to slightly
largerIFvaluethan2,4,6-TCPand2,4,6-TMP,whichmaybeattributedto
the fact that the loss of carbonyl group during hydrolysis of the
carbonate ester provides sufficient space to allow the analyte with
slightly larger structure and vinylphenol-derived hydroxylgroups to
formahydrogenbond.
With regard to the templates, the test phenols arouse larger IF
values in the polymer column with 4-CPC as template than the
other polymers. 4-CPC/4-MPC (1:1, mol/mol) was also served as
the template in order to exploit the dual functional groups of
chloro- and methyl- in 4-CPC and 4-MPC, in an effort to improve
the selectivity of the polymers for both the methyl- and chloro-
substituted phenols [38]. Unsatisfactorily, all the phenols yielded
lower IF values in this polymer than the other two polymers.
Using acetonitrile as porogen, the effect of various templates on
the selectivity of the MIPs was also investigated. Unlike the poly-
mers prepared with chloroform as porogen, the IF values for these
phenols did not show identical trends among these three polymers.
On the contrary, the polymer with the mixture of 4-CPC and 4-MPC
as template shows better selectivity for these phenols. Neverthe-
less, this polymer did not display better selectivity than MIP 2 with
4-CPC and chloroform as template and porogen, respectively.
Hence, MIP 2 was the choice of the optimum polymers for further
investigation.
1.6
4-CPC 4-MPC 4-MPC/4-CPC
1.5
1.4
a
1.3
1.2
1.1
1.0
0.9
0.8
phenol 4-MP 4-CP 2-CP 2,4- 2,4- 2,4,6- 2,4,6-
DCP DMP TCP TMP
1.6
4-CPC
4-MPC 4-MPC/4-CPC
70
60
50
40
1.5
1.4
1.3
1.2
1.1
1.0
0.9
0.8
b
30
20
10
%T
3529.4
3439.7
0
phenol 4-MP 4-CP 2-CP 2,4- 2,4- 2,4,6- 2,4,6-
DMP TCP TMP
1779.0
-5
DCP
4000
3600 3200 2800 2400 2000 1800 1600 1400 1200 1000 800 600 400
Fig. 3. IF values for the phenolic compounds in the columns packed with semi-
covalently imprinted polymers under the chromatographic conditions described in
Section 2.4.1.
Fig. 2. IR spectra of polymers imprinted, before (bottom) and after (top) removal of
template by hydrolysis.

P. Qi et al. / Polymer 51 (2010) 5417e5423
5421
Table 2
3.2. Comparison of imprinting approaches
Physical properties of the semi-covalently imprinted polymer (MIP 2), the non-
covalently imprinted polymer (MIP 1) characterized by nitrogen sorption
porosimetry.
Based on the optimized synthesis condition for semi-covalently
imprinted polymer, the corresponding non-covalently imprinted
polymer was prepared with 4-CP as template, 4-VP, EGDMA, AIBN
and chloroform as functional monomer, cross-linker, initiator and
porogen, respectively. The differences of the polymers were
investigated regarding their physical characteristics, recognition
selectivity and chromatographic behaviors.
Polymers
Surface area (m2 gꢀ1
)
Pore volume (cm3 gꢀ1
)
Pore size (nm)
MIP 1
MIP 2
165.2
286.4
0.223
0.269
5.40
4.55
3.2.2. Difference in molecular imprinting selectivity
Evaluation of the two polymers was performed by the reversed-
phase HPLC and normal HPLC, so as to compare the selectivity of the
polymers for the phenols. In the reversed-phase HPLC system,
acetonitrile was used as the mobile phase at 1 mL minꢀ1. Phenol, 4-CP,
2-CP, 4-MP, 2,4-DMP, 2,4-DCP, 2,4,6-TMP and 2,4,6-TCP were injected
into the MIP columns and the corresponding NIP columns under the
previously described chromatographic conditions. Imprinting factor
was calculated to verify their difference in selectivity and imprinting
effect. As depicted in Fig. 5a, it can be clearly seen that the IF values for
these phenols are higher in MIP 2 column than MIP 1 column, that is,
the semi-covalent imprinting method is more selective. Furthermore,
the IF values for these phenols present different trends in the two MIP
columns. For the non-covalently imprinted MIP 1, 4-CP gave rise to
the largest IF value, followed by 2,4-DCP, 4-MP, phenol, 2,4-DMP, 2-CP,
2,4,6-TCP and 2,4,6-TMP successively. This result explains that the
selectivity of the non-covalently imprinted polymer column for the
test compounds is not only affected by the molecular structure, but
also the acidity of the analytes [19,24]. With regard to semi-covalent
imprinting, the molecular structure is more crucial for the recognition
3.2.1. Physical characteristics
Nitrogen sorption porosimetry measurements were performed
to evaluate the morphology of the imprinted polymers, and the
possible difference resulting from the different imprinting
methods. As illustrated in Fig. 4a, both the semi-covalent MIP 2 and
non-covalent MIP 1 produced type IV isotherm, characterized by its
hysteresis loop, which is associated with capillary condensation
taking place in mesopores [39]. This phenomenon is similar to the
finding by Holland et al. [40]. In the successfully fabricated MIPs,
mesopores predominates and the hysteresis loop is not closed,
implying incomplete removal of the gas adsorbate from narrow
pores [39]. However, MIP 2 had a larger surface area than MIP 1
(Table 2). Higher surface areas indicated that phase separation
occurred at later stages of the polymerization and the formed
polymers were accompanied with smaller pore size distributions
[41]. Fig. 4b also demonstrated that MIP 2 had a narrower pore size
distribution than MIP 1.
1.8
semi-covalently imprinted polymer, MIP 2
1.7
non-covalently imprinted polymer, MIP1
1.6
a
1.5
1.4
1.3
1.2
1.1
1.0
phenol 4-MP 4-CP 2-CP 2,4- 2,4- 2,4,6- 2,4,6-
DCP DMP TCP TMP
2.0
semi-covalently imprinted polymer, MIP 2
1.8
non-covalently imprinted polymer, MIP 1
1.6
b
1.4
1.2
1.0
0.8
0.6
phenol 4-CP 4-MP 2-CP 2,4- 2,4- 2,4,6- 2,4,6-
DCP DMP TCP TMP
Fig. 4. Sorption isotherms (a) and pore volume distributions (b) for the imprinted
polymers prepared by semi-covalent imprinting (MIP 2) and non-covalent imprinting
(MIP 1).
Fig. 5. Comparison of the recognition selectivity of the polymers prepared by semi-
covalent imprinting and non-covalent imprinting approaches in the reversed-phase
HPLC conditions (a) and normal HPLC system (b).

5422
P. Qi et al. / Polymer 51 (2010) 5417e5423
ability of the test analytes in the polymer columns, which has been
extensively illustrated in 3.1.
obtained by Hwang et al. [6]. This fact can be attributed to the
presence of binding sites of different affinity in the non-covalently
imprinted polymer and the uniform distribution of the binding
sites in the semi-covalently imprinted polymers. In the semi-
covalent imprinting, the template molecule is covalently bound to
the monomer and after polymerization the template occupies
exactly the position of the recognition site. Each recognition site
formed after removal of the template was shown to bind one
molecule of phenol through hydrogen bonding. In contrast, the
complexes in the non-covalent imprinting were formed from
template and functional monomer in a somewhat loose manner.
Since the complexes were formed simply by just mixing these two
together and stabilized only by weak interaction like hydrogen
bonding during the polymerization, the construct of future
recognition sites was looser in comparison with semi-covalent
imprinting and contained aggregates of two or more of the
template molecules. This inevitably leads to the heterogeneous
distribution of recognition sites in the imprinted polymer, and
finally peak broadening and tailing when the polymer was used as
the stationary phase in liquid chromatography.
In the normal HPLC, dichloromethane was used as the mobile
phase. As described in Fig. 5b, all the phenols brought to higher IF
values in the MIP column prepared by semi-covalent imprinting
than that prepared by non-covalent imprinting method.
In conclusion, the semi-covalently imprinted polymer showed
better selectivity than the non-covalently imprinted polymer
evaluated by both the reversed-phase and normal HPLC.
3.2.3. Chromatographic behaviors in normal HPLC
Spiked standard solution involving phenol, 4-CP, 2,4-DCP and
2,4,6-TCP were injected into the polymer columns, and the elution
profiles were shown in Fig. 6. It is undeniable that the imprinting
effect exists in the polymers prepared by both non-covalent
imprinting and semi-covalent imprinting approaches. The phenols
have shorter retention times in the NIP columns than in the cor-
responding MIP columns, stemming from the presence of the
imprinted cavities in the MIPs.
As far as the chromatographic separation is concerned, the two
MIP columns are distinguished from each other. The four phenols
had longer retention times in the column packed with non-
covalently imprinted polymer, but they had a much sharper peak
when they were chromatographed on the semi-covalently
imprinted polymer. This result is in accordance with the result
In addition, the phenols at various concentrations were injected
into the two MIP columns, and the chromatograms were plotted.
The phenols at 0.1
packed with semi-covalently imprinted polymer, whilst these
phenols at 1.0
g mLꢀ1 cannot be observed in the column packed
m
g mLꢀ1 can be clearly observed in the column
m
with non-covalently imprinted polymer. This result can mainly be
ascribed to the band broadening and peak tailing in the non-
covalently imprinted polymers, and consequently the poor sepa-
ration efficiency and low resolution. Although the four phenols still
can not achieve complete baseline separation from each other in
the MIP column packed with semi-covalently imprinted polymer,
this result is still inspiring. To the best of our knowledge, the semi-
covalently imprinted polymer has not been used as stationary
phase to separate the priority phenolic compounds.
16000
MIP 1
NIP 1
14000
12000
10000
a
8000
Then the emphasis was paid on the semi-covalently imprinted
polymer MIP 2 column. The four phenols at 10
m
g mLꢀ1 were
2
6000
4000
2000
1
3
injected into the column alone, the retention times were 2.53, 3.18,
5.66, 7.84 min for 2,4,6-TCP, 2,4-DCP, phenol and 4-CP, respectively.
The spiked solution of these phenols at the same concentration was
analyzed under the identical chromatographic conditions, and their
elution profiles were shown in Fig. 7. It can be clearly observed that
every compound has the same retention time both when eluted in
mixture and alone.
4
4
0
0
10
20
30
Time, min
18000
16000
14000
12000
10000
8000
3
3
MIP 2
NIP 2
16000
14000
3
2
12000
1
b
4
1, 2
10000
8000
6000
4000
2000
0
2
6000
4
1
4000
4
8
2000
0
0
2
4
6
10
12
0
2
4
6
8
10
12
Time, min
Time, min
Fig. 6. Elution profiles of phenolic compounds on the HPLC columns packed with the
non-covalently imprinted polymer (a) and the semi-covalently imprinted polymer (b).
Peak designation: (1) 2,4,6-TCP, (2) 2,4-DCP, (3) phenol, (4) 4-CP.
Fig. 7. Chromatograms of phenols on the semi-covalently imprinted polymer (MIP 2)
column. Samples were analyzed by alone (bottom) or mixture (top). Peak designation:
(1) 2,4,6-TCP, (2) 2,4-DCP, (3) phenol, (4) 4-CP.

P. Qi et al. / Polymer 51 (2010) 5417e5423
5423
Acknowledgements
Financial supports from the National Basic Research Program of
China (973 program) (2009CB421602), the National Natural Science
Foundation of China (20775081) and National High Technology
Research and Development Program of China (863 Program)
(2007AA061601) are gratefully acknowledged.
Appendix. Supplementary material
Supplementary data related to this article can be found online at
doi:10.1016/j.polymer.2010.09.037.
References
[1] Mayes AG, Whitcombe MJ. Adv Drug Deliv Rev 2005;57:1742e78.
[2] Whitcombe MJ, Rodriguez ME, Villar P, Vulfson EN.
J Am Chem Soc
1995;117:7105e11.
Fig. 8. Elution profiles on MIP 2 column of phenols with different concentrations, and
the corresponding calibration curves of phenol and 4-CP. Peak designation: (1) 2,4,6-
TCP, (2) 2,4-DCP, (3) phenol, (4) 4-CP.
[3] Lubke M, Whitcombe MJ, Vulfson EN. J Am Chem Soc 1998;120:13342e8.
[4] Boonpangrak S, Prachayasittikul V, Bulow L, Ye LJ. J Appl Polym Sci Symp
2006;99:1390e8.
[5] Petcu M, Cooney J, Cook C, Lauren D, Schaare P, Holland P. Anal Chim Acta
2001;435:49e55.
[6] Hwang CC, Lee WC. J Chromatogr A 2002;962:69e78.
[7] Joshi VP, Kulkarni MG, Mashelkar RA. J Chromatogr A 1999;849:319e30.
[8] Flores A, Cunliff D, Whitcombe MJ, Vulfson EN. J Appl Polym Sci Symp
2000;77:1841e50.
[9] Perez N, Whitcombe MJ, Vulfson EN. J Appl Polym Sci Symp 2000;77:1851e9.
[10] Patel A, Fouace S, Steinke JHG. Anal Chim Acta 2004;504:53e62.
[11] Petcu M, Schaare PN, Cook CJ. Anal Chim Acta 2004;504:73e9.
[12] Patel A, Fouace S, Steinke JHG. Chem Commun; 2003:88e9.
[13] Graham AL, Carlson CA, Edmiston PL. Anal Chem 2002;74:458e67.
[14] Katz A, Davis ME. Nature 2000;403:286e9.
[15] Ki CD, Oh C, Oh S-G, Chang JY. J Am Chem Soc 2002;124:14838e9.
[16] Kirsch N, Alexander C, Lubke M, Whitcombe MJ, Vulfson EN. Polymer
2000;41:5583e90.
The phenols mixture with concentrations from 0.1 to 50 m
g mLꢀ1
was tested in MIP 2 column. The chromatograms were depicted in
Fig. 8. As for phenol and 4-CP, the calibration curve based on the
peak area versus sample concentration was constructed, and the
regression coefficients better than 0.993 were obtained. This result
along with the constant retention implies its potential application
as stationary phase for quantitative analysis, although baseline
separation still needs to be resolved for these two compounds.
It should be noted that with the semi-covalent methods, the
residual template remains covalently bound to the polymer, and is
not subject to leaching under normal conditions of use. In fact,
template leaching was not observed during the analysis. This is very
important to attain the prefect application of molecular imprinting
technique.
[17] Feng QZ, Zhao LX, Yan W, Ji F, Wei YL, Lin JM. Anal Bioanal Chem
2008;391:1073e9.
[18] Li Y, Li X, Li YQ, Qi JY, Bian J, Yuan YX. Environ Pollut 2009;157:1879e85.
[19] Qi P, Wang J, Jin J, Su F, Chen J. Talanta 2010;81:1630e5.
[20] Feng QZ, Zhao LX, Lin JM. Anal Chim Acta 2009;650:70e6.
[21] Feng QZ, Zhao LX, Yan W, Lin JM, Zheng ZX. J Hazard Mater 2009;167:282e8.
[22] Feng QZ, Zhao LX, Chu BL, Yan W, Lin JM. Anal Bioanal Chem
2008;392:1419e29.
[23] Schwarz L, Holdsworth CI, McCluskey A, Bowyer MC. Aust
2004;57:759e64.
[24] Han DM, Fang GZ, Yan XP. J Chromatogr A 2005;1100:131e6.
J Chem
4. Conclusions
[25] Nicholls C, Karim K, Piletsky S, Saini S, Setford S. Biosens Bioelectron
2006;21:1171e7.
[26] Baggiani C, Anfossi L, Giovannoli C, Tozzi C. J Chromatogr B Analyt Technol
Biomed Life Sci 2004;804:31e41.
[27] Luo W, Zhu LH, Yu C, Tang HQ, Yu HX, Li X, et al. Anal Chim Acta
2008;618:147e56.
[28] Zakaria ND, Yusof NA, Haron J, Abdullah AH. Int J Mol Sci 2009;10:354e65.
[29] Masque N, Marce RM, Borrull F, Cormack PAG, Sherrington DC. Anal Chem
2000;72:4122e6.
[30] Caro E, Masque M, Marce RM, Borrull F, Cormack PAG, Sherrington DC. J
Chromatogr A 2002;963:169e78.
[31] Caro E, Marce RM, Cormack PAG, Sherrington DC, Borrull F. J Chromatogr A
2003;995:233e8.
[32] Say R, Ersoz A, Sener I, Atilir A, Diltemiz S, Denizli A. Sep Sci Technol
2004;39:3471e84.
[33] Ersoz A, Denizli A, Sener I, Atilir A, Diltemiz S, Say R. Separ Purif Tech
2004;38:173e9.
[34] JoshiVP, KarodeSK, KulkarniMG, MashelkarRA. ChemEngSci1998;53:2271e84.
[35] Lv YQQ, Lin ZX, Feng W, Tan TW. Chromatographia 2007;66:339e47.
[36] An FQ, Gao BJ, Feng XQ. J Hazard Mater 2008;157:286e92.
[37] Corson BB, Heintzelman WJ, Schwartzman LH, Tiefenthal HE, Lokken RJ,
Nickels JE, et al. J Org Chem 1958;23:544e9.
[38] Krupadam RJ, Bhagat B, Wate SR, Bodhe GL, Sellergren B, Anjaneyulu Y.
Environ Sci Technol 2009;43:2871e7.
[39] Sing KSW. Pure Appl Chem 1982;54:2201e18.
[40] Holland N, Frisby J, Owens E, Hughes H, Duggan P, McLoughlin P. Polymer
2010;51:1578e84.
This research clearly demonstrates the value of the semi-covalent
imprinting approach for fabricating imprinted polymer to deter-
mine the priority environmental pollutants, phenolic compounds
involving phenol, 4-CP, 2,4-DCP, 2,4,6-TCP. As stationary phase, the
semi-covalently imprinted polymer exhibits superior selectivity to
non-covalently imprinted polymer from an application’s point of
view. In this semi-covalent imprinting approach, the carbonyl group
of 4-vinylphenyl carbonate ester acts as a sacrificial spacer. The
covalently-bound template monomer can be easily hydrolyzed with
the loss of CO2, which results in the formation of a non-covalent
recognition site, bearing a phenolic residue capable of recognizing
phenols with homogenous binding sites. Therefore, the use of semi-
covalent imprinting significantly reduced the peak tailing and band
broadening. This advantage along with the constant retention
augurs that the semi-covalently imprinted polymer has the poten-
tial application as stationary phase for quantitative determination of
phenols. Besides with the semi-covalent methods, the residual
template remains covalently bound to the polymer, and is not
subject to leaching under normal conditions of use. This is very
important to attain the applications of molecular imprinting
technique.
[41] Sherrington DC. Chem Commun; 1998:2275e86.