Biosensors and Bioelectronics 21 (2005) 901–907

Size-selective recognition of catecholamines by molecularimprinting on silica–alumina gel

Tzong-Rong Linga,∗, Yau Zen Syua, Yau-Ching Tasib,Tse-Chuan Choub, Chung-Chiun Liuc

a Department of Chemical Engineering, I-Shou University, Ta-Hsu Hsiang, Kaohsiung 84008, Taiwan, ROCb Department of Chemical Engineering, National Cheng Kung University, Tainan 701, Taiwan, ROC

c Department of Chemical Engineering, Case Western Reserve University, Cleveland, OH 44106, USA

Received 1 November 2004; received in revised form 2 February 2005; accepted 11 February 2005Available online 18 March 2005

Abstract

The preparation of a catecholamine receptor was carried out using a molecular imprinting method with silica–alumina gel to form comple-icate fromreasexaminedtsrecognition

–aluminamonitoring

ETtrac-nd

s are

theved

em-lop-asesetc.

g-asedpectnitial

mentary structures for template recognition. The molecularly imprinted polymer (MIP) was synthesized by the condensation of siltetraethyl orthosilictate (TEOS) under hydrothermal conditions at 60◦C. Aluminum chloride was added as a functional monomer to incthe material’s rebinding ability. The selectivity of the MIP receptor prepared with different ratios of template to Si and Al, was ewith seven analytes including: dopamine, epinephrine, norepinephrine, ascorbic acid, homovanillic acid, uric acid, andl-tyrosine. The resulshowed a size selective effect for the receptors with respect to the recognition of the catecholamines. Some factors affecting theability were investigated including: the solution pH of analytes, surface capping on the MIP, and the imprinting pH of the silicasolution. Also, the catecholamine MIP films on quartz crystal microbalance (QCM) electrodes were fabricated as sensors for in situof the analytes in a 2-propanol solution.© 2005 Elsevier B.V. All rights reserved.

Keywords: Catecholamines; Molecular imprinting; Silica–alumina gel

1. Introduction

Dopamine (DA), epinephrine (EP), and norepinephrine(NEP) are collectively known as the catecholamine neuro-transmitters in the brain. These compounds also functionas hormones in the circulatory system and can be excretedin urine (Garrett and Grisham, 1999). The analysis ofcatecholamines in urine is valuable in the diagnosis ofe.g. neuroblastoma. Several methods have been describedfor analyzing these compounds such as the fluorescentmethod ofKim et al. (1999)and various techniques usingion-exchange chromatography (Burtis and Ashwood, 1999;Jeong et al., 1995) which allow for precise quantitativemeasurement. In hospitals, urine samples are analyzed by

∗ Corresponding author. Tel.: +886 7 6577711x3430; fax: +886 7 6578945.E-mail address: trling@isu.edu.tw (T.-R. Ling).

HPLC with electrochemical detectors. A column CAT/Mpacked with ion exchange resins is used after urine extion for isolation of catecholamines including DA, EP, aNEP. However, complicated and expensive instrumentrequired by these methods.

MIP technology is now known as a method fordesign and development of new materials with impromolecular recognition capabilities (Sellergren, 2001). Thesematerials with tailor-made cavities and engineered chical functionalities have great promise for the devement of highly selective sensing materials, stationary phfor HPLC, catalytic supports, separation membranes,(Makote and Collinson, 1998). To achieve selective reconition of catecholamines, the development of an MIP-bsensor having a tailored size and functionality, with resto the target molecule, was considered to be a good iapproach.

0956-5663/$ – see front matter © 2005 Elsevier B.V. All rights reserved.doi:10.1016/j.bios.2005.02.009

902 T.-R. Ling et al. / Biosensors and Bioelectronics 21 (2005) 901–907

Dickey’s experiments provide an early example of the useof silica in imprinting (Dickey, 1949, 1955). This work wasbased on the use of silica solutions in combination with alky-lorange to form silica gel. There is a well documented historyof the use of porous silica as support materials for catalysts.Zeolites, which can be synthesized with a silica–alumina gel,have a variety of uses. They possess a structure similar tothat of a molecular sieve and therefore are able to act as se-lective adsorbants. In other circumstances, they are employedas a result of their ability to catalyze specific reactions (Joneset al., 1998). Structurally, silica–alumina materials are ex-tremely rigid offering very high surface areas and porosities.For a typical silica gel, a high cross-linking density can beachieved by the condensation of the silanol groups, leadingto negligible swelling in organic solvents, even at elevatedtemperatures (Liang et al., 2000; Kresge et al., 1992; Yanoand Karube, 1999; Morihara, 1997; Chang and Yeung, 1995).It has been reported that aluminum ion doped the silica gelcan gave rise to Lewis acid sites formed in complementarycavities (Morihara, 1997). Some papers presented molecu-larly imprinted sol–gel film with good recognition ability(Fireman-Shoresh et al., 2003; Marx and Liron, 2001; Marxet al., 2004), but few papers discuss about surface capping andLewis acid site for the MIP. However, to-date, the body of lit-erature describing the use of silica–alumina gel for molecularrecognition work is not extensive, with very little new workb

per-f andt esss Al/Sir singfi es inn situQ

2

2

thec tedm dP son,2 ds l-l OS)a rredt p them asm ntm P so-l tott y re-

peatedly batch washing (six times) the imprinted material,each wash with aqueous methanol (10 vol.%), was followedby centrifugation to remove the solvent. Finally, the materialwas rinsed with pure methanol and dried under vacuum atroom temperature, after which it was ground and screened togive 200–325 mesh, approximately 74–44�m, particles.

2.2. MIP powder rebinding studies

The prepared molecularly imprinted silica (50 mg) wasplaced in an analyte solution additionally containing 10 mMphosphate buffer (PB), at pH values ranging from 2 to 7,adjusted by the addition of aqueous NaOH, to ensure stablepH conditions during rebinding.

As shown inTable 1, seven common analytes in urineincluding DA, EP, NEP, AA, HVA, UA, andl-tyrosinewere used to examine the adsorption characteristics ofthe molecular-imprinted silica. The silica–alumina particles(50 mg) were added to a 0.2 mM analyte solution (10 ml)and agitated at 25◦C for 30 min. The concentration of an-alyte in solution after adsorption was determined by HPLC(Shimadzu UV-160A,λ = 280 nm) using a C18 column withthe mobile phase being methanol (15 vol.%) in phosphatebuffer(pH 4.0).

TS

A ht

D

N

E

A

H

U

l-Tyrosine 181

eing published in the past few years.In this study, the recognition of catecholamine was

ormed using a molecularly imprinted silica–alumina gelhe optimum conditions for the molecular imprinting procuch as the pH of the solution, surface capping, and theatio were explored. A simple and rapidly responding senlm was prepared for the determination of catecholaminon-aqueous media (2-propanol solution), using an inCM technique.

. Experimental

.1. MIP preparation procedure

Molecularly imprinted silica gel was synthesized byondensation of a silicate compound with an imprinolecule in a homogeneous phase (Dai et al., 1997; Tanev aninnavaia, 1996; Fan et al., 2004; Kanungo and Collin004). A typical preparation of molecularly imprinteilica–alumina gel (with the ratio of DA:Si:Al = 1:10:3) fo

owed the steps: first, 1 ml tetraethyl orthosilictate (TEnd 0.437 ml HCl(1 M) were added into 2 ml water and sti

o achieve complete hydration. Secondly, in order to keeolar ratio of DA:Si:Al at 1:10:3, the TEOS solution wixed with 0.3168 g AlCl3·6H2O and 0.0829 g of the impriolecule (dopamine) to form a homogeneous phase MI

ution, in which the polymerization of silica was allowedake place by keeping the solution at 60◦C for 24 h. In thehird step, the template was extracted. This was done b

able 1tructures and molecular weights of common analytes in urine

nalytes Structure Molecular weig

opamine (DA) 153

orepinephrine (NEP) 169

pinephrine (EP) 183

scorbic acid (AA) 176

omovanillic acid (HVA) 182

ric acid (UA) 165

T.-R. Ling et al. / Biosensors and Bioelectronics 21 (2005) 901–907 903

2.3. MIP film preparation and rebinding studies

The MIP film was coated on QCM electrodes (9 MHz, P-sensor 2000, Asia New Technology) as follows. The QCMelectrodes were immersed in an ethanol solution of mercap-toethanol (10 mM) for 12 h, washed with deionized water anddried with nitrogen. Then, the homogeneous phase MIP solu-tion (Section2.1) was spin coated on to the QCM electrode at2500 rpm for 1 min. The MIP film was dried at 60◦C for 12 h,and then repeatedly washed to remove the template, using anaqueous methanol solution (10 vol.%.). The imprinted filmon the electrode was finally washed twice in pure methanoland then dried at 60◦C.

Based on the Sauerbrey equation (Buttry and Ward, 1992),given in Eq.(1), the mass of adsorbed analytes was evaluated:

�f = −2.3 × 10−6�M

A× f 2 (1)

where �f = change in frequency due to adsorption (Hz),�M = mass of adsorbent (g),A = the adsorption area on theelectrode (cm2) andf = the oscillation frequency of the unab-sorbed quartz crystal (Hz).

3. Results and discussion

3

na-lo ityw alueo ing,t of6 bec y for

F th aD MIPp ss

binding with DA (Darryl et al., 1998). Little, or no, uptakeof dopamine by the control i.e. non-imprinted polymer (NIP)was found using these pH values.

3.2. Capping of surface silanol groups

The capping of surface silanol groups is an important fac-tor which affects the selectivity of rebinding as shown inFig. 2. Although the adsorption of analytes was decreased bysurface capping, the selectivity of DA on the DA-imprintedsilica–alumina gel was significantly improved after treatmentwith the capping agent 1,1,1,3,3,3-hexamethyldisilazane(HMDS). This reagent reacts with surface OH groups and in-troduces a trimethylsilyl terminating group. HMDS cappingwas achieved by immersing the DA-imprinted silica–aluminagel powder in HMDS overnight. After surface capping, thepowder was washed with tetrahydrofuran (THF) and acetoni-trile to remove excess reagents. The trimethylsilyl end cappedMIP’s were extracted with methanol and dried under vacuum.

From Fig. 2, it can be seen that, after surface capping,the DA-MIP exhibits a greater rebinding affinity for DAcompared to the other analytes examined. The Lewis acidsite and surface silanol groups are thought to play animportant play role in the adsorption of analytes (Morihara,1997). Surface capping of silanol groups to give trimethylsi-l rfacec nce,e ebyp tes.I atedi ate,s orc l haves tinct

F th aD DS,r er in1 nw

.1. Effect of solution pH on rebinding analysis

The adsorption of DA is affected by the pH of the ayte solution. This is shown inFig. 1, where DA rebindingn DA-imprinted silica–alumina shows diminishing affinith decreasing pH values. We found the optimal pH vf the analyte solution, giving maximum template rebind

o be 6.5. We have restricted our work to a maximum pH.5. At low pH values the silica surface OH groups willlose to a neutral charge thereby decreasing their affinit

ig. 1. The analyte solution pH effect for DA adsorption on DA-MIP wiA:Si:Al ratio of 1:10:3. The rebinding was tested based on 50 mg DA-owder in 10 ml DA solution (0.2 mM) at 25◦C for 30 min. The analyteolution was in 10 mM PBS, pH = 6.5.

yl residues decreases the adsorption of analytes. Suapped trimethylsilyl residues may generate steric hindraffectively screening the DA-imprinted cavities and therreferentially retarding the entry of non-template analy

nterestingly, non-imprinted materials formed and tren the same way, except for the addition of templhowed negligible rebinding to the imprint moleculeompeting species. The catecholamines as a group, alimilar structures and functional groups, but have dis

ig. 2. Rebinding amount of DA-imprinted on silica–alumina gel wiA:Si:Al ratio of 1:10:3 on surface uncapped and capped with HM

espectively. Each rebinding test was based on 50 mg DA-MIP powd0 ml analyte solution (0.2 mM) at 25◦C for 30 min. The analyte solutioas in 10 mM PBS, pH = 6.5.

904 T.-R. Ling et al. / Biosensors and Bioelectronics 21 (2005) 901–907

features, different molecular weights and sizes. DA, formsimprinted cavities which only allow the entry of the templatemolecule, thus the DA-MIP has the best specific rebindingfor its imprint species. This is possibly due primarily tosize-exclusion effects, as DA has the lowest molecularweight of the group considered and also lacks the hydroxylfunctionality of NEP and EP.

3.3. Effect of Al/Si ratio

Fig. 3 shows the molar ratio of Al/Si able to give thebest adsorption of DA is Al:Si = 3:10 based on one DA im-printed. The DA-imprinted silica–alumina gel was synthe-sized by changing different to aluminum composition of(DA:Si:Al = 0.1:1:x). Little adsorption of DA was found with-out adding aluminum. Gradually, the Lewis acid sites in-creased with Al doping, but there were decreased whenxlarger than 0.3. Using this approach an optimal Al/Si ratiowas found with which to form the DA-MIP, which gave max-imal DA adsorption on rebinding. After the surface cappingof HMDS, the rebinding amounts overall were decreased dueto the presence of surface silanol groups and the Lewis acidsites which were capped or hindered thereby diminishing theDA adsorption.

3.4. Effect of imprinting pH value

reda cap-p ofD t inP g tot d int theAt ed

F IP( . TheD on5T

Fig. 4. The imprinting pH effect on the rebinding quantity of (a) DA on thesurface capped of DA-MIP, and (b) AA on the surface capped of AA-MIP.The imprinting pH values of the silica–alumina solutions were adjusted byusing HCl and NaOH solutions. DA (AA) imprinted silica gel was composedof DA (AA):Si:Al = 1:10:3. Each rebinding test was based on 50 mg DA-MIPpowder in 10 ml DA solution (0.2 mM) at 25◦C for 30 min. The analytessolution is in 10 mM PBS, pH = 6.5.

into silica–alumina gel, which resulted in no observable AArebinding. Probably, the Lewis acid sites had a stronger repul-sive force to the acidic group of AA at the lower imprintingpH. In contrast, the Lewis acid sites become weak at higherimprinting pH values, thus, it can be found that the higherrebinding amount of AA is found at an imprinting pH of 6.

3.5. Comparison between DA and EP imprinting onsilica–alumina gel

The selective adsorption of the catecholamines: DA, NEP,and EP was carried out on both DA, NEP and EP-MIP’s, re-spectively, as shown inFig. 5. All the DA-MIP, the NEP-MIPand the EP-MIP showed good selectivity for their own targetmolecules in comparison to their affinity for other analytesafter surface capping. Overall it was found that DA exhibitedthe best specific binding for the DA-MIP. The DA moleculehas the smallest size of the group examined, so that it canpotentially rebind into the cavities of the DA-MIP more eas-ily than the remaining analytes. In contrast, EP is the largest

Fig. 4(a) shows rebinding on the DA-MIP when prepat different pH values: 0.7, 4.0, and 6.0 after surfaceing. The DA-MIP was synthesised with the molar ratioA:Si:Al being 1:10:3. The rebinding test was carried ouB (pH = 6.5). The result shows a trend for DA rebindin

he DA-MIP decreasing with increasing values of pH usehe imprinting medium. In contrast, the AA rebound toA-MIP increased with the pH value as shown inFig. 4(b). At

he lower imprinting pH value of 0.7, AA is hardly imprint

ig. 3. Effect of the Al/Si ratio on the rebinding of DA on the DA-Msolid line) surface uncapped, and (dashed line) capped with HMDSA-MIP is composed of DA:Si:Al = 0.1:1:x. The rebinding was based0 mg DA-MIP powder in 10 ml DA solution (0.2 mM) at 25◦C for 30 min.he analytes solution was in 10 mM PBS, pH = 6.5.

T.-R. Ling et al. / Biosensors and Bioelectronics 21 (2005) 901–907 905

Fig. 5. Rebinding of the surface capped DA, NEP and EP MIP’s with dif-ferent analytes. The catecholamine imprinted MIP’s were made with DA(AA):Si:Al = 1:10:3. Each rebinding test was based on 50 mg MIP powderin 10 ml DA solution (0.2 mM) at 25◦C for 30 min. The analyte solutionswere made in 10 mM PB, pH = 6.5.

molecule of the group, so the EP-MIP may have bigger cav-ities to allow DA or EP to enter with little hindrance. EP-MIP is, therefore, not so specifically bound with its template.There results indicate that the binding of catecholamines im-printed on silica–alumina gel shows a size selective effect asshown inFig. 6. Compared to NEP and EP, the smallest sizeof DA imprinted cavities only allow to rebinding DA itself.

Fig. 7. In situ QCM frequency decreases due to rebinding of the analytemolecules: DA, NEP, and EP to the DA-MIP films with surface cappingby HMDS treatment. Each run was made by adding 10�l of 1 mM analytesolution into 0.49 ml of 2-propanol at room temperature.

3.6. In situ QCM rebinding study

A molecular imprinted silica–alumina film was preparedand coated on a QCM electrode for the recognition of cat-echolamines. As shown inFig. 7, the in situ QCM fre-quency decreased due to the binding of analytes. The QCMstudy employed 2-propanol as the solvent to avoid hydrol-ysis of the silica gel surface that in aqueous solution that

Ft

ig. 6. Represented diagram of molecular imprinting of dopamine on silica–ahe larger size molecules (norepinephrine and epinephrine). The binding for

lumina gel. The dopamine (with the smallest size) imprinted cavity to screen outce comes from the adsorption on Lewis acid site.

906 T.-R. Ling et al. / Biosensors and Bioelectronics 21 (2005) 901–907

Table 2Relative binding selectivity of imprinted silica–alumina films for different analytes determined by QCM measurement on (a) the uncapped surface and(b) thecapped surface of MIP films

(a) Surface uncapped

Analytes In situ frequency decrease (Hz) Relative binding selectivity

DA-MIP NEP-MIP EP-MIP DA-MIP NEP-MIP EP-MIP

DA 25.0 24.0 8.1 1.00 0.56 0.33NEP 15.9 43.2 21.4 0.64 1.00 0.89EP 6.5 42.5 24.0 0.26 0.98 1.00

(b) Surface capped

Analytes In situ frequency decrease (Hz) Relative binding selectivity

DA-MIP EP-MIP DA-MIP EP-MIP

DA 13.1 3.9 1.00 0.26NEP 0.4 8.4 0.03 0.58EP 0.2 14.6 0.02 1.00

had been found to bring about peeling of the film. In thisway, we obtained a stable QCM signal showing a highlyreproducible frequency decrease. For the rebinding of an-alytes on the DA-MIP film, the response time was found tobe 50–100 s. The rebinding study was carried out with ananalyte concentration of 0.02 mM at room temperature. Thefrequency decreases for coating the DA-MIP, EP-MIP, andNEP-MIP films on the QCM electrode were 9124.2± 1.5,9038.8± 2.1, and 8927.4± 1.6 Hz, respectively (duplicatedeterminations). The rebinding amounts of analytes on theDA-MIP, NEP-MIP, and EP-MIP films were individually cal-culated based on the differences of frequency between ini-tial and final values as shown inTable 2. Before the surfacecapping, the QCM test showed the molecularly imprintedsilica–alumina films to have better recognition abilities andselectivities for their own template molecules in comparisonto their affinities for the other analytes. After surface capping,the DA-MIP film showed a very specific binding ability withDA in Table 2. For rebinding on the EP-MIP, DA or NEP willbe allowed to enter the cavities of the EP imprinted film dueto the fact that the molecular size of EP is bigger than that ofDA or NEP. An interesting observation from our work wasthat both the film and powder forms of catecholamine MIPexhibit the similar results of recognition ability. The QCMresults indicate that the joint effects of shape and size selec-tivity in the imprinted materials have a marked effect whent

4

im-p rma tin-g tinga oupsw re-b nd-i the

size selective effect. The amount of rebound DA increaseswith the solution pH of the DA-imprinted silica–alumina gel.Al-doping can increase the affinity of the DA-MIP to cat-echolamines, an optimum mole ratio of Si:Al was approxi-mately 10:3 based on imprinting with 1 mol DA. At a lowerimprinting pH value (smaller than 0.7), ascorbic acid is hardlyimprinted onto silica–alumina gel. For catecholamine anal-ysis, both of the molecularly imprinted silica–alumina filmsand powder have better recognition abilities and selectivi-ties for their own template molecules that for the competitiveanalytes examined here.

Acknowledgments

We acknowledge the financial support from ROC Ministryof Education (Grant No. Ex-91-E-FA09-5-4). The authorswould also like to express their gratitude to Dr. John F. Rickfor his assistance with the preparation of this manuscript.

References

Burtis, C.A., Ashwood, E.R., 1999. Tietz Textbook of Clinical Chemistry.W.B. Saunders, Philadelphia, p. 1750.

Buttry, D.A., Ward, M.D., 1992. Measurement of interfacial processes atelectrode surfaces with the electrochemical quartz crystal microbal-

C n sin-aser-

D anylolec-

D ate-nates.iumfromring97).

D Natl.

he MIP’s are used in a sensing application.

. Conclusions

Recognition of catecholamines, by forming molecularrinting materials on silica–alumina gel, in both powder fond film forms, was shown to be a viable method for disuishing the templates from a varied selection of compenalytes. Surface capping to generate trimethylsilyl gras found to significantly improve the selectivity of DAinding. The DA-MIP exhibited the most specific rebi

ng for DA among the catecholamines examined, due to

ance. Chem. Rev. 92 (6), 1355–1379.hang, H.T., Yeung, E.S., 1995. Determination of catecholamines i

gle adrenal medullary cells by capillary electrophoresis and linduced native fluorescence. Anal. Chem. 67 (6), 1079–1083.

ai, S., Yhin, Y., Barnes, C.E., Toth, L.M., 1997. Enhancement of uradsorption capacity and selectivity on silica sol–gel glasses via mular imprinting. Chem. Mater. 9 (11), 2521–2525.

arryl, Y.S., et al., 1998. Molecular imprinted receptors in sol–gel mrials for aqueous phase recognition of phosphates and phosphoIn: Bartsch, R.A., Maeda, M. (Eds.), Washington ACS SymposSeries 703, American Chemical Society, pp. 314–323 (reprintthe symposium on recognition with imprinted polymers held duthe ACS National Meeting in San Francisco on April 13–17, 19

ickey, F.H., 1949. The preparation of specific adsorbents. Proc.Acad. Sci. 35, 227–229.

T.-R. Ling et al. / Biosensors and Bioelectronics 21 (2005) 901–907 907

Dickey, F.H., 1955. Specific adsorption. J. Phys. Chem. 59 (8), 695–707.Fan, H., Yang, K., Boye, D.M., Sigmon, T., Malloy, K.J., Xu, H., Lopez,

G.P., Brinker, C.J., 2004. Self-assembly of ordered, robust, three-dimensional gold nanocrystal silica arrays. Science 304 (23), 567–571.

Fireman-Shoresh, S., Avnir, D., Marx, S., 2003. General method for chiralimprinting of sol–gel thin films exhibiting enantioselectivity. Chem.Mater. 15 (19), 3607–3613.

Garrett, R.H., Grisham, C.M., 1999. Biochemistry, 2nd ed. Saunders Col-lege, Fort Worth, p. S-59.

Jeong, S.Y., Kwon, O.Y., Suh, J.K., Jin, H., Lee, J.M., 1995. Preparationof silica-pillared molecular sieve: effects of catalysts on gelation ofintercalated Si(OC2H5)4. J. Colloid Interface Sci. 175 (1), 253–255.

Jones, C.W., Tsuji, K., Davis, M.E., 1998. Organic-functionalized molec-ular sieves as shape-selective catalysts. Nature 393 (7), 52–54.

Kanungo, M., Collinson, M.M., 2004. Fabrication of two-dimensionallyordered macroporous silica materials with controllable dimensions.Chem. Commun., 528–529.

Kim, K.T., Choi, S.Y., Park, T.J., 1999. Neomycin inhibits catecholaminesecretion by blocking nicotinic acetylcholine receptors in bovineadrenal chromaffin cells. J. Pharmacol. Exp. Therapeut. 288 (1),73–80.

Kresge, C., Leonowicz, M., Roth, W., Vartuli, C., Beck, J., 1992. Orderedmesoporous molecular sieves synthesized by a liquid-crystal templatemechanism. Nature 359, 710–712.

Liang, C., Peng, H., Zhou, A., Nie, L., Yao, S., 2000. Molecular imprint-ing polymer coated BAW bio-mimic sensor for direct determinationof epinephrine. Anal. Chim. Acta 415 (12), 135–141.

Makote, R., Collinson, M.M., 1998. Dopamine recognition in templatedsilicate films. Chem. Commun. (3), 425–426.

Marx, S., Liron, Z., 2001. Molecular imprinting in thin films oforganic–inorganic hybrid sol–gel and acrylic polymers. Chem. Mater.13 (10), 3624–3630.

Marx, S., Zaltsman, A., Turyan, I., Mandler, D., 2004. Parathion sensorbased on molecularly imprinted sol–gel films. Anal. Chem. 76 (1),120–126.

Morihara, K., 1997. Recognition over footprint cavities. In: Bartsch, R.A.,Maeda, M. (Eds.), Washington ACS Symposium Series 703, Ameri-can Chemical Society, pp. 300–313 (reprint from the symposium onrecognition with imprinted polymers held during the ACS NationalMeeting in San Francisco on April 13–17, 1997).

Sellergren, B., 2001. Moleculary Imprinted Polymers. Techniques and In-strumentation in Analytical Chemistry, vol. 23. Elsevier, Amsterdam,p. 213.

Tanev, P.T., Pinnavaia, T.J., 1996. Biomimetic templating of porous lamel-lar silicas by vesicular surfactant assemblies. Science 271 (5253),1267–1269.

Yano, K., Karube, I., 1999. Molecularly imprinted polymers for biosensorapplications. Trends Anal. Chem. 18 (3), 199–204.