S

MD

a

ARRAA

KISCCQS

1

omasimmsd

cotpb(tata

bi

1d

Biochemical Engineering Journal 54 (2011) 57–61

Contents lists available at ScienceDirect

Biochemical Engineering Journal

journa l homepage: www.e lsev ier .com/ locate /be j

ensing capability of molecularly imprinted self-assembled monolayer

in Jae Shin, Won Hi Hong ∗

epartment of Chemical and Biomolecular Engineering, KAIST, Daejeon 305-701, Republic of Korea

r t i c l e i n f o

rticle history:eceived 4 October 2010eceived in revised form 7 January 2011ccepted 27 January 2011vailable online 3 February 2011

a b s t r a c t

A molecularly imprinted self-assembled monolayer (SAM) was fabricated on a gold plate by form-ing a monolayer with both 1-hexadecanethiol and the template molecule, and removing the templatemolecules by solvent extraction. Cholesterol, cholic acid, and deoxycholic acid were used as the templatemolecules. Cyclic voltammograms were obtained using these imprinted gold plates as a working elec-trode, with Ag/AgCl reference electrode and Pt counter electrode. Potassium ferricyanide was used as a

eywords:mprintingelf-assembled monolayerholesterolyclic voltammogram

background material for oxidation and reduction. These imprinted monolayers were capable of discrim-inating the substrate that had been imprinted. The stability of the imprinted monolayers was estimatedprecisely and the thickness change of the monolayer was estimated using quartz crystal microbalance(QCM). During repeated detection, 1-hexadecanethiol molecules on the gold plate were tightly adheredto the gold surface. However, the sensing ability was reduced with repeated detection, suggesting that

ue to

uartz crystal microbalanceensor

these phenomena were d

. Introduction

Recently significant attention has been paid to the developmentf the molecular imprinting technique that enables obtaining poly-ers to mimic biological receptors. This technique is a very useful

pproach for the fabrication of a matrix with molecular recognitionites, which are formed by the addition of template molecules dur-ng the matrix formation process, and the removal of the template

olecule after the matrix formation [1–5]. Molecularly imprintedatrixes have been developed over a decade in many fields,

uch as chromatography, catalyst, artificial antibody, and sensingevices.

The detection of the binding molecules in the imprinted matrixan be achieved by electrochemical, impedance, and optical meth-ds [6], with the former being the easiest and most economical wayo fabricate a commercial sensor [7,8]. The imprinting method inolymer science has been focused on three-dimensional networks,ut the creation of a two-dimensional, self-assembled monolayerSAM) on solid surfaces represents a significant difference fromhese materials and could be helpful for technological and scientificpplications. Therefore, SAMs have recently become very impor-ant due to their potential applications to sensor nanotechnology

nd molecular electronics [9–15].

A sensitive electrochemical sensing protocol for the detectiony a surface molecular self-assembly strategy for the molecular

mprinting at the electrode was reported [16]. Polymer nanowires,

∗ Corresponding author. Tel.: +82 42 350 3919.E-mail address: whhong@kaist.ac.kr (W.H. Hong).

369-703X/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.bej.2011.01.008

the movement of hexadecanethiol molecules on the gold plate.© 2011 Elsevier B.V. All rights reserved.

silica nanotubes, and silica nanoparticles could be used in the sur-face molecular self-assembly strategy [17–19].

SAMs are one of the most important systems for investigat-ing the contributions of molecular structure and composition tothe macroscopic properties of materials. They provide organicsurfaces whose structure and properties can be varied. Controlover dimensions and properties make SAMs excellent systemsfor understanding the fundamentals of many natural phenom-ena. Moreover, their end group dependent properties can bemanipulated easily by various chemistries, thereby rendering themcompatible for the desired applications. A SAM is a layer of molec-ular thickness formed by the self-organization of molecules in anordered manner by chemisorptions on a solid surface. It is themost elementary form of nanometer-scale organic thin-film mate-rial, and is compromised of three significant parts: a surface-activehead group that bonds strongly to a substrate, an alkyl chain giv-ing stability to the assembly by van der Waals interactions, and thefunctionality of end group that plays an important role in terms ofcoupling of a biomolecule to a monolayer.

Alkanethiols spontaneously adsorbed onto gold surface havebeen used for SAM preparation. The self-assembly of alkanethiolson a gold surface occurs due to the strong chemical interac-tions between the sulfur end-group and the gold surface which isbelieved to result in chemical adsorption of the molecules as thi-olates [14]. The templating of SAMs is known as two-dimensional

molecular imprinting. When a SAM of alkanethiol is fabricated ona gold surface, the presence of foreign guest molecules and theirremoval result in holes in the SAM, which are complementary insize with the guest molecule, leading to the formation of bindingsites.

5 l Engi

op(icsf

1diu

2

2

ppdom

2

nmAfw1h1tpioprpeea

2

dt

2

s(iortrfitc

8 M.J. Shin, W.H. Hong / Biochemica

A molecular imprinting polymer-quartz crystal sensor basedn 2-aminoethanethiol has been reported for the detection of ter-ene gases up to 1000 ppm [20]. The quartz crystal microbalanceQCM) is a simple, cost effective, and high-resolution mass sens-ng instrument, which can measure the nanogram level of masshange loaded onto the surface of the QCM resonator. As a massensor, QCM has been widely used in biochemistry, environment,ood, and clinical analysis.

In this study, a SAM was formed on a gold plate with-hexadecanethiol, and a molecularly imprinted matrix was pro-uced using this monolayer. The potential of this molecularly

mprinted monolayer for application as a sensor was estimatedsing potentiostat and QCM.

. Experimental

.1. Materials

1-Hexadecanethiol, cholesterol, cholic acid, deoxycholic acid,otassium ferricyanide [K3Fe(CN)6], and sodium perchlorate wereurchased from Aldrich. 1-Hexadecanethiol was purified by doubleistillation before use, and other chemicals were used with-ut further purification. The chemical structures of the templateolecules in this study are shown in Scheme 1.

.2. Molecular imprinting method using cholesterol

The round-shaped pure Au plate (surface area: 1 cm2 × 2, thick-ess: 0.5 mm), which was connected with Au wire (2 cm), wasanufactured by a goldsmith’s shop. Before every experiment, theu plates were cleaned with ‘piranha’ solution, which was made

rom 20 mL of 30 wt% H2O2 and 60 mL of conc H2SO4. The Au plateas placed in the ‘piranha’ solution for 10 min. And then washed in

00 mL of distilled water for 1 h. To obtain a solution of 100 �M 1-exadecanethiol and 1 wt% cholesterol in ethanol, 0.0026 g (3.1 �L,0 �mol) of 1-hexadecanethiol and 0.79 g (2.04 mmol) of choles-erol was placed in 100 mL of ethanol. The cleaned Au plate waslaced in this solution for 12 h for coating, after which it was placed

n 50 mL of ethanol for 40 min in order to remove the excessivelyvercoated materials and adsorbed cholesterol at the surface of Aulate. A cholesterol-shaped molecular cavity was formed by theemoval of the cholesterol from the surface of the gold plate. Thisrocess was repeated three times using 50 mL of freshly preparedthanol each time. Both sides of the Au plate were used in thisxperiment so the Au plate was hung down into the solution duringll the processes.

.3. Molecular imprinting method using other template materials

The same method was used for imprinting cholic acid andeoxycholic acid by using cholic acid and deoxycholic acid, respec-ively, instead of cholesterol.

.4. Electrochemical measurement

Cyclic voltammograms were detected using an Ivium potentio-tat (Ivium Technologies, Netherlands). The round pure Au platesurface area: 1 cm2 × 2, thickness: 0.5 mm) was used as a work-ng electrode. All the electrochemical experiments were carriedut using a standard, one-compartment, three-electrode cell. Theeference electrode was Ag/AgCl (3 M KCl) and the counter elec-

rode was a platinum wire (10 cm). All electrode potentials wereeferred to the reference electrode. These three electrodes werexed accurately and tightly above the 50 mL reactor to ensurehe identical condition for every experiment. In this electrochemi-al experiment, potassium ferricyanide was used as a background

neering Journal 54 (2011) 57–61

material for oxidation and reduction. The main solution was madeby dissolving 0.0494 g (0.150 mmol) of potassium ferricyanide and0.1837 g (1.50 mmol) of sodium perchlorate in 15 mL of water and15 mL of ethanol. A cyclic voltammogram was conducted between−0.5 V and 0.5 V with a scan rate of 50 mV/s in 30 mL of 5 mM potas-sium ferricyanide and 50 mM sodium perchlorate of 50% aqueousethanol solution. The maximum current of the oxidation step andthe minimum current of the reduction step were recorded at everytime by adding the substrates.

The cholesterol standard solution (1.0 mM) was prepared by dis-solving 19.3 mg of cholesterol in 50 mL of ethanol. And 150 �L of thestandard cholesterol solution was added for cyclic voltammogramexperiment, so that the cholesterol concentration of the main solu-tion was increased 5 �M at every addition. The cholic acid standardsolution was prepared by dissolving 20.4 mg of cholic acid in 50 mLof ethanol. The deoxycholic acid standard solution was prepared bydissolving 19.6 mg of deoxycholic acid in 50 mL of ethanol.

The sensing stability of the imprinted monolayer was deter-mined by measuring the current gap between the results of adding1.05 mL of 1 mM cholesterol and 1.05 mL of 1 mM cholic acid.

2.5. QCM experiment

It is very difficult to measure the change of the absorbed amountafter every adsorption step and every cyclic voltammogram experi-ment using the Au plate directly, so a quartz resonator (USI System,Fukuoka, Japan) was used instead of the Au plate in the QCM tech-nique. The resonator was immersed in a solution for 12 h, removedfrom the solution and dried with nitrogen, after which the change infrequency was measured. The quartz resonators were covered withgold electrodes on both faces and their resonance frequency was9 MHz (AT-cut). The surface roughness factor of these electrodeswas previously estimated to be 1.1 (±5%) by scanning electronmicroscopy. The reproducibility was ±2 Hz over 2 h [21].

The increase in mass [M (g)] for adsorption was estimated fromthe QCM frequency shift [�F (Hz)] using the Sauerbrey equation[21]. The following equation was derived considering the resonatorcharacteristics:

�F = −1.832 × 108 × M

A

where A = 0.16 ± 0.01 cm2, which is the apparent area of themicrobalance electrodes. A 1 Hz change in �F corresponds to0.87 ng.

2.6. Measurement of adsorbed amount in the monolayer usingthe QCM technique

The surface of the resonator covered with gold was cleanedwith a piranha solution (sulfuric acid: hydrogen peroxide = 3:1,v/v), washed with distilled water and dried with nitrogen gas, afterwhich the frequency of the resonator was measured. The resonatorwas placed in the solution of 100 �M 1-hexadecanethiol and 1 wt%cholesterol in 100 mL of ethanol in order to coat the resonatorfor 12 h. The coated resonator was immersed in 50 mL of ethanolfor 40 min to remove the excessively overcoated materials andadsorbed cholesterol at the surface of the Au resonator. This processwas repeated three times using 50 mL of freshly prepared ethanoleach time, and dried with nitrogen gas. After measuring the change

in frequency, the resonator was again placed in 50 mL of ethanol for6 h, the resonator was dried with nitrogen gas, and the change infrequency was measured again. This process was repeated severaltimes. All the results were compared with the initially measuredfrequency after cleaning with piranha solution.

M.J. Shin, W.H. Hong / Biochemical Engineering Journal 54 (2011) 57–61 59

ctures of the template molecules.

3

3

haatamavatr

cpwicetparbraatar

Fi

0

20

40

60

80

100

120

140

100 20 30 40

Red

uce

d v

alu

e (

uA

)

Concentration (uM)

Scheme 1. The chemical stru

. Results and discussion

.1. Sensing capability of the molecularly imprinted monolayer

The sensing surface was formed by coating with 1-exadecanethiol and the template molecule on the gold surface,fter which the template molecule was extracted. This surface hadcavity of the same shape as that of the template molecule. The

emplate molecules used in this study were cholesterol, cholic acid,nd deoxycholic acid. The molecular structures of the templateolecules are shown in Scheme 1. Cholesterol was initially used

s a template molecule, and imprinted SAM was formed. A cyclicoltammogram was conducted in 5 mM potassium ferricyanidend 50 mM sodium perchlorate of 50% aqueous ethanol solu-ion using potassium ferricyanide as a background material foreduction and oxidation.

In Fig. 1, showing the cyclic voltammogram using theholesterol-imprinted monolayer, maximum oxidation currenteak was shown at 0.18 V, and minimum reduction current peakas shown at 0.05 V. Using the Au plate with the cholesterol-

mprinted SAM as a working electrode, the sensing system wasonstructed with the Ag/AgCl reference electrode and Pt counterlectrode. For each cyclic voltammogram, 150 �L of 1.0 mM choles-erol standard solution was added repeatedly to 30 mL of 5 mMotassium ferricyanide and 50 mM sodium perchlorate of 50%queous ethanol solution. The reducing value of the minimumeduction current was measured in comparison with the valueefore the addition of the standard cholesterol solution, and all theesults are recorded in Fig. 2. In addition, 150 �L of 1.0 mM cholic

cid standard solution and deoxycholic acid standard solution werelso added repeatedly to the same newly prepared solution andhe results are shown together in Fig. 2. And in order to comparend estimate the selectivity of the molecularly imprinted SAMs, theeducing value of the minimum reduction current was measured

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

-0.6 -0.4 -0.2 0 0.2 0.4 0.6

Cu

rren

t (m

A)

Potential (V)

ig. 1. Cyclic voltammograms for the gold plate electrode with a cholesterol-mprinted monolayer.

Fig. 2. Dependence of the sensor response of the gold electrode, imprinted withcholesterol, on the concentration of cholesterol (�), cholic acid (�), and deoxycholicacid (�).

using the bare gold electrode without SAM. These results are shownin Fig. 3.

In Fig. 2, cholesterol exhibited a larger reduced value than theother substrates, which was attributed to the fact that cholesterol isfitted to the cavity on the SAM. And comparing the reduced valueson the concentration of cholesterol in Fig. 2 with those in Fig. 3,the reduced values in Fig. 2 is almost three times larger than thosein Fig. 3. These results showed that molecularly imprinted SAMscould be worked as a sensing tool.

Using the Au plate with the cholic acid-imprinted SAM as aworking electrode, the sensing system was also constructed withthe reference and counter electrodes. The conducted experimentswere almost identical to those with the cholesterol-imprinted

SAM and the results are shown in Fig. 4. Cholic acid exhibiteda larger reduced value than the other substrates. However, thegaps between the cholic acid result and those of the other sub-strates were smaller than those presented in Fig. 2. Using the Auplate with the deoxycholic acid-imprinted SAM as a working elec-

0

10

20

30

40

50

100 20 30 40

Red

uct

ion

val

ue

(u

A)

Concentration (uM)

Fig. 3. Dependence of the sensor response of the bare gold electrode without SAMon the concentration of cholesterol (�), cholic acid (�), and deoxycholic acid (�).

60 M.J. Shin, W.H. Hong / Biochemical Engineering Journal 54 (2011) 57–61

0

20

40

60

80

100

100 20 30 40

Red

uce

d v

alu

e (

uA

)

Concentration (uM)

Fca

taatrbsgvi

3

ii1uvcvdfcstitt

Fdc

0

10

20

30

40

50

60

120 24 36 48 60 72 84 96 108 120 132 144 156 168

Gab

of

the

red

uce

d v

alu

e (

uA

)

Time (hr)

Fig. 6. The gap of the reduced value between using cholesterol and using cholic acid,at 35 �M of substrate for the cholesterol-imprinted monolayer.

0

10

20

30

40

120 24 36 48 60 72 84 96 108 120 132 144 156 168Gab

of

the

red

uce

d v

alu

e (

uA

)

ig. 4. Dependence of the sensor response of the gold electrode, imprinted withholic acid, on the concentration of cholesterol (�), cholic acid (�), and deoxycholiccid (�).

rode, the sensing system was also constructed with the referencend counter electrodes. Again, the conducted experiments werelmost identical to those with the cholesterol-imprinted SAM andhe results are shown in Fig. 5. Deoxycholic acid exhibited a largereduced value than the other substrates. However, again, the gapsetween the deoxycholic acid results and those of the other sub-trates were smaller than those presented in Fig. 2. Especially theaps between using deoxycholic acid and using cholic acid wereery small. These results were attributed to the very similar chem-cal structure of deoxycholic acid and cholic acid.

.2. Stability of the molecularly imprinted monolayer as a sensor

In order to check the sensing stability of the molecularlymprinted monolayer, after the cyclic voltammogram test was fin-shed, the molecularly imprinted monolayer was kept in ethanol for2 h and the cyclic voltammogram test was repeated. For examplesing the cholesterol-imprinted monolayer, the reduced currentalues as either 1.05 mL of 1 mM cholesterol or 1.05 mL of 1 mMholic acid was added were compared. The gap between these twoalues was recorded, the experiment was repeated every 12 h for 7ays, and all the results are recorded in Fig. 6. The gap was reducedrom 53.8 �A initially to 17.5 �A after 168 h. The reducing trendontinued, but at a slower rate over time. Another experimental

et was conducted using the cholic acid-imprinted monolayer andhe results are recorded in Fig. 7. The gap was reduced from 38.2 �Anitially to 0.7 �A after 168 h. The reducing trend continued even ashe gap approached zero. In conclusion, these results revealed thathe monolayer was not perfectly stable under this experimental

0

20

40

60

80

100

100 20 30 40

Red

uct

ion

val

ue

(u

A)

Concentration (uM)

ig. 5. Dependence of the sensor response of the gold electrode, imprinted witheoxycholic acid, on the concentration of cholesterol (�), cholic acid (�), and deoxy-holic acid (�).

Time (hr)

Fig. 7. The gap of the reduced value between using cholic acid and using cholesterol,at 35 �M of substrate for the cholic acid-imprinted monolayer.

condition. Therefore, we investigated the stability of the mono-layers by QCM experiment. As gold plate that had been used inthis study could not be used directly in the QCM experiment, theresonator was used instead to check the stability of the molecu-larly imprinted monolayer. Due to the gold-plated surface of theresonator, these results were the same as those with the gold plate.

The resonator was cleaned with piranha solution and the fre-quency of the resonator was measured by QCM. The resonatorwas coated with 1-hexadecanethiol and cholesterol to form themolecular monolayer on the resonator. In this procedure, the excess

1-hexadecanethiol attached on the gold surface had to be removedbefore the frequency was measured again and the frequency shiftwas recorded, after which the resonator was placed in ethanol for12 h, and the frequency shift was recorded again. This process wasrepeated several times, and the results are shown in Fig. 8. The 1-

0

50

100

150

200

120 24 36 48 60 72 84 96 108 120 132 144 156 168

Fre

qu

ency

sh

ift

Time (hr)

Fig. 8. The frequency shift of the coated resonator after storing in ethanol.

l Engi

hs7Tpceo

4

homtuata

A

fbRS

R

[

[

[

[

[

[

[

[

[

[

M.J. Shin, W.H. Hong / Biochemica

exadecanethiol molecules on the resonator remained on the goldurface even until the end of the experiment at 7 days. Even afterdays, only a small amount of 1-hexadecanethiol was reduced.

his indicated that the 1-hexadecanethiol molecules on the goldlate were attached tightly on the surface, suggesting that the senseapability was reduced after the imprinted monolayer was kept inthanol due to the movement of the 1-hexadecanethiol moleculesn the gold surface of the resonator.

. Conclusion

A molecularly imprinted SAM was formed with 1-exadecanethiol. The template molecules and substrates consistedf cholesterol, cholic acid, and deoxycholic acid. This imprintedonolayer discriminated among the substrates, with the choles-

erol substrate being better recognized than the other substratessing the cholesterol-imprinted monolayer. However, the sensingbility was reduced with repeated use for substrate detection. Onhe basis of QCM experiments, the reduced sensing ability wasttributed to the movement of the molecules in the SAM.

cknowledgements

This work was supported by the IT Leading R&D Support Projectrom the Ministry of Knowledge Economy (Korea) through KEIT andy WCU (World Class University) program through the Nationalesearch Foundation of Korea funded by the Ministry of Education,cience and Technology (R32-2008-000-10142-0).

eferences

[1] R. Gupta, A. Kumar, Molecular imprinting in sol–gel matrix, Biotechnol. Adv.26 (2008) 533–547.

[2] A.G. Mayes, K. Mosbach, Molecularly imprinted polymers: useful mate-rials for analytical chemistry, Trends Anal. Chem. 16 (1997) 321–332.

[3] S.W. Lee, D.H. Yang, T. Kunitake, Regioselective imprinting of anthracenecar-boxylic acids onto TiO2 ultrathin films: an approach to thin film sensor, Sens.Actuators B 104 (2005) 35–42.

[4] D.H. Yang, N. Takahara, S.W. Lee, T. Kunitake, Fabrication of glucose-sensitiveTiO2 ultrathin films by molecular imprinting and selective detection ofmonosaccharides, Sens. Actuators B 130 (2008) 379–385.

[

[

neering Journal 54 (2011) 57–61 61

[5] D.H. Yang, Y.R. Ham, M.H. Oh, Y.S. Yoon, J.S. Shin, Y.D. Kim, H. Kim, Simplemethod for the fabrication of 1-hydroxypyrene-imprinted TiO2 gel nanofilms,Curr. Appl. Phys. 9 (2009) e136–e139.

[6] T. Takeuchi, T. Mukawa, H. Shinmori, Signaling molecularly imprinted poly-mers: molecular recognition-based sensing materials, Chem. Record 5 (2005)263–275.

[7] S.A. Piletsky, E.V. Piletskaya, A.V. Elgersma, K. Yano, I. Karube, Y.P. Parhometz,A.V. El’skaya, Atrazine sensing by molecularly imprinted membranes, Biosens.Bioelectron. 10 (1995) 959–964.

[8] Y. Wang, Y. Zhou, J. Sokolov, B. Rigas, K. Levon, M. Rafailovich, A potentiomet-ric protein sensor built with surface molecular imprinting method, Biosens.Bioelectron. 24 (2008) 162–166.

[9] L.I. Andersson, C.F. Mandenius, K. Mosbach, Studies on guest selective molec-ular recognition on an octadecyl silylated silicon surface using ellipsometry,Tetrahedron Lett. 29 (1988) 5437–5440.

10] S.K. Arya, A.K. Prusty, S.P. Singh, P.R. Solanki, M.K. Pandey, M. Katta, B.D.Malhotra, Cholesterol biosensor based on N-(2-aminoethyl)-3-aminopropyl-trimethoxysilane self-assembled monolayer, Anal. Biochem. 363 (2007)210–218.

11] S.K. Arya, P.R. Solanki, M. Datta, B.D. Malhotra, Recent advances in self-assembled monolayers based biomolecular electronic devices, Biosens.Bioelectron. 24 (2009) 2810–2817.

12] P.Y. Chen, P.C. Nien, C.T. Wu, T.H. Wu, C.W. Lin, K.C. Ho, Fabrication of a molec-ularly imprinted polymer sensor by self-assembling monolayer/mediatorsystem, Anal. Chim. Acta 643 (2009) 38–44.

13] L.C.S. Chou, C.C. Liu, Development of a molecular imprinting thick film elec-trochemical sensor for cholesterol detection, Sens. Actuators B 110 (2005)204–208.

14] S.A. Piletsky, E.V. Piletskaya, T.A. Sergeyeva, T.L. Panasyuk, A.V. El’skaya, Molec-ularly imprinted self-assembled films with specificity to cholesterol, Sens.Actuators B 60 (1999) 216–220.

15] J. Sagiv, Organized monolayers by adsoption 1. Formation and structure ofoleophobic mixed monolayers on solid surfaces, J. Am. Chem. Soc. 102 (1980)92–98.

16] C. Xie, H. Li, S. Li, J. Wu, Z. Zhang, Surface molecular self-assemblyfor organophosphate pesticide imprinting in electropolymerized poly(p-aminothiophenol) membrane on a gold nanoparticle modified glassy carbonelectrode, Anal. Chem. 82 (2010) 241–249.

17] C. Xie, B. Liu, Z. Wang, D. Gao, G. Guan, Z. Zhang, Molecular imprinting at wallsof silica nanotubes for TNT recognition, Anal. Chem. 80 (2008) 437–443.

18] D. Gao, Z. Zhang, M. Wu, C. Xie, G. Guan, D. Wang, A surface functionalmonomer-directing strategy for highly dense imprinting of TNT at surface ofsilica nanoparticles, J. Am. Chem. Soc. 129 (2007) 7859–7866.

19] C. Xie, Z. Zhang, D. Wang, G. Guan, D. Gao, J. Liu, Surface molecular self-assemblystrategy for TNT imprintinf of polymer nanowire/nanotube arrays, Anal. Chem.78 (2006) 8339–8346.

20] M. Kikuchi, N. Tsuru, S. Shiratori, Recognition of terpenes using molecularimprinted polymer coated quartz crystal microbalance in air phase, Sci. Tech-nol. Adv. Mater. 7 (2006) 156–161.

21] Y. Lvov, K. Ariga, I. Ichinose, T. Kunitake, Assembly of multicomponent proteinfilms by means of electrostatic layer-by-layer adsorption, J. Am. Chem. Soc. 117(1995) 6117–6123.