Conductive polymer composites as gas sensors with size-related molecular discrimination capability

Sensors and Actuators B 124 (2007) 118–126

Conductive polymer composites as gas sensors with size-related molecular discrimination capability Wei Zeng a,b , Ming Qiu Zhang b,∗ , Min Zhi Rong b , Qiang Zheng c a

PCFM Lab, OFCM Institute, School of Chemistry and Chemical Engineering, Zhongshan University, Guangzhou 510275, PR China b Materials Science Institute, Zhongshan University, Guangzhou 510275, PR China c Department of Polymer Science and Engineering, Zhejiang University, Hangzhou 310027, PR China Received 23 July 2006; received in revised form 2 December 2006; accepted 4 December 2006 Available online 19 December 2006

Abstract To prepare a novel gas sensor with selective sensitivity, waterborne ␤-cyclodextrin-block-polydiethylene glycol hexandioic ester copolymer was synthesized and blended with carbon black (CB). It was found that electric resistance of the composites remarkably increased upon exposure to vapors of low permittivity (like chloroform and tetrahydrofuran), while nearly no response can be detected to vapors of high permittivity (like water and methanol). What is more interesting is that the composite maximum responsivity was correlated to the molecular size of vapors of low permittivity on a half-logarithmic scale. With increasing the molecular size of the low permittivity gaseous analytes, the composite responsiveness gradually decreases. On the basis of this feature, molecular discrimination can be easily conducted. Mechanism study indicated that ␤-cyclodextrin rather than polydiethylene glycol hexandioic ester in the matrix polymer played the leading role in this aspect, which was explained from the viewpoint of host–guest chemistry and the hydrophobicity of the ␤-cyclodextrin cavities as well. © 2006 Elsevier B.V. All rights reserved. Keywords: Gas sensor; Molecular discrimination; Beta-cyclodextrin; Carbon black; Electrical properties; Polymer composites

1. Introduction Recently, gas sensors or chemically sensitive materials made from conductive polymer composites have received significant attention for use in detecting, quantifying and discriminating various organic vapors [1–11]. Their working principle is based on the fact that different composites give different levels of response to a vapor of interests in terms of swelling induced resistance variation. For example, carbon black composite sensor arrays can easily distinguish between a chemically diverse set of analytes and mixtures consisting of two chemically similar analytes [12]. The selectivity of these devices is generally tailored via use of a wide range of polymeric materials available. Nevertheless, poor selectivity is still a serious drawback that has to be overcome in application. Development of highly selective materials is one of the most important targets of this area [13–16].

∗

Corresponding author. Tel.: +86 20 84036576; fax: +86 20 84036576. E-mail address: [email protected] (M.Q. Zhang).

0925-4005/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.snb.2006.12.021

It is known that cyclodextrins (CDs) possess hydrophobic cavities, enabling encapsulation of diverse small organic molecules by forming inclusion complexes [17–20]. Quantification of pairwise interactions and ultimate control over them through manipulation of the microenvironment have proved useful not only in the design of new synthetic host–guest complexes but also, for example, in the rational synthesis of biologically active substances, molecular switches and catalysts that function analogously to enzymes [21,22]. As the interactions belong to both chemisorption and physisorption, cyclodextrins can be used to resolve the contradiction between selectivity and reversibility of signal response. Considering the above characteristics of cyclodextrins, it is natural to associate them with selectivity improvement of gas sensors. That is, when cyclodextrins are introduced into a gas sensing material, the specific adsorption habit of their hydrophobic cavities towards organic molecules might provide well-defined response behavior. To examine the feasibility of this idea, carbon black (CB)-filled waterborne ␤-cyclodextrin-blockpolydiethylene glycol hexandioic ester copolymer (abbreviated to waterborne ␤-CD-block-PDEA copolymer) composites were

W. Zeng et al. / Sensors and Actuators B 124 (2007) 118–126

119

prepared in the present work. Our earlier experiments [23] demonstrated that CB/waterborne polyurethane (WPU) composites were sensitive to many organic solvent vapors regardless of their polarities as characterized by the drastic changes in conductivity because of the coexisting nonpolar and polar segments on the polyurethane chains. By replacing the hard segments of the matrix waterborne polyurethane with ␤-CDs, it is believed that the composites would perform novel functions as deduced hereinbefore. For evaluating the gas sensing capability of the composites, solvent vapors with different permittivity were used. The reason for characterizing the solvents in terms of permittivity lies in the fact that the hydrophobicity of cyclodextrins used to be described by permittivity [24,25]. Here permittivity acts as a comprehensive indicator of solvents’ polarity and hydrophobicity (or hydrophilicity). Mostly, lower permittivity suggests lower polarity and higher hydrophobicity (or lower hydrophilicity), and vice versa. It is worth noting that, however, there exist some exceptional cases. This is because the molecular dipolarity is based on vectors and the higher polar characteristics such as quadropoles and octapoles.

For structural study of the copolymer, purification process was conducted as follows. The emulsion was first dried in vacuum at 105 ◦ C for 72 h to remove water, and then the copolymer was extracted with ether for 72 h to yield the purified product. The reactivity of hydroxyl groups on ␤-CD ranks in the following order: 6-OH  2-OH > 3-OH, while its acidity is differently arranged: 2-OH > 3-OH  6-OH. Therefore, suitably controlling the reaction conditions will yield cyclodextrin derivatives with certain selectivity. In fact, here the waterborne ␤-CD-block-PDEA copolymer was synthesized though three steps (Scheme 1). PDEA first reacted with excessive TDI so that the terminal hydroxyl groups of PDEA were replaced by NCO. When ␤-CD was added into the reaction system, it had to react with the excessive TDI because of the high reactivity of its 2,3-hydroxyls, and then the 2,3-hydroxyls of ␤-CD were replaced by NCO. Afterwards, the right amount of DMPA was incorporated and the hydroxyls of DMPA reacted with NCO of PDEA and ␤-CD, producing the anticipated polymer.

2. Experimental

For fabricating conductive composites, a certain amount of carbon black was added to the above waterborne ␤-CDblock-PDEA copolymer emulsion, and then the mixture was ultrasonically agitated for 0.5 h and stirred for additional 4 h. The composite films were obtained by dropping five drops of the composite latex onto epoxy plates with comb electrodes. Prior to further testing, the conductive composite films were dried at room temperature for 2 days, and then put in a vacuum oven at 50 ◦ C for additional 2 days. After that, the composite films of 40–60 ␮m thick (determined by a micrometer) were stored in a desiccator for further electrical measurements. Electrical response of the composites to solvent vapors was detected by measuring dc resistance variation at 30 ◦ C using the homemade equipment described elsewhere [26]. The measurement of electrical resistance of the composites in response to various organic solvent vapors was carried out as follows. The composite resistor was hung in a glass test container filled with organic solvent vapor provided by a glass gas storage holder (30 times larger than the test container in volume) and then the time dependence of composite resistance was recorded by a UT70C digital multimeter, and vapor pressures were calculated by the thermodynamic equation. The responsivity is characterized by (Rt − R0 )/R0 , where Rt denotes the transient resistance and R0 is the initial one in dry air. Besides, the maximum responsivity is given by (Rmax − R0 )/R0 , where Rmax stands for the maximum resistance in organic vapor. Both 1 H NMR and 13 C NMR spectra of the resultant polymer were collected in DMSO-d6 at 500 MHz by a Varian INOVA 500NB spectrometer under ambient temperature. Chemical shifts were referenced by tetramethylsilane. AFM measurements were made on Shimadzu SPM-9500Z3 apparatus, and the tapping mode was used for the measurements. The filmy samples were prepared by spinning–coating on mica. The obtained phase image is viscoelasticity-dominant one, and

2.1. Materials PDEA (molecular weight = 10,000) that acted as the soft segments of the waterborne polyurethane was supplied by Shan Feng Polyurethane Co., China. It was dried at 100 ◦ C in vacuum for 24 h before use. Analytically pure ethylenediamine anhydrous (EDA), triethylamine (TEA) and dibutyltin dilaurate were dehydrated via 0.4 nm molecular sieves for more than 1 week prior to the experiments. Dimethylolpropionic acid (DMPA) was dried in vacuum at 80 ◦ C for 24 h. Toluene diisocyanate (TDI) was used without further purification. Carbon black (type XC72, specific surface area = 254 m2 /g, DPB value = 174 ml/100 g, particle size = 50–70 nm), purchased from Cabot Co., was dried in vacuum at 110 ◦ C for 48 h before use. 2.2. Synthesis of β-CD-block-PDEA copolymer Synthesis of waterborne ␤-CD-block-PDEA copolymer was prepared by stepwise reactions starting from PDEA and ␤-CD. A typical procedure was as follows. PDEA (0.001 mol) was heated up to 120 ◦ C in vacuum for 1 h, and then the system was filled up with nitrogen. When the temperature was cooled to 80 ◦ C, 0.005 mol of TDI were added. After 2 h of reaction, 0.001 mol of ␤-CD was added into the mixture. Along with the unceasingly increase of viscosity, 30 ml N,N-dimethylformamide (DMF) was also added. Then 0.004 mol of DMPA (dissolved by small amount of DMF) was added into the system. The reaction proceeded for additional 3 h. When the system was cooled down to room temperature, the pre-polymer reacted with 0.004 mol of TEA for 1 h. Then, 100 ml distilled water and 0.001 mol of EDA were added to the pre-polymer with high-speed stirring. Twelve hours later, an emulsion of waterborne ␤-CD-blockPDEA copolymer was obtained.

2.3. Preparation and characterization of the composites

120

W. Zeng et al. / Sensors and Actuators B 124 (2007) 118–126

Scheme 1. Possible mechanism of the synthesis of waterborne ␤-CD-block-PDEA copolymer.

elasticity and viscosity cannot be separated. A part with big delay of phase (high elasticity and high viscosity) is displayed brightly. Scanning electron microscopic (SEM) observation of the cryofractured surface of the materials was conducted by a Quanta 400F thermal field emission environmental SEM-EDS-EBSD equipment.

exhibits a typical microphase separation structure. That is, the congregated hard segments (␤-CDs) disperse in the continuous phase of the soft segments (PDEA). The intermolecular hydrogen bonds between the residuary hydroxyl groups at the side of ␤-CDs should account for the congregation of the hard segment. Having been mixed with CB, ␤-CD agglomerates together with

3. Results and discussions Both 1 H NMR and 13 C NMR spectra of the resultant copolymer are shown in Figs. 1 and 2. By comparing the 1 H NMR spectrum of the copolymer with that of ␤-CD, it is found that the signal of 2,3-OH of ␤-CD disappears in the former, while the signal of 6-OH of the copolymer becomes lower than that of ␤-CD. It means that the 2,3-OH of ␤-CD has been replaced by NCO, and 6-OH has not been completely reacted with NCO. In addition, the characteristic signals of TDI and PDEA can be seen in both 1 H NMR and 13 C NMR spectra of the copolymer. These results indicate that the product of interests is the copolymer of ␤-CD and PDEA. AFM phase images of waterborne ␤-CD-block-PDEA copolymer and its composites with CB are shown in Fig. 3. As the dark region in such phase image represents a hard substance, while the bright corresponds to soft one, it is known from Fig. 3(a) that the waterborne ␤-CD-block-PDEA copolymer

Fig. 1. 1 H NMR spectra of ␤-CD and waterborne ␤-CD-block-PDEA copolymer in DMSO-d6 .

W. Zeng et al. / Sensors and Actuators B 124 (2007) 118–126

121

Fig. 4. Electric resistivity, ρ, of CB/␤-CD-block-PDEA copolymer composites as a function of CB content. Fig. 2. 13 C NMR spectrum of waterborne ␤-CD-block-PDEA copolymer in DMSO-d6 .

the filler particles (Fig. 3(b)). Compared to PDEA, ␤-CD seems to have preferential affinity for CB. Fig. 4 illustrates the dependence of the composite resistivity on the CB concentration. Clearly, the percolation threshold is rather low, ∼0.5 wt%. Since our previous studies suggested that the optimum responsiveness appeared at a CB content much higher than the percolation threshold [27,28], the CB content of all the composites used in this work was fixed at 3.0 wt%. On the basis of the above work, it is necessary to check whether the composites prepared in the present work have acceptable gas sensibility. Fig. 5 gives a group of response curves of the composites to dichloromethane vapor at low vapor concentration. It is seen that when the composites are exposed to the solvent vapor, the resistance increases rapidly as expected. Moreover, the composite conductivity is able to quickly return to its initial value as the specimen is removed from the vapor environment to dry air. The data clearly demonstrate that the composites are qualified for acting as gas sensors.

In fact, the structure of a freshly formed sample film is metastable. As the composite is exposed to a solvent vapor during the tests, repeated swelling and de-swelling would help to relax the composite like annealing. Fig. 5 indicates that the present composite approaches the stabilized state only after one cycle of the test. This means the composite is rather easy to be handled for routine usage. To further explore the composite performance, a wide spectrum of solvent vapors was used. As stated in Section 1, the hydrophobic cavities of ␤-CDs tend to encapsulate organic molecules via forming inclusion complexes. Therefore, the composites were first tested by a series of solvent vapors with low permittivity. It is interesting to find from Fig. 6 that the response behavior of the composites is far from what assumed by the authors despite the fact that the composites exhibit different levels of response. In other words, the maximum responsivity of the composites nearly has nothing to do with the solvent permittivity. Although CHCl3 and ethyl ether have similar permittivity, for example, the composites generate completely different magnitudes of resistance variation in the two kinds of vapors. The similar case can also be observed for benzene and cyclohexane

Fig. 3. AFM phase images of the polymer and its composites filled with 1 wt% CB.

122

W. Zeng et al. / Sensors and Actuators B 124 (2007) 118–126

Fig. 5. Electrical responses of CB/␤-CD-block-PDEA copolymer composites to dichloromethane vapor at a concentration of 50 ppt (CB content: 3 wt%). The dash lines define the vapor absorption and desorption zones.

vapors. The rule “like dissolves like” seems to be invalid in the present work. Fig. 7 collects the composite responses against the vapors of aromatic compounds, which again evidence this estimation. The significant differences between the maximum responsivity of the composites should not be controlled mainly by the vapor permittivity. Re-considering the structural feature of ␤-CD in the matrix polymer, its hydrophobic cavity might be a key issue that affects the gas sensing manners of the composites. In this context, the size effect of the analytes should be analyzed. Fig. 8 plots the dependence of the maximum responsivity on both the permittivity and the molecular diameter (that was determined from the actual molecular volume [29]) of the vapors. Evidently, the maximum responsivity is closely dependent on the molecular size of vapors. With a rise in molecular size of the organic solvents, the composite responsivity gradually decreases on a half-logarithmic scale. The vapors with smaller molecular size arouse higher electrical responses of the composites. It explains the data in Fig. 7, where the aromatic

Fig. 6. Electrical responses of CB/␤-CD-block-PDEA copolymer composites to low permittivity vapors at a concentration of 100 ppt (CB content: 3 wt%). The numerals in the parentheses behind the solvent names are the permittivity values of the solvents. The dash line defines the vapor absorption and desorption zone.

Fig. 7. Electrical responses of CB/␤-CD-block-PDEA copolymer composites to vapors of aromatic compounds at a concentration of 100 ppt (CB content: 3 wt%). The numerals in the parentheses behind the solvent names are the permittivity values of the solvents. The dash line defines the vapor absorption and desorption zone.

compounds have similar permittivity but different molecular sizes. The composite response to benzene vapor is the highest because its molecular size is the smallest. In the case of larger substituting groups, the maximum responsivity reduces accordingly. Hence the vapor of ethyl benzene results in the smallest responsivity. In contrast, Fig. 8 shows that there is no perceivable correlation between the responsivity and the vapor permittivity, which coincides with the phenomena reflected by Figs. 6 and 7. Therefore, the size/shape fit between host and guest that contributes to the formation of the inclusion complex in ␤-CD cavities might be the most important factor determining the selective sensitivity of the composites. Effect of the soft segments of the matrix polymer of the composites, which lack molecular size-dependent absorbability towards solvents, can be ignored in the present system.

Fig. 8. Effects of molecular size and permittivity of low permittivity solvent vapors on the maximum electrical responsivity of the composites recorded at a vapor concentration of 100 ppt (CB content: 3 wt%).

W. Zeng et al. / Sensors and Actuators B 124 (2007) 118–126

Fig. 9. Effects of molecular size and permittivity of high permittivity solvent vapors on the maximum electrical responsivity of the composites recorded at a vapor concentration of 100 ppt (CB content: 3 wt%).

For yielding more knowledge about the size-related response behavior of the composites, the vapors of high permittivity solvents were also employed for testing. As illustrated in Fig. 9, the resistance of the composites shows inconspicuous variation in these vapors. Compared to the case of low permittivity vapors (Fig. 8), the maximum responsivity given in Fig. 9 is negligible, especially to water vapor. In fact, for high permittivity vapors except water, the composite responsivity is independent on both the permittivity and the molecular size of the vapors. The extremely low responsivity to water vapor should result from its remarkably high permittivity, which makes the water molecules very difficult to enter those hydrophobic cavities in ␤-CDs. In general, the hydroxyl or carboxyl groups of the high permittivity solvent molecules would form intermolecular hydrogen bonds with the hydroxyls on both sides of ␤-CDs besides interstitial bonding [30]. This means that only when the cavities of ␤-CDs in the composite matrix succeed in encapsulating solvent molecules, the composite resistance would change significantly. Other interactions between the matrix and the solvents are not useful enough. The above discussion clearly shows the importance of the matrix polymer of the composites, or more accurately, the importance of ␤-CD. When the composites meet solvent vapors, the three-dimensional hydrophobic cavities of cyclodextrins would encapsulate neutral and charged organic molecules to form inclusion complex through the size filling effect and/or by various weak forces such as van der Waals, hydrophobic, hydrogen bonding, ion–dipole and dipole–dipole interactions. The including ability of cyclodextrins is decided by guest molecular structure, size and hydrophobicity, etc. Low permittivity compounds with molecular sizes matching cyclodextrin cavity size can enter the hydrophobic cavities and cause localized volume expansion, leading to partial disconnection of the conduction paths built by CB particles. Consequently, the composite resistance drastically increases upon exposure to the vapors of these solvents. Usually a good match leads to stronger van der Waals and hydrophobic interactions because the strengths of these two

123

Fig. 10. 1 H NMR spectra of ␤-CD-block-PDEA copolymer in the presence of benzene or ethanol in DMSO at 30 ◦ C: (a) ␤-CD-block-PDEA copolymer (3 × 10−3 M); (b) ␤-CD-block-PDEA copolymer (3 × 10−3 M) + benzene (3 × 10−3 M); (c) ␤-CD-block-PDEA copolymer (3 × 10−3 M) + ethanol (3 × 10−3 M).

types of interaction are closely related to the distance and contact surface area between host and guest. To study the structure of the inclusion complexes, 1 H NMR spectra of the matrix ␤-CD-block-PDEA copolymer in dimethyl sulfoxide (DMSO) and its solution in benzene (low permittivity solvent) or ethanol (high permittivity solvent) were measured (Fig. 10). The chemical shifts of the protons in the polymer are summarized in Table 1. As a result of the formation of the polymer–benzene inclusion complex, the upfield shift of 3-H is rather prominent, followed by 6-H and 5-H. These upfield shifts mainly result from anisotropic shielding by ring current from the aromatic rings of benzene. Comparatively, the upfield shifts of the signals of 1-H, 2-H and 4-H that lie on the outer surface of the cyclodextrin cavities are not observed, suggesting that benzene has been included into the cavities (Scheme 2). After the low permittivity molecules have been included in the cavities of ␤-CD, the latter would adjust their shape and volume to adapt the guest molecules by expansion, leading to longitudinal stretch [31]. Since CB particles are mainly distributed in the hard segments phase (i.e., ␤-CDs) of the matrix polymer (Fig. 3(b)) and their agglomerates are well broken apart (the CB size is about a couple of tens nanometer as reflected by the Table 1 Chemical shifts of the protons of ␤-CD-block-PDEA copolymer in the presence of benzene and ethanol Proton no.

1 2 3 4 5 6

δ (ppm) in the presence of Benzene

Ethanol

0.001 0.001 −0.0224 0 −0.0005 −0.0012

0 0 0 0 −0.005 −0.004

124

W. Zeng et al. / Sensors and Actuators B 124 (2007) 118–126

Scheme 2. The benzene inclusion model of waterborne ␤-CD-block-PDEA copolymer.

Fig. 11. SEM micrographs of cryofractured surfaces of the polymer and its composites filled with 3 wt% CB.

SEM photo in Fig. 11), the volumetric change of the neighboring ␤-CDs would easily cause small scale redistribution of the conductive particulates and a rapid increase in composite resistance appears accordingly. This process provides measurable signals characterizing the surrounding vapor-phase analytes. On the other hand, the volumes of organic vapor molecules ˚ 3 ) are much smaller than that of the ␤-CD cavity (17.8–28.6 A 3 ˚ ). At least a couple of guest molecules can be encapsu(262 A lated by each cavity of ␤-CD. The amount of the encapsulated guest molecules should determine the maximum responsiveness of the composites. Therefore, the composite responsivity to vapors with smaller molecules is higher than that to vapors with larger molecules (Fig. 8). When the composites are removed from solvent vapors to dry air, the low permittivity molecules would be disengaged from the cavities of ␤-CDs because of the weak interactions

between cyclodextrins and the guest molecules. The structure of the cyclodextrins recovers accordingly, so that the composite resistance easily returns to its initial value. Unlike the low permittivity molecules, the high permittivity ones are difficult to enter the hydrophobic cavities of cyclodextrins because of lack of interaction between the solvent molecules and ␤-CDs, which is supported by the data in Table 1. The upfield shifts of the signals of 5-H and 6-H in the polymer–ethanol inclusion complex are inconspicuous, while those of 3-H lying on the inner surface of the secondary hydroxyl group side are zero. In the concrete, hydroxyl groups of ethanol might form intermolecular hydrogen bonds with the residuary 6OH at the top of ␤-CDs, and the ethyl of ethanol would be slightly embedded in the cavities from the top of ␤-CDs (Scheme 3). This accounts for the marginal upfield shift of 5,6-H revealed in Fig. 10(c). Since the partly included ethanol cannot influence

Scheme 3. The ethanol inclusion model of waterborne ␤-CD-block-PDEA copolymer.

W. Zeng et al. / Sensors and Actuators B 124 (2007) 118–126

the 3-H located in the inner cavity of ␤-CDs, no upfield shift of 3-H is perceived. Besides, because there is no more residuary 2,3-OH at the bottom of ␤-CDs (which were completely consumed during the reaction with TDI), the interaction between ethanol molecules and 2,3-OH could not exist. As a result, the composite electrical response to this vapor is poor.

[8]

[9]

4. Conclusions

[10]

On the basis of above investigation, the following conclusions can be drawn:

[11]

1. Carbon black-filled waterborne ␤-CD-block-PDEA copolymer composites prove to be selectively sensitive to organic vapors. The composite electric resistance remarkably increases when exposed to the vapors with low permittivity. However, nearly no response can be detected to the vapors with high permittivity. 2. The composite maximum responsivity is a function of molecular size of low permittivity solvent vapors. With a rise in molecular size of the vapors, the composite response decreases on a half-logarithmic scale. By calibrating this relationship, one is able to discriminate and screen molecules of the related organic solvents. 3. The size/shape fit between host and guest that results in the inclusion complex between ␤-CD and the vapor-phase analytes, and the hydrophobicity of the cavities in ␤-CD as well, accounts for the selective response behavior.

[12] [13]

[14]

[15]

[16] [17]

[18]

Acknowledgements Financial supports by the Natural Science Foundation of China (grant 50133020) and the Team project of the Natural Science Foundation of Guangdong (grant 20003038) are gratefully acknowledged. References [1] J. Chen, N. Tsubokawa, Electric properties of conducting composite from poly(ethylene oxide) and poly(ethylene oxide)-grafted carbon black in solvent vapor, Polym. J. 32 (2000) 729–736. [2] B.J. Doleman, M.C. Lonergan, E.J. Severin, T.P. Vaid, N.S. Lewis, Quantitative study of the resolving power of arrays of carbon black–polymer composites in various vapor-sensing tasks, Anal. Chem. 70 (1998) 4177–4190. [3] B.J. Doleman, E.J. Severin, N.S. Lewis, Trends in odor intensity for human sind electronic noses: relative roles of odorant vapor pressure vs. molecularly specific odorant binding, Proc. Natl. Acad. Sci. U.S.A. 95 (1998) 5442–5447. [4] J.H. Chen, N. Tsubokawa, A novel gas sensor from polymer-grafted carbon black: responsiveness of electric resistance of conducting composite from LDPE and PE-b-PEO-grafted carbon black in various vapors, Polym. Adv. Technol. 11 (2000) 101–107. [5] F. Zee, J.W. Judy, Micromachined polymer-based chemical gas sensor array, Sens. Actuators B: Chem. 72 (2001) 120–128. [6] J.F. Feller, Y. Grohens, Evolution of electrical properties of some conductive polymer composite textiles with organic solvent vapours diffusion, Sens. Actuators B: Chem. 97 (2004) 231–242. [7] X.M. Dong, R.W. Fu, M.Q. Zhang, B. Zhang, J.R. Li, M.Z. Rong, Vapor induced variation in electrical performance of carbon black/poly(methyl

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

125

methacrylate) composites prepared by polymerization filling, Carbon 41 (2003) 371–374. J.R. Li, J.R. Xu, M.Q. Zhang, M.Z. Rong, Electrical response to organic vapor of conductive composites from amorphous polymer/carbon black prepared by polymerization filling, Macromol. Mater. Eng. 288 (2003) 103–107. S.G. Chen, J.W. Hu, M.Q. Zhang, M.W. Li, M.Z. Rong, Effects of temperature and vapor pressure on the gas sensing behavior of carbon black filled polyurethane composites, Sens. Actuators B: Chem. 105 (2005) 187–193. M. Narkis, S. Srivastava, R. Tchoudakov, O. Breuer, Sensors for liquids based on conductive immiscible polymer blends, Synthetic Met. 113 (2000) 29–34. S.G. Chen, X.L. Hu, J. Hu, M.Q. Zhang, M.Z. Rong, Q. Zheng, Relationships between organic vapor adsorption behaviors and gas sensitivity of carbon black filled waterborne polyurethane composites, Sens. Actuators B: Chem. 119 (2006) 110–117. J.K. Albert, N.S. Lewis, Cross-reactive chemical sensor arrays, Chem. Rev. 100 (2000) 2595–2626. F. Goettmann, A. Moores, C. Boissi`ere, P.L. Floch, C. Sanchez, A selective chemical sensor based on the plasmonic response of phosphinine-stabilized gold nanoparticles hosted on periodically organized mesoporous silica thin layers, Small 1 (2005) 636–639. P.L. Drake, G.J. Price, Crown-ether containing copolymers as selective membranes for quartz crystal microbalance chemical sensors, Polym. Int. 49 (2000) 926–930. A. Hierlemann, A.J. Ricco, K. Bodenhofer, A. Dominik, W. Gopel, Conferring selectivity to chemical sensors via polymer side-chain selection: thermodynamics of vapor sorption by a set of polysiloxanes on thicknessshear mode resonators, Anal. Chem. 72 (2000) 3696–3708. F. Avila, D.E. Myers, C. Palmer, Correspondence analysis and adsorbate selection for chemical sensor arrays, J. Chemometr. 5 (1991) 455–465. T. Tanabe, K. Touma, K. Hamasaki, A. Ueno, Fluorescent cyclodextrin immobilized on a cellulose membrane as a chemosensor system for detecting molecules, Anal. Chem. 73 (2001) 1877–1880. R. Auzely-Velty, C. Pean, F. Djedaini-Pilard, Th. Zemb, B. Perly, Micellization of hydrophobically modified cyclodextrins. Part 2. Inclusion of guest molecules, Langmuir 17 (2001) 504–510. A. Harada, J. Li, M. Kamachi, Formation of inclusion complexes of monodisperse oligo(ethy1ene glycol)s with ␣-cyclodextrin, Macromolecules 27 (1994) 4538–4543. D. Tsiourvas, Z. Sideratou, A.A. Haralabakopoulos, G. Pistolis, C.M. Paleos, Molecular recognition of amphiphilic di(dodecyl)barbituric acid with long-chain alkylated adenine and thymine derivatives, J. Phys. Chem. 100 (1996) 14087–14092. A. Mulder, A. Jukovi´c, F.W.B. van Leeuwen, H. Kooijman, A.L. Spek, J. Huskens, D.N. Reinhoudt, Photocontrolled release and uptake of a porphyrin guest by dithienylethene-tethered ␤-cyclodextrin host dimmers, Chem.-Eur. J. 10 (2004) 1114–1123. Z. Fan, C.-H. Diao, H.-B. Song, Z.-L. Jing, M. Yu, X. Chen, M.-J. Guo, Encapsulation of quinine by ␤-cyclodextrin: excellent model for mimicking enzyme–substrate interactions, J. Org. Chem. 71 (2006) 1244– 1246. J.W. Hu, S.G. Chen, M.Q. Zhang, M.W. Li, M.Z. Rong, Low carbon black filled polyurethane composite as candidate for wide spectrum gas-sensing element, Mater. Lett. 58 (2004) 3606–3609. I. Tabushi, Chiral selection and chiral induction by the use of regiospecifically di-(or poly) substituted cyclodextrins, Pure Appl. Chem. 58 (1986) 1529–1534. B. Vercelli, S. Zecchin, N. Comisso, G. Zotti, Solvoconductivity of polyconjugated polymers: the roles of polymer oxidation degree and solvent electrical permittivity, Chem. Mater. 14 (2002) 4768–4774. S.G. Chen, J.W. Hu, M.Q. Zhang, M.Z. Rong, Q. Zheng, Improvement of gas sensing performance of carbon black/waterborne polyurethane composites: effect of crosslinking treatment, Sens. Actuators B: Chem. 113 (2006) 361–369. J.R. Li, J.R. Xu, M.Q. Zhang, M.Z. Rong, Carbon black/polystyrene composites as candidates for gas sensing materials, Carbon 41 (2003) 2353–2360.

126

W. Zeng et al. / Sensors and Actuators B 124 (2007) 118–126

[28] X.M. Dong, R.W. Fu, M.Q. Zhang, B. Zhang, J.R. Li, M.Z. Rong, Gas sensing materials from carbon black/poly(methyl methacrylate) composites, Polym. Polym. Compos. 11 (2003) 291–299. [29] M. Kodaka, Correlation between molecular size and packing density of solvents, J. Phys. Chem. B 108 (2004) 1160–1164. [30] Y. Liu, C.-C. You, S.-Z. Kang, C. Wang, F. Chen, X.-W. He, Synthesis of novel ␤-cyclodextrin and calixarene derivatives and their use in gas sensing on the basis of molecular recognition, Eur. J. Org. Chem. 4 (2002) 607–613. [31] K.-L. Yin, D.-J. Xu, C.-L. Chen, Molecular dynamics simulation of dynamic structure of the inclusion by ␤-cyclodextrin and p-cresol in vacuum, Chin. J. Inorg. Chem. 19 (2003) 480–484.

Biographies Wei Zeng is presently completing his PhD study in the Materials Science Institute of Zhongshan University, China. His current research interests are synthesizing polymers for conductive composites and investigating their structure and property relationships. Ming Qiu Zhang received his BSc degree in physics from Zhongshan University in 1982, a PhD degree in polymer chemistry and physics from the same university

in 1991. He is now a professor of materials science and acts as the directors of Key Laboratory for Polymeric Composite and Functional Materials of Ministry of Education of China, and Materials Science Institute of Zhongshan University, China. His research interests include structure–property–processing relationships, characterization techniques and applications of polymers and polymer composites. Min Zhi Rong obtained his PhD degree in polymer chemistry and physics in 1994 in Zhongshan University. He is now a professor of the Materials Science Institute, Zhongshan University, Guangzhou. His research interests include thermosetting/thermoplastic blends, polymeric functional materials, structure of polymer networks, polymeric nanocomposites, natural fiber composites, and dynamic mechanical properties of polymers and polymer composites. Qiang Zheng graduated in the Department of Chemistry, Zhejiang University in 1982 and received his PhD degree in Sichuan University in 1994. He is now a professor of the Department of Polymer Science and Engineering, Zhejiang University, Hangzhou. His research activities are focused on rheology of polymers, polymer blends and polymer composites, conducting polymer composites and polymer-based nanocomposites.