Synthetic Metals 144 (2004) 81–88

Influence of processing conditions on sensitivity of conductive polymer composites to organic solvent vapours J.F. Feller a,∗ , D. Langevin b , S. Marais b b

a Polymers & Processes Laboratory, South Brittany University, Saint-Maudé Street, 56 321 Lorient, France Polymers, Biopolymers & Membranes Laboratory, Rouen University, Maurice de Broglie Boulevard, 76 821 Mont Saint-Aignant, France

Received 30 June 2003; received in revised form 21 January 2004; accepted 4 February 2004 Available online 21 March 2004

Abstract Electrical properties of two conductive polymer composites (CPC) not yet studied in the literature, poly(amide12-b-tetramethylene oxide)-carbon black (PEBAX-CB) and poly(ethylene-co-ethyl acrylate)-carbon black (EEA-CB) in the presence of two organic solvent vapours, chloroform and toluene, have been studied as a function of processing conditions (in the melt by extrusion and from solution by casting). The experiments show that the electrical response of the CPC to solvent vapours can be modulated not only by changing the chemical nature of the polymer but also by tailoring the CPC morphology, i.e. the area accessible to solvent molecules both at a micro and nanometric scale and the conductive paths structure. For this purpose, the use of either block or statistical copolymers is interesting in achieving microphase separation or good CB dispersion, respectively. The results show on the one hand that important responses to the detriment of response time. On the other hand, the solvent vapours R/Rmax can be achieved with high conductivity samples but at √ use of the curve shape, i.e. the diffusion mode characterised by R/Rmax versus t curves slope and onset, would allow us to decrease the number of CPC necessary for vapour detection. Moreover, sorption kinetic measurements show that toluene diffusivity is about two times lower in EEA-CB than in EEA. This phenomenon can be explained by a hindrance effect of the carbon black particles and a decrease of plasticization of EEA chains by toluene molecules due to interactions between EEA and CB. Therefore, the sensitivity of the CPC to solvent vapours results from conductive particles disconnection due to volume expansion during the matrix swelling. © 2004 Elsevier B.V. All rights reserved. Keywords: Conductive polymer composites; Solvent vapours; Diffusion coefficient

1. Introduction In the last decade, electrically conductive polymer composites (CPC) were studied by many groups for their sensing ability [1,2]. In fact, CPC exhibit several interesting features due to their resistivity variation with thermal [3–10], mechanical [11,12] or chemical [13–21] solicitations. CPC are obtained by blending an insulating polymer matrix with conductive fillers like carbon black [3–9,13–21], carbon fibres [11], metal particles [10] or conductive polymers [22]. Whatever be the application used, one main parameter determining CPC properties is the conductive pathways structure, depending on many parameters such as filler content (φ), surface free energy of the filler and the matrix [5,6], crystallinity [4,7], reticulation [10] and exclusion volume [9], i.e. zones where carbon black is concentrated [5,18]. The

∗ Corresponding author. E-mail address: [email protected] (J.F. Feller).

0379-6779/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.synthmet.2004.02.006

desired conductivity level will depend on the application: for heating applications CPC resistivity must be between 100 and 104  cm. For sensor applications according to an abundant literature [13,21], the CPC resistance can be high (over 104 ) provided the response time of vapour detection is low and the response R/R0 is high. Thus, most of the CPC used for the latter application are deposited from solutions on insulating substrates in low thickness layers by various processes, spin coating, spray, cast, etc. Few studies are reported in the literature on extruded CPC used for solvent sensing [17,18] as CPC obtained from solutions have the advantage of providing very thin films and inducing some porosity in the film resulting from the solvent evaporation. The necessity to have selective systems toward vapour detection imposes to combine several CPC into arrays, which gives rise to an important diversity in CPC systems. In the present work, we have investigated the electrical properties of CPC not yet studied in the literature poly(amide12-b-tetramethylene oxide)-carbon black (PEBAX-CB) and poly(ethylene-co-ethyl acrylate)-carbon

82

J.F. Feller et al. / Synthetic Metals 144 (2004) 81–88

black (EEA-CB) in the presence of two organic solvent vapours, chloroform and toluene, as a function of the CPC morphology, i.e. the area accessible to solvent molecules both at a micro and nanometric scale and the conductive paths structure resulting from processing conditions.

2. Experimental 2.1. Materials Poly(ethylene-co-ethyl acrylate) filled with 37% w/w (23% v/v) of Cabot Vulcan XC72 carbon black is LE 7704 from Borealis; EEA is a statistical copolymer resulting from radical polymerisation; poly(amide12-b-tetramethylene oxide) filled with 40% w/w (29% v/v) of Cabot Vulcan P carbon black is from Ato-Fina. PEBAX is a copolymer resulting from the condensation of amide and ether blocks. Some of these CPC characteristics are given in Table 1. Chloroform (stabilised +99%), toluene (99%) and xylene (99%) were obtained from Accros Organics. All solvents were used as received without further purification. Some of the solvent characteristics can be found in Table 2. Samples processing in the melt was done using a twin-screw Brabenderextruder (L = 400 mm, Φ = 16 mm) to elaborate the electrically CPC. Next step concerns the process of textile or film material. (i) For electrical measurements, we used a textile obtained by manual weave of the extruded thread, with a small loom. To Table 1 Polymer composite characteristics EEA-CB (◦ C)

PEBAX-CB

−33 ± 3 99.5 ± 0.5 83 ± 0.5 63

Tg Tm (◦ C) Tc,n (◦ C) Hm (J g−1 ) corrected from CB% Density (at 25 ◦ C) Solubility parameter δ (MPa1/2 ) Percentage of carbon black (w/w)

0.925 ± 0.05 7.9 (PE)

−75/−40 165 ± 0.5 145 ± 0.5 16.7 1.09 ± 0.05 23.02 (PA), 19 (POE)

37

40

Table 2 Solvent characteristics

Formula Molar weight (g mol−1 ) Density at 20 ◦ C (g cm−3 ) Solubility parameter δ (MPa1/2 ) Molar volume (dm3 mol−1 )

Chloroform

p-Xylene

Toluene

CH–Cl3 119.37

C6 H4 (CH3 )2 106.17

C7 H8 92.14

1.49 18.7 8.012×10−2

0.866 18.0 1.225×10−1

0.87 18.3 1.059×10−1

prevent movement during the measurement, thread ends at the edge of the sample were soldered together. Threads were spanned using a capillary die (L = 15 mm, Φ = 1 mm) with a weight flow of 0.4 < m ˙ < 1.8 g min−1 , an extrusion speed 1 < Vextr < 5 rad min−1 , a pulling speed of 0 < Vp < 10 m s−1 , a pressure 10 < p < 21.9 MPa. The thread diameters was 180 < d < 780 ± 10 ␮m depending on material and processing conditions. (ii) Films were extruded using a flat die (L = 100 mm × 40 mm × 1 mm) with weight flows 8 < m ˙ < 10.8 g min−1 , extrusion speeds 14 < Vextr < 20 rad min−1 , a pulling speed of Vp = 0 and pressures in the die 3.7 < p < 6.7 MPa; resulting films thickness was about 430 < e < 510 ␮m ± 10 ␮m. The following temperature profiles were used (from feeding zone to die): 200/210/210/230 ◦ C for EEA-CB and PEBAX corresponding to a processing temperature of 230 ◦ C. The main processing conditions of the different samples are summarised in Table 3. Cast films of CPC were obtained from solutions of a 20 g CPC dm−3 xylene by deposition with a pipette on a glass slide. To ensure a good wetting of the CPC film on the substrate, the glass slides were previously etched in a piranha solution (30% H2 O2 + 70% H2 SO4 v/v) for 1 h. Up to six layers were successively deposited, after drying in air for 2 h for each layer, to get appropriate film thickness for further conductivity measurement. 2.2. Characterisation Calorimetric measurements were made on a Perkin-Elmer Pyris 1 differential scanning calorimeter (DSC) with Pyris V3.0 software for data collection and treatment. The calibration was done with indium and zinc. The base line was checked every day. Aluminium pans with holes were used and the samples mass was approximately 10 mg. All the temperatures measured from a peak extremum (Tc , Tm ) are determined at less than ±0.5 ◦ C and from a sigmoid (Tg ) at less than ±1 ◦ C. Electrical resistance was derived from the intensity and voltage measurement with a Keithley 2000 multimeter. Tension was adjusted with a stabilised source. Considering the important resistance of the sample, a two-probe technique was used. Collection and processing of data were done by an acquisition program developed with Visual Designer 4.0 described in a previous work [8]. Silver paint was used to ensure good electrical contacts. Solvent vapour/air cycles were applied two times to the samples at room temperature for 15 min periods. Solvent in liquid state was disposed in the bottom of the cell to insure atmosphere saturation with the corresponding vapour. The cell volume was 40 mm × 20 mm × 80 mm. Gravimetric sorption measurements have been proceeded at 25 ◦ C on an automated electronic microbalance (IGA 002 from Hiden Analytical Ltd., Warrington, UK) equipped with a vapour generator. Film samples of about 100 mg in weight were used and firstly outgassed under high vacuum at 50 ◦ C

J.F. Feller et al. / Synthetic Metals 144 (2004) 81–88

83

Table 3 Processing conditions

Extrusion temperature Extrusion speed (rad min−1 ) Pulling speed (m s−1 ) Die dimensions (mm) Weight flow (g min−1 ) Pressure in the die (MPa) Thickness (␮m)

PEBAX-CB film

PEBAX-CB textile

PEBAX-CB textile

EEA-CB film

EEA-CB textile

230 20 0 100 × 40 × 1 10.8 6.7 510

230 1 10 15 × 1 0.4 10 180

230 5 10 15 × 1 1.75 21.9 780

230 14 0 100 × 40 × 1 8 3.7 430

230 3 10 15 × 1 1.8 17.2 650

then 25 ◦ C until constant weight (reference dry mass of the material). Then, successive equal toluene vapour pressure steps were applied corresponding to 25, 50 and 75% of the saturated vapour pressure of toluene at 25 ◦ C (=3792 Pa). For each pressure, the sample weight variation was recorded as a function of time until sorption equilibrium was reached. These collected data were used to plot the toluene sorption isotherm curves and to calculate the sorption kinetic parameters.

PEBAX-CB thin textile

log R ( Ω )

(◦ C)

PEBAX-CB extruded film PEBAX-CB thick textile

1,E+08

PEBAX-CB cast film Cycle chloroform/air

1,E+07 1,E+06 1,E+05 1,E+04 1,E+03

3. Results and discussion

1,E+02 500 1000 1500 2000 2500 3000 3500 4000

0

3.1. Electrical properties of the CPC in the presence of solvent vapours Since for some of the samples used in this study (especially those resulting from cast process), the thickness is difficult to measure with enough precision, only resistance was expressed in spite of resistivity. Thus, all samples have different sections but identical length exposed to solvent vapour, which explains partly the differences in resistance observed. In Fig. 1, PEBAX-CB samples have been submitted to a saturated atmosphere of chloroform vapour. It can be seen that for the same formulation (matrix and carbon black content), the electrical response is very different depending on the processing conditions. As expected, the cast film obtained from a solution of PEBAX-CB in xylene leads to the higher initial resistance and gives the shorter response time due to its lower thickness. About 300 and 100 s are necessary to reach equilibrium conditions during the first and second cycle, respectively, showing that the delay between cycles was not sufficient to allow full desorption of the vapour, although a rather flat signal was obtained. The film behaves differently once it has already been in contact with solvent molecules. For PEBAX-CB textiles, it can be observed in Fig. 1 that the initial resistance (intermediate between cast and extruded film) is identical for both samples, whatever the thread

t (s)

Fig. 1. Electrical response of PEBAX-CB in the presence of chloroform as a function of sample nature.

thickness, and then diverges over 300 s, the thinner one leading to larger resistance variation. It is also clear that in both cycles (excepted for the cast film), the desorption time is not sufficient to recover the initial resistance level, although the slopes of the curves during the sequence in air are almost identical. The extruded film behaviour is interesting as it gives rise to important resistance amplitude although the sorption equilibrium is not reached at the end of the first cycle. In Fig. 2, the responses of PEBAX-CB

PEBAX-CB thick textile

log R (Ω)

Electrical properties of poly(amide12-b-tetramethylene oxide)-carbon black and poly(ethylene-co-ethyl acrylate)carbon black in the presence of two organic solvent vapours, chloroform and toluene, has been studied as a function of processing conditions (from the melt by extrusion and from solution by casting).

PEBAX-CB extruded film

1,E+07

PEBAX-CB cast film Cycle toluene/air

1,E+06

1,E+05

1,E+04

1,E+03

1,E+02 0

500

1000 1500 2000 2500 3000 3500 4000 t (s)

Fig. 2. Electrical response of PEBAX-CB in the presence of toluene as a function of sample nature.

J.F. Feller et al. / Synthetic Metals 144 (2004) 81–88 PEBAX-CB thin textile PEBAX-CB extruded film PEBAX-CB thick textile PEBAX-CB cast film Cycle chloroform/air

EEA-CB thick textile EEA-CB extruded film

1,E+07

EEA-CB cast film

R/R max

log R (Ω)

84

Cycle chloroform/air

1,2 1

1,E+06 0,8

1,E+05

0,6 Rmax-Ro 0,4

1,E+04

0,2 s

1,E+03 0 0

5

10

15

1,E+02 0

500 1000 1500 2000 2500 3000 3500 4000 t (s)

20 t onset

25

30

35

40

45

t1/2 (s1/2)

√

Fig. 5. Resistance ratio of PEBAX-CB as a function of form/air cycles.

t for chloro-

EEA-CB thick textile EEA-CB cast film

1,E+08

EEA-CB extruded film Cycle toluene / air

1,E+07

PEBAX-CB thick textile PEBAX-CB extruded film PEBAX-CB cast film Cycle toluene/air

1,2 1 0,8

0,6 0,4 0,2 0 0

5

10

15

20

25

30

35

40 1/2

t

Fig. 6. Resistance ratio of PEBAX-CB as a function of cycles.

√

45 1/2

(s )

t for toluene/air

by Fick [23,24], show that the sorbed matter ratio is proportional to the square root of time, we have √ plotted in Figs. 5–8, the resistance ratio R/Rmax versus t for the different CPC/vapour couples. These figures obtained from the first vapour/air cycle show new interesting features resulting from a data treatment which allows the determination

R/ R max

log R (Ω)

samples obtained for toluene vapour can be ranked in the same way as for chloroform vapours, i.e. as a function of the process type (however, if the curves are analogous to those obtained for chloroform, lower resistance amplitude as obtained revealing that in this case the diffusion process is slower). One can imagine that the curves obtained for toluene should correspond to that of chloroform but truncated over 500 s. Therefore, this phenomenon is less marked for the cast film for which the equilibrium conditions are reached in all cases. Fig. 3 shows that EEA-CB cast film provides quasi-instantaneous response to chloroform and that for this sample the two cycles are superimposable. Once more, the pattern differences resulting from the process can be clearly recognised and suggest specific diffusion mechanisms. In the presence of toluene (Fig. 4), EEA-CB samples respond slower than for chloroform but the responses stay very dependent on the sample shape. To go further into curves analysis, it is now necessary to use additional data processing. Considering that resistance variations in the CPC result from the matrix swelling due to solvent diffusion and that diffusion laws as expressed

R/ R max

Fig. 3. Electrical response of EEA-CB in the presence of chloroform as a function of sample nature.

EEA-CB thick textile EEA-CB extruded film

1,2

EEA-CB cast film Cycle chloroform/air

1 0,8

1,E+06 1,E+05

0,6

1,E+04

0,4

1,E+03

0,2 0

1,E+02 0

500 1000 1500 2000 2500 3000 3500 4000

0

5

10

15

20

25

30

35

Fig. 4. Electrical response of EEA-CB in the presence of toluene as a function of sample nature.

40

45

t1/2 (s1/2)

t (s)

Fig. 7. Resistance ratio of EEA-CB as a function of cycles.

√

t for chloroform/air

EEA-CB thick textile EEA-CB extruded film EEA-CB cast film Cycle toluene/air

1,2 1

85

(Rmax-R0)/ R0 25000 20000

0,8

0,4

5000

0,2

0

0 0

5

10

15

20

25

30

35

40

45

t1/2 (s1/2)

Fig. 8. Resistance ratio of EEA-CB as a function of cycles.

√

t for toluene/air

PEBAX/Chloroform EEA/Chloroform PEBAX/Toluene EEA/Toluene

Cast Film

10000

Extruded Film

0,6

Thin Textile

15000

Thick Textile

R/ Rmax

J.F. Feller et al. / Synthetic Metals 144 (2004) 81–88

Fig. 11. Responses of EEA-CB and PEBAX-CB CPC to toluene and chloroform vapours.

23.44 Slope s 2 1,8 1,6 1,4 1,2 1 0,8 0,6 0,4

PEBAX/Chloroform EEA/Chloroform

0,2

PEBAX/Toluene

0

Thick Textile

EEA/Toluene Thin Textile

Extruded Cast Film Film

Fig. 9. Slopes of EEA-CB and PEBAX-CB curves for toluene and chloroform vapours.

of several parameters among which: the slope (s), the onset time (tonset ) and the resistance amplitude (R/R0 ) useful for diffusion behaviour analysis. These parameters have been plotted in Figs. 9–11, respectively. The slope analysis

t

1/2

onset 30 25 20 15 10 PEBAX/Chloroform EEA/Chloroform PEBAX/Toluene EEA/Toluene

5

Cast Film

Extruded Film

Thin Textile

Thick Textile

0

Fig. 10. Onset times of EEA-CB and PEBAX-CB curves for toluene and chloroform vapours.

of the different CPC, Fig. 9, shows that all systems whatever the process and the chemical nature can be analysed except EEA-CB cast film/chloroform which gives a too important slope to be represented in the same plot as the other ones. It can be noticed that for PEBAX-CB/chloroform, the higher slope is obtained with the thin textile and the lower with the thick textile, suggesting that the slope can be adjusted by the thread thickness. Surprisingly for PEBAX-CB whatever the solvent, cast film does not lead to the more important √ slope, as it is the case for EEA-CB. tonset is related to the response time of the systems for a defined vapour, Fig. 10 provide complementary data showing that the slope is not √ necessarily related to tonset . For example, extruded films can give important slopes but also important onset time, whereas cast film leads to the slowest response time but not necessary to the highest slopes. Fig. 12 shows the texture of EEA-CB cast film observed in SEM with a 1000× magnification. The porous morphology of the coating resulting from the solvent evaporation together with the low thickness of the film could explain the weak response times obtained with the cast film process. The third parameter, which can be determined, is R/R0 the electrical response amplitude of the CPC; from Fig. 11 it can be seen that for all systems extruded films provide the most important response due to √ their low initial resistance. The combination of s, tonset to ∆R/R0 appears to be valuable for vapour identification since it couples several phenomena, which do not appear in the sole response. Moreover, for a given system, it is possible √ to change the values of s, tonset and ∆R/R0 by adjustment of the sample processing. 3.2. Sorption properties of EEA and EEA-CB To go further in the understanding of the phenomena behind resistance variations in the CPC due to solvent diffusion, we have performed sorption experiments with EEA and EEA-CB extruded films in the presence of toluene. One main point is to determine the influence of CB in the diffusion process using an independent technique.

86

J.F. Feller et al. / Synthetic Metals 144 (2004) 81–88

Fig. 12. SEM micrograph × 1000 of EEA-CB cast film deposited from a xylene solution.

EEA

1,4 113 µm

EEA-CB

1,2

1

430 µm

0,8 0,6 0,4 0,2

0 0

50

100

150

200

250

300

350

400

t (min)

Fig. 13. Toluene vapour sorption isotherms at 25 ◦ C for EEA-CB and EEA polymer films.

Equilibrium toluene concentration (% Md)

EEA-CB corrected

9

EEA-CB

8

EEA

7

Polynomial (EEA) Polynomial (EEA-CB)

6 5 4 3 2 1 0 0

0,2

0,4

0,6

0,8

Equilibrium toluene vapor activity a t

Fig. 14. Comparison of the toluene sorption rates by EEA-CB and EEA films. Equilibrium toluene vapour activity = 0.25, temperature = 25 ◦ C. Sample characteristics in Table 4.

Experimental diffusion coefficient x 10-12 (m 2 s-1)

Mass regain (mg)

Toluene vapour sorption at 25 ◦ C was studied on EEA (thickness = 113 ␮m; dry mass = 57.151 mg) and EEA-CB (thickness = 430 ␮m; dry mass = 96.430 mg, 37% w/w carbon black) polymer films. Fig. 13 represents toluene sorption at equilibrium (in percent of the sample dry mass) as a function of toluene activity (at = vapour pressure/saturated vapour pressure) for both materials. From toluene sorption isotherms so obtained, the curvature is typical of a Flory–Huggins type sorption mechanism [25] and the concentration at equilibrium indicates that toluene is about two times more soluble in EEA than in EEA-CB. In order to highlight the role played by the CB particles in the CPC (EEA-CB), and on the basis that the vapour solvents are not sorbed in fillers, the toluene sorption isotherm has been calculated (see Fig. 13). Fig. 14 compares the rate of toluene sorption by the two polymers for at = 0.25, the sorption is much slower in EEA-CB film than in the EEA one, according to the respective thickness of the samples. The linearization of the kinetic

EEA

25

EEA-CB

20

15

10

5

0 0

0,2

0,4

0,6

0,8

1

Equilibrium toluene vapor activity a t

Fig. 15. Experimental toluene diffusion coefficient in EEA-CB and EEA films at 25 ◦ C as a function of the equilibrium toluene vapour activity.

J.F. Feller et al. / Synthetic Metals 144 (2004) 81–88

87

Table 4 Sorption measurement features Sample features

EEA-CB film

EEA film

Thickness (␮m) Initial mass (mg) Mass loss during outgassing (mg) Dry mass (Md) after outgassing (mg) (measured under vacuum)

430 96.713 0.4291 (0.44% Md) 96.430

113 57.148 0.0822 (0.14% Md) 57.151

0.885 (0.92% Md) 1.995 (2.09% Md) 3.375 (3.5% Md)

1.156 (2.02% Md) 2.812 (4.92% Md) 4.438 (7.77% Md)

Equilibrium toluene activity Mass regain at equilibrium (mg) 0.25 0.50 0.75 Experimental diffusion coefficient × 1012 (m2 s−1 ) 0.25 0.50 0.75

√ curves versus t, assuming a Fickian diffusion system, allows us to calculate a mean diffusion coefficient of toluene for the first half sorption time following Eq. (1) [26]:
 m D 1/2 √ =4 t (1) m∞ πL where m is the mass regain of the sample during the sorption (m∞ at equilibrium), L the sample thickness. In this case, D can be determined from the slope k = 4(D/πL)1/2 in Eq. (1) for t < t1/2 . These first results, presented in Fig. 15, show that whatever the polymer, D increases with toluene activity indicating a plasticization of the material by the solvent. It is well known that with organic vapour molecules a concentration-dependent diffusion coefficient is generally observed in polymer materials [26]. However, it is interesting to note that the plasticization phenomenon is reduced with the introduction of fillers in the polymer matrix. Secondly, it appears that the toluene diffusivity is about two times lower in EEA-CB than in EEA, which is certainly due to the hindrance effect of the carbon black particles. In other words, the decrease of toluene solubility is the result of the tortuosity effect due to morphology of CPC with a dispersion of carbon black particles, which act as barrier components and increase the path for toluene molecules inside the composite. Moreover, it seems that the presence of CB in contact with solvent molecules leads to an organising effect in the CPC able to bring more cohesion in the polymer matrix suggesting the presence of strong interactions between EEA chains and CB resulting from an adsorption process. The features of the sorption measurements are gathered in Table 4. The fact that the diffusion coefficient is decreased in the presence of carbon black is a valuable result, which allows us to understand better the mechanism leading to resistivity increase in the presence of vapours. In fact, the hypothesis on which conductive particles disconnection results mainly from solvent molecules diffusion through the conductive pathways, due to more important porosity, is unlikely. It is

3.6 4.62 7.1

5.8 9.5 21.7

most probable that CB particles are disconnected by the volume expansion resulting from the matrix swelling.

4. Conclusion Due to their heterogeneous nature resulting from the association at the nanometric scale of insulating polymer chains and conductive particles, conductive polymer composites (CPC) exhibit sensing properties towards variations of temperature, mechanical stress or chemical substance concentration. To benefit from these interesting capabilities of CPC for detection, whatever the application, it is necessary to control some key parameters as processing, filler dispersion or exclusion, macromolecules/particles interactions. In addition, for vapour detection, it is interesting to identify the diffusion mechanisms, plotting R/Rmax versus, which appear to be very different from a CPC to another. These differences in curve shapes (R/R0 , slope, tonset ) resulting from specific polymer/solvent interactions and peculiar morphologies (microphase separation, crystallinity) can be used to obtain discriminating parameters allowing easier vapour identification without using hundreds of CPC. Moreover, the use of CPC textiles allows combining heating and vapour detection even if the vapour saturation conditions are not reached. In addition, sorption measurements show that the toluene diffusivity is about two times lower in EEA-CB than in EEA, which is certainly due to the hindrance effect of the carbon black particles. This suggests that conductive particles disconnection probably results from the volume expansion due to the matrix swelling more than from solvent molecules diffusion through the conductive pathways. In fact, a secondary effect of the matrix plasticization is to increase EEA/CB interactions and thus to decrease the solvent diffusion.

Acknowledgements The authors would like to thank D. Thoby, E. Le Bellego, H. Béllégou, F. Peresse and A. Bourmaud for their

88

J.F. Feller et al. / Synthetic Metals 144 (2004) 81–88

contribution to this work. This project was supported by the French Ministry of Research & Technology. References [1] G. Harsanyi, Polymer Films in Sensor Applications: Technology, Materials, Devices and their Characteristics, Technomic, Lancaster, 1995, pp. 113–149. [2] N. Akmal, A.M. Usmani, Polymers in Sensors: Theory and Practice, ACS Symposium Series 690, Oxford University Press, New York, 1998, pp. 88–257. [3] M. Narkis, A.V. Tobolsky, Chemically crosslinked polyethylene: modulus–temperature relations and heat stability, J. Appl. Polym. Sci. 13 (1969) 2257–2263. [4] J. Meyer, Glass transition temperature as a guide to selection of polymer suitable for PTC materials, Polym. Eng. Sci. 13 (5) (1973) 462–468. [5] F. Gubbels, R. Jérˆome, E. Vanlathem, R. Deltour, S. Blacher, F. Brouers, Kinetic and thermodynamic control of the selective localization of carbon black at the interface of immiscible polymer blends, Chem. Mater. 10 (1998) 1227–1235. [6] C. Zhang, X.S. Yi, H. Yui, S. Asai, M. Sumita, Morphology and electrical properties of short carbon fiber-filled polymer blends: Highdensity polyethylene/poly(methyl methacrylate), J. Appl. Polym. Sci. 69 (9) (1998) 1813–1819. [7] C. Lagrève, J.F. Feller, I. Linossier, G. Levesque, Poly(butylene terephthalate)/poly(ethylene-co-alkyl acrylate)/carbon black conductive composites: influence of composition and morphology on electrical properties, Polym. Eng. Sci. 41 (7) (2001) 1124. [8] J.F. Feller, Conductive polymer composites: influence of extrusion conditions on positive temperature coefficient effect (PTC) of poly(butylene terephthalate)/poly(olefin)-carbon black blends, J. Appl. Polym. Sci. 91 (4) (2004) 2151–2157. [9] K. Cheah, M. Forsyth, G.P. Simon, Conducting composite using an immiscible polymer blend matrix, Synth. Met. 102 (1999) 1232– 1233. [10] G. Boiteux, J. Fournier, D. Issotier, G. Seytre, G. Marichy, Conductive thermoset composites: PTC effect, Synth. Met. 102 (1–3) (1999) 1234–1235. [11] I. Weber, P. Schwartz, Monitoring bending fatigue in carbonfibre/epoxy composite strands: a comparison between mechanical and resistance techniques, Comp. Sci. Technol. 61 (2001) 849–853. [12] J. Vilcakova, P. Saha, V. Kresalek, O. Quadrat, Pre-exponential factor and activation energy of electrical conductivity in polyester resin/carbon fibre composites, Synth. Met. 113 (1–2) (2000) 83–87.

[13] E.J. Severin, B.J. Doleman, N.S. Lewis, An investigation of the concentration dependence and response to analyze mixtures of carbon black/insulating organic polymer composite vapor detectors, Anal. Chem. 72 (4) (2000) 658–668. [14] B.C. Munoz, G. Steinthal, S. Sunshine, Conductive polymer-carbon black composites-based sensor arrays for use in an electronic nose, Sens. Rev. 19 (4) (1999) 300–305. [15] B. Lundberg, B. Sundquist, Resistivity of a composite conducting polymer as a function of temperature, pressure, and environment: application as a pressure and gas concentration transducer, J. Appl. Phys. 60 (1986) 1074–1079. [16] J. Chen, N. Tsubokawa, Novel gas sensor from polymer-grafted carbon black: vapor response of electric resistance of conducting composites prepared from poly(ethylene-block-ethylene oxide)-grafted carbon black, J. Appl. Polym. Sci. 77 (11) (2000) 2437–2447. [17] S. Srivastava, R. Tchoudakov, M. Narkis, Preliminary investigation of conductive immiscible polymer blends as sensor materials, Polym. Eng. Sci. 40 (7) (2000) 1522–1528. [18] M. Narkis, S. Srivastava, R. Tchoudakov, O. Breuer, Sensors for liquids based on conductive immiscible polymer blends, Synth. Met. 113 (2000) 29–34. [19] X. Chen, Y. Jiang, Z. Wu, D. Li, J. Yang, Morphology and gassensitive properties of polymer based composite films, Sens. Actuators B: Chem. 66 (2000) 67–69. [20] J. Chen, N. Tsubokawa, 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 (3) (2000) 101–107. [21] M.C. Burl, B.C. Sisk, T.P. Vaid, N.S. Lewis, Classification performance of carbon black-polymer composite vapor detector arrays as a function of array size and detector composition, Sens. Actuators B: Chem. 87 (1) (2002) 130–149. [22] C.W. Lin, B.J. Hwang, C.R. Lee, Characteristics and sensing behavior of electrochemically codeposited polypyrrole-poly(vinyl alcohol) thin film exposed to ethanol vapors, J. Appl. Polym. Sci. 73 (1999) 2079– 2087. [23] J. Cranck, The Mathematics of Diffusion, 2nd ed., Oxford University Press, London, 1975. [24] I. Linossier, F. Gaillard, M. Romand, J.F. Feller, Measuring water diffusion in polymer films on the substrate by internal reflection Fourier transform infrared spectroscopy, J. Appl. Polym. Sci. 66 (13) (1997) 2465–2473. [25] T.V. Naylor, Permeation properties, in: C. Booth, C. Price (Eds.), Comprehensive Polymer Science, 1st ed., Vol. 2, Pergamon Press, Oxford, 1989. [26] J. Cranck, G.S. Park, Methods of measurements, in: Diffusion in Polymers, Academic Press, London, 1968, pp. 1–39.