Synthetic Metals 217 (2016) 266–275
Contents lists available at ScienceDirect
Synthetic Metals journal homepage: www.elsevier.com/locate/synmet
Acid-dopant effects in the formation and properties of polycarbonate-polyaniline composites Yuriy Noskova , Sergei Mikhaylova,b , Patrice Coddevilleb , Jean-Luc Wojkiewiczb , Alexander Puda,* a b
Institute of Bioorganic Chemistry and Petrochemistry, National Academy of Sciences of Ukraine, 50 Kharkivske Shose, 02160 Kyiv, Ukraine Mines Douai, Département Sciences de l’Atmosphère et Génie de l’Environnement (SAGE), 941 rue Charles Bourseul, F-59508 Douai, France
A R T I C L E I N F O
Article history: Received 19 February 2016 Received in revised form 7 April 2016 Accepted 15 April 2016 Available online xxx Keywords: Aniline polymerization Polycarbonate-polyaniline composites Acid-dopants effects Thermal stability Conductivity Ammonia sensing properties
A B S T R A C T
The significant effect of various aromatic sulfonic acids on peculiarities of chemical polymerization of aniline in the water dispersion of polycarbonate (PC) powder, morphology and properties of the formed PC-polyaniline composites has been found and studied. In particular, we demonstrate that the rate of the aniline polymerization and polyaniline yield strongly depend on the size and surface activity of the aciddopant anions. It is found that in the case of the large dopants (e.g. dodecylbenzenesulfonic acid) the composites unexpectedly have the reduced thermostability while the composites with the small dopants (e.g. p-toluenesulfonic acid or 2-naphtalenesulfonic acid) display the enhanced thermal stability as compared with the pure PC. We confirm that the acid-dopant nature affects the thermal stability of conductivity of the compression molded composite films. The solution-cast composite films have the dopant dependent sensitivity to gaseous ammonia in a wide ppm range of its concentrations. ã 2016 Elsevier B.V. All rights reserved.
1. Introduction Intrinsically conducting polymers (ICP) are very promising materials due to their unique combination of physical and chemical properties and a great potential for practical applications [1]. Polyaniline (PANI) stands out among ICP due to simple synthesis, stability, sensitivity to various substances, catalytic activity, low cost, etc. [2]. Typically PANI is synthesized by chemical or electrochemical oxidative polymerization in acidic media [2]. The polymerization process runs through a few stages resulting in formation of aniline oligomers and PANI in different oxidation states characterized with specific spectral and physical-chemical properties [3]. These properties open, therefore, possibilities to monitor this process by different physical-chemical methods (e.g. UV–vis spectroscopy, measurements of pH, open circuit potential and temperature of the reaction medium [4]. Importantly that one can control the process course and properties of the formed PANI by changing the nature and concentration of an acid-dopant and oxidant, pH and temperature [4,5]. Despite significant advances in PANI chemistry this polymer is still not used on a large scale because of such drawbacks as low
* Corresponding author. E-mail addresses:
[email protected],
[email protected] (A. Pud). http://dx.doi.org/10.1016/j.synthmet.2016.04.015 0379-6779/ ã 2016 Elsevier B.V. All rights reserved.
mechanical properties, poor solubility and processibility. These problems can be bypassed by some ways, but formation of PANI composites with common polymers and using functionalized protonic acids are probably the most effective methods [6–9]. These materials not only combine properties of polyaniline and other components but often show synergetic enhancement of their characteristics [9]. Among common polymers bisphenol A polycarbonate (PC) attracts significant interest due to an excellent combination of high optical and mechanical characteristics, durability and stability [10]. Therefore, its composites with PANI are of particular interest as these excellent properties will be completed with conductivity and other specific properties of PANI. However, to our knowledge synthesis and properties of such composites have been studied in a not enough extent. Specifically, in sparse papers a formation of polyaniline-polycarbonate (PANI/PC) composites was realized through, chemical or electrochemical template syntheses [11], mixing in joint solutions in organic solvents [12] and emulsion polymerization [8,13]. The two last methods use surface active substances and acids-dopants (organic sulfonic acids with long nalkyl group,) [8,10,12] that facilitates formation of PANI structures with good plasticity, electrical, and mechanical properties [14,15]. Nevertheless, the necessity to use organic solvents in these methods is a significant drawback from technological and ecological points of view. The problem was circumvented by
Y. Noskov et al. / Synthetic Metals 217 (2016) 266–275
chemical aniline polymerization in a water dispersion of polycarbonate powder in presence camphorsulfonic acid (CSA) or ptoluenesulfonic acid aqueous medium (TSA) as dopants [16]. Typically, this approach allows formation of a thin layer or shell of PANI at surface of particles (cores) of a dispersion phase [4,17,18]. In case of polymer-polymer composites and their industrial hot temperature processing, existence of these layers (shells) can be postulated as a good prerequisite for formation of a high quality percolation network in the bulk of a final composite. Indeed, it is very likely that during melting or solution treatment of the PANI containing composite powder, this shell (layer) is transformed into PANI (nano)clusters which self-organize into the percolation network thus maintaining conductivity of the ultimate material [19,20]. This postulate agrees well with recently published data, which discover that properties of PANI in the shell (molecular weight, structure, oxidation state and conductivity) and, therefore, in the formed clusters strongly differ from that of neat PANI [21]. Naturally, properties of these PANI clusters as a separate composite phase are very important for the composite material. One of the effective methods is a use of a dopant, which plasticizes and compatibilizes PANI with other polymer component of the composite [12,15]. However, though this approach has long been known [1,9,12,15,22,23], the current understanding of a role of a dopant structure in the case of PANI-common polymer composites and especially of PANI/PC composites is not enough sufficient. In this work we directed our efforts toward creating and study of PANI/PC composites formed by polymerization of aniline in aqueous media containing different aromatic sulfonic acids and dispersed PC particles. An effect of the sulfonic acid-dopant on the aniline polymerization specificity, morphology and properties of the synthesized PANI/PC composites is estimated. Particularly, the effects of the different aromatic sulfonic acids on thermal, electrical and ammonia sensing properties are investigated. 2. Experimental 2.1. Materials Aniline (Merck) was distilled under reduced pressure and stored under argon at 5 C. PC powder was obtained by precipitation method similar to [21]. In short, the 2 wt.% solution of Lexan (General Electric) in chloroform (50 mL) was dropwise added to acetone (500 mL) under stirring. The formed dispersion was filtered after 3 h of sedimentation; the precipitate was washed 3 times with acetone, and dried under vacuum at 50–60 C for 24 h. The separated dried PC powder was sieved through a 0.1 mm sieve that allowed a powdered mixture of PC particles with sizes less than 100 mm. The acids: benzenesulfonic acid (BSA) (Aldrich), p-toluenesulfonic acid (TSA) (monohydrate, Aldrich), dodecylbenzenesulfonic acid (DBSA) (Acros Organics), 2-naphtalenesulfonic acid (NSA) (Aldrich), 1.5-naphtalenedisulfonic acid (NDSA) (Aldrich), dinonylnaphtalenesulfonic acid (DNNSA) (Aldrich) and the oxidant ammonium persulfate (APS) were of analytical or reagent grade and used as received. 2.2. Preparation of the composites The polymerization of aniline in the PC aqueous dispersions was carried out in accord with the method described elsewhere [16] at next ratios of the reaction mixture components: aniline/ PC = 2.5/97.5 wt.%, aniline/acid-dopant = 1/1.5 (mol/mol) and aniline/oxidant = 1/1.25 (mol/mol) at ambient temperature 22–23 E. Typically, at the first stage of the preparation of the polymerization mixture the acid was dissolved in 9.8 ml of water to maintain its concentration at 0.063 M. At the second stage, after 30 min of the acid dissolution, the calculated quantity of aniline (C = 0.043 M)
267
was added to this acid solution followed by stirring for 1 h. At the third stage 2 g of the PC powder (a specific weight 0.344 g/cm3) was added to the anilinium salt solution followed by stirring this mixture for 30 min. At the fourth stage the APS aqueous solution (0.052 M) in 3 ml of distilled water was prepared and then added to the reaction mixture followed by stirring for 24 h at ambient temperature. The final product PC/PANI was filtered out, rinsed with distilled water and dried under vacuum at 60–70 C for 24 h to a constant weight. The obtained PANI/PC powder composites were processed into films both by compression molding technique at 240 C under 5 MPa (using SPECAC press) for 1 min and by casting on glass plates from their 3% chloroform solutions or dispersions prepared under ultrasonication. 2.3. Measurements The real PANI contents in the prepared PANI/PC powder composites were determined by the UV–vis spectroscopy (spectrophotometer M-40) analysis of their solutions in N-methyl-2pyrrolidone (NMP) in accord with the method described elsewhere [17,21]. Briefly, the composite powder was dedoped with ammonium hydroxide aqueous solution and dried to a constant weight. Then a fixed amount of this composite was dissolved in NMP and reduced to leucoemeraldine base (LEB) by a surplus of ascorbic acid. The concentration of LEB was then determined using absorbance at 343 nm of this solution in 1 mm quartz cuvette in comparison with the calibration curve based on UV–vis spectra of different concentrations of LEB solution in NMP. Based on this concentration value, the PANI loading in the PANI/PC composite can be easily calculated [17,20]. The development of the aniline polymerization process in the PC powder dispersions was monitored by open circuit potential (OCP) measurements of the reaction mixture with the help of the redox-electrode Hamilton Liq-Glass ORP attached to pH/redox/ temperature measuring instrument GMH 3530 (Greisinger Electronics).Fourier transform infrared (FTIR) spectra of the samples of the synthesized powder composites and PANI-TSA (obtained after extraction of PC by chloroform from the PANI-TSA/PC composite) in pellets with KBr were measured using Bruker Vertex 70 spectrometer. Scanning electron microscopy (SEM) images were obtained with a help of the HITACHI S-4300 SE/N microscope. Thermal properties of the composites were estimated by thermogravimetric analysis (TGA) in air with a heating rate of 10 C/min using a MOM Q-1500 D (Paulik-Paulik-Erdey) Derivatograph. DC conductivity of the synthesized composites films was measured by standard two-probe (for s < 107 S/cm) and four-probe (for s > 103 S/cm) techniques at ambient conditions. Sensing properties of the PC/PANI nanocomposites in ammoniaair gas mixtures were estimated by changes in their resistances. The sensor responses (SR) were determined as a relative variation of the resistance R of the sensor exposed to the analyte compared to the initial value R0: SR = (R R0/R0) 100% at ambient temperature and relative humidity at ca. 50%. To accomplish these measurements, a 1 mL volume of the ultrasonically treated dispersions of the nanocomposites in chloroform (2% w/v) was drop-cast on the miniature system of gold interdigitated electrodes formed on the glass–ceramic substrate. The thickness of the formed sensing layers was ca. 4 micrometers. The formed sensing elements were installed into the airtight testing chamber (volume = 5.3 L) at ammonia concentrations from 100 to 1000 ppm. 3. Results and discussion It is known that the aniline polymerization process is a specific combination of reactions which differ from most of classical
268
Y. Noskov et al. / Synthetic Metals 217 (2016) 266–275
polymerization processes. Its characteristic features are synchronous changes of some physical-chemical parameters of the reaction medium: redox-potential, pH and temperature [3,4]. Rates of these changes depend on such factors as concentration of reagents, temperature and pH of the reaction medium, nature of acid and of particles of a disperse phase [3–5,18,24]. The last two factors are critically important for the polymerization way of formation of polyaniline composites. Specifically, it is clearly demonstrated that aniline is polymerized much faster if solution of its salt contains particles of an inert phase [4,17,24,25]. This effect can be explained probably by adsorption of the monomer, acid and oxidant on the surface of the particles and, respectively, by running the polymerization at concentrations of these reagents, which are higher than those of in the solution bulk [4,17]. Moreover, a catalytic effect of both the surfaces of these particles and the precipitated pernigraniline (PNA) on the polymerization course cannot be excluded [4,17]. Composite particles with the shell of conducting PANI and the core of the particle of the dispersion phase are a typical result of the aniline polymerization in the dispersion medium. A representative scheme of formation of such a core-shell particle is shown in Fig. S1. Depending on a size and surface activity, charge-compensating anions of a doping acid can affect not only an access of an oxidant to the monomer and its polymerization rate but also determine the morphology of the formed shell/layer of PANI. Specifically, the effect of the acid nature on formation and properties of polyaniline composites has been discovered many years ago [22,26,27]. In particular, due to the phenomenon of counter-ion induced processibility, mixing of PANI doped by functionalized acids with common polymers allows preparation of composites characterized with quite high conductivity [22]. A combination of this approach with the core-shell composites gives a hope on creation of technological PANI containing materials with high conductivity and multifunctional applications. Based on known effects of a doping acid on PANI properties [22], one can postulate that depending on size and surface activity of the anion the doping acid will affect not only peculiarities of the aniline polymerization but also properties of the final PANIcontaining composite [26]. Trying to implement this postulate in PANI/PC composites, we used aromatic sulfonic acids with different substituent groups in benzene or naphthalene rings (see Section 2.1) which can be conventionally reduced to two types: (1) acids with large anions (DBSA and DNNSA) and (2) acids with small ones (BSA, TSA, NSA and NDSA). 3.1. The acid-dopant effect on the aniline polymerization in a presence of the dispersed PC particles We monitored the aniline polymerization by continuous measurements of OCP of the reaction mixture that allowed to
evaluate the rate and effectiveness of this process as well as to observe a development of its stages [17,25,28]. Based on the measured OCP profiles (Fig. 1) we used the time of achieving the maximum value of the potential (Emax) of the polymerization mixture as the relative rate of the aniline polymerization in these systems. At the same time, this Emax value corresponds to the highest concentration of pernigraniline (PNA) in the system and to a complete exhaustion of the oxidant. At the next stage PNA is reduced completely by the unreacted monomer to the emeraldine state and OCP drops [4,17]. In accord with these considerations and the relative rates of the aniline polymerizations in the PC particles dispersion (Fig. 1, Table 1) one can conclude that in the presence of acids with large surface active anions (containing long alkyl chains) the polymerization is 2–5 times slower than in the case of the acids with small anions and less surface activity. The existence of this anion effect matches well with a radicalcation character of the aniline polymerization and specificity of the micellar solution. Thus, whereas intermediates of the aniline polymerization are positively charged (cation-radicals of the monomer and growing PANI chains), there should be a significant effect of charge-compensating anions of the acid-dopant on this process. Indeed, a difference in pH of the polymerization media can be neglected because of the same concentrations of the used aromatic sulfonic acids and of the fact that these ones are strong acids (pKa < 2). Therefore, under these conditions practically complete protonation of aniline and PANI under conditions of the investigation (the starting pH is about 1.2 due to 0.063 M concentration of a strong monobasic acid) should be realized. Respectively, one can assign the observed impacts of the used acids and the differences in these impacts to the anions effects. In particular, anions with bulky hydrophobic groups in the case of DBSA or DNNSA are surface-active and, therefore, are able to structure the solution through hydrophilic-hydrophobic interactions with water and to form a hydration shell around them. As a consequence, at concentrations above critical micelle concentration (CMC) these acids form micelles in the solution that complicates interactions between its components (protonated aniline, growing charged PANI chains, water, oxidant and acid anions, particles of the dispersion phase). This is the case of the system under investigation whereas the acid concentration is 0.063 M which is higher than CMCDBSA 8.4103 M [29], and CMCDNNSA 105 M [30]). After the interaction of these acids with aniline (C = 0.043 M), the free acid (0.02 M) and the anilinium salt (C = 0.043 M) are present in the solution. This fact results obviously in the coexistence of two types of micelles, which are formed of only the acids anions (Fig. 2a) and of the acids anions involving anilinium cations (Fig. 2b). Therefore, the superfluous negatively charged acid micelles located around the anilinium salt micelles can hinder access of the oxidant anions (S2O82) to the monomer. This reason can probably
Fig. 1. OCP profiles of the aniline polymerization medium containing PC dispersion and different sulfonic acids-dopants: (a) acids with large anions; (b) acids with small anions. (1) DBSA; (2) DNNSA; (3) BSA; (4) TSA; (5) NSA; (6) NDSA.
Y. Noskov et al. / Synthetic Metals 217 (2016) 266–275
269
Table 1 The effect of the acids-dopants on the yield of PANI and its contents in the synthesized PC/PANI composites. Theoretical PANI base contents in the composites is 2.5 wt.%. Acid
Real contents of PANI base in the composites, wt.%
Calculated contents of the PANI salts in the Yield of PANI, composites, % wt.%
Sulfonic acids with large anions DNNSA 634 93 DBSA 668 66
2.05 2.17
5.0 4.4
82 87
acids with small anions 706 17 716 35 720 30 724 22
2.2 2.25 2.3 2.4
3.5 3.8 4.4 3.6
88 90 93 96
Sulfonic TSA NSA NDSA BSA
Emax, mV
Time of achieving Emax (relative polymerization rate), min
explain the fact that in the PC dispersions containing sulfonic acids with large surface-active anions (DBSA and DNNSA) the aniline polymerization rate runs slower as compared with media with acids with small anions (BSA, TSA, NSA and NDSA). Probably, variations in the yield of PANI in the prepared composites relates to this phenomenon too. Thus, as one can see in Table 1, the PANI yield is higher in the case of acids with small anions. Moreover, this yield is symbate to the height of the pernigraniline maximum at the OCP profile of the polymerization medium (Table 1). 3.2. Properties of the synthesized PC/PANI composites 3.2.1. The nanocomposites morphology In spite of small contents of the doped PANI (<5 wt.%, Table 1) particles of the formed composites are highly agglomerated as compared with the parent PC ones and form cauliflower-like aggregates (Fig. 3). However, unlike the above discussed polymerization specificity, there is not a specific influence of the type (small or large anions) or structure of the dopants on the sizes and morphology of these aggregates. Thus, the smallest PC/PANI-DNNSA aggregates (3–8 mm, Fig. 3b), as well as the largest ones of PC/PANI-DBSA (20– 30 mm) and PC/PANI-BSA (20–40 mm) display a quite densely packed morphology (Fig. 3c and d). The agglomerates of the PC/ PANI-NSA (2–20 mm) and PC/PANI-NDSA (3–25 mm) look polydisperse and loosely packed. In turn, the agglomerates of the PC/PANITSA (2–20 mm) demonstrate the densely packed morphology, a large part of their surface is highly inhomogeneous and porous unlike the other composites agglomerates. Although the obtained results are not enough to conclusively explain these irregular differences in agglomeration and morphology of the composites, one can assume that this problem is related
to peculiarities of aniline polymerization under the investigation conditions (see Chapter 2.2). Indeed, the polymerization occurs in the dispersion containing high concentration of the PC powder (44,6% vol/vol, calculated in accord with the PC loading in the reaction mixture volume, Chapter 2.2). Obviously, in this medium the PC particle surface is charged by the adsorbed reactive species which are anilinium cations (the small dopants case) or salt micelles (the large dopants case) (see Chapter 3.1). Their charges are compensated by surrounding dopant-anions. After addition of the oxidant the polymerization begins and occurs probably both at the particle surface and in the bulk of the reaction solution. As a consequence, the formed PANI macromolecules both grow into the solution bulk from the surface and precipitate at this surface from the solution. However, in the latter case due to the high concentration of PC particles the PANI can precipitate on a few particles simultaneously and thus forms the composite aggregates. Naturally, because of various physical-chemical interactions (Coulombic, hydrophilic-hydrophobic, intermolecular) with the surroundings, the dopant anions affect degree of the agglomeration and morphology of the aggregates. In particular, the small sizes of the aggregates of the PC/PANI-DNNSA composites could be kept by large DNNSA anions surrounding the positively charged PANI chains and thus separating the neighbor PC particles. However, to find a more complete answer on the question about the direction and way of the dopants influence on the composites aggregates sizes and morphology a separate study with different contents and nature of the composites components is still needed. 3.2.2. FTIR spectroscopy PANI macromolecules contain many amine and imine groups being able to participate in intermolecular interactions specifically in formation of multiple H-bonds between neighboring PANI chains. Naturally, in case of PANI composites a part of these groups
Fig. 2. Schemes of the micelles (combined with the scheme from [31]) co-existing in the aniline polymerization medium and including: (a) the free acid anions micelles; (b) the acids anions micelles involving the anilinium cations.
270
Y. Noskov et al. / Synthetic Metals 217 (2016) 266–275
Fig. 3. SEM images of the as-prepared composite particles: a— parent PC particles; b— PC/PANI-DNNSA; c— PC/PANI-DBSA; d— PC/PANI-BSA; e— PC/PANI-NSA; f— PC/PANINDSA; g— PC/PANI-TSA.
can switch to hydrogen bonding with the relevant counterpart groups of other composite component. In particular, in PC/PANI bulk composites the PANI amine groups can form H bonds with carbonyl groups of PC. We used FTIR-spectroscopy to estimate whether these bonds can be determined in the PC/PANI core-shell composites under study. The normalized by PC carbonyl group (C¼O) band at 1768 cm1 FTIR-spectra of the PC/PANI composites doped with various acids
are very similar and only slightly different from the typical spectrum of pristine PC [12,21] probably because of both the small content of the PANI salts (5 wt.%, Table 1) and overlapping the PC bands with the PANI ones (Fig. 4a). However, at higher PANI salt contents (ca. 9.1 wt.%, on the example of PANI-TSA) we see quite clear changes in the composite spectrum (Fig. 4a,b). Specifically, a structureless absorption band in the range of 1900 cm1–2800 cm1 indicates presence of free charge carriers in
Y. Noskov et al. / Synthetic Metals 217 (2016) 266–275
271
Fig. 4. (a) FTIR spectra of PC/PANI composites doped by different acids dopants (Table 1): (1) pristine PC; (2) PC/PANI-DBSA; (3) PC/PANI-DNNSA; (4) PC/PANI-BSA; (5) PC/ PANI-NSA; (6) PC/PANI-NDSA; (7) PC/PANI-TSA (3.5 wt.% of PANI-TSA); (8) PC/PANI-TSA (9.1 wt.% of PANI-TSA). (b) Comparison of FTIR spectra of (1) pristine PC; (2) PC/PANITSA composite with 9.1% of PANI-TSA; (3) PANI-TSA salt left after the extraction of PC from the composite PC/PANI-TSA.
the conductive PANI salt component of the composite [32]. Moreover, though most of characteristic bands of doped PANI are masked by PC bands in this composite spectrum, one can clearly distinguish the band at 1591 cm1, 1297 cm1 and shoulder at 1130 cm1 (Fig. 4a) assigned to stretching vibrations of quinonoid rings, C N and Q = N+H–B or B–NH+–B vibrations, respectively [33,34]. These band positions are different from the respective ones (1563 cm1, 1300 cm1 and 1136 cm1) in the spectrum of the PANI-TSA obtained after removal of PC from the composite (Fig. 4b, Table 2). The found shifts of the PANI-TSA bands confirm a specific interaction of PANI-TSA with PC for account of hydrogen bonding of carbonyl groups of PC with amine groups of PANI in the composite [21]. Indeed, this interaction agrees well with the shift of carbonyl group stretching vibrations in the pristine PC and PC/PANI-TSA spectra from 1768 cm1 to 1760 cm1 [8,21]. Moreover, this interaction of PANI-TSA with PC suggests that it can be the additional factor affecting the morphological specificity of the composite aggregates formed in the presence of the other used dopants. 3.2.3. Thermal stability One of key properties of polymer composites is their thermostability at processing conditions. Typically it depends on nature and content of the composite components, their interactions among themselves, morphology of this material as a whole, etc. In particular, the PC matrix polymer of the composites under study has high thermostability and is typically processed in conditions of injection molding at temperatures of ca. 300 C [37]. Another component of these composites, doped PANI, is also thermally stable and its backbone does not degrade even up to 400 C [38]. However, it loses a dopant and conductivity under
much lower temperatures [39,40]. Therefore, thermal behavior of the doped PANI can affect thermal behavior of its composites. Indeed, even at the low PANI contents one can see differences in TG curves of pure PC and PC/PANI composites (Fig. 5, Table 3). Depending on the acid-dopant type (large or small anions) the composites demonstrate a quite different thermal behavior, which obviously is effected by a specificity (plasticizing ability, volatility and thermal stability) of the acids-dopants and their complexes with PANI as well as by an interaction of an eliminated dopant and products of its degradation with the matrix polymer PC (Fig. 5, Table 3). In particular, in the case of large dopants with long alkyl tails one can see decreased thermostability of the composites up to ca. 530 C as compared with the pure PC. This thermal behavior of the composites strongly differs from that of the known PC/PANI–DBSA composite (PANI content 16.7 wt.%), which was synthesized through an inverted emulsion polymerization route and was more thermally stable than PC [8]. The difference between known [8] and our composites with dopant DBSA is not completely understood for the moment. However, as one of possible reasons of this difference we consider various methods used to synthesize PANI in a presence of PC that resulted in different thermal stability of our and known composites. Weight losses of the PC/PANI-DBSA and PC/PANI-DNNSA composites after ca. 300 C are higher than contents of the both dopants (DBSA, DNNSA) and, moreover, after 375 C are even higher that real contents of PANI-DBSA and PANI-DNNSA (Fig. 5a, Tables 1 and 3). Therefore, taking into account the known high thermal stability of PANI base [35,38], one can assign these losses before 300 E mainly to elimination of water (typically up to 120– 130 E) and evaporation/degradation of the dopant (DBSA, DNNSA) (unbound or appeared after the PANI salt thermal dissociation
Table 2 Main band positions (cm1) of the doped PANI in the FTIR spectra of the PC/PANI-TSA composite and of the PANI-TSA left after the extraction of PC from the composite. The band assignments are based on the known data [32–36]. Band assignments
PC/PANI-TSA composite
PANI-TSA, after PC extraction
quinonoid ring stretching vibrations benzenoid ring stretching vibrations CN stretching vibrations of the protonated aromatic amine structures CN+ stretching vibrations Q = N+HB or BNH+B vibrations/S¼O stretching mode of sulfonic acid CH in-plane bending vibrations
1589 n/a 1297 n/a shoulder at 1130 n/a
1563 1485 1300 1244 Shoulder at 1136/1118 797
272
Y. Noskov et al. / Synthetic Metals 217 (2016) 266–275
Fig. 5. Thermogravimetric curves of the PC/PANI composites with different sulfonic acids-dopants (composition see in Table 2): (a) acids with large anions; (b) acids with small anions. 1— DBSA (4.4 wt.%); 2— DNNSA (5 wt.%); 3— 1st control sample for the PC/PANI-DBSA composite—mechanical mixture of PC and PANI-DBSA (4.4 wt.%); 4) 2nd control sample for the PC/PANI-DBSA composite—mechanical mixture of PC and DBSA (2.25 wt.%); 5— BSA; 6— TSA; 7— NSA; 8— NDSA; 9— control sample for the PC/PANITSA composite mechanical mixture of PC and PANI-DBSA (3.5 wt.%).
[35,39]). However, the much larger weight losses of the composites above 300 E (Fig. 5a and b, Table 3) suggest a specific influence of these dopants and/or products of their degradation on PC stability at high temperatures. This influence is hardly facilitated by the discussed above hydrogen bonding of doped PANI with carbonyl groups of PC because this type of interaction typically improves thermal stability of PANI based composites [8,40], while thermal stability of model mechanical mixtures of PC with PANI-DBSA (4.4 wt.%) and PC with DBSA (2.35 wt.%) is even lower than that of the polymerization sample of PC/PANI-DBSA (4.4 wt.%). The synthesized composites doped by dopants with small anions have a completely different thermal behavior and increased thermal stability as compared with the case of the dopants with large anions (Fig. 5b). Specifically, these composites demonstrate weak and smooth gradual decline of TG curves and have weight losses less than initial content of a dopant even at 400 E except the case of BSA (Table 3). Therefore, we can assign these weight losses before 400 E to elimination of water and the small dopants [39,40]. However, at higher temperatures one can see more significant weight losses (Fig. 5b, Table 3), most of which is obviously caused by the thermo-oxidative degradation of PC. Importantly that all the synthesized composites are more stable at temperatures above ca. 500/530 E (small/large dopant anions) than pristine PC while the model mechanical mixtures with the same quantities of the doped PANIs and PC are much less stable (Fig. 5a and b, Table 3). This fact agrees well with the earlier found specific state of PANI located as a shell at surface of a core material in the core-shell composites (see Introduction and [21,41]). On the
other hand, some stabilization of the PC core particle by the dedoped at high temperatures PANI shell cannot be excluded. At the same time, taking into account similar functionalities and nature of the used doping acids we cannot explain from chemical positions the discovered difference in the thermal behavior of the synthesized composites with large and small dopants. However, as a possible reason of this difference one can suggest intermolecular interactions (causing a plasticizing effect [42]) of the large dopant anions and/or thermally released dopant molecules with PC chains. Indeed, unlike the small dopants, the large dopants contain long alkyl substituents in aromatic rings, which are able to plasticize PANI [9,43] and probably interfere in the intermolecular interactions between PC macromolecules that in turn can lead to weakening of both these interactions and thermal stability of the PC/PANI-DBSA (DNNSA) composites. 3.2.4. Electrical and sensing properties of the PC/PANI doped composite films The thermally induced changes in the composites (see Chapter 3.2.3) can affect their electrical properties, which strongly depend on the state of the PANI component distributed in the thermally treated material [9,44]. Indeed, if to compare properties of the compression molded and cast films of the synthesized composites (Tables 1 and 4), one can see much higher conductivity of the latter composites. As the main reason for this difference one can consider a thermal dissociation of the PANI salt followed by dissolution of the released acids-dopants in the melt under high temperatures of the films formation [21].
Table 3 Weight losses (in wt.%) of the synthesized PC/PANI doped composites and control mixture samples at different temperatures. Weight contents of the acid-dopants in the composites are given in brackets. Temperature, C PC/PANIDBSA (2.35) 120 240 300 400 500 600 700
0.1 0.6 2.3 4.9 42.7 65.4 85.0
PC/PANI-DBSA (2.35), mixture
PC/ DBSA (2.35), mixture
PC/PANIDNNSA (2.43)
PC/PANIBSA (1.1)
PC/PANITSA (1.3)
PC/PANI-TSA (1.3), mixture
PC/PANINSA (1.55)
PC/PANINDSA (2.1)
PC
0 0.6 1.5 7.3 43.0 90.4 99.7
0.4 1.0 1.4 4.1 60.0 95.1 100
0.1 0.9 2.3 5.6 38.6 65.1 90.8
0 0.9 1.2 1.7 4.6 68.6 80.8
0 0.12 0.2 0.9 3.5 68.1 79.9
0.1 0.4 0.6 1.2 6.4 82.4 100
0 0.12 0.2 1.0 3.1 69.0 86.0
0.1 0.33 0.5 1.5 5.9 72.6 85.5
0 0 0 0 3.6 74.9 98.8
Bold values emphasize weight losses of the composites at the compression molding temperature (at 240 C, see Chapter 2.2).
Y. Noskov et al. / Synthetic Metals 217 (2016) 266–275
273
Table 4 Influence of the formation conditions and dopants on the properties of the synthesized PC/PANI composite films. Composite
Calculated contents of the PANI salts in the composites, wt.%
Electrical conductivity, S/cm Compression molded films
Cast films
PC/PANI-DBSA PC/PANI-DNNSA PC/PANI-BSA PC/PANI-TSA PC/PANI-NSA PC/PANI-NDSA Pristine PC
4.4 5.0 3.6 3.5 3.8 4.4 n/a
1.2 104 1.2 1011 4.5 1010 5.0 105 1.4 105 2.6 1011 1015
1.6 103 4.0 107 2.3 105 9.5 104 5.6 104 4.5 105
At the same time, independently on the preparation method the best conductivity is observed for the composite films with the large dopant DBSA probably due to its plasticizing and compatibilizing properties, which facilitate better quality of the doped PANI percolation network inside the PC matrix. One would expect a similar effect in the case of DNNSA dopant with two large nonyl substituents in the aromatic ring but instead we obtained much worse conductivity of the PC/PANI-DNNSA films (Table 4). Probably, bulky anions of DNNSA hinder contact between conducting PANI macromolecules and clusters that retards charge transport in the PANI-DNNSA network of the material. It should be emphasized here that the dopant type (small or large anions) did not have a specific influence on conductivity both of the compression molded and cast films (Table 4). In particular, in the case of the former practically dielectric conductivity was observed for the synthesized composites PC/PANI-DNNSA (large anions type), PC/PANI-BSA and PC/PANI-NDSA (small anions type). The lowest conductivity of these films in turn matches to some extent with the highest weight losses of the parent composites under temperature of compression molding (240 E, Table 3), which obviously could be a cause of a deterioration of both the doped PANI and its percolating network. On the other hand, additional causes of the total conductivity losses of the compression molded films can vary due to different structures of DNNSA, BSA and NDSA dopants. Indeed, the loss of conductivity of PC/PANIDNNSA composite can be probably caused by DNNSA solubility in the PC melt and, respectively, by the dopant distancing from PANI chains after thermal dissociation of PANI-DNNSA. In the case of PC/ PANI-BSA one the dopant BSA does not contain compatibilizing alkyl substituents and therefore, there can be opposite causes e.g. its insolubility in the melt after thermal dissociation of the PANIBSA salt and precipitation as a separate phase. Unlike BSA, compatibility of NDSA with PC matrix for account of naphthalene rings can be better. However, there can be here a negative impact of reprotonation and cross-linking of adjacent PANI macromolecules by the bifunctional anion of NDSA. As a consequence, this PANINDSA becomes cross-linked and forms larger clusters that results in a degraded percolating network in the PC/PANI-NDSA compression molded film. We have found earlier on the example of the compression molded PC/PANI-TSA films that this composite demonstrates a typical percolation behavior in spite of the thermal dissociation of the doped PANI in the PC melt [21]. However if to compare the percolation curve of these films with that of the cast films one can see that a difference in conductivity between these composites significantly decreases with the increase of the PANI-TSA content from ca. 4.3 orders of magnitude (at 0.8 wt.%) to ca. 0.5 ones (at 18.7 wt.%) (Fig. S2). This phenomenon strongly suggests the existence of a concentration suppression of the reversible thermal dissociation of PANI salt at its high contents in the composite. Due to the conducting PANI component and its physicalchemical properties [45] the synthesized composites can serve as materials sensitive to different pollutants. We estimated this
ability by exposing their drop-cast thin films (excluding the highly resistive PC/PANI-DNNSA one) to ammonia-air mixtures with concentrations of ammonia in the range of 100 ppm–1000 ppm. As one can see from Fig. 6 there is not a specific influence of the type (small or large anions) and structure of the dopants on the sensor responses of these films. In particular, the most conductive PC/PANI-DBSA and the much less conductive PC/PANI-BSA films (1.6 103 S/cm and 2.3 105 S/cm, respectively, Table 4) have practically similar weak responses to ammonia while the other composite films have the significantly better sensitivity up to 0.6 %/ppm in the case of the PC/PANI-TSA one (Fig. 6). Based on the above results and well-known PANI sensing ability it is difficult to explain why we observed these significant differences in the sensitivity of these composites. Indeed, the strong acidic nature of the used sulfonic acids-dopants (pKa < 2) facilitates formation of the PANI salts with similar strength of the acid-base interactions of the acid-dopant and PANI base. Therefore, one would expect similar sensor responses of the composites to such basic substance as ammonia. Nevertheless, the difference in sensitivity between the PC/PANI-DBSA (PC/PANI-BSA) films with PC/PANI-TSA (PC/PANI-NSA and PC/PANI-DNSA) ones suggests additional effects (e.g. hydrophobic and hydrophilic interactions) in the interaction of ammonia with the PANI salts. Indeed, this interaction can be modulated by embedding the doped PANI clusters in the hydrophobic PC matrix after dispersing the composite powders in chloroform followed by dissolution of PC cores and casting of the formed mixture onto electrodes. However, for better understanding of these effects additional study of the PC/ PANI-doped composites is needed.
Fig. 6. Sensor responses (calibration curves) of the cast composite films to different concentrations of gaseous ammonia in the mixtures with air and their sensitivity comparison.
274
Y. Noskov et al. / Synthetic Metals 217 (2016) 266–275
In general, the linear calibration curves of the responses and good sensitivity of the PC/PANI-TSA, PC/PANI-NSA and PC/PANIDNSA drop-cast films suggest their applicability as sensing materials.
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. synthmet.2016.04.015.
4. Conclusion References This study testify that peculiarities of the aniline polymerization in the PC powder dispersion in water containing the aromatic sulfonic acid strongly depend on its anion size and surface activity. In particular, in the case of DBSA and DNNSA with the large surface active anions (containing long alkyl chains) the polymerization runs 2–5 times slower than in the case of the acids with small anions and less surface activity. We assign this difference to the coexistence in the DBSA and DNNSA solutions of two types of micelles formed of only the acids anions and of the acids anions involving anilinium cations that causes some screening of the latter by the former from access of the oxidant anions to the monomer. Variations of the doped PANI yield in the composites obviously relate to this phenomenon too. Thus, this yield is higher in the case of acids with small anions (BSA, TSA, NSA and DNSA). The difference in the effect of the acid-dopant anion type (large or small) on the composite properties was the most conspicuous in the thermal stability. In particular, in the case of the large dopants the composites had the reduced thermostability while the composites with the small dopants displayed the increased thermal stability as compared with the pure PC, respectively. As a possible reason for this difference we suggest intermolecular interactions of the plasticizing large dopant anions and/or thermally released dopant molecules with PC chains that in turn can lead to weakening of both these interactions and thermal stability of the PC/PANI-DBSA(DNNSA) composites. The thermally induced changes in the composites affected their electrical property that was well seen from higher conductivity of the cast composite films as compared with that of compression molded films. As the main reason for this difference we consider the thermal dissociation of the PANI salt followed by dissolution of the released acids-dopants in the melt under high temperatures of the films formation. Independently of the preparation method, the film sample doped with the large dopant DBSA had a higher conductivity as compared with other dopant cases. It can be explained by its plasticizing and compatibilizing effect which improves the quality of the percolation network in the PC matrix. Based on specificity of the used dopants, the differences in conductivity of the all synthesized composites were interpreted by the behavior of the doping agents in the PC matrix during the film preparation process and by its influence on the percolation network. As application, sensing properties of the cast composite films were studied showing the ability of these composites to detect gaseous ammonia in the ppm concentration range. Acknowledgements This work is partially supported in the frames of the project “The formation, properties and interactions of nanocomposites of conducting polymers and bioactive compounds in heterophase systems” of the NASU program of fundamental research and the complex scientific-technical program “Sensing devices for medical–ecological and industrial–technological problems: metrology support and trial operation” of National Academy of Sciences of Ukraine for a partial financial support. Sergei Mikhaylov acknowledges the Mines Douai and University of Lille 1 for PhD scholarship and Armines for the partial financial support. Authors are also thankful to Mrs. E. Fedorenko for assistance with FTIR measurements.
[1] A. Pron, P. Ranou, Processible conjugated polymers: from organic semiconductors to organic metals and superconductors, Prog. Polym. Sci. 27 (2002) 135–190. [2] H.S. Nalwa, Handbook of Organic Conductive Molecules and Polymer, 2, Wiley, 1997. [3] E.T. Kang, K.G. Neoh, K.L. Tan, Polyaniline: a polymer with many interesting intrinsic redox states, Prog. Polym. Sci. 23 (1998) 277–324. [4] A.A. Pud, Yu.V. Noskov, A. Kassiba, K.Yu. Fatyeyeva, N.A. Ogurtsov, M. Makowska-Janusik, W. Bednarski, M. Tabellout, G.S. Shapoval, New aspects of the low-concentrated aniline polymerization in the solution and in SiC nanocrystals dispersion, J. Phys. Chem. B 111 (2007) 2174–2180. [5] J. Stejskal, R.G. Gilbert, Polyaniline, Preparation of a conducting polymer, Pure Appl. Chem. 74 (5) (2002) 857–867. [6] V. Kumar, T. Yokozeki, T. Goto, T. Taka, Synthesis and characterization of PANIDBSA/DVB composite using roll-milled PANI-DBSA complex, Polymer 86 (2016) 129–137. [7] C.R. Martins, P.S. de Freitas, M.-A. De Paoli, Physical and conductive properties of the blend of polyaniline/dodecylbenzenesulphonic acid with PSS, Polym. Bull. 49 (2003) 379–386. [8] B.H. Jeon, S. Kim, M.H. Choi, I.J. Chung, Synthesis and characterization of polyaniline-polycarbonate composites prepared by an emulsion polymerization, Synth. Met. 104 (1999) 95–100. [9] A. Pud, N. Ogurtsov, A. Korzhenko, G. Shapoval, Some aspects of preparation methods and properties of polyaniline blends and composites with organic polymers, Prog. Polym. Sci. 28 (2003) 1701–1750. [10] D. Freitag, U. Grigo, P.R. Müller, W. Nouvertne, Polycarbonates/in Encyclopedia in Polymer Science and Engineering, 11, John Willey & Sons, New York, 1988, pp. 648. [11] F.S. Kim, G. Ren, S.A. Jenekhe, One-dimensional nanostructures of p-conjugated molecular systems: assembly, properties, and applications from photovoltaics, sensors, and nanophotonics to nanoelectronics, Chem. Mater. 23 (2011) 682–732. [12] W.J. Lee, Y.J. Kim, S. Kaang, Electrical properties of polyaniline/sulfonated polycarbonate blends, Synth. Met. 113 (2000) 237–243. [13] P.S. Rao, S. Subrahmanya, D.N. Sathyanarayana, Polyaniline-polycarbonate blends synthesized by two emulsion pathways, Synth. Met. 143 (2004) 323– 330. [14] M.G. Han, S.K. Cho, S.G. Oh, S.S. Im, Preparation and characterization of polyaniline nanoparticles synthesized from DBSA micellar solution, Synth. Met. 126 (2002) 53–60. [15] M. Sniechowski, D. Djurado, B. Dufour, P. Rannou, A. Pron, W. Luzny, Direct analysis of lamellar structure in polyaniline protonated with plasticizing dopants, Synth. Met. 143 (2004) 163–169. [16] A.A. Pud, Yu.V. Noskov, G.V. Dudarenko, G.S. Shapoval, Effect of the nature of acid dopant and oxidizer on the polymerization of aniline in the presence of polycarbonate dispersion, Theor. Experim. Chem 44 (2008) 54–59. [17] A.A. Pud, Yu.V. Noskov, N.A. Ogurtsov, O.P. Dimitriev, Yu.P. Piryatinski, N.M. Osipyonok, P.S. Smertenko, A. Kassiba, K.Yu. Fatyeyeva, G.S. Shapoval, Formation and properties of nano- and micro-structured conducting polymer host-guest composites, Synth. Met. 159 (2009) 2253–2258. [18] J. Stejskal, O. Quadrat, I. Sapurina, J. Zemek, A. Drelinkiewicz, M. Hasik, I. Krivka, J. Prokes, Polyaniline-coated silica gel, Eur. Polym. J. 38 (2002) 631–637. [19] V.N. Bliznyuk, A. Baig, S. Singamaneni, A.A. Pud, K.Yu. Fatyeyeva, G.S. Shapoval, Effects of surface and volume modification of poly(vinylidene fluoride) by polyaniline on the structure and electrical properties of their composites, Polymer 46 (2005) 11728–11736. [20] C. Barthet, S.P. Armes, M.M. Chehimi, C. Bilem, M. Omastova, Surface characterization of polyaniline-coated polystyrene latexes, Langmuir 14 (1998) 5032–5038. [21] N.A. Ogurtsov, Yu. Noskov, K.Yu. Fatyeyeva, V.G. Ilyin, G.V. Dudarenko, A.A. Pud, Deep impact of the template on molecular weight, structure, and oxidation state of the formed polyaniline, J. Phys. Chem. B 117 (17) (2013) 5306–5314. [22] Y. Cao, P. Smith, A.J. Heeger, Counter-ion induced processibility of conducting polyaniline and of conducting polyblends of polyaniline in bulk polymers, Synth. Met. 48 (1992) 91–97. [23] O.T. Ikkala, J. Laakso, K. Väkiparta, E. Virtanen, H. Ruohonen, H. Järvinen, T. Taka, P. Passiniemi, J.-E. Österholm, Y. Cao, A. Andreatta, P. Smith, A.J. Heeger, Counterion induced processibility of polyaniline-conducting melt processible polymer blends, Synth. Met. 69 (1–3) (1995) 97–100. [24] K. Tzou, R.V. Gregory, Kinetic study of the chemical polymerization of aniline in aqueous solutions, Synth. Met. 47 (3) (1992) 267–277. [25] N.A. Ogurtsov, Yu.V. Noskov, A.A. Pud, Effect of multi-wall carbon nanotubes on the kinetics of the aniline polymerization: the semi-quantitative OCP approach, J. Phys. Chem. B 119 (2015) 5055–5061.
Y. Noskov et al. / Synthetic Metals 217 (2016) 266–275 [26] P.J. Kinlen, J. Liu, Y. Ding, C.R. Graham, E.E. Remsen, Emulsion polymerization process for organically soluble and electrically conducting polyaniline, Macromolecules 31 (6) (1998) 1735–1744. [27] I. Sapurina, J. Stejskal, The mechanism of the oxidative polymerization of aniline and the formation of supramolecular polyaniline structures, Polym. Int. 57 (12) (2008) 1295–1325. [28] H.S. Kolla, S.P. Surwade, X. Zhang, A.G. MacDiarmid, S.K. Manohar, Absolute molecular weight of polyaniline, J. Am. Chem. Soc. 127 (2005) 16770–16771. [29] N. Kohut-Svelko, S. Reynaud, J. Francois, Synthesis and characterization of polyaniline prepared in the presence of nonionic surfactants in an aqueous dispersion, Synth. Met. 150 (2) (2005) 107–114. [30] M. Mashimo, H. Sato, M. Ueda, I. Komasawa, Extraction equilibria of aluminum and beryllium from sulfate media by mixture of bis(2-ethylhexyl) phosphoric acid and dinonylnaphthalene sulfonic acid, J. Chem. Eng. Jap. 30 (4) (1997) 706–711. [31] M.G. Han, S.K. Cho, S.G. Oh, S.S. Im, Preparation and characterization of polyaniline nanoparticles synthesized from DBSA micellar solution, Synth. Met. 126 (2002) 53–60. [32] Z. Ping, In situ FTIR reflection spectroscopic investigations on the base-acid transitions of polyaniline, J. Chem. Soc. Faraday Trans. 92 (1996) 3063–3067. nková, J. Stejskal, Spectroscopy of thin [33] M. Trchová, Z. Morávková, I. Šede polyaniline films deposited during chemical oxidation of aniline, Chem. Pap. 66 (2012) 415–445. [34] D.W. Hatchett, M. Josowicz, J. Janata, Acid doping of polyaniline: spectroscopic and electrochemical studies, J. Phys. Chem. B 103 (1999) 10992–10998. [35] X.-G. Li, M.-R. Huang, J.-F. Zeng, M.-F. Zhu, The preparation of polyaniline waterborne latex nanoparticles and their films with anti-corrosivity and semiconductivity, Colloid Surf. A 248 (2004) 111–120.
275
[36] X. Lu, H.Y. Ng, J. Xu, C. He, Electrical conductivity of polyanilinedodecylbenzene sulphonic acid complex: thermal degradation and its mechanism, Synth. Met. 127 (2002) 167–178. [37] G. Montaudo, S. Carroccio, C. Puglisi, Thermal and thermoxidative degradation processes in poly(bisphenol a carbonate), J. Anal. Appl. Pyrolysis 64 (2002) 229–247. jka, J. Brodinová, A. Kalendová, J. Prokeš, J. Stejskal, [38] M. Trchová, P. Mate Structural and conductivity changes during the pyrolysis of polyaniline base, Polym. Degrad. Stabil. 91 (1) (2006) 114–121. [39] J. Stejskal, J. Prokeš, M. Trchová, Reprotonated polyanilines: the stability of conductivity at elevated temperature, Polym. Degrad. Stab. 102 (2014) 67–73. [40] J. Bhadra, N.K. Madi, N.J. Al-Thani, M.A. Al-Maadeed, Polyaniline/polyvinyl alcohol blends: effect of sulfonic acid dopants on microstructural, optical, thermal and electrical properties, Synth. Met. 191 (2014) 126–134. [41] N. Ogurtsov, Yu. Noskov, V. Bliznyuk, V. Ilyin, J.-L. Wojkiewicz, E. Fedorenko, A. Pud, Evolution and interdependence of structure and properties of nanocomposites of multiwall carbon nanotubes with polyaniline, J. Phys. Chem. C 120 (2016) 230–242. [42] E.H. Immergut, H.F. Mark, Principles of plasticization, chapter 1, in: N. Platzer (Ed.), Plasticization and Plasticizer Processes, 148, Advances in Chemical American Chemistry Society, Washington, DC, 1965, pp. 1–26. [43] Q.-H. Zhang, X.-H. Wang, D.-J. Chen, X.-B. Jing, Electrically conductive, meltprocessed ternary blends of polyaniline/dodecylbenzene sulfonic acid, ethylene/vinyl acetate, and low-density polyethylene, J. Polym. Sci. Polym. Phys. 42 (2004) 3750–3758. [44] L.F. Malmonge, L.H.C. Mattoso, Thermal-analysis of conductive blends of PVDF and poly(o-methoxyaniline), Polymer 41 (2000) 8387–8391. [45] J.L. Wojkiewicz, V.N. Bliznyuk, S. Carquigny, N. Elkamchi, N. Redon, T. Lasri, A.A. Pud, S. Reynaud, Nanostructured polyaniline-based composites for ppb range ammonia sensing, Sens. Actuators B Chem. 160 (2011) 1394–1403.