**4. Discussion**

In this work, we show that large sized (4 cm × 5 cm) activated carbon cloths can be easily modified with electrochemically deposited polyaniline by closely attaching the fabric to a stainless steel anode in a filter-press type cell. A simple galvanostatic procedure at a constant current of 100 mA (~14 mA g<sup>−</sup><sup>1</sup> cm<sup>−</sup>2) allowed the electropolymerization process to be carried out at an anodic potential varying in the range 0.6–0.8 V vs. Ag/AgCl. This electrode potential is su fficient to oxidize aniline monomers, to generate enough PAni nucleation sites and to ensure fast chain growth [16,55,62]. A simple control of the PAni loading could be achieved by passing di fferent amounts of charge, i.e., by adjusting the elapsed

electropolymerization time. The highly porous carbon fabric provides the hybrid composite material with high specific surface area, flexibility and mechanical strength to mitigate the stress associated with polymer volume changes accompanying doping/dedoping [11,12,22], while PAni adds some new enhanced characteristics related to its reversible redox properties and electron conductivity [2,12,62].

It is well known that PAni and its composites can be synthesized with a wide variety of microand nanoscale structures, with a surprising number of di fferent sizes and shapes, from smooth thin films to nanofibers, nanorods or nanotubes to globular or granular agglomerates [26,27]. The observed supramolecular organization and structure is strongly dependent on the method of synthesis and the experimental conditions [6,26,27]. As for electrochemical methods, the applied current/potential and the way the electrode is polarized (i.e., potentiostatically, galvanostatically, by cyclic potential sweep or pulses), as well as the nature and surface condition of the electrode support, all have a dramatic e ffect on the morphology of the deposited polymer film [6]. In the present work, the film morphology varied with processing time from dense and uniform thin films covering individual carbon fibers to a loosely adherent and porous, thick coating of nanofibrous or nanowire-shaped structures. Three-dimensional networks of nanofibrilar or nanowire structures seem the most frequently encountered morphology for PAni deposited on high surface area carbon substrates (e.g., carbon fibers or cloths), either by electrochemical [20,21,62,63] or conventional in-situ chemical polymerization [22–24]. However, in agreemen<sup>t</sup> with our results, some authors have obtained smooth dense layers [8,63] or a transition from this morphology to irregular or randomly connected nanofibrilar aggregates upon extending the polymerization time [12,62]. It has also been reported that long reaction time [23] or prolonged galvanostatic oxidation [20] lead to incontrollable growth among carbon fibers and result in a detrimental e ffect via peeling-o ff from the carbon surface.

Apart from the scattered compact deposits seen in Figure 2, no clear evidence of a PAni film uniformly covering the carbon fibers is provided by SEM imaging. However, the increase in the surface N/C ratio measured by XPS at both sides of the carbon cloth suggests that an (ultra)thin film develops within this timespan. Also, the PAni loadings deduced from thermogravimetric analysis (Table 2) indicate that there is more polymer than what is seen by SEM. We believe that during the pre-conditioning step, anilinium cations are pre-adsorbed on the internal porosity of the carbon cloth [11,55], thus acting as nucleation seeds and reacting with liquid-phase aniline to yield ultra-thin films covering the pore walls. This view agrees with the results by Salinas-Torres and coworkers [17], who used position-resolved micro-Small Angle X-Ray Scattering (SAXS) to demonstrate that a PAni layer of sub-nanometric thickness grows inside the microporosity of activated carbon fibers. The particular non-linear increase in the PAni content revealed by both XPS and thermal analysis (Figures 4 and 8, Table 2) might be explained by a decrease in the polymerization rate caused by the depletion of aniline monomers inside the pores and the induced concentration-gradient between the external bulk solution and the solution filling the pores. Note that dye adsorption studies point out that the bare carbon material possesses a large pore mass transfer resistance (see below). Slow internal pore di ffusion provides enough time for self-assembly of intermediate oligomers into thin films. Once PAni reaches the outer carbon surface, external film di ffusion facilitates fast access of solution aniline to the growing chains and the polymerization rate increases again to build a 3D network of nanofibrilar structures and nanowire aggregates.

As long as a thin film develops on the carbon fabric surface, XPS shows mixed spectroscopic characteristics from both the support and PAni. When a thick coating is formed, XPS is typical of pure PAni, close to the semioxidized emeraldinde salt state. The temperature-induced changes in the hybrid composites (Figure 8) are also compatible with the thermal behavior of PAni in its doped emeraldine salt state [50–53]. The evolution with the electropolymerization time of the dry composite total weight loss and the calculated PAni loading (Figure 8 and Table 2) are consistent with the trends shown by XPS analysis. The PAni loadings achieved in our hybrid composites ranged from ~5 (within the first 40 min of treatment) to 18 mg cm<sup>−</sup>2, at the longest polymerization time. These loadings are in the range of those reported by others for carbon fiber substrates [11,23,62,64].

The electrodeposition of PAni is accompanied by a decrease in the apparent BET area and pore volume. Earlier authors also described a remarkable blockage of micro- and mesoporosity in composites of activated carbon fibers [8,11,17] or activated carbon fiber textiles [23,25,64] with PAni films, deposited either by chemical or electrochemical methods. In these reports, the surface area loss ranged from 20% to 40%. In the present investigation, the specific surface area decreases by ca. 20%–25% and remains basically unchanged within the time interval when a uniform and dense thin PAni film covers the carbon surface. The evolution of pore volumes (Table 1 and Figure 7c) indicates that PAni does not significantly grow inside the narrower micropores, in contrast to the results by Salinas-Torres and coworkers [17]. There is no clear evidence why polymerization inside ultramicropores is hampered, but it may arise from too short a contact time in the pre-conditioning step to allow di ffusion of aniline into such narrow pores. Also, we can tentatively argue that carbon fibers may possess some kind of hierarchical pore structure, with ultramicropores preferentially distributed near the fiber core. Thus, PAni films blocking the outer micro- and mesopores would hinder e fficient monomer supply to sustain polymerization inside the inner pores. In fact, position-resolved micro-SAXS studies revealed a heterogeneous radial distribution of electrosynthesized PAni [17], which preferentially accumulates in the outer shells of carbon fibers. The thick nanofibrous PAni coating grown at prolonged electropolymerization time virtually blocks all the porosity, thereby causing an abrupt decrease in the specific surface area to reach a typical value of pure PAni precipitates [65].

The reversible redox transitions of PAni doping/dedoping processes contribute to moderately enhance the capacitive response of hybrid PAni-ACC composites. The extra pseudocapacitive features observed in the CVs of the hybrid composites (Figure 9) arise from the leucoemeraldine-to-emeraldine state transition (0.2–0.4 V) and a reversible electron-transfer (~0.0 V) probably involving phenazine- or phenoxazine-like segments in the polymer backbone. These latter redox features appear in polymeric materials of the PAni family, like oligomers of o-aminodiphenylamine [41,56], ladder polymers like poly-o-phenylenediamine [57] or poly-o-aminophenol [56], PAni- and poly-o-anisidine-clay nanocomposites [66,67] or reduced graphene oxide-PAni-ACC composites [63]. More recently, it has been claimed that even chemically synthesized PAni contain constitutional phenazine and N-phenylphenazine segments [68] that play a key role in the self-assembly of polymer chains to build di fferent supramolecular structures [27].

The specific capacitance of the hybrid composites increases moderately in hybrid composites with thin dense PAni films. (Table 3). The reported specific capacitance for pure Pani-modified electrodes lies within a broad range from 160 to 815 <sup>F</sup>/g [69], while the theoretical capacitance for Pani at a 50% doping level is 750 <sup>F</sup>/g [70]. Upon considering this latter value, the expected capacitance for a 20 wt.% PAni-ACC composite would be about 250 <sup>F</sup>/g, but if one considers the lowest reported PAni capacitance, an estimated value of about 134 <sup>F</sup>/g would be obtained, which is close to the capacitances listed in Table 3. Then, the slight increase in composite capacitance could be due to the electrodeposition of PAni with a small specific capacitance. Alternatively, it can also be a consequence of the deleterious effect of decreasing the available surface area, which counteracts both the PAni extra capacitance and the reduced sheet resistance. Composites with the highest PAni loading densities show a strongly distorted CV and a decreased capacitance, as a consequence of the increased sheet resistance and the abrupt diminution of the surface area caused by the thick nanofibrous coating. Note that VB analysis revealed that the nanofibrous PAni layers have a higher DOS near the Fermi level (Figure 6). Then, the increased *R*s and lessened electrochemical properties imparted by nanofibrous PAni layers should not be caused by poor intrinsic electron conductivity of PAni, but related to high PAni–PAni and/or PAni-carbon intraparticle resistance to charge transport.

The evolution of the specific capacitance observed in this work is in agreemen<sup>t</sup> with the results reported earlier by other researchers [12,16,21,23,62]. It has been generally shown that the capacitance is enhanced by either smooth homogeneous PAni coatings of small thickness [12] or nanofiber-shaped assemblies with a certain degree of 3D order and small diameter [22,23] obtained at low PAni loading densities. Such thin nanostructures are believed to provide fast access of electrolyte and shortened

path lengths for ion and electron transport. When the loading is raised at longer polymerization times and/or higher monomer concentration, thicker polymer films build up with a high degree of particle agglomeration [12,23,24]. This results in an increase of the di ffusion resistance of electrolyte ions and also in a less e fficient electrical contact between the conducting polymer and the textile carbon [22].

Finally, we studied the e ffect of the modification of the ACC by PAni on the adsorption of Acid Red 27. The dye uptake on both bare ACC and PAni-ACC composites obeys PSO kinetics, with the involvement of mass transport on the adsorption rate control. Our results point out that PAni thin films of moderate loading densities provide accelerated PSO adsorption kinetics of Acid Red 27 from aqueous solution, while keeping the maximum adsorption capacity unchanged. However, the values of *k*2 (Table 5) and the apparent intraparticle di ffusion coe fficients from Boyd plots (Table S4) are still below the range of those commonly reported for dye adsorption on many di fferent adsorbents (in the order of 10−2–10−<sup>4</sup> g mg<sup>−</sup>1min−<sup>1</sup> and ~10−<sup>17</sup> m<sup>2</sup> s<sup>−</sup>1, respectively) [14,65,71–73]. These low values signify that the internal mass transfer resistance is very high and is dominated by the carbon porous structure, thereby explaining the very long time taken by the adsorption system to attain equilibrium. The adsorption of dyes on PAni powder [10,74] or its nanocomposites with other conducting polymers [65], carbonaceous materials, metal oxides, and low-cost bioadsorbents [9,14,19] has been extensively studied to date, with a general consensus that the adsorption rate is governed by PSO kinetics. The uptake of dyes on PAni-based adsorbents seems to be driven by di fferent types of binding forces, namely, π–π attraction, hydrogen bonding or electrostatic interaction between charges residing on both the PAni backbone and the dye functional groups [9,14]. In our case, the electrostatic interaction between negatively-charged sulfonate groups in Acid Red 27 and positively-charged N sites in the backbone of the acid-doped PAni films overrides the loss in the specific surface and leads to adsorption improvement. Further, the exchange of some (bi) sulfate ions with dye molecules as dopant anions should not be ruled out as an additional mechanism to promote their incorporation to the hybrid adsorbent. A similar explanation was put forward by other authors for the promoted adsorption of anionic sulfonated dyes on PAni [9,10] or PAni-based nanocomposites [19,65]. Despite the fact that the same interaction forces are operative when a thick nanofibrous network of PAni is formed at the longest treatment time, the strong blockage of carbon pores and the abrupt decrease in the specific surface area resulted in an important adsorption rate slowdown and a dramatic loss in the maximum adsorption capacity.
