Porous Poly(3,4-ethylenedioxythiophene)-Based Electrodes for Detecting Stress Biomarkers in Artiﬁcial Urine and Sweat

When danger is perceived, the human body responds to overcome obstacles and survive a stressful situation; however, sustained levels of stress are associated with health disorders and diminished life quality. Hence, stress biomarkers are monitored to control stress quantitatively. Herein, a porous sensor (4l-COP/p) composed of poly(3,4-ethylenedioxythiophene) (PEDOT) and poly(3,4-ethylenedioxythiophene-co - N -methylpyrrole) (COP), which is prepared in a four-layered fashion to detect dopamine (DA) and serotonin (5-HT), is presented. Speciﬁcally, the detection is conducted in phosphate-buﬀered saline (PBS), as well as artiﬁcial urine and sweat, by applying cyclic voltammetry. The limit of detection values obtained are as low as 5.7 × 10 − 6 and 1.4 × 10 − 6 m for DA and 5-HT, respectively, when assessed individually in artiﬁcial urine. When mixed in PBS, 4l-COP/p detects both biomarkers with a resolution of 0.18 V and a sensitivity of 40 and 30 μ A mm − 1 for DA and 5-HT, respectively. Additionally, by theoretical calculations, the interaction pattern that each stress biomarker establishes with the PEDOT outer layer is elucidated. Whereas DA interacts with the 𝝅 -system of PEDOT, 5-HT forms speciﬁc hydrogen bonds with the conducting polymer chains. The resolution value obtained depends upon such interactions. Overall, 4l-COP/p electrodes display potential as stress sensing devices for healthcare technologies.


Introduction
Stress has become a constant element in our daily life. For some, stressful situations imply the opportunity to overcome a challenge and learn from it; however, for the majority, stress involves biomarkers are generally found in fluids, as well as being mixed with other substances (i.e., proteins), their detection has been accomplished by applying a wide range of sensing methods, such UV-vis spectroscopy, [5] electrochemical techniques, [6][7][8][9] high pressure liquid chromatography, liquid chromatographymass spectrometry tandem, [10] or microfluidic devices, [11,12] in addition to immunoassays, luminescence, and fluorescence assays, [13] which are routinely conducted in clinical laboratories. For instance, cortisol, a widely studied steroid hormone secreted in response to stress at the adrenal cortex, has been measured in human blood, [14] serum, [14,15] urine, [14] saliva, [6,14,15] sweat, [10,12] and human interstitial fluid. [8] For more examples of publication highlights, we direct the reader to the comprehensive review by Steckl and Ray, [4] which summarizes detection mechanism, materials, and biofluids, among other parameters, of stress biomarkers sensors.
Recently, wearable electronics and smart textiles have been also exploited as platforms for the detection of stress biomarkers. [12,16] For example, stretchable electronics have been applied to quantify those key parameters indicative of the performance and stress level of athletes. [17] Among the materials available for this purpose, the remarkable physical and electrochemical properties of conducting polymers (CPs) provide the basis for an excellent sensing response. The selective detection of troponin in a wide physiological range (1-100 ng mL −1 ) was achieved by electrochemically coating polyaniline onto paper with a sensitivity of 5.5 μA ng −1 mL −1 cm −2 ). [18] Martins et al. optimized their biosensor design by adding nanostructured carbon materials and poly(3,4-ethylenedioxythiophene) (PEDOT), which enhanced the electrocatalytic properties of the paper-modified sensor, thus yielding a low detection limit (14.4 ng mL −1 ). [9] Kim et al. reported a rapid and sensitive cortisol biosensor based on a polypyrrole nanotube field-effect transistor. [19] To the best of our knowledge, and within the context of stress monitoring, little work has been done to detect dopamine and serotonin with CP-based devices. [20] Although CPs have been extensively used to detect dopamine, it has been done mainly as the neurotransmitter involved in Alzheimer's disease, whereas few works report the detection of serotonin but by employing other electrode materials, such as glassy carbon, boron-doped diamond, carbon nanotube networks, or complex nanocomposites. [21][22][23][24] Therefore, in this work, we propose a four-layered allconducting polymer system as electrode able to detect those other stress biomarkers (i.e., DA and 5-HT). During the last decade, we have designed nanostructured conducting systems based on PEDOT and poly(N-methylpyrrole) (PNMPy), [25][26][27][28][29][30][31] as well as the corresponding copolymer, poly(3,4-ethylenedioxythiopheneco-N-methylpyrrole) (COP), which presents a polymer network that organizes in NMPy-and EDOT-rich domains. [28,32] Among the parameters optimized for this system, we focused on physical characteristics (i.e., varying thickness of the layers, from micro- [25][26][27][28] to nanometric layers, [28][29][30][31] as well as the introduction of interface porosity [33] ), composition (i.e., modification with clays [27] or an octanethiol self-assembled monolayer [31] ), and/or polymerization conditions (i.e., dynamic conditions rather than static solutions [26] ).
Herein, the electrodes, which were prepared by the layer-bylayer electrodeposition technique, contain two consecutive COP layers, which act as a double-dielectric sheet, in between of two electroactive PEDOT layers. Such arrangement promotes the synergy between the CPs at the interfaces, thus enhancing their electrochemical properties and increasing their ability to store charge in comparison to the individual CP. [33] Indeed, the key feature of this system is the presence of an ultraporous nanostructured interface in the middle of the dielectric (i.e., PEDOT/COP/p/COP/PEDOT, where p refers to porous). Such interface porosity, obtained by growing NaCl crystals and then eliminating them with water etching, causes structural distortions in the polymer network that, in conjunction with the phase-segregated structure of COP, improved its supercapacitor response. [33] Hence, in order to exploit such enhanced performance, the four-layered all-conducting polymer system has been applied as the main component in a sensing device to monitor stress biomarkers levels (DA and/or 5-HT) in phosphate-buffered saline (PBS), as well as in simulated body fluids, i.e., artificial urine and sweat. For the sake of comparison, sensors with an intermediate region made of PNMPy were also investigated (i.e., PEDOT/PNMPy/p/PNMPy/PEDOT). Finally, density functional theory (DFT) calculations were conducted on model systems to provide understanding of the interactions between the studied biomarkers and the external PEDOT layer, as well as of their influence in the oxidation potential. 4l-COP/p was very successful for the electrochemical detection of DA and 5-HT, either individually or mixed, in spite of the complexity of the two chosen biomarkers.

Synthesis of Four-Layered CP Films
All electrochemical experiments were conducted on a PG-STAT302N AUTOLAB potentiostat-galvanostat connected to a PC computer and controlled through the NOVA 2.1 software. Polymerizations were conducted in a three-electrode onecompartment cell under nitrogen atmosphere (99.995% in purity) at 25°C. Steel AISI 316 sheets of 1 × 1.5 cm 2 area were employed as working and counter electrodes. To avoid interferences during the electrochemical analyses, before each trial the www.advancedsciencenews.com www.mame-journal.de working and counter electrodes were cleaned with ethanol, after that with acetone, and dried in an air-flow. The reference electrode was a Ag|AgCl electrode containing a KCl saturated aqueous solution (E°= 0.222 V vs standard hydrogen electrode (SHE) at 25°C). All the potentials reported in this work were referred to the Ag|AgCl electrode.
CP single-films were prepared filling the cell with 40 mL of an acetonitrile solution of the corresponding monomer (10 × 10 −3 m) with 0.1 m LiClO 4 as doping electrolyte. In the case of COP, the concentration of EDOT and NMPy in the reaction medium was 5 × 10 −3 m each. Films were electrogenerated by chronoamperometry (CA) under a constant potential of 1.40 V and adjusting the polymerization charge to 0.55 C. To render them porous, NaCl crystals were grown by immersing the prepared films in a 20% w/v salt aqueous solution for 5 s and, subsequently, leaving them in a desiccator overnight for drying. Films with a porous surface were obtained by removing the grown salt crystals from CP/NaCl by simple water-etching. For this purpose, CP/NaCl films were immersed in water overnight and, subsequently, maintained in a desiccator 24 h for drying, the resulting films being denoted CP/p, where p refers to porous.
To produce four-layered systems, films containing two PEDOT layers separated by COP/COP, COP/NaCl/COP, or COP/p/COP were prepared by immersing the steel AISI 316 electrode coated with the corresponding single-layered PEDOT film in a cell filled with a 5 × 10 −3 m NMPy-, 5 × 10 −3 m EDOT-, and 0.1 m LiClO 4containing acetonitrile solution. The COP layer was electrogenerated onto the PEDOT film by CA under a constant potential of 1.40 V and adjusting the polymerization charge to 0.55 C. The same process was used to coat the resulting two-layered film, PE-DOT/COP, with another COP layer directly or after growing NaCl crystals, as described previously. Finally, another PEDOT layer was added applying the same procedure. The resulting multilayered films were named PEDOT/COP/NaCl/COP/PEDOT, respectively, the latter transforming into PEDOT/COP/p/COP/PEDOT after removing the inorganic crystals with water. The same procedure was used to prepare PEDOT/PNMPy/p/PNMPy/PEDOT four-layered films, although in this case the monomer concentration for NMPy was 10 × 10 −3 m.

Surface Characterization
The surface morphology of the prepared single-layered films was studied by scanning electron microscopy (SEM). Dried samples were placed in a focused ion beam Zeiss Neon 40 scanning electron microscope operating at 3 kV and equipped with an energydispersive X-ray spectroscopy system.

Electrochemical Characterization of Four-Layered COP-Based Electrodes
All electrochemical experiments were run in triplicate using an acetonitrile solution with 0.1 m LiClO 4 as supporting electrolyte. Cyclic voltammetry (CV) was carried out to evaluate the electroactivity, specific capacitance (SC), and the electrochemical stability of the prepared electrodes. The initial and final potentials were −0.50 V, and the reversal potential was 1.60 V. The number of oxidation-reduction cycles applied was 50. A scan rate of 50 mV s −1 was used in all cases.
The electrochemical activity was evaluated through the voltammetric charge corresponding to the 2nd oxidation-reduction cycle (Q 2nd ; in C). The SC (in F g −1 ) was determined from the registered voltammograms using the following equation where Q is the voltammetric charge, ΔV is the potential window (in V), and m is the mass of CP on the surface of the working electrode (in g). Finally, the electrochemical stability was estimated as the reduction of the SC after 50 consecutive oxidation-reduction cycles (ΔSC; in %).

Stress Biomarkers Detection using Porous Four-Layered COP-Based Electrodes
Electrodes for stress biomarker detection were prepared as described before and aged by applying 50 CV cycles from −0.5 to 1.6 V at 50 mV s −1 in acetonitrile with 0.1 m LiClO 4 as supporting electrolyte to stabilize the CP structure. Detection was conducted by CV (from −0.4 to 0.8 V at 50 mV s −1 ) in artificial fluids with increasing concentration of the biomolecule of interest at room temperature using steel AISI 316 sheets of 1 × 1.5 cm 2 area as working and counter electrodes. Data were extracted from the 2nd cycle. All detection experiments and characterization tests were run in triplicate (n = 3), data and error bar values were reported corresponding to the mean and SD values, respectively.

Computational Studies
DFT calculations at the B3LYP/6-311++G(d,p) level were performed in vacuum and in aqueous solution considering both DA and 5-HT stress biomarkers in their oxidized and reduced forms. In addition, DA···CP and 5-HT···CP model complexes were also calculated in those two environments, where CP is a model PE-DOT chain represented by four repeat units. Optimizations in solution were performed by applying the SMD universal solvation model developed by Truhlar and co-workers. [36] Not only does the SMD model consider the nonhomogenous Poisson equation for electrostatic component in terms of the integral equationformalism polarizable continuum model, but also includes an approximate description of the short-range interactions in the first solvation shell.
www.advancedsciencenews.com www.mame-journal.de Scheme 1. Schematic preparation of 4l-COP/p multilayered system used in this work for the detection of stress biomarkers DA and 5-HT.

Electrochemical and Morphological Characterization of Porous 4l-COP/p Films
The design of advanced components for biosensors to measure stress-indicative biomarkers in a noninvasive, cost-effective, and rapid approach would greatly benefit the field of healthcare technology. With this aim in mind, PEDOT/COP/p/COP/PEDOT electrodes, which were prepared following an already optimized protocol, [33] have been applied to quantify stress biomarkers.
In this four-layered system, two layers of PEDOT are separated by two layers of nanosegregated COP with a porous interface in the middle (Scheme 1). The interface porosity was obtained by immersing the prepared electropolymerized films in a 20% w/v salt aqueous solution and let the crystals grow overnight. After that, they were removed by simple water etching. The choice of PEDOT/COP/p/COP/PEDOT multilayered system as sensor for stress biomarkers was based on our previous results, where the organization of COP in NMPy-and EDOT-rich domains, in combination with interface porosity introduced in the middle of the dielectric, increased its charge storage capacity and overall electrochemical performance. [28,33] Hereafter, PEDOT/COP/p/COP/PEDOT and PEDOT/PNMPy/p/PNMPy/PEDOT multilayered systems are going to be referred as 4l-COP/p and 4l-PNMPy/p, respectively. Before testing the suitability of 4l-COP/p electrodes for the detection of stress biomarkers, their electrochemical performance was studied. Control voltammograms for 4l-COP/p, which were recorded in acetonitrile 0.1 m LiClO 4 between −0.50 V (initial and final potential) and 1.60 V (reversal potential), showed a smooth oxidation shoulder at 0.9 V (Figure 1). As it generally happens for multilayered systems, this oxidation peak was weak, while no reduction peak was detected. In good agreement with previous work, [33] the voltammetric charge (Q 2nd ) and specific capacitance (SC 2nd ) values determined after two consecutive oxidationreduction cycles were of 99 ± 14 mC and 45 ± 6 F g −1 , respectively. Moreover, the electrochemical stability was evaluated by submitting the four-layered system to 50 consecutive oxidationreduction cycles between −0.50 and 1.60 V, which is an assay that produces structural changes in the polymer network. After such amount of cycles, both the cathodic and anodic areas for the 50th cycle recorded were decreased in comparison to the 2nd cycle ( Figure 1). Specifically, the SC decreased to 23 ± 4 F g −1 , which indicates a reduction of 49% with respect to SC 2nd .
SEM characterization of the CP layers that constitute the fourlayered electrodes was conducted to aid in understanding the correlation between porosity and superficial features with the biosensor performance. The morphology of each of the pristine layers, mainly PEDOT, PNMPy, and COP, was in good agreement with previous observations (Figure 2). [33] Briefly, PEDOT was characterized by clustered small aggregates connected by dense networks of thin fiber-like structures (Figure 2a), while PN-MPy displayed a more compact globular morphology (Figure 2b). Being a copolymer, the surface morphology of COP layers displayed features characteristic of both of the individual homopolymers ( Figure 2c): homogenous globular and compact structures were identified for PNMPy, whereas the clustered and more heterogenous ones for PEDOT. Hence, the NMPy and EDOT blocks of the copolymer appeared segregated in two separate phases. [28] For COP/p films, the growth and elimination of NaCl crystals induced an increase of the surface porosity in comparison to pristine films, which is ascribed to the stress exerted by the growing inorganic crystals in the already open polymer network of EDOT-rich phases. [33] Hence, throughout the process of introducing interface porosity, EDOT-rich phases behaved as an adaptive surface that accommodated the crystals. Consequently, once the NaCl crystals were dissolved, EDOT-rich phases appeared much more porous than NMPy-rich phases, which remained more compact. Indeed, the growth of NaCl crystals and their following dissolution yielded COP/p films with a remarkable open surface (Figure 1c). In contrast, despite being submitted to the same procedure, the surface morphology of PNMPy/p films (Figure 1d) remained similar to that shown by pristine PNMPy. In this case, the surface of PNMPy, which was already closed and compact without pores, hindered the penetration of salt crystals during the growing steps, which were simply deposited onto the polymeric surface and removed during the water etching without modifying the surface. A porous morphology is beneficial in that it facilitates the access and escape of dopant ions during oxidation and reduction processes, respectively, thus improving the electrochemical features of the CP-based electrode. Hence, with the aim of exploiting the adequate morphological and electrochemical features of 4l-COP/p films, their performance as biosensor to detect stress biomarkers was evaluated.

Stress Biomarkers Detection in PBS Solution
Among the studied stress-related biomarkers found in the literature, [4] we have placed our attention in detecting DA and 5-HT, two neurotransmitters that are generated by the nervous system and are related to mood/appetite regulation, as well as feelings of happiness. Both biological molecules display redox activity and, in healthy individuals under normal conditions, are found in blood (0 × 10 −9 to 0.25 × 10 −9 m and 0.6 × 10 −6 to 2 × 10 −6 m for DA and 5-HT, respectively), urine (0.3 × 10 −6 to 3 × 10 −6 m and 0.03 × 10 −6 to 0.13 × 10 −6 m for DA and 5-HT, respectively), and sweat (under study for DA). [4] Hence, to determine and optimize the performance of the four-layered systems as detection units, several experiments were carried out, first in 0.1 m PBS, and later considering other simulated body fluids. As a comparative system, 4l-PNMPy/p electrodes, where COP has been replaced by PNMPy thus displaying the same layer distribution but different morphology, have been also analyzed.
CV, which is one of the most-important and most-widespread applied electrochemical techniques when designing electrochemical sensors, allows for simple measurement procedures and short response time, as well as in situ measurements based on direct detection. Figure 3a compares the voltammetric response of 4l-COP/p and 4l-PNMPy/p electrodes for the detection of DA (100 × 10 −6 m) in 0.1 m PBS from −0.2 to 0.8 V at 50 mV s −1 . Both systems are able to detect the oxidation of DA, the anodic peak potential (E p ) being 0.21 and 0.19 V for 4l-COP/p and 4l-PNMPy/p, respectively. Although for 4l-COP/p the anodic peak intensity (I p ) associated with such process (2.5 mA) is lower than that displayed by 4l-PNMPy/p (5.1 mA), the oxidation peak for DA is very well resolved.
The variation of the anodic peak intensity (I p ), as determined by CV, against the DA concentration yields a calibration curve with a linear range from 5 × 10 −6 to 100 × 10 −6 m (Figure 3a). The slope of the linear region denotes a sensitivity of 35 μA mm −1 (R 2 = 0.9678) for 4l-PNMPy/p, and of 19 μA mm −1 (R 2 = 0.9942) for 4l-COP/p. Detection parameters for the four-layered electrodes studied in this work are listed in Tables 1 and 2. Moreover, the resolution of detection can be expressed as 3· , where is the standard deviation of the blank. From this, the LOD value can be calculated as resolution/sensitivity. Thus, the LOD was determined to be 15 × 10 −6 and 8 × 10 −6 m for 4l-COP/p and 4l-PNMPy/p, respectively. Finally, the limit of quantification (LOQ), which is expressed as 10· /sensitivity, resulted in a value of 52 × 10 −6 and 26 × 10 −6 m for 4l-COP/p and 4l-PNMPy/p, respectively. Even though 4l-PNMPy/p electrodes display slightly better sensitivity and lower LOD/LOQ values in comparison to 4l-COP/p, the detection parameters for the latter are in good agreement with previous sensors based on CPs for the detection of DA, [9,18,19] and were steadier with DA concentration. We ascribe such response to the more open morphology of 4l-COP/p that enhanced the flow of dopant ions from the solution to the polymer film, and the other way around.
In contrast, the response of 4l-COP/p and 4l-PNMPy/p systems toward the detection of 5-HT was hindered at concentrations higher than 40 × 10 −6 m, possibly by the adsorption-limited nature of 5-HT in electrochemical detection [21,23,37] that creates a concentration dependency in electrode sensitivity. When increasing 5-HT concentration, I p increased until saturation (Figure 3b), which indicates that 5-HT molecules adsorb on the electrode up to a point where its sensitivity is compromised. Hence, the detection of 5-HT was assessed considering a narrower concentration range, from 5 × 10 −6 to 35 × 10 −6 m.
The voltammetric response of 4l-COP/p and 4l-PNMPy/p electrodes for the detection of 5-HT (35 × 10 −6 m) in 0.1 m PBS is very similar, the differences between them being almost negligible ( Figure 3b). Indeed, for both 4l-COP/p and 4l-PNMPy/p, the oxidation of 5-HT is detected at an E p of 0.38 V, while the I p associated with such process is ≈2.3 μA. In comparison to the detection of DA, the sensitivity values for the detection of 5-HT are slightly higher, but similar for both four-layered systems (Table 1). Specifically, the values are 52 μA mm −1 (R 2 = 0.9901) and 56 μA mm −1 (R 2 = 0.9913) for 4l-COP/p and 4l-PNMPy/p, respectively. Indeed, at both electrodes, the slopes of their linear ranges are approximately the same. Besides, both systems achieve comparable LOD (3.2 × 10 −6 and 2.9 × 10 −6 m for 4l-COP/p and 4l-PNMPy/p, respectively) and LOQ values (11 × 10 −6 and 10 × 10 −6 m for 4l-COP/p and 4l-PNMPy/p, respectively), which were calculated considering only the linear range of the calibration curve (from 5 × 10 −6 to 35 × 10 −6 m, Figure 3b and Table 1). Hence, regardless of the composition and interface porosity, both systems display almost identical detection parameters. Therefore, we hypothesize that the binding phenomena of 5-HT on the electrodes plays a major role in the detection response; however, a more comprehensive study on this factor falls beyond scope of this work.
Finally, in order to fully characterize the performance of 4l-COP/p and 4l-PNMPy/p as stress biomarkers detectors, the electrochemical behavior of a mixture containing DA and 5-HT in 0.1 m PBS was studied. Figure 4 displays the recorded voltammograms from −0.2 to 0.8 V at 50 mV s −1 with 30 × 10 −6 m of DA and 30 × 10 −6 m of DA. The oxidation peaks of the two species, which are clearly identified in the cyclic voltammograms, appeared slightly shifted in comparison to those displayed in Figure 3. For instance, for 4l-COP/p, the oxidation peaks of DA and 5-HT were detected at an E p of 0.06 and 0.26 V, respectively, while the I p associated with such processes are 1.04 and 0.95 μA, respectively. On the other hand, for 4l-PNMPy/p, the oxidation peaks of DA and 5-HT were detected at an E p of 0.12 and 0.29 V, respectively, while the I p associated with such processes are 3.3 and 2.1 μA, respectively. Interestingly, with increasing concentration, the E p for the oxidation peaks assigned to DA and 5-HT was shifted to higher values only for 4l-COP/p electrodes, thus ranging from 0.03 to 0.08 V and from 0.22 to 0.27 V, respectively.
Comparing the detection parameters for each of the fourlayered systems (Tables 1 and 2), we observe that their resolution, which is the difference between the peak oxidation potentials of  the two species studied (i.e., DA and 5-HT), was almost exactly the same (0.18 and 0.19 V for 4l-COP/p and 4l-PNMPy/p, respectively), which indicates that no overlapping was produced and both systems distinguished unequivocally the oxidation peak of each of the studied species. As it happened before, 4l-PNMPy/p electrodes were more sensitive for both stress biomarkers (i.e., 127 μA mm −1 (R 2 = 0.9875) and 146 μA mm −1 (R 2 = 0.9771) for DA and 5-HT, respectively); however, in this case, the detection of 5-HT at low concentration values was also precluded and, therefore, the linear range considered for the calibration varied from 20 × 10 −6 to 45 × 10 −6 m, which limits the application of this type of sensor ( Figure 4). In contrast, the calibration curves ob- Table 2. Sensitivity, LOD and LOQ, and resolution obtained for the determination of DA, 5-HT, and DA + 5-HT in different media using 4l-COP/p as electrodes.  6.4 × 10 −6 and 10 × 10 −6 m for DA and 5-HT, respectively, while LOQ values were 21 × 10 −6 and 34 × 10 −6 m for DA and 5-HT, respectively (Table 2). Accordingly, 4l-COP/p were selected as the only working electrodes in the following section.

Theoretical Calculations on the Electrochemical Detection of Stress Biomarkers
Theoretical calculations were carried out at this point to compare the interactions of DA and 5-HT on the PEDOT outer layer of 4l-COP/p and 4l-PNMPy/p as a model of their electrochemical detection. However, before that, the redox behavior of each stress biomarker as isolated compound was studied (Scheme 2). The thermodynamic cycles displayed in Figure 5 have been used to estimate the oxidation potential of DA to DQ and of 5-HT to T45DO referred to the SHE. For this purpose, the conformation of lowest energy found for each structure under vacuum (included in Figure 5) were used as starting points. In order to calculate the Gibbs free energy in the gas-phase, ΔG 0 (gp), frequency calculations were performed for each conformation of lowest energy. Such structures were re-optimized in aqueous solution using the method described in the Experimental Section.
The thermodynamic cycles depicted in Figure 5 indicate that the free energy associated with the reduction of stress biomarkers in aqueous solution at 298 K and 1 atm, ΔG 0 (aq), is where ΔG sol and ΔG sol-Ox are the solvation energies of reduced (DA and 5-HT) and oxidized (DQ and T45DO) species, respectively, and ΔG sol,H+ is the experimentally reported solvation energy of the H 3 O + specie (ΔG sol,H+ = −261 kcal mol −1 ). [38] The value of ΔG 0 (aq) was corrected to change from 1 atm to a 1 m solution, ΔG 0 (aq; 1 m) [39] ΔG 0 (aq; 1M) = ΔG 0 (aq) + Δn ⋅ R ⋅ T ⋅ ln (24.46) The value of ΔG 0 (aq; 1 m) for the lowest energy conformations of DA/DQ and 5-HF/T45DQ redox pairs was −239.4 and −251.1 kcal mol −1 , respectively. This was transformed into the reduction standard potential, E 0 , and referred to the SHE, E 0′ , Figure 5. Thermodynamic cycle used to predict the free energy associated with the reduction of a) DQ to DA and of b) T45DO to 5-HT in aqueous solution at 298 K and 1 atm, ΔG 0 (aq). The minimum energy structure calculated for each species in vacuum is displayed. Calculations in vacuum and aqueous solution (SMD model) were performed at the B3LYP/6-311++G(d,p) level.
using the following equations ΔG 0 (aq; 1M) = −n ⋅ F ⋅ E 0 (4) where n is the number of involved electrons (n = 2), F is the Faraday constant (23.061 kcal (mol V) −1 ), and E SHE is the reduction potential of the SHE in water, which was taken from the literature (E SHE = 4.44 V). [40] The E 0' predicted for DA/DQ from theoretical calculations is 0.751 V, which is in excellent agreement with experimental value (0.75 V). [41] The theoretical prediction for the 5-HT/T45DO pair is 1.00 V. According to these values, the difference between the reduction potentials referred to the SHE, 0.25 V, is around 25-30% higher than the resolution experimentally determined for 4l-PNMPy/p and 4l-COP/p electrodes referred to Ag|AgCl (0.19 and 0.18 V, respectively). Considering that the outer PEDOT layer directly interacts with the stress biomarkers during the detection process for both 4l-PNMPy/p and 4l-COP/p electrodes, such interactions are responsible for the reduction in the resolution of the peaks.
In order to corroborate that the interactions with the surface of the PEDOT outer layer affect the redox behavior of DA/DQ and 5-HT/T45DO pairs during the electrochemical detection, DFT calculations were conducted on DA···CP and 5-HT···CP model complexes, where CP refers to a PEDOT model molecule containing four repeat units. A total of 26 starting geometries were constructed for each model complex, considering not only different relative arrangements but the different types of intermolecular interactions: conventional and nonconventional hydrogen bonds (i.e., N-H···O, O-H···O, C-H···O, and N-H···S), dipole··· -cloud (i.e., N-H··· ), and -stacking interactions. Geometry optimizations, which were performed in vacuum and aqueous solution, led to 10 and 14 minimum energy structures for DA···CP and 5-HT···CP complex, respectively. The lowest energy complex for DA···CP (the rest were higher in energy by at least 2 kcal mol −1 ), which is displayed in Figure 6a, contains the more representative intermolecular interactions found in the whole set of optimized geometries. As it can be seen, the DA neurotransmitter and the PEDOT chain form specific interactions in which the -conjugated system of the CP plays a crucial role (i.e., O-H··· and T-shaped -stacking interactions). It should be noted that because of surface roughness, aromatic rings of PEDOT chains can be exposed at some regions of the film's surface, as it was recently demonstrated by studying a nonidealized surface by combining atomistic molecular simulations and atomic force microscopy. [42] In the case of 5-HT···CP model complexes, the two minima displayed in Figure 6b were the most stable and almost isoenergetic. Interestingly, such complexes, as well as many of the other minima, were stabilized by intermolecular hydrogen bonds (i.e., O-H···O and N-H···O), which evidences divergence with respect to the DA···CP complexes. Such different behavior should be attributed to the flexibility of the 5-HT alkyl chain that, in comparison to DA, exhibits an additional methylene unit and to the steric interactions induced by the two adjacent hydroxyl groups of DA in comparison to the single hydroxyl group of 5-HT. Thus, these structural variations facilitate and preclude the participation of 5-HT and DA, respectively, in the formation of intermolecular hydrogen bonds.
The E 0' predicted from DFT calculations using the model complexes displayed in Figure 6a,b is 0.81 and 0.98 V for DA and 5-HT, respectively, which yields a peak resolution of 0.17 V, that is 0.08 V lower than that obtained for isolated molecules (0.25 V), and consistent with the peak resolution obtained for 4l-PNMPy/p and 4l-COP/p electrodes. In summary, based on theoretical calculations, we ascribe the peak resolution of the examined stress biomarkers to be determined by their interactions with the outer PEDOT layer.

Stress Biomarkers Detection in Synthetic Body Fluids: Artificial Urine and Sweat
As mentioned earlier, DA and 5-HT are found in blood and urine, whereas the presence of DA in sweat is currently being under study. [4] Hence, the suitability of the 4l-COP/p system as detection unit was verified in artificial body fluids. Specifically, its performance was evaluated in artificial urine (for both DA and 5-HT) and artificial sweat (only for DA).
Using 4l-COP/p as electrodes, cyclic voltammograms were recorded from −0.4 to 0.8 V at 50 mV s −1 in artificial urine with increasing concentration of DA (from 10 × 10 −6 to 100 × 10 −6 m, Figure 7). Depending on DA concentration, the oxidation peak of DA was identified at E p between 0.27 and 0.30 V (Figure 7a), thus slightly shifted in comparison to PBS. Again, the variation of I p against the DA concentration yielded a linear calibration curve (Figure 7c), even though two regions were observed in this case. A linear range from 10 × 10 −6 to 50 × 10 −6 m (with a sensitivity of 26 μA mm −1 and R 2 = 0.9855), and a second one from 60 × 10 −6 to 100 × 10 −6 m (with a sensitivity of 6.5 μA mm −1 and R 2 = 0.9376). For the curve at lower concentrations, the LOD and LOQ values were determined to be 5.7 × 10 −6 and 19 × 10 −6 m, respectively. This artificial urine fluid contains ascorbic acid (AA), for which a peak associated with its electro-oxidation process is detected at ≈0.2 V. When DA is present at low concentrations (10 × 10 −6 m, for instance), the oxidation peaks for both AA and DA are observed in the CV. However, at higher concentrations of DA, the only peak detected is that associated with the oxida-tion of DA, which shifts to higher potential values, most probably as a consequence of a secondary oxidation reaction taking place. Peak potential values depend both on the thermodynamics and on the kinetics of the redox reaction, whereas the scan rate affects the kinetics of charge transfer and, consequently, the intensity and potential of the redox wave. [43] Therefore, at low DA concentrations, our system does detect simultaneously the oxidation of DA and AA, thus discriminating the target from interference molecules and rendering the device selective to DA in front of AA, in good agreement with our previous experience. [44] Hence, based on that, we also expect 4l-COP/p electrodes to be able to discriminate DA from uric acid, another main interfering species in urine.
Conversely, 4l-COP/p electrodes better allowed for the selective detection of 5-HT in artificial urine. Cyclic voltammograms, which were obtained from −0.4 to 0.8 V at 50 mV s −1 with increasing concentration of 5-HT (up to 60 × 10 −6 m), displayed an oxidation peak at E p between 0.45 and 0.5 V (Figure 7d). The calibration curve, which resulted in a linear regression only for concentrations in the range from 0 × 10 −6 to 35 × 10 −6 m, displayed a sensitivity of 85 μA mm −1 and R 2 = 0.9867. For higher 5-HT concentrations, the current was saturated at 5 μA. Such response is not foreseen as a problem considering the concentration values at which 5-HT is found in urine (i.e., 0.03 × 10 −6 to 0.13 × 10 −6 m). LOD and LOQ values were determined to be 1.4 × 10 −6 and 14.7 × 10 −6 m, respectively. In terms of selectivity, when interfering AA is present, 4l-COP/p electrodes do detect 5-HT with adequate resolution since the oxidation peaks of both molecules are well-resolved.
Finally, considering that the presence of DA in sweat is currently under study, [4] we evaluated whether 4l-COP/p electrodes could detect this stress biomarker in artificial sweat. Cyclic voltammograms, which were obtained from −0.4 to 0.8 V at 50 mV s −1 , displayed an oxidation peak E p at 0.45 V for a DA concentration value of 5 × 10 −6 m, which shifted slightly to higher voltage for 100 × 10 −6 m (Figure 8). In this region, no oxidation peak was observed for the blank solution (without DA), thus validating the sensor performance. The calibration curve, which resulted in a linear regression for the whole concentration interval studied (from 5 × 10 −6 to 100 × 10 −6 m), displayed a sensitivity of 12 μA mm −1 and R 2 = 0.973. It is worth noting that 4l-COP/p electrodes were also applied to detect 5-HT in artificial sweat. However, in this case, interferences derived from interactions between the fluid components and 5-HT precluded the correct peak identification and, therefore, detection of the neurotransmitter was not feasible. Overall, an electrochemical sensor for detecting dopamine and serotonin within the context of stress monitoring has been designed by exploiting porous CPbased electrodes and their advantages, such as ease of fabrication, economic value, biocompatibility, possibility of blending to reach stretchability, among others. Optimization of the final device would include improvement of detection values by advanced electrode modification.

Conclusions
In this work, two four-layered electrodes with an ultra-porous nanostructured interface have been prepared for the electrochemical detection of two important stress biomarkers, DA and 5-HT. More specifically, such electrodes, which contain both PE-DOT at the inner and outer layers, differ in the central layers that act as a dielectric and were manufactured considering two layers of PNMPy or two layers of COP with the ultra-porous interface in between. The morphology and properties of the layered dielectric were very different, which was attributed to the already proved phase segregation effect in the COP layers. Despite the complexity of the two chosen biomarkers and the difficulties in the electrochemical detection of 5-HT, the 4l-COP/p was very successful for the voltammetric detection of DA and 5-HT, either individually or mixed. In addition, by theoretical calculations, the different interaction patterns between the stress biomarkers and the outer PEDOT layer were elucidated, which explained why the resolution of the identified peaks was slightly lower than expected for both 4l-PNMPy/p and 4l-COP/p electrodes. Thus, DA mainly interacted with the -system of PEDOT, while 5-HT, which exhibits more flexibility and less steric hindrance than DA, formed specific hydrogen bonds with PEDOT chains. Finally, the detection of the chosen stress biomarkers in artificial biofluids, urine and sweat, was also evaluated. Although the field of stress biomarkers detection is still in a development stage, our results demonstrate that 4l-COP/p should be considered as promising electrodes for their electrochemical detection using noninvasive biological fluids.