3.1. Manufacture of Conductive Filaments and CB-PLA Electrodes
The filaments manufactured in this work were obtained using the solvent-based method (SM). This method was selected due to its simplicity and greater application when compared to thermal methods, which generally require more-sophisticated equipment and greater infrastructure [
24,
25,
26,
27].
In the SM, the choice of one or more solvents is crucial for the effective dispersion of the polymer and conductive material. In the literature, several solvents used for this purpose are reported [
39]. In this context, Stefano et al. [
24] used acetone and chloroform for PLA and graphite. The main problems involved in the use of these solvents are related to the high toxicity of chloroform and the difficulty in obtaining both solvents. In view of this, there has been an increase in studies concerning the search for alternative solvents, which have less toxicity and are easily accessible. To select suitable alternatives to chloroform and acetone, the Hildebrand–Burke solubility parameter theory [
39] was applied, using the PLA solubility parameter as a reference. This theory postulates that optimal solubility occurs when the difference between the total Hildebrand solubility parameter (δt) of the solvent (δ
1) and the solute (δ
2) is less than or equal to 1.7 (cal cm;
−3)
1/2, as described by Equation (1):
Using this approach, it is possible to select solvents that exhibit greater affinity for PLA, to maximize solubilization. The Hansen solubility parameter (δt) of PLA is 21.9 [
40]. Based on this, pyridine, acetic acid, and dichloroethane were initially considered as potential solvents, as shown in
Table S1. Individual solubility tests revealed that, despite the similarity in δt, pyridine did not solubilize PLA, probably due to the steric impediment of the benzene ring. Acetic acid showed partial solubilization, while dichloroethane completely solubilized PLA over a 24 h period.
To optimize the incorporation of CB and reduce the solubilization time, a 3:1 mixture of acetic acid and dichloroethane was proposed. The ratio was defined by using the geometric method on the Teas graph (
Figure S1), which positioned the PLA in a line between the two solvents, resulting in a ratio of 3:1. This mixture promoted the complete solubilization of PLA in less than 1 h. The incorporation of CB into the PLA was conducted under reflux at 70 °C in an oil bath, under constant agitation for total homogenization. The resulting solution was poured into a glass dish and dried at 50 °C in an oven for 12 h to allow for the complete evaporation of the solvent and for the formation of agglomerates (
Figure 2). The dried material was cut into small fragments to optimize extrusion and ensure the homogeneity of the filament.
The filament extrusion process was carried out in a Filmaq3D CV Extruder (Filmaq 3D, Curitiba, Brazil) at 170 °C. At this temperature, it was possible to obtain a more homogeneous filament, with an approximate thickness of 1.75 mm, while maintaining the characteristic of polymer flexibility necessary for an efficient printing process. At temperatures above 170 °C, the extruded filament has a more malleable consistency, generating a filament with an irregular thickness. At lower temperatures, it was not possible to extrude the material, due to the clogging of the extrusion nozzle. After obtaining the filaments, a simple qualitative study was carried out to evaluate electrical conductivity. For this, a system was assembled by connecting the filament to a 3 V battery and an LED, and it was possible to observe the emission of light, indicating the passage of electric current (
Figure 3).
3.2. Electrochemical Characterization of Sensor
After fabrication, the sensors were electrochemically characterized by cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) for comparative evaluation before and after surface treatments. The selected techniques investigated changes in parameters such as the resistance to charge transfer, electroactive area, and heterogeneous kinetic constant, evaluating the influence of the treatments on the electrode–electrolyte interface. Four treatment methods were selected based on the literature: electrochemical activation and chemical polishing with an acid, base, and solvent.
Electrochemical activation is normally used to reduce the graphene present in graphene and PLA filaments, generating functional groups on the electrode surface. Santos et al. [
41] demonstrated that this type of treatment also partially removes the polymeric material, exposing the conductive sites and leaving a more homogeneous surface. Solvent treatment has a similar effect to basic treatment and can help remove the remaining non-conductive material that shields the conductive structures. It is worth noting that the distinction between the two treatments is determined by the fact that DMF can remove the polymer through its dissolution [
42,
43]. Finally, acid treatment aims to oxidize and form oxygenated functional groups on the electrode surface, allowing greater interaction with analytes [
42,
44,
45].
To determine the electroactive area and the kinetic constant, the influence of the CV scan rate on the faradaic process, i.e., in the presence of the redox ferricyanide/ferrocyanide pair, was studied. This process consists of applying the CV technique by varying the scan rate. The voltammograms obtained are illustrated in
Figure 4.
Based on the data obtained from the voltammograms, the anodic peak potential (E
pa) and cathodic peak potential (E
pc) were analyzed, along with the peak potential difference (ΔE
P), anodic peak current (I
pa), cathodic peak current (I
pc), and the ratio between peak currents (I
pa/I
pc) for different scan rates (υ) in each sensor. The results of these analyses are presented in
Tables S2–S6 (Supplementary Materials).
The redox ferricyanide/ferrocyanide pair has a reversible character, through which it is possible to observe the presence of symmetrical anodic and cathode peaks and a ratio between the cathodic and anodic peak currents that is close to 1 [
46,
47,
48]. However, in a preliminary analysis of the voltammograms demonstrated in
Figure 4, it can be observed that as υ increases, both the response current and the distance between the peaks also increase. The observed increase in Ep values reduces the time available for the reaction to reach dynamic equilibrium. Thus, at high scan rates, the initial reversible reactions manifest as
quasi-reversible or irreversible. It is noteworthy that at low sweep rates, the system’s dynamic equilibrium is more easily achieved, resulting in reversible or quasi-reversible systems, which indicates the proper performance of the manufactured electrochemical cell.
Skoog et al. propose that one of the conditions for a reaction to be considered reversible is to have a ΔE
p ≅ 59 mV (25 °C) [
47]. The ΔE
p values for the untreated sensor (
Table S2) ranged from 139 to 225 mV, higher than the ideal value. The ratio i
pa/i
pc presented an average of 0.96 ± 0.02. Thus, despite the peaks being symmetrical, the high ΔE
p indicates a
quasi–
reversible system, as demonstrated by Trachioti et al. [
49]. In a complementary way, the value of α was determined using Equation (2), since in cases where α is between 0.3 and 0.7, the system is quasi-reversible [
47].
The quasi-reversibility characteristic of the system is possibly due to mass transport and interactions with the electrode surface. The same behavior was observed for the 3D-S
E sensors (
Table S3), 3D-S
Acid (
Table S4), 3D-S
Basic (
Table S5), and 3D-S
Solvent (
Table S6).
When considering the behavior of the systems as quasi-reversible, it is possible to describe them using the modified Randles–Ševčík equation (Equation (3)), thus making it possible to estimate the electrochemically active area of the sensors, as proposed by Trachioti et al. [
49] and Washe et al. [
50].
where ip is the peak current (A), A is the electrochemically active area (cm2), n is the number of electron moles involved in the redox reaction, D is the diffusion coefficient of the reduced species (cm2/s), CR∞ is the concentration of the reduced species within the solution (mol/cm3), υ is the scan rate (V/s), and K(Λ, α) is the modified dimensionless parameter.
The electrochemically active area of the sensors (
Table 1) shows that the 3D sensor has an electrochemically active area approximately 12.3 times smaller than the geometric area of the sensor (12.56 mm
2). The small electrochemically active area is possibly due to the high presence of PLA in the conductive filament since PLA is an insulating material that suppresses the number of active sites originating from the CB.
For the sensor treated with HNO3, it was possible to observe an increase in the electroactive area of 6.20 mm2, in comparison to the 3D sensor. This increase is likely due to the formation of oxygenated functional groups on the electrode surface that have contributed to the increase in available active sites. For the NaOH-treated sensor, an increase in the electroactive area of 9.01 mm2, in comparison to the 3D sensor, was observed, which may be related to the fact that the NaOH solution hydrolyzes the PLA present in the 3D sensor, exposing the conductive CB fibers. DMF treatment also exposes nanostructured conductive fibers, but less efficiently when compared to acid and basic treatment.
After estimating the areas, the velocity constant of an electrochemical reaction (k°) was investigated, which upholds the velocity law for chemical reactions. In electrochemistry, k° involves the transfer of electrons through an electrode/solution interface. Each electron transfer process at the electrode/solution interface occurs in a specific way, and depending on this form, there is more-appropriate mathematical modeling to describe the system and find the value of k°. The type of electrochemical system is determined with the help of factors such as its degree of reversibility, the type of redox reactions, and the investigation technique.
As the electrochemical systems formed for the 3D, 3D-S
E, 3D-S
Acid, 3D-S
Basic, and 3D-S
Solvent sensors have
quasi-reversible characteristics and the analysis technique used was CV, Nicholson’s interpretation of mathematics, Equation (4), is the one that best suits the systems.
where Ψ is the dimensionless parameter that expresses the degree of reversibility, k° is the heterogeneous kinetic constant (cm/s), R is the gas constant (8.314 J/mol*K), T is the temperature (298 K), D is the diffusion coefficient (cm2/s), F represents Faraday’s constant (96.485 C/mol), and υ is the scan rate (V/s).
The value of Ψ is related to ΔE
p (
Table S6), and
Figure S2 illustrates a graphical representation of this relationship between Ψ and ΔE
p. Ψ values were determined using data interpolation based on the values tabulated by Bard and Faulkner [
46]. With the values of Ψ, the graphic correlation of these values was performed as a function of the inverse square root of the scan rate and the linear regression of the values
(Figures S3–S7). Using the linear regression of these data, the slope of the regression line was determined. This slope reflects how systems respond to changes in scan rates. An increase in slope indicates a greater influence of the scan rate on the value of Ψ, which can be interpreted as greater non-reversibility. By replacing the square root of the inverse of the sweep velocity (√1/υ) in Equation (4) with the slope of the line, the calculation for the k° of electrochemical systems can be developed, obtaining the k° values shown in
Table 1.
The k° value for the 3D sensor was shown to be large enough when compared to the values reported in the literature, so that the system is prone to behave reversibly, particularly at a lower scan rate, as observed in
Figure S8 [
51]. For the 3D-S
E, 3D-S
Acid, 3D-S
Basic, and 3D-S
Solvent sensors, despite showing a relationship between peak currents close to the ideal for the system that is reversible and α values close to that of the 3D sensor, the k° values were small when compared to the literature, indicating that even at low speeds the systems for these sensors do not tend to show reversibility.
To investigate the interfacial physicochemical processes, we used EIS, which provides information on the resistivity to charge transfer and the presence of secondary reactions. The analysis of the Nyquist diagrams and the proposition of an equivalent circuit (
Figure 5) allowed us to determine the resistance of the systems. In the Nyquist diagrams, the x-axis represents the real (resistive) part, and the y-axis represents the imaginary part (capacitive) [
46,
52]. The diameter of the semicircle corresponds to the load transfer resistance.
Two equivalent circuits were proposed: the Randles circuit in
Figure 5F was used for the 3D sensors, 3D-S
E, 3D-S
Acid, and 3D-S
Solvent, while the equivalent circuit in
Figure 5G was used for 3D-S
Basic. Using the circuits, the polarization resistance (Rp), uncompensated resistance (RΩ), and non-ideal capacitance (Q) were determined, as represented in
Table 1. Rp indicates the sensitivity of the sensor (lower values correspond to higher sensitivity), Q describes the capacitance of the electric double layer (higher values indicate higher charge accumulation), and RΩ represents the charge transfer resistance of the solution [
53,
54].
For the 3D-SE sensor, the values of polarization resistance, ohmic resistance, and phase constant admittance were higher than those of the untreated sensor. This result suggests that the electrochemical treatment alone did not improve the electrochemical performance of the sensor. The values found for the 3D-SAcid sensor were higher than those for the untreated sensor. 3D-SBasic, the equivalent circuit, presented lower values than the untreated sensor, with the appearance of a capacitive element (C) of 5.08 mF. The increase in Q and the presence of C indicate an accumulation of charges on the surface, which is consistent with the extended voltammograms, and which can be mitigated by increasing the ionic strength of the medium. Finally, for the 3D-SSolvent sensor, the equivalent circuit presented a higher RΩ when compared to the 3D-SBasic sensor, but a lower value than that of the 3D sensor, demonstrating the potential of 3D-SSolvent.
In general, we can observe that this study investigated different surface treatments on 3D-printed sensors to optimize their electrochemical performance. Electrochemical activation proved ineffective, while acid polishing produced properties inferior to those of the original sensor. Polishing with DMF showed similar performance to the untreated sensor. In contrast, basic polishing with NaOH (3D-SBasic) stood out as the most promising treatment, promoting a substantial increase in the electroactive area, a significant reduction in the resistance to electron transfer, and an increase in the reaction speed constant (k°). Although 3D-SBasic shows more capacitive behavior, the benefits regarding its electroactive area and load transfer resistance consolidate it as the best option. Therefore, NaOH polishing is the most effective treatment to improve the electrochemical performance of 3D-printed sensors. Future studies should be conducted to optimize the treatment parameters to minimize capacitive effects.
3.3. Evaluation of the Potential of 3D-SBasic
To evaluate the potential of the 3D-
SBasic sensor in developing electrochemical sensors, the sensor was applied to the detection of acetaminophen (APAP). The analyses were performed using cyclic voltammetry (CV) in a 0.1 mol L
−1, pH 7.00 phosphate buffer solution, in the presence and absence of APAP (10 μmol L
−1). A sensor produced with commercial CB-PLA filament (Protopasta), named 3D-
SBasicProtopasta, was also evaluated for the determination of APAP. The results obtained are illustrated in
Figure 6A. It is possible to observe through the voltammograms (
Figure 6A) that the filament developed in this work presents characteristic peaks for APAP, while for the sensor produced with commercial filament, it is not possible to observe the peaks. Therefore, this result suggests the potential application of the proposed sensor.
Sensor stability studies were also conducted. For this purpose, the sensor was stored for 7 days. The results are illustrated in
Figure 6B. Through this study, it was observed that there was no significant signal loss in this time interval, indicating that the sensor was stable during this period. Additional studies should be performed to evaluate the long-term stability of the sensor.
A comparative study was carried out for the determination of APAP using conventional electrodes (glass carbon, carbon paste, and platinum disc) and 3D-S
Basic. The electrochemical performance of the sensors was evaluated according to their determination of APAP in 0.10 mol L
−1 phosphate buffer, pH 7.0, by cyclic voltammetry. For this, successive additions of different concentrations of APAP (5.0–50.0 μmol L
−1) were performed. In
Table 2, the sensitivity values for each electrode are shown. From the data obtained, it is possible to observe that the 3D-S
Basic sensor showed better electrochemical performance compared to the conventional ones. This may be related to the larger area of the 3D-S
Basic sensor due to the NaOH treatment.
The electrochemical performance of the proposed sensor was evaluated according to its ability to determine APAP in 0.10 mol L
−1 phosphate buffer, pH 7.0, for different APAP concentrations (5.0–50.0 μmol L
−1) using the differential pulse voltammetry (DPV) technique. The results are illustrated in
Figure 7A. Through the voltammograms, it is possible to observe a gradual increase in the peak current, indicating that the sensor can detect different concentrations of APAP. The anodic peak current was correlated as a function of the APAP concentration to obtain an analytical curve (
Figure 7B).
The limits of detection (LOD) and quantification (LOQ) were three and ten times the standard deviation of a blank solution divided by the sensitivity, respectively. The values found were 0.198 μM and 0.659 μM. The sensitivity and LOD found in this study were compared with some sensors reported in the literature for APAP detection (
Table 3).
Although the proposed sensor does not have a lower LOD than those found by some studies in the literature, the performance of the 3D-S
Basic is promising, since the sensor developed in this work is a disposable sensor, which uses alternative materials in its composition, in addition to not involving superficial modification. Furthermore, it is worth noting that 3D-S
Basic has shown promise, indicating that it can be applied in the concentration range of possible APAP poisoning, since 24 h after the ingestion of the drug, the toxic concentration is above 40 μM [
66].
Therefore, the results obtained in this work show promise, but further studies are needed to improve the sensor’s performance. The filament production conditions can be optimized by evaluating the ratio between the conductive material and PLA. Although the basic treatment with NaOH significantly improved the electroactive area and resistance to electron transfer, evaluating the treatment time and concentration may further enhance the electrochemical response. To study modifications on the sensor’s surface due to the treatment, characterizations using such methods as scanning electron microscopy, atomic force microscopy, and Raman will be evaluated. Finally, it is expected that this work can contribute to advancements in the development of 3D-printed sensors obtained from laboratory-fabricated filaments, with their low cost and alternative materials to commercial ones.
