1 Introduction

The search for clean and renewable energy sources capable of replacing the current energy matrix (based on fossil fuels) has driven the research and development of technologies aimed at the efficient hydrogen production, a highly versatile and environmentally friendly fuel [1,2,3,4,5]. Water electrolysis has emerged as a promising route for hydrogen production because it utilizes electricity generated from renewable sources, such as solar and wind energy, to split water into oxygen and hydrogen [6,7,8,9]. However, the efficiency of this process depends largely on the quality of the electrodes used.

Recent advances in 3D printing technology have provided new opportunities for the design and fabrication of custom electrodes, enabling the precise optimization of their geometry and surface properties [10, 11]. This opens the door to enhancing the efficiency and cost-effectiveness of water electrolysis [11, 12]. In recent years, there has been a significant increase in the use of Fused Deposition Modeling (FDM) to produce 3D-printed electrodes for hydrogen production. However, obtaining high-quality filaments with good conductive properties is a central challenge in the production of 3D-printed electrodes for hydrogen generation through water electrolysis [13].

Therefore, the use of commercial conductive filaments in the manufacturing of 3D-printed electrodes has become popular [14,15,16,17,18,19,20,21,22,23]. Although their use is common and offers advantages in terms of convenience and reduction of experimental steps, these filaments have some limitations. Studies have shown that conductive filaments often do not exhibit good conductivity and require activation to expose their electrochemically active sites [15, 18, 24,25,26,27,28,29]. Moreover, de Farias et al. [30] confirmed in their work that commercial filaments are highly versatile and indispensable to the scientific landscape. However, the authors mentioned that despite these materials being used for various applications, they were not exclusively produced for this purpose. As a result, sophisticated and time-consuming surface-activation procedures are required. Another crucial point is that manufacturers only partially disclose the composition of conductive filaments, significantly impacting research development. A study conducted by Pumera et al. [31] identified undisclosed metallic impurities (Fe and Ti) that directly affect electrocatalytic performance. This underscores the need for caution when using these filaments and the importance of transparency in the composition of commercial conductive filaments. A fundamental aspect that can be highlighted regarding commercial filaments (Black Magic® and Proto-Pasta®), investigated by Kalinke et al. [32], is their tendency to have their properties and electrochemical performance affected over time. This study found that the aging phenomenon compromises the conductivity of the filaments and, consequently, of the 3D-printed electrodes from these materials. In this context, the ability to produce filaments offers a substantial advantage. In light of this, by mastering the production of filaments, users can ensure a material with good properties and performance.

Recent studies have focused on the production of conductive filaments for the 3D printing of electrodes [33,34,35,36,37,38,39]. For example, Banks et al. [40] proposed a method for producing conductive filaments. Consequently, the electrodes printed using these filaments exhibited good efficiency for the water electrolysis reaction without the need for prior surface treatment. However, despite its good efficiency, the filament was prepared using Pt, which increased the production cost of these electrodes. Similarly, Pumera et al. [35] developed a filament for the 3D printing of electrodes composed of 2D MoS2, Carbon, and PLA (polylactic acid). This material was used for photo and electrochemical energy conversion and storage. This study highlights the innovative application of this filament in creating high-performance energy-storage devices and its feasibility for 3D printing. However, despite its high potential, researchers have reported difficulty in achieving a composition where the material properties are suitable for the 3D printing of electrodes.

Considering this, we noticed that preparing conductive filaments for 3D printing electrodes with good electrochemical performance without requiring prior surface activation, for example, compared to commercial filaments and their respective electrodes, is a challenge. Therefore, this study aims to present the preparation of contaminant-free conductive filaments for the 3D printing of electrodes and investigate their performance in comparison to a commercial filament and its corresponding 3D-printed electrode for hydrogen production. The 3D-printed electrodes underwent a Cobalt Phosphate (CoPi) electrodeposition procedure, and their potential for hydrogen gas production in the absence and presence of a catalyst on their surface was investigated.

2 Materials and methods

2.1 Materials

All chemicals were analytical grade and were used without any further purification. PLA pellets were acquired from 3DLAB (Minas Gerais, Brazil). Polyethylene glycol (4000) (PEG), from Synth (São Paulo, Brazil). Carbon black (CB) powders (VULCAN® XCMAX™ 22—CBMAX22, and VULCAN® XC72R—CB72R) were obtained from CABOT™ (Massachusetts, EUA). Protopasta®, obtained from Proto Pasta (Vancouver, USA). Cobalt (II) nitrate hexahydrate (Co(NO3)2.H2O, 98%), Sodium hypophosphite monohydrate (NaH2PO2.H2O, ≥ 99%), acetone (P. A.), chloroform (99.8%), ethanol, Hydrochloric acid (HCl, 36.46%), Sulfuric acid (H2SO4, 95.0–98.0%), Sodium hydroxide (NaOH, 99.0%) were obtained from Merck (Brasil). Potassium nitrate (KNO3, ≥ 99%), hexaammineruthenium (III) chloride ([Ru(NH3)6]Cl3, 98%) were purchased from Sigma-Aldrich (USA).

2.2 Manufacture of composite filament and CB-PLA electrodes

The method adapted by Janegitz et al. [41] was used for the fabrication of conductive filaments. Thus, 7.5 g CB powder was dispersed in 200 mL of a mixture of acetone and chloroform (3:1 v/v) for solubilization of the thermoplastic material, under magnetic stirring and heating to temperatures up to 70 °C during 30 min, in a reflux system. Next, 13.5 g of PLA and 9.0 g of PEG (totalizing 30 g of PLA, PEG, and CB) were added to the mixture and kept under constant stirring and heating for 3 h. After that, the obtained homogeneous mixture was precipitated by transferring all the content into a recipient containing 800 mL ethanol, and a uniform PLA-PEG@CB composite was obtained. The composite was filtered, washed with ethanol, and left to dry at 50 °C for approximately 12 h (Fig. S1a). After the material was completely dry, it was cut into small pieces, which were placed in an extruder, Filmaq3D®, operating at 200 ºC and with the maximum extrusion speed achieved using the extrusion equipment mentioned. The prepared filaments were identified according to the starting CB, PLA@CBMAX22 and PLA@CB72R, respectively (Fig. S1b). The filaments obtained was used in characterization analysis and for the 3D printing of electrodes.

The 3D electrode model to be printed was created using TinkerCad® design software, with the dimensions described in Fig. S2a. The 3D electrodes were printed using Repetier-Host software on the FDM-type 3D printer, which was assembled by our research group in the laboratory [42], as shown in Fig. S2b. The main parameters for 3D printing and their respective values were as follows: nozzle temperature (200 °C), heated bed temperature (85 °C), layer height (0.4 mm), nozzle diameter (1.0 mm), and fill density (100%).

This same process was carried out to obtain filaments using the two types of CB mentioned above as additive material, which after preparation, were identified as PLA@CBMAX22 and PLA@CB72R. In the present study, we chose two different carbon blacks to evaluate the performance of the polymeric matrix. According to the manufacturer [43], CBMAX22 carbon black has a better conductivity performance than CB72R, with smaller weights (Fig. S3). To assess the performance of the lab-made filaments and their corresponding 3D-printed electrodes compared with the commercial filament and its 3D-printed electrode, these materials underwent the same characterization and application analyses.

2.3 Scanning electron microscopy (SEM) and energy dispersive spectrometer (EDS)

The morphologies of the samples were examined with a Quanta FEG 250 field emission scanning electron microscope (FEI Co., USA) equipped with an Oxford X-MAX50 energy dispersive spectrometer (EDS) (Oxford, UK). To analyze the surfaces of the materials, filaments and 3D-printed electrodes samples were mounted on a conductive carbon tape followed by iridium sputtered coating in a Bal-Tec MD020 instrument (Balzers). To analyze the internal parts of the materials, filaments and 3D-printed electrodes samples were freeze fractured in liquid nitrogen and mounted on a conductive carbon tape followed by iridium sputtered coating in a Bal-Tec MD020 instrument (Balzers). Note that SEM characterization aims to obtain information about the morphology of both the surface and the internal part of the prepared materials (filaments and 3D-printed electrodes). For the images generated by EDS, we expected to identify the chemical elements related to the catalyst that was electrodeposited on the surface of the 3D-printed electrodes.

2.4 Thermal characterization

Thermogravimetric analyses were carried out using a Thermogravimetric SDT Q600 V8.3 analyzer (TA Instruments, USA) between 27 and 1000 °C, at 10 °C min−1, under inert atmosphere (N2 atmosphere). Differential scanning calorimetry studies of the PLA, conductive filaments and 3D-printed electrodes were carried out using a DSC Q10 V9.9 (TA Instruments, USA) at a heating and cooling rate of 10 ºC min−1 in an inert atmosphere (N2 atmosphere). Standard aluminum pans were used to analyze the samples. The experiment was conducted from 25 ºC to 230 ºC (first heating run), then followed by cooling (second run) from 230 ºC to 25 ºC and third run from 25 ºC to 230 ºC, for all the samples. Glass transition temperatures (Tg) and melting behavior were observed from the second heating run of the DSC plot. The data from the second heating cycle were used to eliminate the thermal history. Where, χc is the degree of crystallinity 29, ∆Hm is the filament melting enthalpy (endothermic), ∆Hc is the crystallization enthalpy (exothermic) and ∆Hºm is the melting enthalpy for pure crystalline PLA (93.7 J g−1) according to the Eq. (1) follow:

$${\chi }_{c}=\frac{\Delta {H}_{m}-\Delta {H}_{c}}{\Delta {H}_{m}^{0}}*100$$
(1)

2.5 Electrochemical characterization

The electrochemical cell was composed of the 3D-printed working electrodes, a saturated calomel reference electrode, and a platinum auxiliary electrode. The filaments and electrodes were electrochemically characterized by cyclic voltammetry (CV) in the presence of 1.0 mmol L−1 [Ru(NH3)6]3+ as the electrochemical probe. 0.10 mol L−1 KNO3 was used as a supporting electrolyte. CV measurements were performed from − 0.60 to + 0.30 V, at a scan rate of 25 mV s−1. Measurements were carried out using an Autolab PGSTAT 101 potentiostat/galvanostat (Metrohm Autolab, Utrecht, Netherlands), controlled by NOVA 2.1.4 software. The heterogeneous electron transfer rate constant (k°obs) was determined using the Kochi method [44] and calculated using the Eq. (2) bellow, where D is the diffusion coefficient (9.10 × 10−6 cm2 s−1 for [Ru(NH3)6]3+), α is assumed to correspond to 0.5, n is the number of electrons transferred in the electrochemical process, F is the Faraday constant, ν is the scan rate, R is the gas constant, T is the temperature and ΔEp is the peak-to-peak separation.

$${k}_{obs}^{0}=2.18 {\left(\frac{D\alpha nFv}{RT}\right)}^\frac{1}{2}exp\left[-\left(\frac{{\alpha }^{2}nF}{RT}\right)\Delta {E}_{p}\right]$$
(2)

The electrochemical active surface area (ECSA) were evaluated by CV measurements performed at different scan rates, from 5 to 50 mV s−1 (5, 10, 20, 30, 40, and 50 mV s−1), according to the Randles–Ševčíck Eq. (3) [45], where Ip is the peak current, A is the electrode area, C is the concentration of the redox probe, D is the diffusion coefficient and n is the number of electrons transferred.

$${I}_{p}=2.69\times {10}^{-5}AD{C}^\frac{1}{2}{n}^\frac{3}{2}{\nu }^\frac{1}{2}$$
(3)

The electrodes were characterized by electrochemical impedance spectroscopy (EIS) in presence of 1.0 mmol L−1 [Ru(NH3)6]3+ in 0.10 mol L−1 KNO3 as the electrochemical probe. EIS measurements were recorded from 500 kHz a 0.05 Hz, and an amplitude of 0.01 V. The equivalent circuit was used to adjust the experimental data and determine the charge transfer resistance (Rct) related to the redox reaction of the probe.

2.6 Stability of filaments and electrodes at different pH’s

A stability test was carried out with the filaments and electrodes at three different pH values (approximately 1, 7 and 14) to verify the resistance of the materials in each medium (acidic, close to neutrality, and alkaline). Solutions of 1 mol L−1 NaOH (pH 14) and 0.1 mol L−1 HCl (pH 1) were prepared. After preparing the solutions, approximately 3.5 cm of each filament and electrode were placed (separated) in vials with lids. Each flask was then filled with approximately 5 mL of each previously prepared solution at pH 1 and 14. For pH 7 (or close to neutral), deionized water was used, and approximately 5 mL was added to the flasks. The vials were photographed at time intervals (0, 0.5, 1, 2, 3, 4, 5, 6, 9, 12 and 24 h), and the monitoring was recorded at these intervals. At each cycle time, the flask containing the filament and electrodes was stirred. Fig. S4 presents a summary of the experimental procedure used to evaluate the stability of filaments and 3D-printed electrodes as a function of pH.

2.7 Preparation of cobalt phosphate (CoPi)

Co(NO3)2.6H2O (1 mmol) and NaH2PO2.H2O (1 mmol) was dissolved in 100 mL of a mixture of ethanol and deionized water (v/v = 1:1) in a 100 mL beaker. A 3D-printed electrode was used as the working electrode, Ag/AgCl as the reference electrode, and a Pt counter electrode. Cyclic Voltammetry was performed at a scan rate of 5 mV s−1 for 10 cycles within a voltage range of − 1.3 to 0.3 V. The 3D-printed electrodes were washed with water and dried for approximately 12 h, Fig. S5.

2.8 Hydrogen production

For the hydrogen production using 3D-printed electrodes, 400 mL of solution of H2SO4 0.5 mol L−1 was prepared. We assembled a hydrogen production system using 3D-printed electrodes. Fig. S6 in the supporting information provides more information about how to assemble this system. The electrodes were then immersed in this solution, which was previously connected to a source adjusted to 12 V for one hour. The Protopasta® electrode, as a standard, was connected to the positive pole, whereas the negative pole was connected to the electrodes of interest for the production of hydrogen gas. The gases produced during the process were collected and measured using a syringe. Three hydrogen production measurements were performed for each electrode.

3 Results and discussion

3.1 Morphological characterization of the materials

The surface of the electrode is the key factor in the electrochemical reactions between chemical species and substrates [26]. Because of this, the morphological analysis of the filaments and 3D-printed electrodes were performed by scanning electron microscopy (SEM). The obtained images (Fig. 1a–c) revealed that all the filament surfaces are smooth. However, the filament prepared with CBMAX22 shows some bubbles on the surface that are probably caused by residual water. The porosity and mixture between the polymer and carbon black of the surface of the filaments become evident with the magnification of the images, as can be seen Fig. 1d–f.

Fig. 1
figure 1

SEM images were obtained for conductive filaments and 3D-printed electrodes. a–c surface of Protopasta® filament, PLA@CBMAX22 filament and PLA@CB72R filament, respectively; d–f fracture of Protopasta® filament, PLA@CBMAX22 filament and PLA@CB72R filament, respectively; g–i surface of Protopasta® 3D-printed electrode, PLA@CBMAX22 3D-printed electrode and PLA@CB72R 3D-printed electrode, respectively; fracture of j, m Protopasta® 3D-printed electrode, k, n PLA@CBMAX22 3D-printed electrode and l, o PLA@CB72R 3D-printed electrode

Full size image

After the printing processes is possible to observe some irregularities on the surface of the 3D-printed electrodes (Fig. 1g–i). However, the Protopasta® filament (Fig. 1a) and the Protopasta® 3D-printed electrode (Fig. 1g) exhibited smoother surfaces compared to the other samples analyzed. In addition, all filaments and 3D-printed electrodes were fractured to analyze the internal morphological structure of each material (Fig. 1j–l).

From the fractures performed, it is observed that PLA, a thermoplastic polymer shown in the composition of the samples, is present in smooth phases both in the filament and in the 3D-printed electrode. On the other hand, the filament and 3D-printed electrode prepared with CBMAX22 (Fig. 1b and h) shows the formation of morphological irregularities both on the surface of the filament and on the electrode. These irregularities resemble bubbles, which are attributed to the presence of water during the filament preparation or electrode printing, as mentioned before. This is evident when we analyze Fig. 1k, which is related to the fracture of the PLA@CBMAX22 3D-printed electrode after the 3D electrode printing holes were formed, possibly caused by the evaporation of water during the printing process.

It is worth mentioning that similar to the Protopasta® filament and Protopasta® 3D-printed electrode, the PLA@CBMAX22 filament and PLA@MAX22 3D-printed electrode (Fig. 1e and n) present smooth phases inside the material, albeit on a smaller scale. Similar to the other samples, the PLA@CB72R filament and PLA@CB72R 3D-printed electrode shown a smooth surface (Fig. 1c), while the electrode (Fig. 1i) has some irregularities that suggest the presence of water in the analyzed material, but not so intense. As in the others, through the fracture (Fig. 1f and o), it was possible to observe a smooth phase of PLA, thus revealing its distribution within the material. In Fig. 1j–l, we can also observe that the printed layers of the electrode have adhered well to each other. Another point that should be mentioned is related to the smooth phase of PLA observed both in the filaments (Fig. 1d–f) and in the electrodes (Fig. 1m–o). These figures show that, compared to Protopasta® filament and Protopasta® 3D-printed electrode, the other materials do not have a good distribution of PLA (noted by the smooth phase present in the materials). This lack of uniformity may be a justification for the preparation of filaments and electrodes with low mechanical resistance.

3.2 Thermal characterization of the materials

Thermal analysis experiments were performed on PLA, conductive filaments, and 3D-printed electrodes to investigate the differences among various conductive materials. The TGA curves obtained for the filaments and 3D-printed electrodes are illustrated in Fig. 2a and b and organized in Table S1. All samples exhibited the first mass loss process attributed to the decomposition of the PLA polymer [46], occurring between 302 and 327 °C, except for the PLA@CBMAX22 filament and PLA@CBMAX22 electrode, which exhibited 290 and 252 °C, respectively. The final thermal process (587–651 °C) can be attributed to the combustion of carbon, specifically carbon black (CB). Similarly, it was observed that the filament and 3D electrode lab-made from CBMAX22 differed from the other materials by exhibiting a lower temperature range for CB decomposition (502 and 520 °C, respectively). As previously stated, lower thermal process temperatures were recorded for the PLA@CBMAX22 filament and electrode, indicating that the thermal stability of these analyzed materials is lower than that of the others studied [32].

Fig. 2
figure 2

Thermogravimetric Analysis (TGA) obtained for PLA, filaments (a) and 3D-printed electrodes (b) from 25 to 1.000 ºC

Full size image

The DSC curves depicted two endothermic peaks related to the glass transition temperature (Tg) and PLA melting temperature (Tm), and an exothermic peak associated with the crystallization temperature (Tc), as presented in Fig. S7. Table 1 displays the DSC processes, enthalpy energies, and degrees of crystallinity, Eq. (1). Tg is related to the amorphous phase of polymers, which can be observed in amorphous and semicrystalline polymers such as PLA [47]. It is noteworthy that, according to Table 1, for the filaments and their corresponding electrodes that were prepared, Tg was lower than that of pure PLA and the commercial filament. Tc of a polymer refers to the temperature at which the polymer solidifies and organizes itself into a structured form. PLA is a semicrystalline polymer, which means that it has a slow crystallization rate. As shown in Table 1, the prepared filaments exhibited lower Tc values. Concerning the degree of crystallinity (χc), Protopasta® filament showed the highest degree despite its Tc being intermediate, and its enthalpy of crystallization was the lowest. The filaments prepared in the laboratory showed a higher degree of crystallinity, which is in agreement with the values of ΔHc and ΔHm, since for these the enthalpy variation was greater, taking the 3D-printed electrode from the commercial filament for comparative purposes.

Table 1 DSC analysis obtained for PLA, conductive filaments and 3D-printed electrodes
Full size table

3.3 Electrochemical properties of the filaments and 3D-printed electrodes

The first part of this work involved a study of the effect of the scan rate of the filaments and 3D-printed electrodes. The voltammograms obtained in this study are shown in Fig. 3. As can be observed, all cyclic voltammograms show the profiles of a system that presents good reversibility with defined anodic and cathodic peaks, both oxidation and reduction [45]. The only exception was the Protopasta® filament, which did not show defined peaks as a function of the scan rate.

Fig. 3
figure 3

Cyclic Voltamogram of a Protopasta® filament, b PLA@CBMAX22 filament, c PLA@CB72R filament, d Protopasta® 3D-printed electrode, e PLA@CBMAX22 3D-printed electrode, f PLA@CB72R 3D-printed electrode. All experiments used 1.0 mmol L−1 [Ru(NH3)6]3+ as the electrochemical probe. Supporting electrolyte: 0.10 mol L−1 KNO3

Full size image

Based on the registered data from the influence of the scan rate, it was possible to obtain parameters for the electrochemical processes for each material tested. The data are presented in Tables S2, S3, S4, S5, S6, and S7 of the Supporting Information. From these parameters, it was also possible to verify the reversibility of the redox process studied: the difference between the anodic and cathodic peak potential values (∆Ep) and the ratio between the anodic and cathodic peak current (Iap/Icp). The reversibility criteria were calculated for the differences in the potential measured as a function of the scan rate.

Figure 4 shows the variation in ∆Ep as a function of the potential scan rate for the electrochemical cell using the studied materials. Figure 4a and b indicate that, in general, both the Protopasta® filament and its electrode provided higher ∆Ep values, especially at scan rates greater than 50 mV s−1. Interestingly, for the PLA@CBMAX22 and PLA@CB72R filaments (Fig. 4a) there is not a large variation of ∆Ep as there is an increase in scan rate, supporting the idea that these materials behave as more reversible systems, because according to the literature, for a fully reversible process, that is, the one for which the electron transfer rate is high, the anodic and cathodic peak potentials are independent of the potential scan rate [48]. When there is some kinetic limitation to electron transfer, it is manifested by the increase in ΔEp values with increasing potential scan rate, indicating that the studied process approaches an irreversible process [49, 50]. Therefore, we can infer that the high ΔEp values for the Protopasta® materials reveal that their surfaces are less active than the other materials, thus the lower activity of the electrode surface manifests itself as a kinetic limitation to electron transfer that, in turn, causes an increase in the ΔEp values. This result was confirmed by the heterogeneous electron transfer constant (kºobs) of filaments and 3D-printed electrodes (Table S8, in the Supporting Information), where kºobs values of 6.77 × 10–5, 5.71 × 10–4, 7.13 × 10–4 were determined for filaments Protopasta®, PLA@CBMAX22, and PLA@CB72R, respectively in relation to the electrodes, the PLA@CB72R 3D-printed electrode presented the highest kºobs determined, 1.77 × 10–3. According to the literature, an increase in kºobs results in a better contribution to the redox responses [26, 32].

Fig. 4
figure 4

Variation of ΔEp as a function of potential scan rate for a filaments and b 3D-printed electrodes, and variation of Iap/Icp ratio as a function of potential scan rate for c filaments and d 3D-printed electrodes

Full size image

The variation in the Iap/Icp ratio with increasing potential scan rate for filaments and 3D-printed electrodes is shown in Fig. 4c and d, respectively. The criterion for the reversibility of an electrochemical process is the ratio between the anodic and cathodic peak currents, which must be close to 1. The results presented in Fig. 4c and d demonstrate that for all scanning speeds evaluated, the PLA@CB72R filament and 3D-printed electrode provided Iap/Icp ratios closer to 1 than those of the other prepared materials. Thus, we concluded that these materials exhibit better reversibility.

As shown in Table S8, the filaments prepared in the laboratory presented larger electroactive areas than the commercial filament. It is important to highlight that the PLA@CB72R filament and its 3D-printed electrode were those with the highest electroactive area (ECSA), 0.8613 and 0.8639 cm2, respectively, among the materials studied (in the supporting information, Section S10, a comment is presented on the method used to determine the ECSA in this work). These results corroborate the results presented so far, which indicate the better electrochemical performance of the materials prepared in the laboratory compared to the commercial one. Notably, the electroactive area of each electrode was much greater than that of the filament used for printing. This difference may be attributed to the 3D printing process, specifically the FDM method, as shown in Table S8. Despite the similarity in geometric areas between the filaments and electrodes, the FDM printing method [51] results in a somewhat uneven surface on the electrodes due to the layer-by-layer printing process (as illustrated in Fig. S8, which pertains to the filaments and 3D-printed electrodes, and Fig. S9, which relates to the 3D-printed electrode design and its dimensions, as well as in Fig. 1a–c and g–i), as opposed to the smoother surface of the filaments. This unevenness likely increased the contact surface area of the materials, leading to an observed increase in the ECSA of the electrodes. Therefore, we propose that the FDM method for 3D printing may be the cause of this increase.

All the prepared filaments and 3D-printed electrodes were subjected to Electrochemical Impedance Spectroscopy (EIS) studies. This technique is essential for determining the significant parameters and measuring the charge-transfer resistance (Rct) to establish activity trends. In EIS, the semicircular portion at higher frequencies and lower frequencies represent Rct. Figure 5 shows the resulting Nyquist plots for all materials. Nyquist plots (Fig. 5) were analyzed using the equivalent Randles circuit to determine the electrical component of the Rct associated with the [Ru(NH3)6]3+ redox reaction. A decrease in the size of the semicircle was observed for the electrodes printed from the filaments. According to the literature [52], the semicircle diameter, which corresponds to Rct, represents the difficulty of electron transfer between the solution and electrode for a given redox probe ([Ru(NH3)6]3+). Thus, a larger semicircle indicates great difficulty in charge transfer. As shown in Fig. 5, all filaments had larger semicircle diameters than the 3D-printed electrodes, indicating higher Rct values (Table S9). The Rct values decreased for all electrodes (Table S9), suggesting that the printing process exposed more carbon black and enabled faster electron transfer in all 3D-printed electrodes.

Fig. 5
figure 5

Nyquist plot for all materials obtained. a Protopasta® filament (χ2 = 0.0121), b PLA@CBMAX22 filament (χ2 = 0.0162), c PLA@CB72R filament (χ2 = 0.0061), d Protopasta® 3D-printed electrode (χ2 = 0.0066), e PLA@CBMAX22 3D-printed electrode (χ2 = 0.0043) and f PLA@CB72R 3D-printed electrode (χ2 = 0.0051). Insert: proposed electrochemical circuit

Full size image

An equivalent circuit was proposed for each material, as shown in Fig. 5, to help interpret the EIS data. The quality of fit was represented by the value of χ2 (chi-squared). The better the fit, the closer the value of χ2 is to 0. The values of χ2 are presented in the caption of Fig. 5. As we can see, for all adjustments low values were obtained, showing a good fit. Among the elements that were used (which consequently were the ones that were most adapted during the data adjustment process) are Rs (electrolyte solution resistance), Rct (charge transfer resistance), CPE (constant phase element), and W (Warburg impedance). It is important to note that when the CPE is inserted into the equivalent circuit, the material discussed does not behave like an ideal capacitor and therefore needs a different element that better portrays the reality of the material [53]. In this context, the insertion of the CPE in the circuit indicates that the surface of the material is inhomogeneous; therefore, the materials present irregularities at the interface between the material and the solution [50]. In addition, W is used in the interpretation of impedance data in the region where low frequencies predominate and is associated with the diffusion of ions from the bosom of the solution to the electrode/solution interface [54].

3.4 Stability at different pH’s

The stabilities of the PLA, filaments and 3D-printed electrodes were studied at different pH values (1, 7, and 14) using NaOH and HCl solutions. We observed no difference in the ongoing experiment between Protopasta® and PLA@CBMAX22 filaments during the first 4 h. However, when the flask containing the PLA@CB72R filament in an alkaline solution was shaken, small fragments were observed in the suspension, which we believe to be carbon black due to its dark color. The experiment was continued for 12 h, and when all the flasks were shaken, nothing was observed in any other container except the one with the PLA@CB72R filament in an alkaline medium. In Fig. S10, it is possible to observe a dark color at the bottom of the container, which originates from the particles present in the filament and are detached from the material. The same situation was observed at the end of 24 h, but it was much more evident, as shown in Fig. S10c. Regarding the PLA and 3D-printed electrodes (Fig. S11), no visible changes were observed with the naked eye during the 24 h of observation. Thus, these electrodes showed good stability against the different pH values studied.

According to the literature [26, 55], PLA can undergo saponification reactions in both acidic and alkaline media. However, it is known that the breaking of ester bonds is slower in acidic media, whereas it is more susceptible to undergo more effective hydrolysis in alkaline media (Fig. S12). There was no visible change in any of the filaments in the acidic medium, including the PLA@CB72R filament, after 24 h. However, it was possible to observe that the theory was proven for the mentioned filament, indicating that these bonds were indeed broken.

This phase of the research served as an inquiry into the stability of the materials analyzed under extreme pH conditions, as the subsequent stage pertains to hydrogen generation. Therefore, we sought to determine the optimal pH range in which the materials used for fabricating 3D-printed electrodes exhibited superior performance.

3.5 Electrodeposition of cobalt phosphate onto 3D-printed electrodes for hydrogen production

CoPi catalyst was electrodeposited on the 3D-printed electrodes by cyclic voltammetry. Fig. S13 shows voltammograms related to the CoPi electrodeposition process. The electrochemical characteristics of the 3D-printed electrodes were evaluated by cyclic voltammetry at various scan rates, as shown in Fig. S14. The CoPi#Protopasta® and CoPi#PLA@CBMAX22 electrodes presented well-defined oxidation and reduction peaks compared to the CoPi#PLA@CB72R electrode. From these voltammograms, it was possible to determine the electroactive areas of these materials, the values of which are presented in Table S10. A decrease in the electroactive area was observed when comparing the values obtained after electrodeposition (Table S10) with those of the electrodes without CoPi (Table S8).

As shown in Fig. 6a–c present the morphological characteristics of the CoPi electrodeposited on the electrodes. Several sphere-like aggregates were observed on the surfaces of the investigated electrodes after electrodeposition. A greater abundance of these spherical structures and a larger size were noted on the PLA@CB72R 3D-printed electrode.

Fig. 6
figure 6

SEM and EDS images were obtained for the 3D-printed electrodes after the CoPi electrodeposition. a and d CoPi#Protopasta® electrode, b and e CoPi#PLA@CBMAX22 electrode, c and f CoPi#PLA@CB72R electrode and g amount of hydrogen produced by each electrode

Full size image

Furthermore, elemental mapping images obtained by EDS (Fig. 6d–f and Fig. S15, S16 and S17) reveal that Co, O, and P are homogeneously dispersed on the electrode surface. Notably, the concentrations and distributions of these elements were higher in the PLA@CB72R electrode. A higher concentration of mapped elements can also be observed in the formed spherical structures, indicating that their origin is from CoPi electrodeposition. More information can be found in Supporting Information, S15.

The performance of 3D-printed electrodes in hydrogen production was investigated before and after CoPi electrodeposition through water electrolysis in acidic media, as described in the experimental procedure. The values of hydrogen produced and data treatment are available in Tables S11 and S12, present in the supplementary material. The efficiency of 3D-printed electrodes in hydrogen production was remarkable, especially after the CoPi electrodeposition procedure, as shown in Fig. 6g. The average value of hydrogen produced by Protopasta®, PLA@CBMAX22, PLA@CB72R, CoPi#Protopasta®, CoPi#PLA@CBMAX22 and CoPi#PLA@CB72R 3D-printed electrodes was 5.67, 6.93, 10.08, 8.50, 10.08, and 10.20 μmol min−1, respectively. These results indicate that 3D-printed electrodes with filaments prepared in our laboratory showed better performance in hydrogen production. Furthermore, it was possible to verify that among the 3D-printed electrodes investigated in this study, the one that presented the best performance was the PLA@CB72R electrode, both before and after the CoPi electrodeposition process.

4 Conclusion

This work proposes the production of new filaments prepared from PLA and different sources of Carbon black for use in the 3D printing of electrodes with interest in applications in hydrogen production. Thermal characterization was performed to determine the stability of the materials. Electrochemical characterization demonstrated that the prepared filaments and their respective 3D-printed electrodes exhibited better performance. The studies electrodes showed good stability at acid and neutral pH. In alkaline medium the PLA is hydrolyzed. The studied electrodes were modified by CoPi electrodeposition, and the procedure used was ideal, with confirmation growth of spherical structures by morphological and compositional analyses. Also analyzed was the performance of the 3D-printed electrodes for hydrogen production, and the electrodes printed with filaments prepared in the laboratory showed better results than those printed with commercial filaments. It is important to highlight that the PLA@CB72 and CoPi#PLA@CB72R electrodes show promising performance in hydrogen production, with 10.08 and 10.20 μmol min−1 values, respectively. Therefore, the filaments prepared by us for the 3D printing of electrodes present a better alternative for hydrogen production when compared to commercial filament. Additionally, our study paves the way for the conductive filaments development that can be used in energy conversion.