3.1. Fabrication and Characterization of Ni-G/SPE via Laser-Assisted Synthesis Using the Nelder–Mead Optimization Approach
The Nelder-Mead [
34] method was used to optimize the laser fabrication process by systematically adjusting the synthesis parameters to achieve a high sensor response of the fabricated electrodes to the target analyte. The Nelder-Mead method works by maintaining a simplex - a geometric figure consisting of n + 1 vertices in an n-dimensional parameter space. Through iterative processes of reflection, expansion, contraction, and shrinkage, the algorithm explores the parameter space to find the extremum of the objective function, which in this case was the maximization of the electrode response in 0.1 M PBS containing 100 µM DA.
The optimized parameters were laser power and graphene (G) suspension concentration in the DES. The laser power controls the temperature in the reaction zone at a given scanning speed, while the graphene suspension concentration affects the composition of the final structure. Therefore, these two parameters are essential for the one-step fabrication of the Ni-G composite on the SPE surface and allow to effectively tune the electrochemical properties of the fabricated electrodes. The feasible ranges for output laser power and graphene concentration were determined to be 500-1500 mW and 3-50 mass%, respectively. The parameter limits for laser power were chosen based on previous results for laser-assisted nickel patterning from DES [
32], while the graphene concentration was chosen over a wide range to ensure comprehensive experimentation. The proposed approach, based on the Nelder-Mead algorithm, provides a means for interactive experimentation. Here, the functional properties of the material are monitored directly during the development of the synthesis methodology, in contrast to the trial-and-error approach, where the investigation of functional properties and the determination of optimal synthesis conditions often occur after obtaining a large number of samples under different conditions. Such an approach allows for a significant reduction in the number of required experiments for material optimization tailored to specific applications.
Based on a data on nickel pattern fabrication from DES [
31,
32], the starting point for Nelder-Mead optimization was chosen with a laser power of 540 mW and a graphene suspension concentration of 10 mass%. Two additional points, located at a small distance along each of the parameter axes, were determined to form the initial simplex. At each point, laser synthesis of the electrode was performed, followed by measurement of the electrochemical response to dopamine.
Table S1 presents the values of laser power and graphene suspension concentration, along with the measured electrochemical signal for the simplex coordinates. SEM analysis of the samples produced at the simplex vertices (samples 1-8 on
Figure 2 and sample 9 on
Figure 3) revealed that structures fabricated from DES with low graphene concentration (up to 10%) exhibited a morphology typical of materials synthesized by laser-assisted deposition from DES [
35], consisting of nickel nanoparticles forming a continuous conductive pattern. As the graphene concentration in DES increased, distinct nanoparticle-decorated graphene sheets became visible. The much larger surface area of graphene compared to bare SPE provides numerous interfacial nucleation sites for heterogeneous crystallization of nickel nanocrystals. This large surface area facilitates the uniform dispersion of nanoparticles and promotes the formation of numerous small nuclei rather than a few larger ones, resulting in efficient and uniform decoration. However, an excessive amount of graphene led to an inhomogeneous modification of the electrode surface, which compromised the sensor performance. As a result of optimization using the Nelder-Mead method, an optimal point with coordinates of 973 mW laser power and 19.7 mass% graphene suspension concentration in DES was obtained. The response of the electrode at this point 9 was significantly enhanced, reaching a current of 90.1 µA (
Figure 2).
The electrode fabricated under the optimized conditions (sample 9) was characterized by scanning electron microscopy (SEM), X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), and Raman spectroscopy to investigate its morphological and structural properties. XRD patterns (
Figure S1) show reflections corresponding to graphite from the original non-modificated screen-printed electrode itself [
36], as well as a broad peak corresponding to the polyethylene terephthalate (PET) substrate of the SPE [
37]. In addition, reflections of metallic nickel were observed, confirming the reduction of nickel ions and the formation of metal nanoparticles due to laser-assisted pyrolysis. The diffraction peaks for Ni at 2θ values of 44.5° and 51.8° correspond to the characteristic (111) and (200) planes, which are indicative of face-centered cubic (fcc) structures [
38,
39].
The elemental composition and chemical states of the Ni-graphene/SPE were thoroughly investigated by XPS. The survey spectra (
Figure S2) show the presence of Ni, C, N and O. The high resolution XPS spectrum of Ni 2p (
Figure 3d) can be deconvoluted into two spin-orbit doublets and two shake-up satellites, with the main peaks observed at 873.4 and 855.8 eV, corresponding to a spin-orbit separation of 17.6 eV. According to literature data, these spectral features are characteristic of nickel nanoparticles, which readily oxidize upon exposure to air, resulting in XPS spectra dominated by oxide species. The surface sensitive nature of XPS, combined with the propensity of transition metal nanoparticles to form surface oxide layers and adsorb impurities, typically results in the presence of Ni²⁺ peaks in the spectra [
38,
40]. XPS analysis of oxygen reveals two distinct peaks, located at binding energies of 531.3 eV and 532.8 eV. The peak at 531.3 eV is typically associated with OH-groups and nickel-oxygen bonds, indicating the presence of nickel oxide or surface hydroxide species [
41]. This suggests that some nickel nanoparticles on the composite surface have undergone partial oxidation. The second peak at 532.8 eV corresponds to surface C-O bonds, which are characteristic of oxygen-containing functional groups on graphene sheets [
42]. These groups may include epoxy or hydroxyl functionalities introduced during the laser-assisted synthesis process. The high-resolution XPS scan of N1s revealed the presence of three nitrogen species: the N–Ni peak centered at 399.6 eV [
38,
43], graphitic nitrogen at 401.4 eV [
38,
43], and a peak at 402.5 eV assigned to N2 trapped within the material during laser processing, as well as N-O surface contamination [
44]. The formation of N–Ni and graphitic N species is highly favorable for electrochemical sensing applications and often results from high-temperature synthesis procedures [
43,
45]. The C 1s scan shows a spectrum typical of graphene structures and is consistent with the O 1s spectrum [
46,
47] The main peak at 284.6 eV can be attributed to the sp² carbon state, while the peaks at 286.3 eV and 288.9 eV indicate the presence of oxygen-containing functional groups.
The collected Raman spectra for the SPE, Ni-G/SPE, and graphene suspension all exhibit the characteristic D and G bands typical of carbon-based materials (
Figure 3f) [
48,
49]. Additionally, the 2D band is observed, which consists of several components. The multi-component nature of the 2D band can be attributed to oxidized forms of multilayer graphene and graphite [
50,
51]. The as-received SPE shows peaks specific to partially oxidized, highly oriented graphite (
Figure 3f). Considering the XRD data, the reasonable suggestion would be surface oxidation during manufacture and storage, while the main component of the SPE is graphite. As long as the width of this 2D band corresponds to layers non-uniformity, i.e., the presence of lattice disorder and defects, it can be concluded that laser irradiation promotes the formation of defects in the graphene layers. Therefore, laser heating not only initiates the pyrolysis of DES but also modifies the carbon component of the composite. The increase in defect sites and the formation of oxygen-containing groups can lead to a higher number of electrochemically active sites of graphene, thereby accelerating electron transfer during the electrochemical oxidation of dihydroxybenzenes, including catechol [
52]. Since the target biogenic amine, dopamine, is a catecholamine compound and its electrooxidation involves the oxidation of hydroxyl groups (
Figure S3), it is reasonable to assume that the formation of more defective graphene may also promote its electrooxidation.
3.2. Electrochemical Characterization of Ni-G/SPE and Dopamine Detection
Cyclic voltammetry (CV) measurements were performed in 0.1 M PBS supporting electrolyte in the presence of 50 µM DA to investigate the electrochemical behavior of the fabricated electrodes (
Figure 4a). The Ni-G/SPE exhibited a significantly higher current response in CV compared to the bare SPE. This improvement can be attributed to the incorporation of graphene on electrode surface, resulting in a composite with a larger active surface area compared to unmodified SPE, which allows for a greater number of simultaneous electrochemical reactions per unit of geometric surface area. In addition, the decoration of graphene with Ni NPs leads to an increase in the electron transfer rate compared to the bare SPE, resulting in higher peak currents observed in CV. Faster electron transfer kinetics are also indicated by smaller potential differences between the anodic (Epa) and cathodic (Epc) peaks for the Ni-G/SPE. The smaller potential difference suggests that redox reactions on the modified electrode are more reversible. The Ni-G/SPE shows a well-defined oxidation peak corresponding to the oxidation of DA to dopamine quinone (DAQ) and a reduction peak corresponding to the reduction of DAQ back to DA (
Figure S3).
Differential pulse voltammetry (DPV) was used for DA analysis because of its high sensitivity (
Figure 4b). The anodic peak current (Ipa) dependence on DA concentration was linear in the range of 0.25 to 100 µM, indicating an analytical sensitivity of 0,91 µA µM
-1 (
Figure 4c). The limit of detection (LOD) was determined to be 0.095 µM based on a S/N = 3, where S is the standard deviation of the blank response and N is the slope of the calibration curve. The analytical performance of the Ni-G/SPE for DA detection was listed along with sensors previously described in the literature [
53,
54,
55,
56,
57,
58,
59,
60]. The properties of the developed electrode are comparable with many existing analogues (
Table S2). However, it is worth highlighting the advantages of the Ni-G/SPE over similar electrodes, such as the precise modification of the electrode with localized Ni-G deposition on the SPE surface controlled by laser beam movement, while keeping the fabrication procedure rather simple. In addition, the well-known advantages of SPEs, including cost effectiveness, versatility, scalability, and flexibility, make them indispensable tools in modern electrochemical research and commercial sensor development. The developed modification strategy for SPEs significantly improves their electroanalytical performance in DA sensing. Using DES, this methodology facilitates the effective incorporation of advanced composite materials on the electrode surface, resulting in improved sensitivity, linear range, and LOD. Moreover, this approach is highly versatile and can be extended to the detection of a wide range of analytes as well as to the fabrication of other advanced composite electrodes.
In addition, the ability of Ni-G/SPE to detect DA in small volume samples was demonstrated (
Figure 4d). Aliquots of 50 µL containing three different concentrations of DA were used as model probes. The Ni-G/SPE electrode showed a prominent DPV response, and the peak current values fell within the linear regression, demonstrating the feasibility of the small volume sensing. However, the electrochemical response differs from data collected in bulk cells. In bulk electrochemical cells, mass transport is primarily driven by both convection and diffusion. This combination ensures a consistent and uniform supply of analytes to the electrode surface, maintaining a stable electrical double layer and consistent charge transfer kinetics. In contrast, in small volume probe measurements rely primarily on diffusion for mass transport because convection is minimal or absent [
61]. The limited volume and potential for evaporation in droplets affects the peak currents and the shape of the DPV curves. In this respect, and considering the promising initial results, further in-depth investigations are needed in the future to optimize the analytical protocol.
Further studies were aimed at investigating the selectivity, reproducibility and repeatability of the analysis as well as testing the electrodes for DA detection on real samples. The repeatability of the experiment was evaluated by eight successive measurements in a solution containing 50 µM DA. The relative standard deviation (RSD) of the peak current was 3.3%, indicating excellent repeatability of the electrode. Reproducibility was investigated based on measurements with three independently fabricated electrodes, and the RSD was found to be 4.5%. Such favorable analytical performance of the electrode can be attributed not only to the electrode composition, but also to the optimized fabrication conditions that allow the modification of the SPE with Ni-G having good adhesion to the SPE surface. Adhesion was tested by performing a scotch tape test (
Figure S5), confirming the robustness of the modified layer.
The most relevant coexisting substances in biological fluids are uric acid (UA) and ascorbic acid (AA), which are typically present at concentrations approximately 100 times higher than that of DA [
62]. In this context, the selectivity of the fabricated sensor was tested by DPV in the presence of 50 µM DA and 100-fold excess concentrations of AA and UA, as well as other possible interferents and electrolytes. The results are presented in
Table S3. All interferents tested, except UA, did not produce a significant analytical signal, and the signal deviation for DA did not exceed the calculated relative standard deviation for DA analyses without the interferents. For UA, an analytical signal was observed around 0.375 V. However, it does not affect the DA response, and the signal deviation for DA also did not exceed the RSD. The results obtained provide strong evidence for the excellent interference resistance of the developed electrode to the most abundant coexisting species.
The practical applicability of the Ni-G/SPE was confirmed by measurements on pharmaceutical samples of dopamine hydrochloride injections available at a local drugstore. The results are shown in
Table S4. According to the manufacturer, the concentration of dopamine hydrochloride in the injection is 5 mg/mL (26 mM). Sample #1 was prepared by diluting a 9,6 µL aliquot of as-received injection solution in 50 mL 0.1 M PBS, suggesting the 5 µM of analyte in the final solution. Sample #2 was prepared similarly, except that the aliquot volume was increased to 38,4 µL. In addition, a protocol based on the standard addition method was also used (samples #3 and #4), where the dopamine hydrochloride injection solution in PBS was spiked with a DA solution of known concentration prepared in the laboratory. The developed sensor showed acceptable RSDs (N = 3) and recoveries, indicating satisfactory accuracy for real sample analyses.