The materials used for the study were of analytical grade. Epinephrine hydrochloride, iron II sulphate heptahydrate (99% FeSO₄·7H₂O), and iron III chloride hexahydrate (99% FeCl₃.6H₂O) were purchased from Sigma Aldrich. Ammonium hydroxide (25% NH₄OH), zinc chloride (99% ZnCl₂), nitric acid (70% HNO₃), sodium hydroxide (99% NaOH), hydrochloric acid (38% HCl), sodium dihydrogen phosphate (99% NaH₂PO₄), disodium hydrogen phosphate (99% Na₂HPO₄), dimethylformamide (99% DMF) from Sigma Aldrich Chemical, South Africa. The screen-printed silver electrodes (SPSE) were from Metrohm, South Africa. The spleen, blood serum, breast muscle, kidney, and liver of Buff Orpington Rooster and Rhodes Island White Broiler used for real sample analysis were collected from a local butchery immediately after slaughter. Ethical practices were adhered to in all the sample collection, processing, and analysis procedures.

The magnetic iron oxide nanoparticles (Fe₃O₄ NPs) were synthesized using a technique previously adopted by Gorospe et al. (Gorospe et al., 2019) with slight modification. The FeCl₃·6H₂O and FeSO₄·7H₂O precursors were dispersed into a conical flask containing nitrogen gas and purged distilled water at a respective molar ratio of 2:1. This mixture was then subjected to vigorous stirring on a magnetic stirrer. Approximately 5 ml of ammonium hydroxide was transferred into a burette and discharged in drops into the flask containing a mixture of the iron salts. The resultant black suspension was immediately covered and subjected to stirring for almost 10 min. The magnetic nanoparticles were separated using a bar magnet and dispersed in distilled water. These nanoparticles were washed with sufficient distilled water for 6-times and subsequently with ethanol 4-times. The black nanoparticles obtained were dried under vacuum at room temperature.

The drop-dry technique was used to fabricate the Fe₃O₄ NPs modified SPSE (SPSE/Fe₃O₄). The as-synthesized Fe₃O₄ NPs was dispersed in DMF and sonicated for 24 h to obtain a homogenous paste. About 8 μl of this paste was dropped on the screen printed silver working electrode and allowed to dry at room temperature. Scheme 2 shows the schematic representation of the synthesis of Fe₃O₄ NPs and subsequent modification of SPSE with the as-synthesized Fe₃O₄ NPs. The modified electrode (SPSE/Fe₃O₄) was applied for the electroanalysis of epinephrine.

The OR and RIWB blood samples were collected in vials and kept in ice to prevent coagulation. The sample pre-treatment was done according to a method presented by Zhao et al. (Zhao et al., 2015) with minor modification. About 0.05 ml concentrated nitric acid (70% HNO₃) was added to the blood sample in a centrifuge tube and allowed to stand for 2 h. Neutralization of the mixture was carried out with 0.05 ml of 1 M sodium hydroxide. The resultant solution was centrifuged for 20 min at 4,000 rpm. The clear serum solution at the top was separated by decantation and stored at 4°C for further analysis. The treated OR and RIWB blood samples were labelled Serum 1 and Serum 2, respectively. The sample preparation for the breast muscle, spleen, kidney and liver from both chicken species (OR and RIWB), were collected similarly to the blood samples before electrochemical analysis followed a method by Kahlouche et al. (Kahlouche et al., 2018) with slight modification. A reasonable amount of 0.1 M PBS (pH 7) was added to the samples before pulverization to a homogeneous paste. The resultant paste was dispersed in a solution of 1 M ZnCl₂ and allowed to stand for 30 min. Afterwards, this solution was centrifuged at 4,000 rpm for 15 min. The supernatant was separated from the lower solid masses and stored at 4°C for analysis. The treated kidney samples obtained from RIWB and OR were labelled RIWKID and ORKID, respectively. While the liver samples from RIWB and OR were tagged RIWLIV and ORLIV, respectively. RIWB and OR breast muscle extracts were coded RIWBM and ORBM, respectively. RIWBBS represents the spleen extract from the spleen of RIWB.

Fourier transform infrared (FT-IR) Opus Alpha-P spectrophotometer supplied by Brucker Corporation, United States, UV-visible, and X-ray diffraction (XRD) spectroscopy were used to characterize Fe₃O₄ NPs. The XRD and UV-visible data were collected using the Rontgen X’Pert Pro diffractometer and Uviline 9400 spectrophotometer (Germany). Microscopic characterization of nanoparticles was achieved with the scanning electron microscope (Quanta FEG 250 supplied by ThermoFisher Scientific, United States) and transmission electron microscope. The scanning electron microscopy (SEM) and transmission electron microscopy (TEM) images obtained with this equipment provided details about the morphology and ultrastructure of nanoparticles. All electrochemical analyses were done using portable DropSens potentiostat (MetrohmR) connected to a personal computer for data processing. This workstation includes a screen-printed silver electrode (SPSE) with embedded working, counter, and reference electrodes. The supporting electrolyte used for all electroanalyses (0.1 M phosphate buffer solution, PBS) was prepared using the appropriate concentration of Na₂HPO₄ and NaH₂PO₄.

The XRD diffractogram of Fe₃O₄ NPs (Figure 1) showed peaks at 30.1°, 35.7°, 43.3°, 53.9°, 57.5°, and 63.0°, corresponding to Miller indices of 220, 311, 400, 422, 511, and 440, respectively. This diffraction pattern suggests that the Fe₃O₄ NPs have a spinel structure (according to JCPDS No. 19-629) (Pîslaru-Dănescu et al., 2017). A similar diffraction pattern has been reported for Fe₃O₄ NPs in various studies on Fe₃O₄ NPs (Loh et al., 2008; Shagholani et al., 2015; Pîslaru-Dănescu et al., 2017). Using the Scherrer equation (Eq. 1), the crystallite particle size using the peak at 35.7° (311) was found to be 10.91 nm. D = 0.9 λ B c o s θ     (1)

D, λ , B and, ɵ in Eq. 1 represent the crystallite particle size, X-ray’s wavelength (0.154 nm), full width at full maximum (FWHM), and half of the diffraction angle (2 theta/2), respectively.

The FT-IR spectrum of Fe₃O₄ NPs (Figure 2) shows peaks at 565, 1,048, 1,348, 1,571, and 3,174 cm⁻¹. The peak at 565 cm⁻¹ suggests the presence of Fe-O bond in Fe₃O₄. Also, the absorption band at 565 cm⁻¹ confirms the magnetite phase in the Fe₃O₄ NPs (Zavareh et al., 2017). The absorption band at about the same wavenumber has been reported for Fe-O in previously synthesized Fe₃O₄ NPs (Bertolucci et al., 2015; Pham et al., 2016; Zavareh et al., 2017). The absorption band at 3,174 cm⁻¹ is probably due to the–OH stretching vibration from the water adsorbed on the surface of the Fe₃O₄ NPs (Pham et al., 2016). The absorption bands at 1,571 cm⁻¹ is due to the–NH bending vibration of the–NH bond in the residual unreacted ammonia.

The UV-visible spectra of the Fe₃O₄ precursors depicted in Figure 3A show two peaks at 292 and 738 nm for FeCl₃·6H₂O and a single peak at the visible region (738 nm) for FeSO₄·7H₂O. The UV-visible spectrum of the Fe₃O₄ NPs presented in Figure 3B showed two strong absorptions at 386 and 634 nm. Although these absorption bands (in Fe₃O₄ NPs) are similar to FeCl₃·6H₂O, they emerged at different wavelengths (Figures 3A,B). The significant difference in the absorption wavelengths of the precursors and the Fe₃O₄ NPs suggests the successful synthesis of the nanoparticles. Noteworthy, the emergence of strong absorptions at the ultraviolet and the visible region has been reported for Fe₃O₄ NPs (Cao et al., 2015). Also, absorption peaks at about 386 nm have been reported for Fe₃O₄ NPs prepared using various precursors (Hu et al., 2014; Cao et al., 2015; Taka et al., 2019). The absorption band at 634 nm represents the plasmon resonance band.

Figures 4A,B shows the SEM image of the Fe₃O₄ NPs obtained at two different magnifications. The images obtained at high magnification (100, 000 X) (Figure 4A) show that the Fe₃O₄ particles are made of aggregates of spherical masses. At lower magnification (×20,000) (Figure 4B), the spherical particles appeared to form aggregates that give the entire Fe₃O₄ NPs a collective rough surface morphology. Interestingly, magnetic nanoparticles of similar surface morphology have been reported (Pîslaru-Dănescu et al., 2017). Using the SEM image obtained at higher magnification, an average particle size of 63.6 nm was calculated for the Fe₃O₄ NPs. The internal structure of the Fe₃O₄ NPs depicted on the TEM micrograph of these nanoparticles (Figure 4C) also confirmed the spherical nature of the individual particles and the aggregation of the Fe₃O₄ NPs. The average particle size of the nanoparticles obtained from the analysis of the TEM image (Figure 4D) is 11.8 nm. This value is close to the average crystallite particle size calculated from the X-ray diffractogram. The disparity between the particle sizes obtained from SEM and TEM is due to the difference in resolution of the microscopes used for both techniques.

Electrochemical characterization of the bare SPSE and SPSE/Fe₃O₄ was done using these electrodes for EIS studies in the presence of 1 mM EP (pH 7). Figure 5A shows the Nyquist plot of the bare and modified electrode in EP. Figure 5B depicts the equivalent circuit for fitting the EIS data. The values of the solution resistance (R_s), charge transfer resistance (R_ct), constant phase element (CPE), and the Warburg impedance (W) used for the EIS circuit fitting have been itemized in Table 1. The R_ct of the bare SPSE (8,005 Ω) dropped drastically after modification with Fe₃O₄ (145 Ω), suggesting an improvement in the ability of the electrode to support charge transfer after surface modification. This indicates that the Fe₃O₄ in SPSE/Fe₃O₄ improved the conductivity of the bare electrode, resulting in the better electrocatalytic activity of SPSE/Fe₃O₄ towards EP oxidation. A similar decrease in R_ct and the attendant increase in the conductivity of electrodes after modification with nanoparticles has been reported in several studies (Zangeneh Kamali et al., 2014; Arani et al., 2019). The difference in R_ct values of the electrodes also confirms the superior EP current response recorded at SPSE/Fe₃O₄ (with CV) relative to that of the bare electrode (Figure 6B). Also, the Warburg impedance (W) obtained at the low frequency end of the EIS spectra of the bare and modified electrode suggests a diffusive behavior around the electrodes (Vedalakshmi et al., 2009). Noteworthy, Y_o and N in Table 1 represent the magnitude of the CPE and exponent of the CPE, respectively.

The electron transfer rate constant (ks) was calculated using Eq. 2 (Adekunle et al., 2010). R, T, n, A, and F in Eq. 2 represent molar gas constant (J mol⁻¹ K⁻¹), absolute temperature (K), number of electrons Figure 6B transferred, electrode surface area and Faraday’s constant (C mol⁻¹), respectively. The ks values for bare and SPSE and SPSE/Fe₃O₄ were calculated as 1.01 and 21.28 s⁻¹. This implies that the modification of bare SPSE with Fe₃O₄ drastically raised the electron transfer rate constant, resulting in the significantly higher current response recorded at SPSE/Fe₃O₄ relative to bare SPSE (Figure 6B). k s = R T n 2 F 2 A R c t C (2)

The electrochemical analysis of EP was done with bare SPSE and Fe₃O₄ NPs modified SPSE (SPSE/Fe₃O₄) in 0.1 M PBS (pH 7) at a scan rate of 25 mVs⁻¹. Figure 6A shows the cyclic voltammogram of 0.1 M PBS (pH 7) at the bare and modified electrodes. These voltammograms revealed that no peak emerged after the analysis of the blank solution. In the presence of EP, the bare electrode showed only the anodic peak at 0.3 V. On the other hand, the modified electrode showed a prominent anodic peak at 0.32 V and a weak cathodic peak at −0.12 V (Figure 6B). The anodic peak current (I_ap) obtained with bare and modified electrodes were 30 and 79 μA, respectively. The electroactive surface area of the bare and modified electrodes calculated from Randle Sevcik equation (Eq. 3) were 0.79 and 2.08 cm², respectively. D, C, n, v, A, and i in Eq. 3 represent diffusion coefficient (cm² s⁻¹), concentration of EP (mol cm⁻³), number of electrons, scan rate (Vs⁻¹), electrode surface area (cm²) and peak current (A), respectively. D was taken as 1.01 × 10^–5 cm² s⁻¹ for EP in 0.1 M PBS (pH 7) and n = 2 as earlier reported (Wang et al., 2006). Noteworthy, the current response recorded at SPSE/Fe₃O₄ is about 2.6 times higher than that of the bare electrode. This significant increase in I_ap at SPSE/Fe₃O₄ could be because the magnetic Fe₃O₄ nanoparticles increased the electroactive surface area of SPSE (as evident in the values of A for both electrodes). Such improved current response of screen-printed electrodes in the presence of neurotransmitters after modification with nanomaterials have been reported (Valentini et al., 2014; Jahani and Beitollahi, 2016; Beitollahi et al., 2018; Samie and Arvand, 2019). The increase in k_s value of the electrode after modification with Fe₃O₄ NPs could have also contributed to the drastic increase in EP oxidation peak current at SPSE/Fe₃O₄ relative to bare SPSE.

Also, the CV of the electrodes in EP showed that a redox process was recorded at the modified electrode while only the oxidation of EP was achieved with the bare electrode (Figure 6B). This confirms that the magnetic nanoparticles on the bare SPSE offered a catalytic effect that supports both the EP’s oxidation and the reduction. The ratio of I_ap and the cathodic peak current (I_cp) recorded at the modified electrode gave a value of 3.43, which suggest a quasi-reversible redox process (since I_cp is 23 μA). i = 2.69   x   10 5 n 3 / 2 A D 1 / 2 C v 1 / 2 (3)

The mechanism of EP redox process involves the two electrons and protons process depicted in Scheme 3. The quasi-reversible process in Figure 6B shows the formation of epinephrine quinone after the loss of two protons and electrons by the EP molecule for the oxidation reaction while the reverse occurs for the reduction process (Scheme 3).

The effect of scan rate on the EP current response at SPSE/Fe₃O₄ was investigated using cyclic voltammetry at a scan rate range of 25–400 mV s⁻¹ at a pH of 7. There was an increase in I_ap and I_cp with every increase in scan rate (Figure 7A). In addition, the anodic peak potential (E_ap) increases with the increase in scan rate (at lower scan rate) while the E_ap slightly changes at higher scan rates (Figure 7C). Instead, the cathodic peak potential (E_cp) was almost constant with an increase in scan rate except for a few changes at lower scan rates. This peak potential and scan rate relationship is entirely different from the steady increase in E_ap with an increase in scan rate reported in various studies (Valentini et al., 2014; Beitollahi et al., 2018). A linear relationship represented as I_ap = −50.2933 + 16.4391 v^1/2 was observed between the square root of the scan rate and the anodic peak current (I_ap) (Figure 7B). The fact that this linear fit was obtained, especially at a higher scan rate (>200 mV s⁻¹) suggests that the redox process at the modified electrode is a diffusion-controlled process. A similar mechanism has been reported for EP oxidation at modified electrodes (Wierzbicka and Sulka, 2016).

The effect of EP concentration on current response recorded at the modified electrode was investigated using SPSE/Fe₃O₄ for the analysis of EP with the square wave voltammetry (SWV) techniques at a pH of 7 (scan rate of 25 mV s⁻¹). The SWV parameters used for this analysis include a potential step of 10 mV, 10 mV amplitude, and a frequency of 10 Hz. The voltammogram in Figure 8A represent the response of the modified electrode to change in EP concentration. The current response increased with an increase in concentration over a range of 9.99–83.8 μM (Figure 8B). The relationship between (I) and [EP] was linear over a range of 9.99–60.9 μM (Figure 8B). The regression equation representing the linear relationship between current (I) and [EP] using SWV is I = 17.4612 + 0.1376 [EP].

The detection limit obtained with the modified electrode is 19.3 μM. This detection limit is comparable to sensors previously fabricated for EP detection (Table 2). Specifically, the detection limit of the sensor is lower than the value reported for a sensor fabricated through the modification of GCE with poly ferulic acid and MWCNTs composite (da Silva et al., 2017). On the other hand, this sensor gave a wider linear range than some sensors earlier fabricated (Mphuthi et al., 2017). Noteworthy, the detection limit of the sensor was calculated using the formula 3.3 ɛ/m where ɛ and m represent the standard deviation of the intercept, and the slope of the regression equation, respectively.

The ability of the fabricated SPSE/Fe₃O₄ sensor to simultaneously detect epinephrine and ascorbic acid (AA) in a solution containing a mixture of these biomolecules was investigated. Essentially, the presence of AA in extracellular fluids (where EP can equally be found) is several orders of magnitude higher in concentration than EP. As a result, a mixture containing 1 mM AA and 0.4 mM EP at a pH of 7 was subjected to electroanalysis by the fabricated sensor using SWV (potential step of 10 mV, 10 mV amplitude, and a frequency of 10 Hz). The sensor was applied for the electroanalysis of AA to identify its oxidation potential. A peak representing AA was detected at 0.189 V. The analysis of the mixture of AA and EP showed peaks at 0.167 and 0.446 V, respectively (Figure 9). The peak separation between the peak potentials of these analytes (0.279 V) is wider than the value reported for some previously fabricated sensors (Salimi et al., 2004; Ren et al., 2006; Ghanbari and Hajian, 2017). This outcome suggests that the sensor is capable of discriminatory EP detection in the presence of AA.

The analysis of EP in chicken blood serum, breast muscle, spleen, kidney, and liver from chickens was done using the standard addition method. About 0.4 ml of the real sample was diluted with 15 ml of PBS. The EP in this unspiked sample and real sample solution was analyzed using the fabricated sensor. The data obtained (in triplicates) from the analysis of all real samples was highlighted in Table 3. The percentage recovery obtained (n = 3) for the Serum 1 (OR), Serum 2 (RIWB), ORBM, ORKID, ORLIV, RIWKID, RIWBM, RIWLIV and RIWBBS were 102.13, 100.38, 102.15, 107.10, 101.71, 108.09, 100.48, 106.79 and 98.33%, respectively. The % RSD for each recovery has been itemized in Table 3. Generally, the % RSD obtained from the analysis of the real samples is satisfactorily low. The data obtained from the real sample analysis also revealed that the fabricated sensor could be applied to detect EP in the blood serum and body fluid of an animal. The data from this analysis also confirm that the portable SPSE/Fe₃O₄ sensor can be used for EP real time point-of-care analysis. In addition, this sensor has the potential of serving as a reliable tool for EP detection in various extracellular fluids without elaborate sample preparation.

The stability of SPSE/Fe₃O₄ in the presence of EP was investigated using 1 mM EP in PBS (pH 7) at 25 mV s⁻¹. Figure 10 A depicts the current response of the modified electrode recorded after 16 cyclic voltammetry scans. The % RSD of the peak currents recorded for the 16 scans is 22.88%. The analysis of the voltammogram (Figure 10A) also shows that the electrode retained only 47.8% of its initial current after 16 scans. Compared to some other sensors that lost much smaller percentage of their initial current response after higher CV scans (Taleb et al., 2018; Immanuel and Sivasubramanian, 2020), this electrode is considered unstable. The repeatability of the sensor’s output was investigated by independent analysis of EP with the modified electrode (under the same condition deployed for the stability study) in three separate EP solutions (1 mM, pH 7). The electrode shows a high level of repeatability after the second attempt. This evident in the closeness of the current response recorded for the second and third attempt (Figure 10B) as well as the low % RSD recorded (11.9%) for the anodic current obtained after the three attempts.