Silver nitrate (AgNO₃, 99.9%) was used in the synthesis of the nanoparticles and the electrolyte used throughout this study was pH 7.4, 0.2 M phosphate buffer (PBS). PBS was prepared using sodium phosphate monobasic dihydrate (NaH₂PO₄·2H₂O, 99%) and sodium phosphate dibasic dihydrate (Na₂HPO₄·2H₂O, 99.5%). All were purchased from Sigma Aldrich, Kempton Park, Johannesburg, South Africa. Fresh bananas and grapes used as capping and stabilizing agents in the synthesis of the nanoparticles were purchased from Checkers Hypermarket, Cape Town, South Africa. The deionised ultra-purified water prepared with a Milli-Q water purification system was procured from Merck KGaA, Darmstadt, Germany. When not in use, the prepared electrolyte solutions were kept refrigerated at 4˚C.

High-Resolution Transmission Electron Microscopic (HR-TEM) analyses of the nanoparticles were performed with a Tecnai G2 F20 X-Twin HRTEM purchased from FEI Company, Hillsboro, OR, United States of America. Cyclic Voltammetry (CV) studies were performed with a PalmSens3 Potentiostat from PalmSens3 BV, Houten, The Netherlands. Furthermore, the structural and optical properties of the synthesised silver nanoparticles were studied with Ultraviolet-Visible (UV-Vis) Spectroscopy (with a Nicolett Evolution 100 from Thermo Electron Corporation, Johannesburg, South Africa) and Fourier Transform Infrared Spectroscopy (FT-IR) using a PerkinElmer Spectrum 100-FT-IR Spectrometer from PerkinElmer (Pty) Ltd., Midrand, South Africa. The internal structure of the nanoparticles was studied with Small-Angle X-ray Scattering (SAXS) (Anton Paar, GmbH, Australia).

Banana peels were separated from bananas and then finely cut before being washed repeatedly together with 50 g grapes using ultra-purified water to remove any dust and organic impurities. Thereafter, the grapes and cut banana peels were placed into different beakers containing 300 mL distilled water where each mixture was heated at 80˚C and stirred at 1000 rpm for 30 min. The extracts were then kept at room temperature and allowed to cool followed by centrifugation at 1000 rpm for 30 min. The obtained products were filtered twice to remove any insoluble fractions and macromolecules. The two extracts were then combined at various ratios, however, the ratio of 1 (banana):3 (grape) was found to synthesize the most stable nanoparticles with shelf-lives in excess of 3 months.

In this study, the natural reducing agents present in grapes and banana peels were used to produce silver nanoparticles. A stock solution of 2.0 × 10⁻⁵ M AgNO₃ was prepared in 100 mL ethanol. Then, 50 mL of the AgNO₃ stock solution was mixed with 50 mL of the grape-banana peel extract (GBPE). The mixture was boiled at a reaction time of 18 min at 60˚C with stirring maintained at 1000 rpm. The formation of a dark brown-coloured solution was an indication of the successful synthesis of the nanoparticles. Figure 1 clearly indicates the successful reduction of silver (Ag³⁺) (Figure 1(b)) to Ag-NPs (Figure 1(c)) using GBPE (Figure 1(a)).

Prior to analysis, glassy carbon electrodes were cleaned by polishing the electrodes on different alumina surfaces followed by thoroughly rinsing with distilled water. Thereafter, 4 µL of the synthesized nanoparticles was drop cast onto the electrodes and allowed to dry for 24 h. Analyses were initiated by placing the modified glassy electrode in a 3-electrode configuration cell together with Ag/AgCl reference and Pt wire counter electrode prior to analyses using cyclic voltammetry (CV) to study the electrochemical properties of the nanoparticles. The chosen potential window was from −1.5 V to 1.5 V and the analysis was achieved at different scan rates (10 - 100 mV/s).

The antibacterial activities of Ag-NPs were tested against E. coli 0157:H7 using

the traditional disk diffusion method widely known as the Modified Kirby-Bauer susceptibility assay [25]. A single colony of glycerol stock E. coli 0157:H7 strain was grown overnight in a 20 mL Luria Broth (LB) medium. The media was then placed on a rotary shaker at 220 rpm at 37˚C to grow bacteria for 24 hours. Thereafter, the prepared bacterial suspension was serially diluted between 10 - 1016 CFU/mL with LB media in test tubes and a loop full of bacterial culture was placed on Thema Muller-Hinton agar (MHA) platets before placing 6 mm diameter disks impregnated with interventions on the plate (3 per plate). Three sets of plates were incubated with 10 µg/L concentration Ag-NPs in the following E. coli 0157:H7 volumes 100 µL, 150 µL and 200 µL, respectively and incubated at 37˚C for 24 hours. Thereafter, the average diameters of the zones of inhibition around the disks were determined by a Vennier calliper.

Studies have focused on the functionalization of silver nanoparticles using banana peel extract. However, there is currently no literature search that has thus far revealed the combination of grape and banana peel extracts (GBPE) for the synthesis of Ag-NPs. The attraction to these compounds is attributed to the fact that independently and in combination with silver metal precursors, banana peel and grapes have the ability to form biocompatible nanoparticles [26]. This study revealed increased stability and shelf life of the synthesized nanoparticles as opposed to the synthesis of the nanoparticles using the respective biological compounds. We hypothesize that it is the oxygen atoms from phenolic acid-type biomolecules found in both banana peels and grape-fruit that are responsible for the biosynthesis of Ag-NPs in this study [15]. Literature reports phenolic acid-type biomolecules to be responsible for the formation and stabilization of Ag-NPs [27]. This mechanism occurs according to Pearson’s acid-base concept which elucidates that when a soft metal such as Ag+ specifically binds with a hard ligand, the chosen soft metal undergoes a reduction reaction since no complex compound is produced but instead the desired Ag-NPs are produced [16]. UV-Vis was used to follow the reduction of aqueous AgNO₃ ions during the reaction with GBPE. A strong absorption band at 440 nm was observed confirming the presence of Ag-NPs as shown in Figure 2, literature reports confirmed that silver nanoparticles absorb radiation in the visible regions of 400 - 500 nm [28]. Recently Masum and co-workers also confirmed the maximum wavelength (λ_max) of silver nanoparticles to be approximately 436 nm [29] while in another study by Kumar and co-workers it was obtained at 440 nm [30].

The visual observation of the reaction media changed, forming a dark brown colour, as shown previously in Figure 1 illustrating the formation of Ag-NPs [31]. The colour of the reaction mixture started to change from pale brown to light brown within 9 min forming a dark brown solution after 12 min, as shown in Figure 3. As the reaction mixture became darker, more Ag-NPs were formed

[32]. The brown colour arises due to the excitation of surface plasmon resonance (SPR) vibrations in the silver nanoparticles [33]. This is because silver nanoparticles have free electrons which give rise to SPR absorbance bands [34].

The intensity of SPR peaks was also seen to gradually increase during the synthesis of the nanoparticles with increased reaction time from 3 min until 18 min as shown in Figure 4. At 18 min, the maximum SPR peak was observed due to the depletion of silver ions (Ag⁺) in the GBPE. The SPR band of silver nanoparticles shifts to the higher wavelength as the reaction time increases [35] [36] [37] indicating a redshift [38]. Redshift shows that there is a gradual increment in the mean diameter of GBPE-Ag-NPs [32]. The spectra show the reduction of Ag⁺ to Ago specifying the formation of Ag-NPs with a completed time of 18 min. The optimal time in this study was 18 min, which is relatively less than in other reports. For instance, Kumar and co-workers reported 60 min for Ag-NPs synthesised using Hydnocarpuspentandra leaves [39].

FT-IR was employed to validate the major functional groups in the extract and their involvement in the synthesis and stabilization of the synthesized silver nanoparticles. Figure 5 represents the spectra of GBPE (black curve) and the

synthesized Ag-NPs (red curve) where they were compared in order to investigate the changes before and after the reaction with peak assignment. Similar spectra between the extract and colloidal solution were obtained as shown in indicating that the same compounds existed before and after the reaction. Although the similarities were observed, some of the absorption peaks are seen to be shifted in their position after the addition of AgNO₃.

From the FT-IR spectra, the presence of the sharp O-H peak at 3366 cm⁻¹ was due to the stretching vibrations of hydroxyl groups (-OH) of alcohol and phenols [40]. The sharp peak at 2922 cm⁻¹ in GBPE represents the O–H stretching vibration bands characteristic of carboxylic acids and C-H stretches from alkanes. Hence, the synthesised Ag-NPs are stabilized and coated by functional groups that are present in the GBPE [41]. The functional groups that are present within GBPE include polysaccharides such as cellulose as well as the phytochemical carotenoid [42] [43]. The reduced peak intensities in the spectrum indicate that the affected groups are involved in the formation of the nanoparticles.

The FT-IR spectrum of Ag-NPs showed a band at 3270 cm⁻¹ for hydrogen-bonded stretching vibrations attributed by phenol and alcohol groups. The weak peak at 2093 cm⁻¹ is due to thiols S-H stretching vibrations indicating the presence of the banana peel in the GBPE extract. The peak observed at 1725 cm⁻¹ was assigned as the stretching vibration of C=O from carboxylic acids (found in bananas), ketones (found in grapes) and aldehydes (found in grapes). The peak at 1633 cm⁻¹ is characteristic of C=C stretching vibrations attributed to alkene groups (found in grapes). The bands at 1407 and 1253 cm⁻¹ are due to the presence of aromatic rings and nitro compounds while the C=C, N-O and C-N stretching vibrations are due to amine groups, the intense peaks at 1407 and 1253 cm⁻¹ were assigned due to nitriles in the GBPE [44]. These peaks were reduced in intensity as compared to the Ag-NPs spectrum revealing the involvement of nitrile groups from the extracts as the capping agents for the nanoparticles. The weak peak at 636 cm⁻¹ corresponds to asymmetric stretching vibrations of C-H groups in the GBPE as shown in the spectrum and seen to disappear in the spectrum of Ag-NPs due to its participation in the synthesis process. The Ag-NPs spectrum also indicates another peak at 585 cm⁻¹ caused by the stretching vibrations of the nanoparticles thus confirming the formation of nanoparticles [45]. Finally, that amine, hydroxyl and carboxyl groups were responsible for the formation of the novel silver nanoparticles while the phenolic compounds that were present in the extract were the driving force in the formation of the nanoparticles. The obtained results are similar to a recent report by Singh and co-workers who reported the synthesis of Ag-NPs using RhodiolaRosea rhizome extract [46]. Hence from the above analysis, we concluded that amine, hydroxyl and carboxyl groups present in the extract played a significant role in the reduction of Ag⁺ ion into Ag-NPs.

HRTEM was used to determine the size and the shape of the bio-synthesized Ag-NPs. Figure 6(a) shows the HRTEM image with a scale view of 200 nm shows the GBPE capped Ag-NPs to be predominantly poly-dispersed and spherical in shape with sizes ranging from 25 to 39 nm, which agrees with the shape of the SPR band in the UV-vis spectrum [47]. Similar results of spherical Ag-NPs shape were obtained using different fruit extracts such as Phoenix dactylifera extract and Watermelon rind extract [48]. Furthermore, the lattice fringes visible in Figure 6(b) indicates good crystallinity and the lattice spacing (d-spacing) was determined to be 0.231 nm. This spacing is known as inter-planar spacing (d-spacing) which is in between the vertical lattice fringes, indicating the space between the XRD Braggs reflection. According to Liang and co-workers, an

interlayer spacing of 0.231 nm corresponds to the lattice spacing of the silver (111) planes [49] [50]. Gomati and co-workers also reported lattice spacing of 0.23 nm for Ag-NPs synthesised using Daturastramonium leaves [51].

Figure 6(c) confirms the crystalline nature of the nanoparticles and the five intense diffraction patterns obtained in the spectrum of nanoparticles correspond to the standard metallic silver JCPDF NO:04-0873 [52]. Aruna and Jarayaman reported similar results for green synthesised silver nanoparticles using Gum tragacanth [53] while Dong and co-workers also reported similar results using wolfberry fruit [47]. Figure 6(d) shows the selected area electron diffraction (SAED) illustrating the following five Braggs reflection rings (111), (200), (220), (311), and (222) planes an indication that the nanoparticles are highly crystalline [54]. The SAED pattern is in good agreement with XRD (Figure 6(c)) findings where intense Bragg reflection peaks at 38.17, 44.43, 64.58 and 77.64 corresponding to the silver crystal planes (111), (200), (220), (311) and (222) an indication of face centered cubic crystalline structured metallic silver [55]. The average particle size of the nanoparticles was determined to be 30 nm based on the full-width half maximum (FWHM) data using Scherer’s equation (Equation 1.1,). Similar reflections were also observed in the XRD spectrum of the green synthesised silver nanoparticles using VitexNegudo [56], sufron waste [57], and Azadirachtaindica leaf [58]. The particle size distribution histogram of the synthesized Ag-NPs shown in Figure 6(e) corresponded well with values obtained from HRTEM and XRD measurements.

The energy-dispersive x-ray spectroscopy (EDS) of the nanoparticles illustrated in Figure 6(f) shows the presence of the silver elements at 3 KeV, which confirms the formation of silver nanoparticles [59]. The presence of carbon and oxygen in the sample was also found due to the presence of polyphenol groups originating from the extract and are bound to the surface of silver nanoparticles thus indicating the reduction of silver ions (Ag⁺) to elemental silver (Ag). The analysis further shows the presence of copper contributed by the copper grid on to which the sample was coated. No other peaks were observed in the spectrum, confirming the complete reduction of silver compounds to silver nanoparticles [60]. Previously obtained reports similar to these results were also reported by [61] [62].

d = K λ β cos β (1.1)

where: d = particle/crystal size, K is a constant and can be defined as the dimensionless shape factor and is equal to 0.94 K usually range from 0.5 - 1, and λ is the X-ray wavelength (0.154 nm), β is the full-width half-maximum (FWHM in radians), θ is the angle of reflection [63].

Small Angle X-Ray Spectroscopy (SAXS) was employed to determine the precise size distribution, shape, and internal structure of the synthesized Ag-NPs. As can be seen in Figure 7(a), the SAXS free model pair-distance distribution function (PDDF) of the GBPE capped Ag-NPs revealed a spherically characteristic profile typically observed for poly-dispersed nanoparticles with the largest particle size detected to have a diameter of 39 nm [26]. Agglomerated particles also appeared at 35 nm which is illustrated by a hump. [64] [65]. Figure 7(b) reveals the size distribution weighted by number (black line) and intensity (red line) where the Ag-NPs highly revealed poly-dispersed particles with average radii of (13 nm, 20 nm and 29 nm), respectively [27]. Most of the particles appeared at 4 nm for weighted by number and 29 nm for intensity distribution functions. The size of 29 nm is close to the determined values as reported by HR-TEM. However, the appearance of an enhanced peak at 29 nm was also ascribed to distribution by intensity. Since, the colloidal suspensions consist of poly-dispersed particles and the measured scattering intensity represents the sum of the scattering intensities from particles of various sizes [28]. This is given by Equation (1.2):

I ( q ) = N V ( p − p 0 ) 2 ∫ 0 ∞ ( π d 3 6 ) 2 f ( d ) P ( q , d ) d d (1.2)

where f ( d ) is the particle size distribution by number.

In this study, the electrochemical analyses of Ag-NPs were performed in order to monitor the characteristic features of the novel GBPE-AgNPs. Figure 8(a) shows the CV voltammograms of the bare GCE as well as the GBPE-Ag-NPs/GCE modified electrode in 0.2 M PBS, pH 7.4 electrolyte solution. As can be seen, the GBPE-Ag-NPs exhibited one pair of redox peaks (E_pc = 0.75 V and E_pa = 0.16 V), which are attributed to the oxidation and reduction of Ag ions. Similar observations were reported by Blake and core-workers [66]. On the other hand, Lima and co-workers also observed an oxidation peak in the same region which (0.24 V) which was associated with the conversion of Ag⁺/Ag^o and also indicated the

presence of Ag-NPs [67]. The voltammograms illustrated in Figure 8(b) reveal increased currents with increasing scan rates (10 - 100 mV/s) during the electrochemical analysis of Ag-NPs. The redox peak currents increased with increasing scan rates while a shift in the potential was observed to more positive values [29]. This is indicative of electro-active nanoparticles with potential applications in sensor development Additionally, the diffusion coefficient denoted as (D_e) of the GBPE-Ag-NPs/GCE electrode was determined to be 5.22 × 10⁻⁵ cm² s⁻¹ using the Randel-Sevcik equation (Equation (1.3)) while the surface concentration was determined to be 2.31 × 10⁻⁵ mol cm⁻² using the Brown-Anson method (Equation 1.4). The obtained D_e was larger compared to the previously reported in literature and therefore larger values of D_e indicates a faster motion of analyte through the solution whereas small values of D_e indicates slower motion. Khan and co-workers reported a D_e value of 7.6 × 10⁻⁶ cm²∙s⁻¹ using Ag-NP-MWCNT-GCE which was lower compared to this study [68].

I p = 0.4463 n F A C ( n F v D e A u − N P s R T ) 1 / 2 (1.3)

I p = n 2 F 2 Γ A u − N P s * A ν 4 R T (1.4)

In this study we investigated the potential detection effects of Ag-NPs synthesized using a mixture of banana peel and grapes. To confirm the effect of the nanoparticles, a comparative study of the Ag-NPs before and after the interaction withE. coli 0157:H7 was performed using UV-vis spectrophotometer. Figure 9(a), shows the UV-Vis spectra of silver nanoparticles before and after their interaction with E. coli 0157:H7 at different concentrations. As seen from Figure 9(a), silver nanoparticles in the absence of E. coli 0157:H7 show an SPR

band at 440 nm in water (black spectrum). In the presence of E. coli 0157:H7 in water, we observed a reduction in the peak intensity with a redshift taking place at the same time. The decrease in peak intensity is the evidence of the interaction between Ag-NPs and bacteria. The redshift was observed by the change in SPR to more extended wavelength regions from 440 nm to 450 nm [69] which is due to many factors such as capping agent replacement, surface plasmon coupling or aggregation or agglomeration between closely spaced nanoparticles [70]. Similar results were reported by Boken and co-workers who also observed changes in SPR bands from 395 nm before the interaction to 408 nm after the interaction with E. coli 0157:H7 when using robust silver nanoparticle [71]. Figure 9(b) show a calibration plot with a good linear relationship plotted using the SPR response versus the different volumes of E. coli 0157:H7 from 80 - 130 µL. The plot shows a linear equation of (y = 0.02547 - 0.454) with a correlation coefficient of (0.9997). Similar results were reported by Wang and co-workers who observe a linear relationship with correlation coefficient of (0.9938).

In this research study, the limit of detection was determined using (Equation 1.5). The limit of detection (LOD) is the lowest concentration that can be measured or detected, and it was found to be 1 × 10² CFU/mL [72] [73]. A study by Yaghubi and co-workers reported similar results using Au-NPs conjugated with antibodies [74]. Our results were much lower or better compared to the reported literature, for example, Haicao and co-workers reported LOD of 2 × 10⁸ CFU/mL using mercaptoethylamine modified Au-NPs for E. coli O157:H7 [75]. Table 1 shows detection limits obtained in studies where E. coli 0157:H7 bacteria were detected using Surface Plasmon Resonance (SPR) sensors. From the table, it is evident that this study illustrated data far better in comparison to the previously reported studies from the literature. Therefore, the developed sensor can detect E. coli 0157:H7 in water using green synthesised nanoparticles without

the support of antibodies. The developed sensors can achieve detection within 3 min, during which the bacterium has sufficient time to bind and for analysis to be achieved. The obtained detection limit in this study falls within the range 1 × 10⁻³ to 1 × 10⁷ CFU/mL of the currently accepted methods reported in the literature [76] [77].

L O D = 3.3 S M (1.5)

From the equation, 3.3 is the coefficient, S represents the value of the standard deviation of the blank samples, and M is the slope of the standard curve within the concentration range.

Various E. coli concentrations from 1 ×10¹ to 1 ×10⁶ CFU/mL were prepared through serial dilution method as described in the previous studies. The electrochemical detection of E. coli was performed immediately after 24 h incubation period. The electrochemically active GBPE-Ag-NPs were electrodeposited on GCE. The choice of this electrode was made due to the cost-effective nature of the electrode as well as its reliable and stable properties. Panhawar and co-workers studied three different electrodes namely (glassy carbon electrode, platinum electrode and screen-printed electrode) for the detection of E. coli, and they reported GCE as the better electrode for the detection of E. coli [82] . All the experiments were performed at optimum conditions at the potential window of −1.5 to 1.5 V, using a scan rate of 50 mV/s under continuous magnetic stirring as, and a sample of 100 μL was analyses from each bacterial concentration. The green synthesised nanoparticles analyses were used to detect targeted bacteria by measuring the electrochemical response of the nanoparticles before and after their interaction with the bacteria. The functionalised nanoparticles show enhanced current responses before their interaction with E. coli. Figure 10(a) shows GBPE-Ag-NPs|GCE detecting different concentrations of E. coli 0157:H7 in PBS solution (0.2 M pH 7.4) and the (black) represents GBPE-Ag-NPs|GCE before interaction with E. coli 0157:H7. Following the interaction, the movement of electrons on GCE is reduced due to the presence of other materials, indicating that the bacteria detection has occurred. Shin and co-workers reported similar

behaviour, in the detection of E. coli in river water using Prussian blue coated with titanium dioxide particles [83]. From the plots, a change of electrochemical signal was observed due to the interaction of nanoparticles with E. coli. The change in current response or electrochemical signal, anodic (I_pa) and cathodic (I_pc) peak current response before the interaction of GBPE-Ag-NPs|GCE with E. coli is compared with the current response after the interaction with E. coli 0157:H7. The change in the current response for the anodic peak (I_pa1) = 18.777 μA and cathodic peak currents (I_pc1) = 0.738 μA peaks before the interaction of GBPE-Ag-NPs|GCE with E. coli 0157:H7 compared with the current response after the interaction with anodic (I_pa1) = 6.851 μA and cathodic (I_pc1) = −8.235 μA indicate the presence of E. coli. Similar results were reported by Klabunde and co-workers when using highly reactive metal oxide nanoparticles to treat E. coli [84] . As the concentration of E. coli increases the current response also increases due to the interaction between E. coli and nanoparticles. Thus, the concentration of the E. coli has a linear relationship with the current. Figure 10(b) is the calibration curve (N = 3) of the GBPE-Ag-NPs|GCE with a good linear relationship between the current response and the logarithms of different concentrations of E. coli 0157:H7. The linearity that was accomplished from the graph was used to calculate linear regression equation which was (r² = 0.9997) and the slope was used to calculate the LOD. The slope of the calibration also gives information about the sensitivity of the sensor, the higher the slope of the calibration curve the higher the sensitivity of the sensor for that specific component [85]. Therefore, the slope was determined to be 0.61071 μA/CFU/mL. The sensitivity of GBPE|Ag-NPs|GCE was more improved compared to the reported sensors in literature, for example our sensitivity was higher than the recently reported study by Cimafonte and co-workers using a Impedimetric Immunosensor based on a screen printed electrode for detection of E. coli [86]. The sensing potential

of GBPE-Ag-NPs|GCE for our experimental results showed LOD of 3.5 × 10¹ CFU/ mL Table 2 shows the limit of detection of more relevant electrochemical studies reported in the literature for the detection of E. coli 0157:H7 compared to this study. From the table, it is evident that the results obtained in this study were better compared to previous reports. Therefore, the developed sensor can detect E. coli 0157:H7 in water using green synthesised nanoparticles without any further modifications. The electrochemical detection process takes 45 min, which is less than 1 hr for the bacterium to bind and for analysis to commence. Therefore, our research findings show a better approach for the detection of E. coli 0157:H7.