Synthesis of silver nanoparticles using a modified Tollens’ method in conjunction with phytochemicals and assessment of their antimicrobial activity

Materials

Silver nitrate (Fischer Scientific, Guangzhou, China, 99.8% analytical reagent grade), polyvinylpyrrolidone (ACROS, Morris Plains, NJ, USA, MW = 58,000, PVP), nitric acid (Merck, Darmstadt, Germany, 69%), nutrient agar, nutrient broth, ammonium hydroxide aqueous solution (Tedia, Fairfield, OH, USA, 25% w/w, ACS grade), and sodium hydroxide pellets (Merck, Darmstadt, Germany, 99%) were used without any further purification. Deionized (DI) water was used for all experiments. Lastly, all glassware was periodically washed with diluted nitric acid (25%) and then dried in a hot-air oven overnight at 40 °C.

Antibacterial tests were then performed using three representative pathogenic species obtained from Princess Haya Biotechnology Center, Jordan University of Science and Technology (JUST) (Irbid, Jordan.): one Gram-positive strain of Staphylococcus aureus (ATCC 25923) and two Gram-negative strains of Escherichia coli (ATCC 12900) and Salmonella enterica (CIP 104220).

OLE and RLE preparation

Leaves of olive and rosemary were collected in June 2014 from the campus of the JUST in Irbid, Jordan. The collected leaves were transported to the laboratory and left to dry at room temperature (25–30 °C) for 10 days, following the procedure described by Dipankar & Murugan (2012). The leaves were then washed, and subsequently dried in a hot-air oven at 40 °C for 5 days to reduce the loss of the leaves’ constituents, as recommended by Pessoa et al. (2007). The dried leaves were then pulverized into a very fine powder by grinder and stored at 30 °C. The extract was prepared by adding 0.1 g plant powder to 100 ml heated water on a hot plate without stirring, and then leaving the mixture to boil for 10 min to obtain an extract of 0.1 wt% concentration. After that, extracts were cooled to 30 °C and the supernatant was slowly filtered with 0.45 μm polyamide membranes (Sartorius Biolab products; Sartorius AG, Göttingen, Germany) via a pump-filter apparatus to remove any remaining solid residues. The plant leaf extracts (PLEs) were kept at −4 °C to be used later.

AgNPs Synthesis via Tollens’ method

Silver nanoparticles were prepared using the well-known Tollens’ method. OLEs and RLEs, rather than saccharides, were used as reducing agents (Kvítek et al., 2008) to reduce the Tollens’ reagent Ag(NH₃)₂⁺ (aq) into AgNPs.

In addition to the previously prepared extracts, stock solutions of AgNO₃ (10⁻³ M), sodium hydroxide (1.25 × 10⁻² M), and PVP (8.4 × 10⁻⁵ M) were prepared. To prepare AgNP dispersions, a tin-foiled 250 ml Erlenmeyer flask container was used as a reaction vessel and placed on a stirrer plate while a syringe pump apparatus was fitted to feed the vessel with the reducing agent (PLEs). Initially, 12.5 ml AgNO₃ stock solution was added to the vessel at a stirring speed of 600–700 rpm. Then 38.5 μl of ammonia solution was added dropwise, followed by 12.5 ml of PVP stock solution. Finally, 25 ml of 1:1 PLE:NaOH mixture was slowly added to the vessel at a rate of 75 ml/h using an automatic dosing syringe through a plastic slip. The synthesized AgNP suspension was then ultrafiltrated using a 10 kDa nominal molecular weight cut-off membrane via ultrafiltration stirred cell (Model 8200; Millipore, Burlington, MA, USA, NMWCO:10,000) for concentration, purification, and pH adjustment. Any nitrate, PVP, Ag⁺, or PLE not bounded to the nanoparticles was therefore removed from the solution. The ultrafiltration process consumed up to 600 ml of DI water for each sample, in order to achieve a final purified concentrate of 50 ml in volume. The AgNP nanosuspension was kept in tin-foiled covered containers at 4 °C for a period of 6 months for further characterization and bactericidal experiments.

Instrumentation for characterization

Optical properties of the prepared AgNP nanosuspension were determined using a UV–Vis spectrophotometer (UV-2550; Shimadzu, Kyoto, Japan). The shape and size of the AgNPs produced were identified using an environmental scanning electron microscope (ESEM) Quanta 450 FEG-USA/EEU that operated at an electron gun power of 30 kV. In order to obtain the ESEM images, the synthesized AgNP dispersions were sonicated for 10 min in an ultrasonic bath (100 watts) that generates ultrasonic waves at a 35 kHz frequency (Ultrasonic LC20H; Elma, Singen, Germany). Then the sonicated dispersions were diluted (1:250,000) and re-sonicated for 30 min. After that, a drop of this dilute nanosuspension was placed on an ESEM pin stub specimen mount for measurement. Average silver core diameter was calculated by averaging 50 particles from the ESEM image. Average hydrodynamic diameter (h_d), size distribution, polydispersity index (PDI) and zeta potential (ζ) were determined by dynamic light scattering analysis (DLS) using a Malvern Zetasizer (Nano-ZS; Malvern Instruments Ltd., Worcestershire, UK). Measurements were determined three times for each sample, and uncertainties are given as standard deviations. Raw data were subsequently correlated to the mean hydrodynamic size by cumulants analysis (Z-average mean), according to ISO 22412:2017 (ISO, 2017). The dispersions were sonicated for 30 min using a 200 watt ultrasonic bath (Jeio Tech, Geumcheon-gu, Korea). No dilution was required before measurement. Fourier Transform Infrared spectra were obtained using a Spectrophotometer (IR_Affinity, Shimadzu, Kyoto, Japan) using potassium bromide pellets at 1:10 dilutions and in the ranges between 400 and 4,000 cm⁻¹. Fourier transform infrared spectroscopy (FTIR) measurements were carried out to identify representative functional groups of possible molecules on the surface of the nanoparticles.

AgNPs synthesis conditions

The effect of the reaction conditions such as PLE concentration, temperature, and pH was evaluated by varying one parameter while the others remained constant; the tested parameters and ranges are shown in Table 1. The UV–Vis spectra were determined in triplicate.

Evaluation of AgNPs yield

The OLE-AgNP and RLE-AgNP dispersions were analyzed for their remaining Ag⁺ concentration using a Shimadzu AA-6200 atomic absorption spectrophotometer (AAS). The nanosuspensions were centrifuged at a maximum relative centrifugal force of 4,185×g (6,000 rpm) for 15min (Z 200 A; HERMLE Labortechnik, Wehingen, Germany). The centrifuge separates the AgNPs at the bottom of the centrifuge tube, leaving the remaining Ag⁺ in the supernatant. After proper dilution, the obtained supernatants were analyzed. The difference in Ag⁺ concentration between the supernatant of the AgNP suspension and the standard AgNO₃ stock solution thus represents the amount of silver transformed to AgNPs (Singhal et al., 2011; Raffi et al., 2010).

Antibacterial susceptibility experiments

Three measures of bacterial growth and viability were used to evaluate the antimicrobial properties of the synthesized AgNPs: The broth microdilution method was used to determine the minimum inhibition concentration (MIC) and minimum bactericidal concentration (MBC), whereas the Kirby–Bauer method was used to assess the sensitivity of bacteria to nanoparticles.

A stock of the AgNP suspension was prepared at a concentration of 100 mg/l and, as recommended by Clinical and Laboratory Standards Institute (CLSI) guidelines (Barry et al., 1999), was tested by broth microdilution using a 96-well microplate. Twofold dilutions of the synthesized AgNPs (e.g., 100, 50, 25, 12.5, 6.25, and 3.125 μg/ml) were prepared in a Mueller–Hinton broth using a 96-well microtitration plate (microdilution), in triplicate. Each well was inoculated with 50 μl of the respective bacterial suspension (of Mueller–Hinton broth) and the concentration was matched against a 0.5 McFarland Standard to obtain a concentration of 1 × 10⁶ CFU/ml. After well-mixing, the inoculated 96-well microtitration plate was incubated at 37 °C for 24 h. Both the MIC and MBC were then detected during this incubation period. Mueller–Hinton broth was used as a negative control.

The Kirby–Bauer method (Bauer et al., 1966) was used to determine bacterial susceptibility to OLE-AgNPs and RLE-AgNPs. A triplicate of 100 μl of the respective bacterial suspension (of Mueller–Hinton broth) with a turbidity of 1 × 10⁵ CFU/ml were spread out on MH agar plates. The cultured MH agar on each plate was perforated with holes (wells) using a sterilized glass tube and labeled appropriately. Each well was filled with 50 μl of either OLE-AgNPs or RLE-AgNPs, and MH agar plates were incubated at 37 °C for 24 h. Finally, the zones of inhibition (ZI) were measured and reported. In all tests, silver nitrate was used as a positive control and a blank agar plate was incubated to detect any contamination that might have occurred during testing.

Optimum AgNPs synthesis

Optical spectra shown in Fig. 1 confirmed AgNP formation at all PLE concentrations by the detection of the peak of absorption between 410 and 420 nm. This range is characteristic of AgNP spectra, due to excitation of surface plasmon vibration. Resultant colors depend on the particle type, size, morphology, and solvent chemical composition (Bhui et al., 2009; Hao et al., 2004). The color of our synthesized dispersions ranged from brownish-yellow to deep browns, which is also an indicator of AgNP formation (Huang & Xu, 2010). The synthesis depended on the PLE concentration, since absorbance intensity increased by 65% (OLE) and 61% (RLE) by increasing PLE concentration from 20 to 250 mg/l. For both PLEs, the highest concentration tested provided the narrowest absorbance spectra, which indicates a more monodisperse nanosuspension (Bhui et al., 2009). Therefore, the 250 mg/l PLE concentration was selected to investigate the effects of temperature and pH on the AgNP synthesis.

As shown in Fig. 2, the UV–VIS absorbance intensity increased sharply as temperatures rose from 20 to 80 °C for 81% (OLE) and 78% (RLE). Such results illustrate the dependence of the synthesis process on temperature. Therefore, our results suggest that the simultaneous use of Tollens’ reagent and PLEs is effective only at high temperatures (i.e., up to 80 °C). As the temperature increases, the reaction rate increases, and more Ag⁺ is consumed for the formation of nuclei that grow into controlled nanoparticles (Yong & Beom, 2009). Therefore, the temperature of 80 °C was fixed for the next pH effect investigation.

Favorable pH basic conditions were shown for the reduction process using both OLE and RLE extracts (Fig. 3) with an increase in absorbance intensity from pH 2 to 11 with 27% (OLE) and 50% (RLE). This may be due to the PLE organic compounds, specifically those with carbonyl functional groups which can act as reducing agents only under basic conditions. Higher pH values increased the number of functional groups that were available to bind with silver ions and therefore increased the production of AgNPs. Under our testing conditions, pH 11 ± 0.2 produced the highest yields in terms of nanoparticles production as shown by the UV–VIS spectra. These findings are similar to Khalil et al. (2013), after using OLE as a reducing and stabilizing agent. However, it was decided that reducing the pH during the ultrafiltration process to 7 ± 0.3 is important to the selected three bacteria strains in the antibacterial susceptibility experiments (the optimum pH of the three species ranges between 7.4 and 7.6). Neutral pH is expected to reduce the AgNP agglomeration that was found to be connected to basicity (Veerasamy et al., 2011).

The time for AgNP formation was also investigated for the best obtained OLE-AgNP and RLE-AgNP synthesis conditions. As shown in Fig. 4, for both nanoparticles, the peak intensity and spectral stability demonstrated that the reduction of silver ions and the formation of stable AgNPs was approximately completed within 3 h. After reduction completion, the absorption peak was found at a wavelength of 410 nm, which implied a AgNP core size of 69 nm and represented the highest production of AgNPs (Huang & Xu, 2010). These formation times are very rapid in comparison with some previously reported plant-mediated synthesis routes with reaction times of 24 h (Chandran et al., 2006) and 4 h (Vilchis-Nestor et al., 2008).

AgNP characterization

At favorable conditions, the nanosuspensions were analyzed for AgNP concentration by AAS. The measured concentrations for OLE-AgNPs and RLE-AgNPs were approximately 50 and 45 mg/l, respectively. The percentages of Ag⁺ conversion into AgNP (Ag⁰) for OLE-AgNPs and RLE-AgNPs were 53% and 48%, respectively.

The morphology of synthesized AgNPs by PLE was examined using ESEM. The ESEM images of AgNPs showed spherical particles with the average core sizes of 45 ± 2 and 38 ± 3 nm for OLE-AgNPs and RLE-AgNPs, respectively (Fig. 5). DLS measurements were performed to determine the average hydrodynamic size, size distribution, and PDI of the AgNPs. The particle size distribution curves of three consecutive measurements for both OLE-AgNPs and RLE-AgNPs are shown in Fig. 6. Measurement uncertainties are given as standard deviations. The average sizes (Z-average size) were 70.27 nm for OLE-AgNPs and 64.16 nm for RLE-AgNPs. The PDI was found to be 0.295 and 0.424 for OLE-AgNPS and RLE-AgNPs, which are both less than 0.7 indicating monodispersed particles (Honary et al., 2013). It was expected that the hydrodynamic diameter would be larger than the core because it includes surface coating materials and a solvent layer attached to the surface of the particle as it moves under the influence of Brownian motion (Hess, Frisch & Klein, 1986).

The AgNPs have negative zeta potentials (ζ) of −43.15 ± 3.65 mV for OLE-AgNPS and −33.65 ± 2.88 mV for RLE-AgNPs that indicate a high colloidal stability of the particles (O’Brien et al., 1990; El Badawy et al., 2010; Edison & Sethuraman, 2012). This negative zeta potential reflects the surface charge of negatively charged AgNPs after being functionalized with the extract compounds (Hunter, 2013). From the zeta potential value, it was evident that the synthesized nanoparticles were found to be stable using leaf extracts and the stability was even enhanced by adding PVP as a secondary capping agent.

Fourier transform infrared spectroscopy

The dual action of leaf extracts as both reducing and capping agents was investigated using FTIR spectroscopy. In general, there are two regions in IR spectra: the functional group region (4,000–1,500 cm⁻¹) and the fingerprint region (1,500–400 cm⁻¹). The intense absorption peaks for both the extracts and the nanoparticles were used to evaluate the surface of the AgNPs produced. Because of the adsorption and modifications occurring at the surface, the functional group region is the most informative. Therefore, our analysis focused on that region for both the extracts and the synthesized AgNPs. The IR spectra of OLE and RLE show only minor differences (Table 2). RLE is mainly composed of isocarnosol (diterpene) dihydronormorphinone (alkaloid) and camphor (terpenoid) (Genena et al., 2008). The main chemical constituents of OLE are oleuropein, quercetin, rutin, and luteolin (all polyphenolic compounds) (Khalil, Ismail & El-Magdoub, 2012).

The spectra for both RLE-AgNPs and OLE-AgNPs (Fig. 7) were also similar to each other, as well as to the IR peaks of the plant extracts used for their synthesis. This is strong evidence that compounds present in the extracts not only participated in the reduction of silver ions, but also were adsorbed on the surface of the nanoparticles produced. The OLE-AgNPs show the presence of the following peaks 3,371, 3,304, 3,238, 3,221, 2,960, 2,875, 1,687, 1,598, 1,456, and 1,384 cm⁻¹ at the functional group region, and the following peaks 3,437, 3,257, 3,219, 2,960, 2,887, 1,727, 1,687, 1,598, and 1,456 cm⁻¹ were present in the IR spectrum of the RLE-AgNPs. The peaks at ~3,437, 3,371, and 3,257 were due to the −NH stretching of amine or −OH stretching of alcohols and phenols, or bending and stretching of hydrogen-bonded alcohols and phenols in the leaf extract. This small shift is an indication of adsorption on the surface, especially for the C=O (more than 10 cm⁻¹). Shanmugam et al. (2014) suggested that these bonds could be due to the stretching of −OH in proteins, enzymes, or polysaccharides present in the extract. In addition, the peak at 2,966 cm⁻¹ is due to C–H stretching and indicated the presence of alkanes, while the peaks at (2,930, 2,875) and (1,687), and (690) cm⁻¹ corresponded to C–H stretching, C=C bond aromatic, and C–H bending respectively, and implied the presence of aromatic compounds. The peak at 1,687 cm⁻¹ also could correspond to amide C=O stretching. The observed O–H and C=O modes for the OH and C=O groups might be attributed to oleuropein, apigenin-7-glucoside and/or luteolin-7-glucoside which, as suggested by Khalil, Ismail & El-Magdoub (2012), are flavonoid compounds present in the olive leaf. It can be concluded from the FTIR that the presence of organic functional groups, such as alkanes, aromatic compounds, and amide linkages of protein and amine, played a major role in the production and stability of AgNPs.

Antimicrobial effect

To explore the antibacterial activity of OLE- and RLE-AgNPs, both the broth microdilution method (as recommended by CLSI protocol) and the Kirby–Bauer method were used. The MIC and MBC of OLE- and RLE-AgNP values against selected bacteria are listed in Table 3.

Table 3 shows that the OLE- and RLE-AgNPs exhibited good bactericidal activity against the three tested bacterial strains after 24 h of incubation. The MIC results for OLE-AgNPs when used against Staphylococcus aureus and Escherichia coli bacteria were 9.38 μl/ml, with MBC of 12.5 μl/ml. The MIC and MBC against Salmonella were 18.75 and 25 μl/ml, respectively. On the other hand, the MIC results for RLE-AgNPs against Staphylococcus aureus and Escherichia coli bacteria were 4.69 μl/ml, with MBC of 6.25 μl/ml; against Salmonella the MIC was 18.75 μl/ml and MBC of 25 μl/ml.

In the Kirby–Bauer method, silver nitrate was chosen as a positive control because of its well-documented antibacterial effects (Liau et al., 1997). The ZI in the antibacterial susceptibility experiments showed that OLE- and RLE-AgNPs have an inhibitory effect toward both gram-negative (Escherichia coli and Salmonella) and gram-positive (Staphylococcus aureus) bacteria, as shown in Table 4. The results showed that the synthesized AgNPs provided inhibition comparable to the control solution of AgNO₃. The PLEs had a negligible inhibitory effect, most likely due to the low concentrations used (Ahmed et al., 2016a).

Silver (or what we now know as silver ions Ag⁺) was well-known for its antimicrobial properties even during ancient times (Sharma, Yngard & Lin, 2009). Zero-valent silver Ag⁰ (AgNPs) slowly releases silver ions via oxidation under aerobic conditions (Xiu, Ma & Alvarez, 2011). For this reason, the Kirby–Bauer method was selected, since its protocol is suitable for maintaining AgNP oxidation, and Ag⁺ is continuously released.

Table 5 compares our synthesis approach with some previously reported biological routes of AgNP synthesis using different natural plant extracts. As can be observed, most of the AgNPs were spherical in shape, in the range of 5–500 nm, with varied antibacterial potency. Huang et al. (2011), for example, studied the synthesis, formation mechanism, and antibacterial activity of biogenic AgNPs by Cacumen Platycladi Extract. They found that the MIC and MBC against Escherichia coli were 1.4 and 27 μl/ml, respectively, while the MIC against Staphylococcus aureus was 5.4 μl/ml. Our MBC results, however, were better than those reported by Huang et al. (2011), while their MIC results were better than ours. In addition, Tippayawat et al. (2016) also studied the antimicrobial activity of AgNP produced with Aloe vera plant extracts: the MIC of their AgNPs against gram-positive Staphylococcus epidermidis was 10 μl/ml, which is comparable with our MIC results. However, their AgNPs were fabricated at a higher temperature (100 °C for 6 h and 200 °C for 12 h).

These results offer insights into the potential of scaling up the production of AgNPs using leaf extracts from olive and rosemary plants for commercial implementation.

The antimicrobial mechanisms of AgNPs are still not completely understood. One mechanism proposed is that AgNPs are able to interact with the bacterial cell wall, alter its properties via decaying lipopolysaccharide molecules, and form “pits” that increase wall permeability (Sondi & Salopek-Sondi, 2004). In addition, antibacterial properties have been reported to be size-dependent, with higher antimicrobial performance at smaller sizes of nanoparticles (Ajitha et al., 2015; Martinez-Castanon et al., 2008).