UV–Vis spectroscopy is a sensitive method for detecting colloidal silver, because AgNPs exhibit a characteristic absorption peak at about 400 nm, which is attributed to surface plasmon excitation. Metal NPs have free electrons, which give the surface plasmon resonance (SPR) absorption band due to the combined vibration of electrons of metal NPs in resonance with the light wave. Figure 1 shows the UV–Vis absorption spectra of AgNPs synthesized using sodium citrate (SC), tannic acid (TA) and a mixture of both sodium citrate and tannic acid (SC–TA) at room temperature and at 100 °C.
The absorption band maximum for Colloid SC and Colloid SC–TA was observed at λ = 430 nm and λ = 432 nm, respectively. For the same syntheses but carried out at 100 °C (Colloid SC* and Colloid SC–TA*), the maximum of the absorption band is shifted to shorter wavelengths and equals λ = 408 nm, which is characteristic of smaller NPs. No characteristic peak was observed in colloids synthesized using tannic acid (Colloid TA and Colloid TA*).
The colloidal state and the hydrodynamic size of AgNPs in the case of all samples were investigated using DLS (Fig. 2).
The size and size distribution of synthesized AgNPs were determined using STEM (Fig. 3). The particle size histogram was obtained by measuring at least 100 NPs.
AgNPs with different sizes and a diversity of shapes were obtained in syntheses using sodium citrate. In syntheses using tannic acid, mostly spherical particles in a large variety of sizes (5–140 nm) were obtained. AgNPs of homogenous shape and size were obtained in syntheses carried out using a mixture of citrate and tannic acid (Colloids SC–TA and SC–TA*). A summary of AgNP colloid characterization is shown in Table 2.
Table 2 The overall results obtained for the syntheses of AgNPs using sodium citrate (SC) tannic acid (TA) and a mixture of sodium citrate and tannic acid (SC–TA) at room temperature (Colloid SC, TA, SC–TA) and heated to 100 °C (Colloid SC*, TA*, SC–TA*)
The use of a single reagent for the synthesis of AgNPs (either sodium citrate or tannic acid) results in particles with a broad range of sizes and geometries. AgNPs synthesized with sodium citrate alone are inhomogeneous in both size and shape. The sole use of citrate (a weak reducing agent and a strong complexing stabilizer) does not allow the control of AgNP morphology. Similarly, AgNP syntheses using just tannic acid lead to polydisperse colloids. Even a change in reaction conditions, that is, the increase of the synthesis temperature to 100 °C, does not produce uniform AgNPs. The reduction of a silver salt using a mixture of sodium citrate and tannic acid at 100 °C gives particles of ~30 nm diameter with a narrow size distribution. The combined use of sodium citrate and tannic acid enables control over the nucleation, growth and stabilization processes, which leads to reproducible monodisperse AgNPs. The overall results indicate that the presence of both tannic acid and sodium citrate in the synthesis of AgNPs, as well as a reaction temperature of about 100 °C, promote the homogenous growth of particles.
The obtained data indicate that a sodium citrate and tannic acid mixture has better reducing and stabilizing properties in AgNP synthesis than either sodium citrate or tannic acid alone. This may suggest that the formation of a complex of citrate and tannic acid is involved in NP synthesis. To confirm this hypothesis, we performed voltammetric measurements.
Voltammetric measurements were taken to determine the potentials of reduction and oxidation and the changes of these potentials under the influence of substances in the solution. At the same time, the pH of all solutions was measured. In the first stage of electrochemical research, voltammetric measurements were carried out on a solution of sodium citrate (c = 10−3 M), tannic acid (c = 10−3 M) and a solution containing both sodium citrate and tannic acid (c = 10−3 M each). All the solutions contained sodium nitrate as a basic electrolyte (c = 10−3 M). The obtained voltammetric curves are presented in Fig. 4.
Tannic acid in the solution containing sodium citrate undergoes oxidation the most readily. Oxidation starts at a potential of 0.195 V. This value is 0.115 V smaller than in the case of the solution containing tannic acid alone. The reason for such a change in the potential might be the interaction between the particles of tannic acid and citrate anions. Probably, these are interactions of the type of hydrogen bonds between the hydroxyl groups of tannic acid particles and the oxygen atoms of carbonyl groups of citrates. The formation of stable citric acid (CA)–TA complexes was also confirmed by quantum calculations (Table 3 and Fig. 8).
Table 3 The interaction enthalpies in the CA–TA and CA–TAox adducts shown in Fig. 8
The next stage of the research was to carry out voltammetric measurements of solutions containing silvers ions. Measurements were carried out on the following aqueous solutions: silver ions, silver ions and sodium citrate, silver ions and tannic acid and silver ions and both sodium citrate and tannic acid. The concentration of each substance was c = 10−3 M. All the solutions contained sodium nitrate as a basic electrolyte (c = 10−3 M). As seen in the voltammetric curves presented in Fig. 5, the peak potential of the reduction of silver ions in a basic electrolyte is 0.244 V.
In the solution containing silver ions and sodium citrate, the peak potential is shifted to 0.134 V. The reduction of silver ions in this case is difficult, which is due to the complexation of silver ions by the citrate. The formation of silver complexes with citrates has already been described in the literature (Dadosh 2009). In the solution containing silver ions and tannic acid, the peak potential of the reduction of silver is0.173 V. In this case, the reduction of silver is also difficult. The probable cause is the complexing of the silver ions by tannic acid. Similar interactions of tannic acid with other metals have already been described (Zhang et al. 2004, 2016). In the case of the solution containing silver ions, sodium citrate and tannic acid, the peak potential of the reduction of silver is 0.252 V. Moreover, the peak current in this case is much higher than for other solutions. The value of the reduction indicates that the reduction in this solution proceeds as easily as in the solution containing only basic electrolyte and that it proceeds twice as fast, which is suggested by the value of the current. A comparison of the potential of the oxidation of tannic acid (0.195 V) and the potential of the reduction of the silver ions (Fig. 6) in solutions containing sodium citrate and tannic acid shows that the reduction of silver should proceed spontaneously at room temperature. This conclusion is in accordance with what was observed in the reactions.
The voltammetric measurements revealed that the reduction potential of the CA–TA mixture is lowered compared to the potentials of CA and TA separately. This indicates that the complex acts as a reducing agent in the synthesis of AgNPs. In the case of syntheses carried out at room temperature, the rate of reaction is slower compared to syntheses carried out at 100 °C; hence, the size of the obtained AgNPs is larger and more polydisperse (Table 2).
To investigate the possibility of the formation of stable complexes of CA with tannic acid (TA) in aqueous solution, we applied a theoretical approach involving computer simulations and the semi-empirical PM7 method. Considering how large and complex the TA and CA–TA structures are, the configurational space of each is vast and its exhaustive exploration would require extensive computations. In this work, an attempt to find the lowest energy geometries in solution was made by performing relatively short conformational searches, thus giving only a preliminary insight into the structural and energetic properties of these systems. A detailed description of the calculation method is given in Supplementary material.
First, the structure of tannic acid in water was studied. In Fig. 7, we present the lowest energy structure of TA (shown from three different sides with the approximate dimensions) found from simulations. This structure was used to construct the CA–TA adducts in a further study of their structures and the strength of interaction between TA and CA. However, it must be understood that this structure is just one of the many various conformers that can be present in the solutions.
Considering the steric hindrance in the tannic acid structure, one could expect that only external gallic units (those most distant from glucose) may be involved in the oxidation reaction. However, the structure presented in Fig. 7 shows that such an assumption may be incorrect. The π electrons of aromatic rings belonging to different gallic units interact, and the ‘branches’ are arranged into groups containing three and two parallel elements (Fig. 7b, c). Hence, the internal gallic units (directly bound to glucose) are exposed and can also be involved in the complexation reaction.
In the next step, a series of simulations was performed to evaluate the strength of interactions between TA and CA. Both TA and TA with one oxidized gallic unit (to quinone) that may be formed during AgNP synthesis were included in the analysis. The intermolecular interactions of CA with either internal or external gallic units of TA were evaluated by placing CA near the unit of interest and performing the simulation to find its energetically favoured orientation. The enthalpies of interaction shown in Table 3 were calculated for the most stable structures obtained from the simulations (Fig. 8). Our preliminary results predict that the intermolecular interactions between CA and TA and TAox are of similar magnitude, but in the CA–TA complex, it is several kcal/mol stronger than in CA–TAox. The results also suggest that the interactions in the structures ‘ex’ shown in Fig. 8b, d are weaker than in the structures ‘in’ (Fig. 8a, c).
The structures of possible CA–TA adducts with the lowest energies formed in the aqueous solution are shown in Fig. 8. The size of the CA–TA adducts is comparable with the tannic acid molecules (i.e. maximum 1.2 × 2.4 nm).
Considering the colloid composition, tannic acid, citrate as well as adducts of TA–CA can be present on the surface of NPs. However, taking into account the DLS results (hydrodynamic size of AgNPs containing the metallic core of NPs and the hydration coating on NPs), compounds of larger sizes are present on the surface of NPs (see Table 1). The sizes of CA (0.86 × 0.40 × 0.35 nm), simulated TA (1.22 × 2.36 × 1.02 nm) structures and CA–TA complexes (~2.4 nm) indicate that TA molecules and adducts of CA–TA are adsorbed on AgNPs over citrate ions. However, identifying which adducts are actually present on the surface of AgNPs following synthesis is not trivial and requires further investigation.
On the basis of the results of our voltammetric measurements and computer simulations, we propose a scheme of silver ion reduction by a sodium citrate–tannic acid complex (Fig. 9). In the first stage of the reaction, the CA–TA complex is formed. The adduct shows better reducing properties than either CA or TA alone. In the next stage, the CA–TA complex reduces silver ions to metallic Ag0, which is then stabilized with CA–TAGin, CA–TAGex, CA–TAox Gin and CA–TAoxGex present in the colloid solution.
The CA–TA complex is being formed before AgNP formation starts (the mixture of CA and TA is prepared a few minutes before the reaction in a separate vessel, see Synthesis of AgNPs). Hence, the rate of complex formation should not influence the final formation of AgNPs. Moreover, free CA and TA molecules not included in complexes have a negligible effect on the obtained AgNPs because of their weaker reducing properties (voltammetric measurements). The formation of CA–TA adducts as well as the oxidation of gallic units in tannic acid during the reduction of Ag+ may result in lowering the solubility of the complex and result in precipitate formation (which is not mentioned in the literature). The precipitate appears not only in the AgNP colloid but also in the CA–TA mixture (after about a week of storage). Precipitate formation results from the reduction reaction in the case of the AgNP colloid and from the partial oxidation of TA (which is involved in the complex) with the oxygen in the air in the case of the CA–TA mixture. The storage of TA does not result in precipitation; hence, precipitate formation must be closely associated with the CA–TA complex. To confirm the presence of partially oxidized TA in the complex and to identify the compounds adsorbed on the surface of AgNPs, FT-IR measurements were examined.
IR spectroscopy was performed on the evaporated purified AgNPs and purified sedimented precipitate after AgNP synthesis, and the results are compared with the spectra of sodium citrate and tannic acid (Fig. 10). There are many IR spectra of tannic acid in the literature (Cowen and Al-Abadleh 2009; Falcao et al. 2014). Tannic acid is a substance extracted from plants; hence, the IR spectrum of this natural compound (band intensity and/or its position) may be slightly different depending on the origin of the tannic acid. To correctly identify the bands characteristic of tannic acid, we compared the measured spectrum of tannic acid with the computed spectrum obtained from the vibrational analysis performed using the PM7 and PM6 methods (see Fig. 1S and Table 1S in Supplementary material).
All IR spectra were obtained using the GATR accessory. This method allows us to obtain a more intense signal at lower wavelengths compared to standard transmission spectra obtained using KBr. The analysis of the IR spectrum of TA and CA is crucial in describing processes taking place during AgNP synthesis. The characteristic bands for TA were observed at 1710 (C=O stretching), 1612 (Car–Car stretching), 1317 (Car–OC stretching, Car–O–H in-plane bending, Car–Car stretching), 1197 cm−1 (O–CO and Car–CO stretching, Car–O–H in-plane bending) and for sodium citrate at 1582 cm−1 (C=O stretching asymmetric in COO–), 1387 cm−1 (C–OH stretching), 1274 cm−1 (C–H stretching deformation). The analysis of the characteristic regions marked in colour in Fig. 10 (blue and red areas) allows us to identify the compounds present on AgNPs as well as to determine the composition of the precipitate sedimented after the AgNP synthesis. A band characteristic of quinone units in the region 1750–1650 cm−1 (Car=O stretching) was observed in the spectrum of both precipitate (two maxima at 1700 and 1675 cm−1,Car = O stretching in ortho and para positions) and AgNPs (1690 cm−1). This indicates that tannic acid was oxidized to quinine (Socrates 2001) during the synthesis of metal NPs and confirmed the role of TA as a reducing agent in the reaction. The intensity of the band characteristic of O–H (at 1240 cm−1) is smaller in the precipitate compared to the band for TA which indicates that OH groups are oxidized to C=O in the precipitate, which is further proof of the oxidation of TA. This indicates that a precipitate contains oxidized forms of tannic acid (quinone units). The precipitate spectrum is significantly changed compared to TA and also contains the band characteristic of citrate at 1581 cm−1. This proves that the precipitate consists of sodium citrate–tannic acid complexes, the structures of which were simulated and are shown in Fig. 8c, d. The spectrum of AgNPs contains bands characteristic of both TA (1690 and 1196 cm−1) and citrate (1583 and 1397 cm−1). However, the bands’ position is shifted compared to the precipitate, which may be because the spectrum was collected from the surface of AgNPs. The intensity of the OH band versus C=O band changed proportionally as in the case of the precipitate. This indicates that the part of the OH groups was oxidized to quinone. However, the band from the quinones is shifted from 1700 cm−1 observed in the precipitate to 1690 cm−1 in the case of AgNPs. In the spectrum of AgNPs, the bands characteristic of citrate are also present at 1583 cm−1 (C=O stretching asymmetric in COO–) and a weak signal at 1397 cm−1 (C=O stretching symmetric in COO–) was also observed. The intensity of these bands is higher than those present in the precipitate. This indicates that on the surface of AgNPs, it is likely that sodium citrate–tannic acid complexes are present as well as citrate and tannic acid, while the precipitate contains mostly complexes of sodium citrate–tannic acid (see Fig. 8c, d).