Journal of Nanoparticle Research

, Volume 11, Issue 6, pp 1453–1463

The mechanism of metal nanoparticle formation in plants: limits on accumulation


    • School of Engineering and Advanced TechnologyMassey University
  • A. T. Marshall
    • School of Engineering and Advanced TechnologyMassey University
Research Paper

DOI: 10.1007/s11051-008-9533-6

Cite this article as:
Haverkamp, R.G. & Marshall, A.T. J Nanopart Res (2009) 11: 1453. doi:10.1007/s11051-008-9533-6


Metal nanoparticles have many potential technological applications. Biological routes to the synthesis of these particles have been proposed including production by vascular plants, known as phytoextraction. While many studies have looked at metal uptake by plants, particularly with regard to phytoremediation and hyperaccumulation, few have distinguished between metal deposition and metal salt accumulation. This work describes the uptake of AgNO3, Na3Ag(S2O3)2, and Ag(NH3)2NO3 solutions by hydroponically grown Brassica juncea and the quantitative measurement of the conversion of these salts to silver metal nanoparticles. Using X-ray absorption near edge spectroscopy (XANES) to determine the metal speciation within the plants, combined with atomic absorption spectroscopy (AAS) for total Ag, the quantity of reduction of AgI to Ag0 is reported. Transmission electron microscopy (TEM) showed Ag particles of 2–35 nm. The factors controlling the amount of silver accumulated are revealed. It is found that there is a limit on the amount of metal nanoparticles that may be deposited, of about 0.35 wt.% Ag on a dry plant basis, and that higher levels of silver are obtained only by the concentration of metal salts within the plant, not by deposition of metal. The limit on metal nanoparticle accumulation, across a range of metals, is proposed to be controlled by the total reducing capacity of the plant for the reduction potential of the metal species and limited to reactions occurring at an electrochemical potential greater than 0 V (verses the standard hydrogen electrode).




Metal nanoparticles have many current and potential technological applications including as catalysts, in optical materials, in medical treatments, in sensors, and in energy storage and transmission (Mohanpuria et al. 2008). The function of these materials depends on their composition and structure.

A possible route to the synthesis of metal nanoparticles is by biological production including from vascular plants such as may occur in phytoextraction or phytomining (Brooks et al. 1998) and phytoremediation (Pilon-Smits 2005; Arthur et al. 2005). Metal nanoparticles in plants have been observed for gold (Anderson et al. 1998; Sharma et al. 2007; Gardea-Torresdey et al. 2002), silver (Harris and Bali 2008; Brown et al. 1962), copper (Manceau et al. 2008), and of an alloy of gold–silver–copper (Haverkamp et al. 2007). Similarly, gold and silver nanoparticles have been formed by other organisms (Mohanpuria et al. 2008). Plants offer a green route to nanoparticle production and may perhaps also enable the preparation of difficult to synthesize materials. The authors could find no studies conclusively demonstrating metal nanoparticle formation for many of the elements commonly studied for phytoremediation and phytoremediation including nickel, copper, cadmium, lead, chromium, and zinc.

While many studies have looked at metal uptake by plants, particularly with regard to phytoremediation and hyperaccumulation, and a few have distinguished between elemental metal deposition (M0) and metal salt accumulation (Mx+), quantification of the proportion of the salts converted to metal nanoparticles has only rarely been addressed (Marshall et al. 2007; Gardea-Torresdey et al. 2005a).

In order to produce metal nanoparticles in the tissue of plants the metal of interest must be available (soluble) in the growing medium, the metal must be transported across the root membrane and translocated in the plant. There is some debate as to whether nanoparticles are formed outside the plant and then transported through the root membrane and into the plant or whether nanoparticles are formed within the plant by the reduction of the metal salt. One proposal has been that nanoparticles may be formed on the roots and then be transported in the plant (Gardea-Torresdey et al. 2003; Sharma et al. 2007). This hypothesis is briefly tested in this study by using a suspension of silver nanoparticles. However, the prevailing view is that the ionic form must be transported across the root membrane, the salt must be translocated in the plant and the salt must then be reduced to the element (Gardea-Torresdey et al. 2005a).

The mechanism of transport across the root membrane depends on the species of plant involved. Some plants have complex processes to control ion transport into the roots while others, such as Brassica juncea studied here, may have less robust mechanisms to control the passage of ions (Wright and Diamond 1977). The details of the transport of ions into the root are not the subject of this work.

Once the metal ions have been transported into the roots, the question to be answered is what determines the amount of metal that is deposited as elemental nanoparticles and is there an upper limit on this.

Silver was chosen as a model compound because silver can form metal nanoparticles in plants (Brown et al. 1962), high levels of silver have been achieved in plants (Harris and Bali 2008), silver nanoparticles exhibit good catalytic properties (Bocquet and Michaelides 2006) and other useful properties (Nair and Laurencin 2007), silver has a high electrochemical reduction potential (i.e., it is easily reduced), and complexes of silver with lower (more cathodic) reduction potentials are available.

The total concentration of silver in the plants was determined by atomic absorption spectroscopy (AAS). The chemical form that this silver takes in the plant was determined by X-ray absorption spectroscopy (XAS), including both XANES and extended X-ray absorption fine structure (EXAFS). XAS is a powerful technique for determining the chemical nature of elements in materials. It is able to distinguish between different compounds formed by a particular element, including being able to easily differentiate between elemental metal and metal salts. It has sufficient sensitivity to be useful at the metal concentrations obtained in plants (Gardea-Torresdey et al. 2005b; Chen et al. 2006; Hannemann et al. 2006).

The work reported here, investigates the factors controlling the quantity of silver nanoparticles accumulated in plants and the relationship between the proportion of silver accumulated as metal nanoparticles and as ionic silver. From these findings it is possible to draw broader inferences.

In the work reported here, silver complexing agents are used to alter the ease with which the silver ion may be reduced and therefore to assist in the investigation of the reduction process. Three compounds were chosen: silver nitrate in which the silver is not complexed; Ag(S2O3)23− and Ag(NH3)2+ both of which are stable complexes in solution. This is in contrast to the use of complexing agents including EDTA, DTPA, cyanide, NTA, and a range of other organic acids, used on a number of metals, in order to achieve better solubility from the soil, particularly for toxic or undesirable heavy metals in phytoremediation (Meers et al. 2004; Evangelou et al. 2007; Schwab et al. 2008; Luo et al. 2005).

Experimental methods

Seeds of Brassica juncea were germinated in deionised water, and then transferred to 100 mL pots filled with glass wool. These were mounted in a custom built hydroponic system in which a standard Hoagland solution (pH = 5.5 and conductivity of 2,000 μS cm) was continuously circulated. This solution was replaced every 2–3 days. A high-pressure sodium lamp (600 W Lucagrow, General Electric) on a 16 h light/8 h dark cycle provided the necessary light for the plant growth. After 5 weeks of growth, the Hoagland solution was replaced with deionised water and the system flushed for 2 days. After this time, a silver solution (AgNO3, [Ag(NH3)2]NO3, or Na3[Ag(S2O3)2]) was circulated through the system. Randomly selected plants were harvested at selected time intervals. Multiple plants (1–4 but usually 2) were harvested for each experimental condition. The harvested plants were freeze dried and finely ground for AAS and XAS.

Samples for AAS were prepared by dissolving 0.1 g of dried plant in 2 mL of concentrated HNO3 and 2 mL of deionised water at 90 °C for 2 h. This solution was filtered through a glass frit and made up to 25 mL in a volumetric flask. Absorption was measured at 328.1 nm with the lamp operating at 4 mA with a GBC 903 atomic absorption spectrophotometer. AgNO3 in 0.1 M HNO3 was used to determine the standard curve. Total silver concentrations are reported here as a weight percentage on a dry plant basis.

For XAS dried plant samples were pressed into 5-mm thick aluminum holders and sealed with Kapton adhesive tape. Ag foil (40-μm thick) was used a reference compound in addition to reference compounds of AgNO3, Ag2O, [Ag(NH3)2]NO3, Na3[Ag(S2O3)2] diluted with boron nitride to give thickness of approximately 2 absorption lengths. Ag foil was also used as an internal energy calibration standard by measuring its absorption simultaneously with the sample by placing it between the second and a third ion-chamber. XAS was performed at the Photon Factory (Advanced Ring) on beamline NW-10A (Nomura et al. 2007). The ring was operated at 6.5 GeV and approximately 50 mA with the beam monochromated using a Si(311) crystal and focused using a Pt coated mirror. The spot size was approximately 0.5 × 1.1 mm (FWHM) at the sample. Ag K-edge (25.514 keV) XAS were recorded at room temperature in transmission mode with Argon filled ion-chambers. The XAS data were analyzed using the Athena and Artemis software package in the normal manner (Ravel and Newville 2005). Two measurements were taken of each plant sample and there was no evidence in the recorded spectra of beam damage.

For TEM the plant material was fixed with 3% glutaraldehyde, 2% formaldehyde, 0.1 M phosphate buffer at pH 7.2. There was no secondary fixing with osmium. This was followed by a buffer wash, acetone series dehydration and set in a Procure 812 epoxy resin. TEM sections were cut on a diamond knife, mounted on Cu grids and imaged without staining. A Philips CM10 TEM was used with an acceleration voltage of 60 kV. Images were recorded with a SIS Morada high-resolution camera.

Results and discussion

The three silver compounds, simple Ag+ (from silver nitrate) and the complexes Ag(NH3)2+, and Ag(S2O3)23−, were all taken up by the hydroponically grown Brassica juncea, although to different extents. In each case the uptake into the plant was approximately proportional to time, after an initiation period (Fig. 1). This indicates an approximately constant rate of transport of silver into the plants for each of the ion species.
Fig. 1

Silver concentrations determined by AAS in leaf and stem of Brassica juncea grown from silver containing solutions: 10 g/L Ag as AgNO3, 10 g/L Ag as Ag(NH3)2NO3, 10 g/L Ag as Na3Ag(S2O3)2

There is a clear difference in the amount of silver taken up by the plants between the three forms of silver in solution. With Ag+ and Ag(NH3)2+ 1.1 and 0.9% total silver, respectively, accumulate in the plant after 8 h, while with Ag(S2O3)23− the silver is slower to accumulate, reaching 0.3% after 8 h. The transport of ions into a plant depends on the charge of the ions with large anions normally less readily transported (Wright and Diamond 1977) and this factor may contribute to the lower concentration of total silver in the plant with the thiosulfate complex.

It is apparent from plant growth on silver nitrate solutions at different concentrations that the rate of accumulation of silver is approximately independent of solution concentration, i.e., the rate is zero order in concentration (Fig. 2). This suggests that the rate of transport through the root membrane or within the plant at these relatively high-solution concentrations has an absolute limit. This is to be expected at these high concentrations as the ion channels in the membrane may be operating at their maximum capacity. However, at lower concentrations where the transport mechanisms of the plant are not saturated, as is often the case in phytoremediation, the transport may be limited by the concentration difference between the plant and its environment.
Fig. 2

Silver concentrations determined by AAS in leaf and stem of Brassica juncea grown in silver nitrate solutions: 2 g/L Ag as AgNO3, 4.5 g/L Ag as AgNO3, 10 g/L Ag as AgNO3

Of primary interest in this work is the rate of formation of silver metal nanoparticles, not just the rate of total accumulation of silver. Previously it has been observed that only a portion of the metal present in plants is present as M0 (Marshall et al. 2007). In the present work (Gardea-Torresdey et al. 2005b), Ag K-edge XANES has been used to determine the silver species present in the plant samples studied. Clear differences between the spectra of the silver reference compounds are apparent (Fig. 3) enabling confidence in the fitting of the spectra from the plant material.
Fig. 3

XANES standards for silver. a AgNO3, b Ag(NH3)2NO3, c Ag2O, d Na3Ag(S2O3)2, e Ag foil

All the plants, after silver accumulation, exhibit Ag K-edge XANES spectra, which can be fitted using a linear combination of the silver reference compounds XANES spectra (Fig. 4).
Fig. 4

XANES spectra for plants after 8 h in solution of a Na3Ag(S2O3)2, b Ag(NH3)2NO3, and c AgNO3

For the plants subjected to silver nitrate, three forms of silver appear to be present: metal, nitrate, and a third component which may be an oxide. It may be that a different oxide is present to the standard used or a complex of silver with an organic component for which a standard was not run.

In the silver nitrate treated plants the proportion of silver as metal gradually reduces with time while the proportion as nitrate increases with time (Fig. 5). The ratio of silver metal to oxide, if oxide is really present, is approximately constant with time. The absolute amount of each of the three components increases with time (Fig. 6). Ag2O has been observed previously on the surface of colloidal silver (Chen et al. 2006), so in this case the Ag2O present may represent silver which was initially reduced (from Ag+) then subsequently oxidised.
Fig. 5

Fraction of species of silver in Brassica juncea grown from silver containing solutions obtained from XANES: a 10 g/L Ag as AgNO3, b 10 g/L Ag as Ag(NH3)2NO3, and c 10 g/L Ag as Na3Ag(S2O3)2
Fig. 6

Quantification of silver in Brassica juncea according to chemical species obtained from combined XANES and AAS analysis: a 10 g/L Ag as AgNO3, b 10 g/L Ag as Ag(NH3)2NO3, and c 10 g/L Ag as Na3Ag(S2O3)2

For the plants subjected to silver ammonium complex only two forms of silver are present: metal and ammonium complex. As with the nitrate, the proportion of silver as metal decreases with time while the proportion as the complex increases with time (Fig. 5). The absolute amount of each of these increases with time (Fig. 6).

For the plants subjected to thiosulfate again two forms of silver are present: metal and thiosulfate. Only a small proportion of the salt is reduced to the metal. The proportion of metal increases with time while the proportion of thiosulfate is approximately constant (Fig. 5). The absolute amount of each of these increases with time (Fig. 6).

EXAFS on the Ag+ and Ag(NH3)2+ material supports the XANES observations, with Ag and Ag2O, or a similar component, observed for the AgNO3 plant samples and just Ag0 for the Ag(NH3)2+ plant samples (Fig. 7). Ag(S2O3)23− absorbed plants had a too low a level of Ag for useful EXAFS. A coordination number of 3–5 for Ag0 in AgNO3 and Ag(NH3)2+ plant samples was determined from EXAFS. These coordination numbers are lower than for bulk silver (which has a coordination number of 12, although the EXAFS fit here gives 13.5) and are indicative of nanoparticles of silver metal, although it is not possible to give the particle size with precision from these measurements. The bond length determined by EXAFS for the Ag0 in the AgNO3 and Ag(NH3)2+ plant samples is 0.286 nm (less than the bulk Ag value of 0.289 nm) also pointing to nanoparticle formation.
Fig. 7

EXAFS for Ag metal standard and 8 h plant samples from 10 g/L Ag as AgNO3 and 10 g/L Ag as Ag(NH3)2NO3. a EXAFS raw data, b EXAFS fourier transforms. The peak at 2.86 Å corresponds to the Ag–Ag bond in Ag metal

Silver nanoparticles were observed in the TEM images (Fig. 8) with particle size ranges of 4 to 35 nm for AgNO3, 3 to 7 nm for Ag(NH3)2NO3, and 2 to 7 nm for Na3Ag(S2O3)2. The complexed forms of silver result in the formation of smaller metal particles. A similar relationship between complexed silver and smaller silver particles has been observed previously for silver nanoparticles formed in micelles by UV-photoactivation (Ghosh et al. 2003).
Fig. 8

TEM images of silver nanoparticles from plant samples with a 10 g/L Ag as AgNO3, b 10 g/L Ag as Ag(NH3)2NO3, and c 10 g/L Ag as Na3Ag(S2O3)2

The concentration and composition of silver in different parts of the plant have also been measured. For plants placed for 8 h in 10 g/L silver nitrate, measurements have been made on a section of the lower stem and the upper portion of the plant. The total silver was 6.2 and 0.14%, respectively. However, the Ag0 metal concentration was 0.30 and 0.04%, respectively. In the lower stem, even when the total silver concentration within the plant tissue is high, the deposited silver metal levels are not high. After longer times (12 and 24 h), the reduced silver in the AgNO3 grown plants reached a level of 0.35%.

An experiment where the plants were placed in a solution containing silver nanoparticles, not silver ions, resulted in no take up of silver, suggesting that if silver nanoparticles are formed outside of the plant, for example, on the roots (Manceau et al. 2008; Gardea-Torresdey et al. 2003), the nanoparticles do not readily enter the plant.


These observations can be explained by the following reasoning. First, the silver ions are transported into the roots. At these high concentrations this occurs at zero order with respect to solution concentration (Fig. 2). Once in the plant two processes take place: translocation of the ions (Hall and Williams 2003) and reduction of the ions to metal nanoparticles. The detailed mechanisms of the reduction processes in plants are complex and discussed in detail elsewhere (Schutzendubel and Polle 2002).

It is proposed here that plants have a limited capacity for reducing metal ions, and that this capacity depends, in part, on the reduction potential of the ion. For an easily reduced ion (i.e., an ion having a large positive electrochemical potential, such as Au+), the effective reducing capacity of the plant may be larger than for a more difficult to reduce ion because of the range of mechanisms or components in a plant that may be involved in reduction of metal ions. The total capacity of the plant to carry out reduction may depend, for the most part, on the immediately available reducing agents, but also to minor extent on the ability of the plant to manufacture additional reducing agents. This has two consequences for metal nanoparticle formation.

The first consequence is that if there is a limited capacity of the plant to carry out reduction (a limited store of reducing agent) then there will be a limit to the quantity of metal ions that may be reduced from solution and therefore a limit on the amount of metal nanoparticles that may be formed regardless of the concentration of the metal ions in the plant.

The second consequence is that the amount of metal nanoparticles that may be formed depends on the reduction potential of the ion or complex to be reduced. More easily reduced ions will be amenable to reduction by a wider range of reducing compounds and therefore the effective reducing capacity of the plant will be greater for this ion. Likewise, ions or complexes that are more difficult to reduce will have a smaller upper limit on the quantity of nanoparticles that may be formed.

The results reported here for the three silver complexes are consistent with this proposition. Ag+ is the most easily reduced of the three complexes (Table 1) and this generates the most silver nanoparticles at 0.30% dry weight of the plant, followed by the ammonium complex with 0.26%.
Table 1

Standard electrochemical reduction potentials, relative to the standard hydrogen electrode (SHE) (Aylward and Findlay 2002; Bard et al. 1985; Hubin et al. 2004), approximating reactions observed to occur in plants


Eθ (V) SHE


Reactions observed

    CrIV → CrIII


Bluskov et al. (2005)

    CrV → CrIV


Aldrich et al. (2003)

    CrV → CrIII


Aldrich et al. (2003)

    AuCl4 → Au0


Gardea-Torresdey et al. (2002)

    Ag+ → Ag0


This work, Harris and Bali (2008), Gardea-Torresdey et al. (2003)

    Na2SeO3 → Se–S (approx)


Montes-Bayon et al. (2002)

    Au(CN)2 → Au0


Marshall et al. (2007)

    AsIV → AsIII


Brooks (1992), Bard et al. (1985)

    Ag(NH3)2+→ Ag0


This work

    Cu2+→ Cu0


Haverkamp et al. (2007), Manceau et al. (2008)

    Cu2+→ Cu+


Polette et al. (2000)

    Ag(S2O3)23− → Ag0


This work

On the other hand, the silver thiosulfate has a more cathodic reduction potential and is therefore harder to reduce and this results in less silver metal being deposited, at around 0.03%. There may also be a concentration limiting effect involved with the thiosulfate as a result of the poor uptake of the Ag(S2O3)23− ion since the total levels of silver in the plant are low. This is unsurprising since large anions are not as readily transported through root membranes (Wright and Diamond 1977). Continuing the uptake for a longer period of time would determine the extent to which concentration contributes to this limitation.

In this study, the maximum level for reduced silver (metal nanoparticles) in Brassica juncea is found to be of the order of 0.35% Ag by dry weight. Concentrations of total Ag reported that are higher than this (Harris and Bali 2008) are likely to contain the same absolute quantity of nanoparticles, but higher levels of unreduced metal ions in solution. These higher totals can therefore only be obtained from high levels of ions in solution, after allowing time for an equilibrium concentration to be established in the plant. It is proposed here that this will not, however, increase the quantity of nanoparticles produced. It is suggested that the amount of this metal that is present as nanoparticles has an upper limit, which depends on the reducing capacity of the plant for the ease of reduction of the metal. A plot silver metal concentrations obtained in the plants for all three silver species used in this work illustrates this principle (Fig. 9).
Fig. 9

Reduced silver metal and total silver in Brassica juncea grown with AgNO3 for a variety of solution concentrations and growth times

Metallic forms have only been observed for Au, Ag, and Cu. It may be that the less easily reduced metals cannot form metal nanoparticles in plants so that this method of nanoparticle formation is limited to metals with reduction potentials of at least 0.04 V (the value of Ag(S2O3)23−) or higher. The metals that may be formed therefore include Cu, Ag, Au (these elements have been observed in plants in their metallic state) and as well as elements such as Ru (0.46 V), Rh (0.5 V), Pd (0.64 V), Ir (0.86 V), Pt (0.74 V) and where the electrode potential given in parentheses is the reduction to the metal from their common oxidation state as the chloride complex (Bard et al. 1985). These elements have not been detected in plants in their metallic state yet.

The limit on the reduction potential of plants can be investigated by observing, not only the formation of metal nanoparticles, but also other reduction reactions. Some of these have been tabulated (Table 1). These electrode potentials are standard potentials for the reactions and therefore only approximate the electrode potentials in the systems containing these ionic species. The concentration of the species involved in the reaction (which may include H+, for example) also have a significant affect on the actual potential. These numbers are therefore an approximate guide only, and in general do not correspond precisely to the conditions in the cited references.

Other reduction reactions have been investigated, and metal formation would have been detected by the methods used, but were not observed. These include Pb2+ → Pb0 at −0.13 V (Sharma et al. 2004), Ni2+ → Ni0 at −0.24 V (Kramer et al. 2000), Zn2+ → Zn0 at −0.24 V (Salt et al. 1999), TlI → Tl0 at −0.336 V (Scheckel et al. 2004), and Cd2+ → Cd0 at −0.40 V (Pickering et al. 1999), none of which was observed (De La Rosa et al. 2004).

The limit in plants of the effective reduction potential to reduce metal ions appears to be at around 0 V. This indicates that the nanometal particle production will be limited to the precious and semiprecious metals, i.e., those elements whose salts have a potential for reduction to metal above about 0 V (on the SHE scale).


By studying the uptake of AgNO3, Na3Ag(S2O3)2, and Ag(NH3)2NO3 solutions by hydroponically grown Brassica juncea using XAS and AAS it has been possible to determine the factors controlling the amount of silver accumulated. It is found that there is a limit on the amount of metal nanoparticles that may be deposited, of about 0.35 wt.% Ag by dry plant weight, and higher levels of silver are only obtained by the concentration of metal salts within the plant, not by deposition of metal. The limit on metal nanoparticle accumulation is proposed to be controlled by the total reducing capacity of the plant for the reduction potential of the particular metal species reacting. This is extended more generally by proposing that plants have a limited capacity for reducing oxidants such as metal ions, and that this capacity is depends, in part, on the electrochemical potential to be established. This necessitates that there is a limit on the amount of metal nanoparticles that may be formed regardless of the concentration of the metal ions in solution. From a review of the electrochemical potentials observed for reactions which do and do not take place it is proposed that metal nanoparticle formation is restricted to precious and semiprecious metals.


The authors wish to thank the Photon Factory Advanced Ring, Tsukuba, Japan, for beam time access under proposal 2008G207; Dr Masaharu Nomura and Garry Foran, Photon Factory, Tsukuba, for their assistance; NZ Synchrotron Group Limited for a travel grant; Prof Clive Davies for advice on hydroponics and Doug Hopcroft, Manawatu Microscopy Centre, for assistance with the TEM work.

Copyright information

© Springer Science+Business Media B.V. 2008