Hyperstoichiometric interaction between silver and mercury at the nanoscale.
ABSTRACT Breaking through the stoichiometry barrier: as the diameter of silver particles is decreased below a critical size of 32 nm, the molar ratio of aqueous Hg(II) to Ag(0) drastically increases beyond the conventional Hg/Ag ratio of 0.5:1, leading to hyperstoichiometry with a maximum ratio of 1.125:1. Therein, around 99% of the initial silver is retained to rapidly form a solid amalgam with reduced mercury.
Hyperstoichiometric Interaction Between Silver and Mercury at the
Kseniia V. Katok, Raymond L. D. Whitby,* Takahiro Fukuda, Toru Maekawa, Igor Bezverkhyy,
Sergey V. Mikhalovsky, and Andrew B. Cundy
Extraction of heavy metals (e.g. silver, gold) through the
formation of amalgams with mercury has been utilized in
jewellery productionand miningfor over 2500 years. This
has been contemporarily applied in a nanoformulation for the
opposite process with the abstraction of mercury from waste
sources using zero-valent nanoparticles of noble metals.The
interaction of metal nanoparticles with other species has
shown great promise for a number of applicationsand in
particular silver nanoparticles (AgNPs) have been used as an
anti-microbial agent in textiles and compositesand also (less
frequently) for the destruction of pesticides or the removal of
mercury from industrial effluents and other waters.With the
formation of silver through a silicon hydride reduction, we
expect that as the size of the particle is reduced to the
nanoscale its interaction with aqueous mercury will be
dominated by surface forces.
It is well known that at the bulk scale, aqueous mercu-
ry(II) interacts with silver metal(0) with a stoichiometric ratio
of 1:2 [Eq. (1)], resulting in zero-valent mercury.
ðaqÞþ 2AgðsÞ! HgðsÞþ 2Agþ
Herein we report that as the diameter of AgNPs is
reduced below 32 nm, mercury(II) is reduced from water onto
AgNPs with the mercury-to-silver ratio reaching 1.125:1 for
11 nm AgNPs. Moreover, silver is not oxidized into solution,
rather, Ag–Hg solid amalgams are rapidly formed, immobi-
lized on an inert silica substrate. This effect promises new
insights to nanoscale chemistry and significant advancements
in applications for wastewater purification, enhanced chem-
ical catalysis, and toxicity of nanoscale systems.
Current approaches to the generation of AgNPs use
chemical reducing agents and stabilizers, which result in
residual groups on the surface of AgNPs, for example,
carboxylic or citrates.Such groups promote electrostatic–
ionic attractions between the nanoparticle surface and heavy
metal ions.Metal particles have high surface energy and the
reduction of AgNP size incurs a further increase in their
surface energy causing their agglomeration and thus limiting
the available surface area for sorption. To control their size,
we generated AgNPs discretely separated on the surface of
a modified silica substrate containing silicon hydride groups?
SiH. Silicon hydride groups anchored to the surface of silica
particles possess weak reducing properties,which are
sufficient for generating “chemically pure” zero-valent
silverby the reduction of silver cations according to
Equation (2) (see Supporting Information, Figure S2).
?SiH þ Agþþ 2H2O ! ?SiOH þ Ag0þ H3Oþþ1=2H2
By adjusting the surface density of silicon hydride groups
and the incubation time with a solution of silver nitrate, it was
possible to control the size of the AgNPs formed on the silica
surface (Figure 1a–c). Although the primary interaction
between the silicon hydride groups and silver cations yields
zero-valent Ag atoms, the final metal deposition occurs in
a form of much larger nanoparticles, thus we postulate that
the mechanism of AgNP formation incorporates nucleation,
growth, Ostwald ripening (also called coarsening), and
Their separation from neighboring AgNPs on the silica
substrate ensures that a greater proportion of the surface of
the silver is available for interaction with its local environ-
ment, as compared with an agglomeration of AgNPs.
Extensive washing of silver-modified silica resulted in no
leaching of the AgNPs, confirming that they are stably
immobilized on the substrate. Moreover, according to FTIR
analysis all silicon hydride groups have reacted and oxidized
to silanol groups (Figure S2).
The silicon hydride groups have a significant advantage
over use of chemical agents in that the final silver is free of
[*] Dr. K. V. Katok, Dr. R. L. D. Whitby, Prof. S. V. Mikhalovsky,
Prof. A. B. Cundy
Nanoscience & Nanotechnology Group, Faculty of Science and
Engineering, University of Brighton
Lewes Road, Brighton, BN2 4GJ (UK)
Prof. S. V. Mikhalovsky
53, Kabanbay Batyr Avenue, Astana (Kazakhstan)
Dr. T. Fukuda, Prof. T. Maekawa
Bio-Nano Electronics Research Centre, Toyo University
2100 Kujirai, Kawagoe Saitama 350-8585 (Japan)
Dr. I. Bezverkhyy
Laboratoire Interdisciplinaire Carnot de Bourgogne UMR
5209 CNRS-Universite de Bourgogne
9 av A. Savary, BP 47870, 21078 Dijon Cedex (France)
[**] This research has been supported by the FP7-Marie Curie Program
2009-255635-HYREM and Leverhulme Trust Fellowship. We thank
the RCUK Academic Fellowship programme (R.W.) and Dr. Chris
Dadswell (University of Sussex), Dr. Martin Smith (University of
Brighton), and Prof. Yury Gogotsi (Drexel University) for useful
Supporting information for this article is available on the WWW
Re-use of this article is permitted in accordance with the Terms and
Conditions set out at http://angewandte.org/open.
? 2012 Wiley-VCH Verlag GmbH & Co. KGaA, WeinheimAngew. Chem. Int. Ed. 2012, 51, 2632–2635
any residual groups and lends
itself to the direct formation of
amalgams with mercury. Once
removed from solution. TEM-
dispersive X-ray) revealed that
mercury is only located on the
silica at the sites of AgNPs
(Figure 1d–f and S2) confirm-
ing the Ag–Hg interaction. In
the absence of AgNPs, no mer-
cury was detected on silica
across the pH range tested,
demonstrating that silanol or
account for the vast volume of
therms were obtained for all
materials (Figure 2a) and the
maximum adsorption mercury
loading (A) was calculated from
the linearized Langmuir equa-
tion [Eq. (3)],
where Amand klare the char-
acteristic Langmuir parameters
related to the maximum adsorption capacity and the intensity
of adsorption, respectively. The adsorption profile fits the
Langmuir equation with a high degree of accuracy (Fig-
ure 2b), indicating that the adsorption mechanism does not
change for AgNPs when varying their size from 11 to 45 nm.
When analyzing the ratio of mercury removed from solution
to the amount of silver present, it was found that the smaller
sized AgNPs on silica possessed a higher capacity for mercury
binding and a higher rate of mercury removal from solution
(Figure 2c), which corresponds to a previously unreported
and unusually high mercury-to-silver stoichiometric ratio
We confirmed the bulk-scale effect using a silver rod.
Examining the solution after the reaction using ICP-MS, we
found 4.3?10?3mmol of AgIin solution compared with 2.9?
10?4mmol of HgIIremoved from solution, which demon-
strates that the system behaves according to the conventional
understanding of a redox reaction between HgIIand Ag0
[Eq. (1)] with a ratio of 0.07:1.0 (Figure 2d under the
horizontal broken line). However, when the Ag particle size
is reduced below a critical size, ca. 32 nm (according to linear
extrapolation in the hyperstoichiometry zone, Figure 2d), the
quantity of HgIIremoved from solution is far greater than the
amount of AgIreleased into solution.
Figure 1. Modification of C-120 type silica (specific surface area
114 m2g?1) using triethoxysilane generates silicon hydride groups.
These are subsequently used to reduce a silver nitrate solution to
AgNPs, which are affixed to the top of silica. TEM images reveal the
distribution of near-spherical AgNPs, appearing as darker contrast
particles (white arrows) against the fused silica substrate, with particle
sizes averaging a) 11 nm, b) 31 nm, and c) 45 nm. d) EDX mapping
analysis reveals the corresponding location of silver and mercury in
the TEM image for the e) Ag (Lapeak) and f) Hg (Mapeak) after their
reaction and shows that the distribution of Hg correlates only to the
location of AgNPs.
Figure 2. a) HgIIadsorption isotherms on silica with AgNPs were performed in batch with 0.1 g of the
three silver-containing silica samples per 40 mL of Hg(NO3)2solution with concentrations ranging from
0.15 to 312 mgL?1in the pH range of 4 to 7. The time allowed for establishing equilibrium was 120 min
with constant shaking. The solution was separated by filtration and the residual mercury concentration
was determined using inductively coupled plasma mass spectrometry (ICP-MS). b) Adsorption isotherms
for the maximum adsorption of mercury onto different sized and quantities of AgNPs on silica were fitted
to a Langmuir adsorption equation with R2values above 0.99. c) The kinetics of HgIIuptake, starting with
a 312 mgL?1stock solution, was studied for three different AgNP sizes on silica over 30 min. d) Ag rods,
0.01 mm in diameter and 5 mm in length, exhibited removal of HgIIwith a stoichiometric ratio of HgIIto
Ag0below 0.5, but when using AgNPs this ratio increases from 0.29:1 up to the maximum of 1.125:1
when the Ag particle size is reduced from 45 nm to 11 nm. The change from conventional to hyper-
stoichiometry occurs around 32 nm.
Angew. Chem. Int. Ed. 2012, 51, 2632–2635? 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
In the system containing silica with 11 nm AgNPs, the
solution at equilibrium contained 7?10?4mmol of AgIions
released from the silica substrate after reaction with mercu-
ry(II), compared with the adsorption of 0.056 mmol of HgII
from solution. This effect, which we call hyperstoichiometry,
is solely dependent on the size of AgNPs, therefore the ratio
of HgIIcoming out of solution is compared with the initial
quantity of Ag0available that drives the reaction rather than
compared with the resulting AgIin solution, otherwise the
stoichiometry ratios would be far higher. We have eliminated
various parameters that might affect stoichiometry, namely
pH, residual silicon hydride andsilanol groups, light reduction
of hydrated silver, and contaminants within the materials and
solutions. The hyperstoichiometry effect has been confirmed
for both mercury nitrate and mercury acetate, the latter
exhibited a 1.7:1 ratio (Hg to Ag) for ca. 10 nm AgNPs, and
will be evaluated for other anion species. It has been found
that Ag+can adsorb to AgNPsand would therefore be
undetectable by ICP-MS. Therein, Ag+could be reduced by
the anion (see Supporting Information) to Ag0and partake in
further reduction of Hg2+to Hg0. It is apparent that the
hyperstoichiometric ratio is linked to the size of the AgNPs
and it is therefore surmised that the smaller AgNPs have
superior release and catalytic recycling of the Ag+released
into solution, possibly driven by the greater surface energy of
nanosized silver particles over its bulk scale counterparts.
The powder X-ray diffraction (XRD) profile of AgNP on
silica is consistent with that of silver metal possessing a face-
centred cubic lattice (Figure 3a, line 1). During its interaction
with mercury, no crystal phase transformation of silver was
observed (Figure 3a, lines 2 and3), but aprogressive decrease
of the diffraction intensity of the silver peaks was found with
an increasing amount of mercury adsorbed by the system.
Additional diffraction peaks emerged (Figure 3a, lines 3 and
4 asterisks), which correspond to the formation of the
Schachnerite amalgam (Ag1.1Hg0.9), thus demonstrating that
HgIIhas reduced to Hg0and Ag0is still present to facilitate
a direct interaction between the metals. At higher loadings of
mercury, no crystal phases were detected (Figure 3, line 5)
indicating the formation of a non-crystalline state. This can be
clearly seen under TEM observations where the average
particle size of AgNPs has increased on contact with mercury
(Figure 3c,e and S3–4), which accompanies the disappear-
ance of the crystal lattice fringes of silver (Figure 3d,f).
The final composite was analyzed by using X-ray photo-
electron spectroscopy (XPS) (Figure S5) and high-resolution
TEM, but the results revealed that the levels of mercury were
lower than calculated from the ICP-MS measurements, which
we assign to its evaporation under the application of high
vacuum required by these characterization methods. EDX
analysis within TEM (Figure 1e,f) indicated an elemental
composition ratio of Hg to Ag close to 1:1. Therefore, we
dissolved the metal fractions from the final composite sample
of 11 nm AgNPs–Hg amalgam in concentrated nitric acid and
the liquid was then analyzed by ICP-MS, which revealed
a ratio of Hg to Ag of 0.7:1. Herein, the slightly lower than
expected ratio arises as the level of dissolved silver
approached the lower detection limit of the ICP-MS, but
demonstrates that without the application of a vacuum, the
level of HgIIadsorbed into AgNPs through the hyperstoichi-
ometry effect is high.
At present it is difficult to suggest the exact structure of
the Ag–Hg composite in the AgNPs. Taking into account that
the atomic radii of Ag (160 pm) and Hg (150 pm) are similar,
it is possible to estimate the “kissing”, or the Newton number,
Figure 3. Silica samples containing 11 nm AgNPs were shaken with
40 mL of Hg(NO3)2solution of increasing concentration for 120 min at
pH 7. The solution was separated by filtration and the solid washed
with deionized water then dried. a) Powder XRD patterns were
recorded for the final Ag–Hg amalgams on silica for concentrations
0.15 (line 1), 1.56 (line 2), 15.6 (line 3), 39 (line 4) to 78 mgL?1
(line 5). b) TEM reveals 11 nm AgNPs on silica before interaction with
mercury and c) the crystal lattice structure revealed under high
magnification. d) After interaction with mercury, TEM images show an
increase in size of AgNPs due to the formation of a Ag–Hg amalgam,
which e) exhibits no crystallinity.
Table 1: Summary of analysis of mercury adsorption on different sized
(mmol of Ag/
(mmol of Hg/
ratio of HgII
[a] The size of the Ag rod was 0.01 by 5 mm. [b] This ratio was taken from
the quantity of HgIIreduced from solution with that of AgIoxidized into
? 2012 Wiley-VCH Verlag GmbH & Co. KGaA, WeinheimAngew. Chem. Int. Ed. 2012, 51, 2632–2635
of spheres that can simultaneously touch the central sphere of
the same radius. For a 2-D system the number is 6 and for a 3-
D system the number is 12.As AgNPs would have both 2-D
(adjacent to the surface) and 3-D dimensions, the maximum
number of Hg atoms around each Ag atom could be between
6and 12.In bulksystems such ahigh ratio between Hgand Ag
can be only achieved for liquid amalgams. The fact that the
Ag–Hg composites obtained using AgNPs are solid suggests
that certain properties of the system have changed signifi-
cantly at the nanoscale dimensions. The diffusion coefficients
(D) of Hg in its amalgams vary widely from 10?10cm2s?1to
10?13cm2s?1.Assuming that D is 10?10cm2s?1, it is possible
to estimate the diffusion path of a mercury atom within
aspherical solidAg particletobe ca. 5 nm within3 min, which
is the time chosen according to the experimental data on
adsorption kinetics presented in Figure 2c. This result qual-
itatively explains the fastest adsorption kinetics for the
smallest AgNPs and the reduction of the hyperstoichiometric
ratio with the increase of NP size.
The AgNPs generated herein are deemed to be free of
residual chemical groups, which enable surface effects to
predominate in their physicochemical properties. Further
experimentswill becarried out toconfirm theanion reduction
effect by use of mercury chloride, which would prevent the
recycling of Ag+through the precipitation of AgCl. Ulti-
mately, we show that at the nanoscale a hyperstoichiometry
effect between the aqueous mercury and AgNPs occurs.
Silica matrices were modified with triethoxysilane in the presence of
acetic acid under reflux (2 h, 908 8C) in order to graft silicon hydride
groups(?SiH) onto the silicasurface (C-120-H). Driedmodifiedsilica
was then stirred with AgNO3at room temperature, which generate
AgNPs on the silica surface. Samples were then dried for 24 h at
1508 8C. All batch sorption experiments were performed with silica
containing either 0.05, 0.10, or 0.44 mmol of silver, where typically
0.1 g of silver-containing silica was placed in a conical flask and 40 mL
of a pre-determined concentration of Hg(NO3)2was added. The
mixture was shaken for 120 min at room temperature to allow
mercury to adsorb on the Ag-loaded silica. After soaking, the
resultant solid was filtered and dried in an air environment. Solutions
were analysed using ICP-MS.
Received: September 23, 2011
Revised: November 23, 2011
Published online: February 3, 2012
redox chemistry · silver
Keywords: hyperstoichiometry · mercury · nanoparticles ·
 R. W. Hesse, Jewelrymaking through History, Greenwood, Con-
 L. Lacerda, W. Salomons, Mercury from Gold and Silver Mining:
A Chemical Time Bomb?, Springer, Berlin 1998.
 M. S. Bootharaju, T. Pradeep, J. Phys. Chem. C 2010, 114, 8328–
 a) C. T. Yavuz, J. T. Mayo, W. W. Yu, A. Prakash, J. C. Falkner, S.
Yean, L. L. Cong, H. J. Shipley, A. Kan, M. Tomson, D.
Natelson, V. L. Colvin, Science 2006, 314, 964–967; b) D. T.
Wasan, A. D. Nikolov, Nature 2003, 423, 156–159; c) M. A.
Shannon, P. W. Bohn, M. Elimelech, J. G. Georgiadis, B. J.
Marinas, A. M.Mayes,Nature2008,452, 301–310;d) J. K. Yuan,
X. G. Liu, O. Akbulut, J. Q. Hu, S. L. Suib, J. Kong, F. Stellacci,
Nat. Nanotechnol. 2008, 3, 332–336.
 a) V. Alt, T. Bechert, P. Steinrucke, M. Wagener, P. Seidel, E.
Dingeldein, E. Domann, R. Schnettler, Biomaterials 2004, 25,
4383–4391; b) H. J. Lee, S. H. Jeong, Text. Res. J. 2005, 75, 551–
556; c) X. Y. Xu, Q. B. Yang, Y. Z. Wang, H. J. Yu, X. S. Chen,
X. B. Jing, Eur. Polym. J. 2006, 42, 2081–2087.
 T. Pradeep, Anshup, Thin Solid Films 2009, 517, 6441–6478.
 G. Q. Luo, H. Yao, M. H. Xu, X. W. Cui, W. X. Chen, R. Gupta,
Z. H. Xu, Energy Fuels 2010, 24, 419–426.
 “?ber Silicone (I). Zur angewandten Chemie der Silicone
(Synthese)”: R. M?ller, Chem. Tech. 1950, 2, 7–13.
 V. A. Tertykh, K. V. Katok, V. V. Yanishpolskii, Russ. J. Phys.
Chem. A 2008, 82, 1438–1441.
 J. J. Reed-Mundell, D. V. Nadkarni, J. M. Kunz, C. W. Fry, J. L.
Fry, Chem. Mater. 1995, 7, 1655–1660.
 M. Harada, E. Katagiri, Langmuir 2010, 26, 17896–17905.
 J. Liu, R. H. Hurt, Environ. Sci. Technol. 2010, 44, 2169–2175.
 J. H. Conway, N. J. A. Sloane, Sphere Packings, Lattices, and
Groups, 2nd ed., Springer, New York, 1993.
 K. H. Lee, M. C. Shin, J. Y. Lee, J. Mater. Sci. 1986, 21, 2430–
 M. L. Malhotra, L. Reynolds, R. E. Barker, J. Chem. Phys. 1974,
Angew. Chem. Int. Ed. 2012, 51, 2632–2635? 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim