Benchtop electrochemical liquid-liquid-solid growth of nanostructured crystalline germanium

Article abstract of DOI:10.1021/ja205299w

An electrochemical liquid-liquid-solid (ec-LLS) process that produces large amounts of crystalline semiconductors with tunable nanostructured shapes without any physical or chemical templating agent is presented. Electrodeposition of Ge from GeO₂(aq) solutions followed by dissolution into a liquid Hg electrode, saturation of the liquid alloy, and precipitation can yield polycrystalline Ge(s) under ambient conditions. A unique advantage of ec-LLS is that it involves precipitation under electrochemical control, where the applied bias precisely defines the flux of Ge into the liquid electrode. Fidelity of the saturation and precipitation of Ge from liquid electrodes affords a variety of material morphologies, including dense films of oriented nanostructured filaments with large aspect ratios (>10³). Electrodeposition involving a liquid electrolyte, a liquid electrode, and a solid deposit under ambient conditions represents a conceptually unexplored direct wet-chemical route for the preparation of bulk quantities of crystalline group-IV semiconductors without the time- and energy-intensive processing steps required in traditional preparations of semiconductor materials.

Full text of DOI:10.1021/ja205299w

COMMUNICATION
pubs.acs.org/JACS
Benchtop Electrochemical LiquidꢀLiquidꢀSolid Growth of
Nanostructured Crystalline Germanium
Azhar I. Carim,† Sean M. Collins,† Justin M. Foley,‡ and Stephen Maldonado*,†,‡
†Department of Chemistry and ‡Program in Applied Physics, University of Michigan, 930 North University Avenue, Ann Arbor,
Michigan 48109-1055, United States
S Supporting Information
b
indicating that at this formal concentration, the Hg electrode is
more electroactive toward reaction 1 than H2 evolution. The
ABSTRACT: An electrochemical liquidꢀliquidꢀsolid (ec-
LLS) process that produces large amounts of crystalline
semiconductors with tunable nanostructured shapes with-
out any physical or chemical templating agent is presented.
Electrodeposition of Ge from GeO2(aq) solutions followed
by dissolution into a liquid Hg electrode, saturation of the
liquid alloy, and precipitation can yield polycrystalline Ge(s)
under ambient conditions. A unique advantage of ec-LLS is
that it involves precipitation under electrochemical control,
where the applied bias precisely defines the flux of Ge into
the liquid electrode. Fidelity of the saturation and precipitation
of Ge from liquid electrodes affords a variety of material
morphologies, including dense films of oriented nanostruc-
tured filaments with large aspect ratios (>103). Electrodeposi-
tion involving a liquid electrolyte, a liquid electrode, and a solid
deposit under ambient conditions represents a conceptually
unexplored direct wet-chemical route for the preparation of
bulk quantities of crystalline group-IV semiconductors without
the time- and energy-intensive processing steps required in
traditional preparations of semiconductor materials.
voltammetric response was nominally independent of scan rate,
and exchange of the borate buffer with KHCO3(aq) effected no
change in the current magnitudes or amount of electrodeposited
Ge. Two anodic voltammetric waves were observed at applied
potentials (E) of ꢀ0.5 and 0.0 V for the oxidation of electro-
deposited Ge, as previously described in polarographic reports
on electroanalytical Ge detection.5,6 Chronocoulometric experi-
ments conducted at potentials more negative than ꢀ1.2 V on the
time scale of hours showed stable and continuous passage of
charge for reaction 1 (Figure 1b). A dark, dull particulate film
completely coated the electrode interface after polarization at
E = ꢀ1.9 V for 20 min (Figure S1 in the Supporting In-
formation). By comparison, electrodeposition of Ge onto Cd
and Zn, two other group-IIB metal electrodes with similarly poor
electrocatalytic properties for H2 evolution, yielded significantly
less Ge (the faradaic efficiencies for Ge electrodeposition at Hg,
Cd, and Zn at E = ꢀ1.5 V were 89 ( 6, 24 ( 3, and 18 ( 11%,
respectively).
The as-collected black film electrodeposited at Hg pool
electrodes showed strong evidence of crystallinity. The first-
order Raman spectra obtained for Ge electrodeposited at several
applied biases exhibited a common, single pronounced signature
near 300 cmꢀ1, characteristic of crystalline Ge(s) (Figure 2a).7
The absence of a mode centered near 270 cmꢀ1 indicated no
substantial content of amorphous Ge collected at Hg,8 in contrast
to Raman spectra of Ge electrodeposited at Cd and Zn electrodes
(Figure S2). The Raman spectra also showed that the potential
used for electrodeposition influenced the observed crystallinity.
The phonon mode near 300 cmꢀ1 was red-shifted slightly and
broadened for Ge electrodeposited at increasingly more negative
potentials, indicating crystalline domain sizes approaching or
below the 24.3 nm Bohr exciton radius for Ge.9,10 Corresponding
FTIR and Raman spectra of these Ge samples bore no detectable
evidence of hydrogenation (Figure S3). Separate powder X-ray
diffractograms of the electrodeposited Ge (Figure 2b and Figure
S4) exhibited reflection patterns consistent with the premise that
the as-prepared electrodeposited Ge possessed only a diamond
cubic lattice. The relative intensities of the peaks in the diffracto-
grams also indicated no net Ge crystalline orientation at any of
the applied potentials. However, the line widths in the diffracto-
grams were sensitive to the electrodeposition potential, with a
perceptible broadening of each reflection for Ge electrodeposited
ew methods that do not innately require large energy inputs
N
to prepare functional optoelectronic materials from raw
sources are crucial for minimizing the environmental impact of
the manufacture and processing of group-IV semiconductor-based
technologies.1 No unabated electrodeposition process that produces
crystalline rather than amorphous group-IV semiconductors at solid
electrodes from oxidized, dissolved precursors has been identified to
date. More specifically, strategies for electrodeposition of large
amounts of Ge continuously from aqueous solutions in a single
step have been previously investigated but proved difficult to
develop because of poorly understood self-limiting/surface-fouling
processes at solid metal electrodes during Ge electrodeposition.2ꢀ4
Accordingly, we report data demonstrating the use of liquid
electrodes as platforms for electrodeposition of crystalline Ge
directly from aqueous electrolytes without cessation. All potentials
reported herein are with respect to a Ag/AgCl (saturated KCl)
reference electrode. Without dissolved GeO2, only a voltammetric
response for H2 evolution was observed at Hg pool electrodes in
alkaline solution (Figure 1a). In the presence of 50 mM GeO2, the
cathodic current increased markedly at potentials more negative
than ꢀ1.2 V at pH 8.5 as a result of reaction 1,
HGeOꢀ3 ðaqÞ þ 4eꢀ þ 2H2OðlÞ f GeðsÞ þ 5OHꢀðaqÞ
ð1Þ
Received: June 14, 2011
Published: August 10, 2011
r
2011 American Chemical Society
13292
dx.doi.org/10.1021/ja205299w J. Am. Chem. Soc. 2011, 133, 13292–13295
|

Journal of the American Chemical Society
COMMUNICATION
Figure 1. (a) Voltammetric responses for a Hg pool electrode immersed
in deaerated 10 mM Na2B4O7 (dotted line) without and (solid line) with
50 mM GeO2 at a scan rate of 0.01 V sꢀ1. The dashed line indicates the
standard potential for the HGeO3ꢀ/Ge redox couple at pH 8.5 and 298 K.
(b) Chronocoulometric response for a Hg pool electrode immersed in
10 mM Na2B4O7 and 50 mM GeO2 and biased at E = ꢀ1.5 V vs Ag/AgCl
for 8 h.
Figure 3. SEM micrographs of as-prepared Ge electrodeposited from
the same electrolyte as in Figure 1 under various constant applied biases
(E): (a) ꢀ1.2 V; (b) ꢀ1.5 V; (c) ꢀ1.8 V; (d) ꢀ2.7 V (the inset shows a
higher-magnification view of the underside of a filament film section);
(e) ꢀ2.7 V; (f) ꢀ2.1 V; (g) ꢀ2.4 V; (h) ꢀ2.7 V. Scale bars: (aꢀc)
0.5 μm; (d, e) 2 μm; (fꢀh) 0.5 μm.
indicating uniform morphology across the entire electrode sur-
face area. The length of the dense mat of filaments depended
on the total time of electrodeposition and was ∼10ꢀ5 m for long
(t g 2.5 h) electrodeposition times. Figure 3e presents a side view
showing the high density, uniformity, and length of the Ge filaments.
The average diameter of the as-collected Ge filaments in Figure 3f-h
was a function of the bias used for electrodeposition, with average
diameters of 53 ( 16, 45 ( 14, and 26 ( 9 nm, respectively.
Figure 3fꢀh also shows that the surfaces of the Ge filaments
possessed a nodular texture with diameters that fluctuated slightly
along their entire lengths.
The bright-field transmission electron microscopy (TEM)
image of a representative Ge filament prepared via electrodeposi-
tion at E = ꢀ2.7 V (Figure 4a) shows the local fluctuation in the
diameter of a filament. TEM micrographs also showed variations
in the local contrast of individual filaments, indicating multiple
and distinct crystalline domains. Phase-contrast TEM exhibited
lattice fringes from multiple nonaligned crystallites as well as
moirꢀe fringes indicating partially misoriented crystalline domains
with the same lattice parameter through the projected volume of
the filaments (Figure S5). The selected-area electron diffraction
(SAED) pattern collected for the same isolated filament (Figure 4b)
confirmed the polycrystalline nature of the Ge filament and indicated
a local structure consisting of multiple nonaligned domains of crystal-
line Ge. Figure 4cꢀe shows dark-field diffraction-contrast TEM
micrographs collected using three distinct (220)-diffracted elec-
tron beams. In each of these dark-field images, the bright sections
identify local regions within the filament at an orientation com-
mensurate with the particular (220)-diffracted beam selected.
The absence of any regular or consistent pattern in the contrast
of the filaments across Figure 4cꢀe shows the absence of long-
range ordering and preferred orientation of the Ge crystalline
domains along the filament length, indicating that each Ge filament is
a collection of randomly fused Ge crystallites. The electrical proper-
ties of the Ge filaments were assessed through single-filament
Figure 2. (a) First-order Raman spectra for Ge electrodeposited at a
constant potential on Hg pool electrodes immersed in the same electrolyte
as in Figure 1 and biased at several discrete potentials with respect to a Ag/
AgCl reference electrode. The spectra have been offset for clarity. (b) Powder
X-ray diffractograms of Ge electrodeposited on Hg pool electrodes under the
same conditions as in (a). Specific reflections expected for crystalline Ge with
a diamond cubic lattice structure are indicated at the top. (c) Observed mean
crystallite size (as determined through the line broadening in the powder
X-ray diffractograms) in as-prepared Ge as a function of the applied potential
used for electrodeposition at Hg pool electrodes. The dashed line indicates
the least negative potential that facilitated Ge filament formation.
at more negative potentials. The diffraction patterns in Figure 2b
indicated that the average crystalline domain size ranged between 53
and 8 nm for Ge electrodeposited between ꢀ1.2 and ꢀ2.7 V. The
dependence was nonmonotonic, with a large decrease in average
crystalline domain size between ꢀ1.2 and ꢀ1.9 V and a crystalline
domain size of ∼8 nm between ꢀ1.9 and ꢀ2.7 V (Figure 2c).
Representative scanning electron microscopy (SEM) images
of Ge electrodeposited at E = ꢀ1.2, ꢀ1.5, and ꢀ1.8 V are shown
in Figure 3aꢀc. At E = ꢀ1.2 V, the deposit exhibited an irregular,
particulate film morphology without any discernible arrange-
ment. At more negative applied potentials, the grains thinned and
appeared to be loosely interconnected. At E = ꢀ1.8 V, the as-
deposited Ge films showed a dense leaflike structure. The texture
of the Ge films changed markedly at potentials more negative
than ꢀ1.9 V. Figure 3d shows three-dimensional (3D) mats
of vertically aligned Ge filaments observed for films prepared at
E = ꢀ2.7 V. The Figure 3d inset highlights the individual filament
tips at the bottom of each film section. Bundles of filaments with
footprints as large as 2.5 ꢁ 103 μm2 were routinely observed,
13293
dx.doi.org/10.1021/ja205299w |J. Am. Chem. Soc. 2011, 133, 13292–13295

Journal of the American Chemical Society
COMMUNICATION
Figure 5. (a) Proposed description of the formation of crystalline Ge at
Hg electrodes by ec-LLS: (1) HGeO3ꢀ(aq) is reduced at the Hg
electrode surface; (2) Ge is dissolved into the Hg electrode; (3) the
resultant GeꢀHg amalgam reaches the saturation point (dictated by the
solubility of Ge in Hg and the flux of Ge into Hg), after which Ge
crystallites (denoted as cubes) precipitate within the Hg pool; (4) the
initial Ge crystallites serve as subsequent crystallization centers that
produce polycrystalline aggregates. (b) SEM micrograph of an angled
view of the top of a Ge filament film section (scale bar: 500 nm).
(c) Proposed scheme for Ge filament growth. An initial dense layer of
larger Ge particles provides nucleation points for unidirectional spheru-
litic filament growth. New extensions to the filament length occur within
the GeꢀHg amalgam at the bottom of the film.
Figure 4. (a) Bright-field TEM micrograph of an individual Ge filament
prepared via electrodeposition at E = ꢀ2.7 V in the same electrolyte as in
Figure 1. (b) SAED pattern for the filament in (a). (cꢀe) Dark-field
diffraction-contrast TEM images collected with three distinct electron
beams diffracted along the <220> direction. (f) Voltageꢀcurrent
response for a single Ge filament device. The inset shows an electron
micrograph of a Ge filament with four patterned metal contacts. Scale
bars: (a, cꢀe) 20 nm; (f) 500 nm.
devices prepared via electron-beam patterning.11 Contact resis-
tances between Ge and Pt were nominally 107 Ω. The measured
resistivities of the as-prepared Ge filaments were on the order of
0.4 Ω cm (Figure 4f),12 which is much lower than expected for
either amorphous Ge or undoped crystalline Ge.13,14 Trace incor-
poration of Hg (a p-type dopant in Ge) is a likely contributor to the
relatively low resistivity (Figure S6). Gated currentꢀpotential
measurements showed no field dependence of the measured
resistivity for gate voltages of (20 V, precluding the determination
of n- or p-type character. The measured resistivities were higher than
expected for degenerate doping of crystalline Ge, but grain boundary
scattering along the filament length could increase the resistance
through a filament.15 The present analyses do not rule out the
possibility of degenerate doping under these conditions but do show
that the as-prepared materials are of sufficient quality for electrical
applications.
The cumulative data describe a three-step process for Ge electro-
deposition at Hg involving reduction of HGeO3ꢀ(aq) to Ge, disso-
lution of Ge into the bulk Hg pool as an amalgam, and supersaturation
of the Ge amalgam resulting in the precipitation of crystalline Ge
(Figure 5a). The low solubility [(2 ( 0.5) ꢁ 10ꢀ7 M] and diffu-
sivity [(1.3 ( 0.1) ꢁ 10ꢀ5 cm2 sꢀ1] of Ge in Hg5 support the
contention that under high-current densities for Ge reduction,
rapid saturation of the near-surface region of the Hg pool with Ge
is achieved. In effect, Ge ec-LLS has stronger parallels specifically
to systems involving precipitation from supersaturated solutions,
such as polymer blends,16 dissolved inorganic minerals,17 and
supercooled molten metal alloys,18 rather than conventional electro-
deposition at solid electrodes. For example, electrodeposition of
ramified metal filaments (Figure S6)19ꢀ21 is characterized by the
presence of a strong electric field (101ꢀ103 V cmꢀ1) between
two closely spaced electrodes,21 astronginfluenceofthecounterions
on the electrolyte morphology and growth rate,22 the formation
of filament networks in a 2D plane,20,21 and the lengthening of a
metal filament via reduction of metal cations at the tip of an
existing filament.19,22 For electrodeposition of Ge filaments at
Hg, the applied electric field between the working and counter
electrodes was small (<1 V cmꢀ1); moreover, a change in the identity
of the buffer electrolyte did not produce any perceptible change
in the electrodeposition process or product, and the filaments
formed extended 3D mats that covered the entire surface of the
Hg pool with a thickness dictated by the electrodeposition time.
SEM (Figure 5b) indicated that the tops of the Ge filament films
had leaflike morphologies similar to that shown in Figure 3c and
that discrete thin filaments emanate from the underside of this
top layer. These aspects suggest that Ge filaments are produced
through the formation of an initial dense layer of crystalline Ge
followed first by rapid nucleation/precipitation of Ge in the
supersaturated amalgam and then unidirectional spherulitic fila-
ment growth (i.e., dense polycrystalline filament growth from a
plane rather than a central point) (Figure 5c). Spherulitic growth
of filaments can occur in supersaturated solutions when the rota-
tional and translational motion of nuclei during crystal formation are
independent and prevent crystallization along a specific lattice
direction.23 Spherulitic filament growth in saturated mixtures re-
quires a large gradient in the chemical potential of a precipitate
(typically effected through temperature) to drive nucleation and
directional solidification.24 The depth-dependent Ge concentration
within the liquid electrode, which is set by the rate of dissolution of
Ge at the electrodeꢀelectrolyte interface, produces a chemical
potential gradient for Ge in the amalgam. Although unidirectional
spherulitic growth is not common,24 the observed morphology and
randomized orientation of the crystalline domains within the
filaments are consistent with the predictions of spherulitic models.
13294
dx.doi.org/10.1021/ja205299w |J. Am. Chem. Soc. 2011, 133, 13292–13295

Journal of the American Chemical Society
COMMUNICATION
The presented ec-LLS process for Ge preparation has several
salient features. First, this electrodeposition scheme is a single-
step benchtop process that is inherently less time- and energy-
intensive than the standard routes for producing bulk quantities
of crystalline Ge. The conventional industrial method for pre-
paring crystalline Ge involves a multistep thermal reduction of
GeO2 (energy input .500 kJ molꢀ1).25ꢀ28 The energy input
required to drive reaction 1 is only ∼270 kJ molꢀ1 at an
overpotential of 0.6 V and a faradaic efficiency of 85%. Coupling
Ge electrodeposition with a useful anodic half-reaction (e.g.,
anodic Cl2 evolution) could yield an electrolysis process for
producing valuable chemical/material products at both electro-
des. Second, ec-LLS represents a new and controllable experi-
mental design for studying polycrystalline nucleation from
saturated mixtures. The ability to regulate precisely the condi-
tions governing the flux of the soluble species into the liquid
solvent (i.e., Hg) affords control and tunability that are not
readily achievable in saturation/precipitation systems of general
interest such as polymer blend crystallization,16 gelation,29 and
mineral formation.17 ec-LLS could prove useful for testing and
validating evolving models of crystal nucleation.30,31 Third, ec-
LLS does not utilize any templating agent to produce various
nanostructured morphologies. The filaments shown herein were
free of residual organics and did not have to be removed from a
porous template prior to electrical device incorporation, aspects
that are atypical of wet-chemical syntheses of nanostructured
semiconductors.32,33 Furthermore, the conditions used for
supersaturation/precipitation allow a myriad of distinct crys-
talline morphologies.17 Fourth, ec-LLS is a general process.
Although the data shown here are results specifically for Hg as
the electrode material, we have performed analogous Ge
electrodeposition experiments with liquid Ga electrodes and
obtained similarly high levels of crystallinity. Differences in
the physicochemical (e.g., surface tension, density) and elec-
trochemical (i.e., electrocatalysis) properties of distinct liquid
metal electrodes could substantially affect the morphology
of the electrodeposited material and are currently under
investigation. Initial results with Ga pool electrodes showed
alternative morphologies of high-aspect-ratio Ge filaments.
Other metals with low melting temperatures could also be
envisioned as possible liquid electrodes under moderate (T <
300 °C) conditions. To date, such metal electrodes have
not been ardently investigated specifically for semiconduc-
tor electrodeposition. This report provides new impetus for
such work.
’ ACKNOWLEDGMENT
The authors gratefully acknowledge generous startup funds
from the University of Michigan. The FEI Nova Nanolab Dual-
Beam FIB-SEM, JEOL 3011 TEM, and JEOL 2010F (S)TEM
instruments used in this work are maintained by the University of
Michigan Electron Microbeam Analysis Laboratory through NSF
support (DMR-0320740, DMR-0315633, and DMR-9871177,
respectively). J.M.F. acknowledges a Rackham Merit Fellowship
from the University of Michigan.
’ REFERENCES
(1) Williams, E. D.; Ayres, R. U.; Heller, M. Environ. Sci. Technol.
2002, 36, 5504.
(2) Liang, X.; Zhang, Q.; Lay, M. D.; Stickney, J. L. J. Am. Chem. Soc.
2011, 133, 8199.
(3) Hall, J. I.; Koenig, A. E. Trans. Electrochem. Soc. 1934, 65, 215.
(4) Fink, C. G.; Dokras, V. M. J. Electrochem. Soc. 1949, 95, 80.
(5) Karpiꢀnski, Z. J.; Kublik, Z. J. Electroanal. Chem. 1977, 81, 53.
(6) Karpiꢀnski, Z. J.; Pozosak, A.; Kublik, Z. Anal. Chim. Acta 1980,
120, 55.
(7) Parker, J. H.; Feldman, D. W.; Ashkin, M. Phys. Rev. 1967, 155, 712.
(8) Wihl, M.; Cardona, M.; Tauc, J. J. Non-Cryst. Solids 1972, 8ꢀ10, 172.
(9) Fuji, M.; Hayashi, S.; Yamamoto, K. Jpn. J. Appl. Phys. 1991, 30.
(10) Maeda, Y.; Tsukamoto, N.; Yazawa, Y.; Kanemitsu, Y.; Masumoto,
Y. Appl. Phys. Lett. 1991, 59, 3168.
(11) Hanrath,T.;Korgel,B.A.Proc. Inst. Mech. Eng., Part N 2004,218,25.
(12) Walton, A. S.; Allen, C. S.; Critchley, K.; Gꢀorzny, M. y.;
McKendry, J. E.; Brydson, R. M. D.; Hickey, B. J.; Evans, S. D.
Nanotechnology 2007, 18, No. 065204.
(13) Walley, P. A. Thin Solid Films 1968, 2, 327.
(14) Hung, C. S.; Gliessman, J. R. Phys. Rev. 1954, 96, 1226.
(15) Baccarani, G.; Riccꢁo, B.; Spadini, G. J. Appl. Phys. 1978, 49, 5565.
(16) Magill, J. H. J. Mater. Sci. 2001, 36, 3143.
(17) Beck, R.; Andreassen, J.-P. Cryst. Growth Des. 2010, 10, 2934.
(18) Li, D.; Herlach, D. M. Phys. Rev. Lett. 1996, 77, 1801.
(19) Chazalviel, J.-N. Phys. Rev. A 1990, 42, 7355.
(20) Fleury, V.; Kaufman, J. H.; Hibbert, D. B. Nature 1994,
367, 435.
(21) Fleury, V.; Rosso, M.; Chazalviel, J.-N.; Sapoval, B. Phys. Rev. A
1991, 44, 6693.
(22) Bazant, M. Z. Phys. Rev. E 1995, 52, 1903.
(23) Grꢀanꢀasy, L.; Pusztai, T.; Tegze, G.; Warren, J. A.; Douglas, J. F.
Phys. Rev. E 2005, 72, No. 011605.
(24) Lovinger, A. J. J. Appl. Phys. 1978, 49, 5003.
(25) Adams, J. H. In ASM Handbook, 10th ed.; ASM International:
Materials Park, OH, 1990; Vol. 2, p 733.
(26) Andon, R. J. L.; Mills, K. C. J. Chem. Thermodyn. 1971, 3, 583.
(27) Sommelet, P.; Orr, R. L. J. Chem. Eng. Data 1966, 11, 64.
(28) As noted in ref 25, preparation of polycrystalline Ge from GeO2
requires thermal reduction of GeO2 by H2 at 760 °C followed by
subsequent zone refining of the resultant Ge at 1000 °C. A lower bound
for the energy input required by this process was calculated by
considering the energy required to heat GeO2 to 760 °C and then
further heat the Ge to 1000 °C on the basis of the heat capacities
presented in refs 25 and 26.
(29) Kumar, S. K.; Douglas, J. F. Phys. Rev. Lett. 2001, 87, No.
188301.
(30) Grꢀanꢀasy, L.; Pusztai, T.; B€orzs€onyi, T.; Warren, J. A.; Douglas,
J. F. Nat. Mater. 2004, 3, 645.
(31) Kawasaki, T.; Tanaka, H. Proc. Natl. Acad. Sci. U.S.A. 2010,
’ ASSOCIATED CONTENT
S
Supporting Information. Experimental methods; electro-
b
chemical data characterizing H2 evolution and Ge electrodeposition
at Hg, Cd, and Zn electrodes; Raman spectra of electrodeposits
obtained at Cd and Zn; FTIR spectra, powder X-ray diffractograms,
and high-resolution TEM images of Ge electrodeposited at Hg; a
scheme depicting the mechanism of ramified metal filament elec-
trodeposition. This material is available free of charge via the
Internet at http://pubs.acs.org.
107, 14036.
(32) Trindade, T.; O’Brien, P.; Pickett, N. L. Chem. Mater. 2001,
13, 3843.
’ AUTHOR INFORMATION
(33) Trentler, T. J.; Hickman, K. M.; Goel, S. C.; Viano, A. M.;
Gibbons, P. C.; Buhro, W. E. Science 1995, 270, 1791.
Corresponding Author
smald@umich.edu
13295
dx.doi.org/10.1021/ja205299w |J. Am. Chem. Soc. 2011, 133, 13292–13295