XAS/EXAFS studies of Ge nanoparticles produced by reaction between Mg 2Ge and GeCl4

Article abstract of DOI:10.1016/j.jssc.2011.06.020

We present results of an XAS and EXAFS study of the synthesis of Ge nanoparticles formed by a metathesis reaction between Mg₂Ge and GeCl₄ in diglyme (diethylene glycol dimethyl ether). The progress of the formation reaction and the products formed at various stages in the processing was characterised by TEM and optical spectroscopy as well as in situ XAS/EXAFS studies using specially designed reaction cells.

Full text of DOI:10.1016/j.jssc.2011.06.020

Journal of Solid State Chemistry 184 (2011) 2345–2352
Contents lists available at ScienceDirect
Journal of Solid State Chemistry
journal homepage: www.elsevier.com/locate/jssc
XAS/EXAFS studies of Ge nanoparticles produced by reaction
between Mg2Ge and GeCl4
n
Andrew J. Pugsley a, Craig L. Bull b, Andrea Sella a, Gopinathan Sankar a, Paul F. McMillan a,
a Department of Chemistry and Materials Chemistry Centre, University College London, 20 Gordon Street, London WC1H 0AJ, UK
b SUPA, School of Physics and Astronomy and Centre for Science at Extreme Conditions, University of Edinburgh, Mayfield Road, EH9 3JZ, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
We present results of an XAS and EXAFS study of the synthesis of Ge nanoparticles formed by a
metathesis reaction between Mg2Ge and GeCl4 in diglyme (diethylene glycol dimethyl ether). The
progress of the formation reaction and the products formed at various stages in the processing was
characterised by TEM and optical spectroscopy as well as in situ XAS/EXAFS studies using specially
designed reaction cells.
Received 18 February 2011
Received in revised form
2 June 2011
Accepted 20 June 2011
Available online 28 June 2011
& 2011 Elsevier Inc. All rights reserved.
Keywords:
Nanoparticles
Metathesis reaction
X-ray absorption spectroscopy
1. Introduction
ether) has been described [35–40]. The nanoparticle formation
proceeds via metathesis/comproportionation reactions between
Semiconducting nanoparticles are well developed for applica-
tions ranging from biomedical and environmental sensors to
energy-producing devices. Most work to date has focused on II–VI
semiconductor systems such as CdS and CdSe that have bandgaps
and luminescent properties that vary systematically with nanopar-
ticle size [1–8]. One main reason for the rapid development of these
materials is the existence of convenient solution chemistry routes to
produce the luminescent nanoparticles via arrested precipitation
reactions, combined with the ready availability of soluble precursors
[9–13]. Following synthesis, the nanoparticles are ‘‘capped’’ with
organic groups in solution to terminate, stabilise and functionalise
the surface and facilitate their incorporation into devices. Less
progress has been made in developing routes to analogous nano-
particles based on well-known group 14 semiconductors such as Si
or Ge. In bulk form these materials provide indirect-gap semicon-
ductors that only exhibit significant optical absorption or lumines-
cent properties well above the bandgap energy. However,
nanocrystalline as well as amorphous varieties exhibit light absorp-
tion and photoelectric energy conversion throoughout the IR-visible
range [14–24]. Because of the compatibility of these materials with
existing semiconductor technology it is important to develop routes
to the nanoparticle synthesis [25–34].
soluble molecular SiCl4 or GeCl4 that provide a source of electro-
positive Si4þ or Ge4þ, and anionic Siꢀ, Ge4ꢀ species present
within solid state Zintl compounds such as NaSi or Mg2Ge
containing polyanionic units (e.g., Si44ꢀ tetrahedra) [41–43]. Once
in colloidal suspension the resulting Si or Ge nanoparticles are
transferred to organic solvents such as hexane to carry out
capping reactions [35–38]. This type of reaction is being devel-
oped as a general approach to provide various capped and core–
shell Si/Ge semiconductor-oxide nanoparticles [37,38,44]. It is
important to characterise the reaction conditions leading to
production of the nanoparticles and their chemical and physical
state. Here we used X-ray absorption spectroscopy (XAS) imple-
mented at the Ge K edge [45–50] to study the reaction between
Mg2Ge and GeCl4 in diglyme, that produces luminescent Ge
nanoparticles within a 2–10 nm size range [35–40]:
Mg2Ge (s)þGeCl4 (l)-2Ge (nano)þ2MgCl2 (s)
(1)
2. Experimental methods
2.1. Synthesis
A convenient route to produce Si/Ge nanoparticles suspended
in organic liquids such as diglyme (diethylene glycol dimethyl
All chemical manipulations were carried out in dry N2 using
glovebox and Schlenk line techniques (o1 ppm O2/H2O). Liquid
GeCl4 (Acros, 99.99%), was used as received. Crystalline Mg2Ge
was prepared by heating a stoichiometric mixture of Mg (Aldrich,
99þ%) and Ge (Aldrich, 99.999þ%) contained in Ta foil and sealed
n Corresponding author.
E-mail address: p.f.mcmillan@ucl.ac.uk (P.F. McMillan).
0022-4596/$ - see front matter & 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jssc.2011.06.020

2346
A.J. Pugsley et al. / Journal of Solid State Chemistry 184 (2011) 2345–2352
in a silica tube at 650 1C (3 days): the material was confirmed by
powder X-ray diffraction (Siemens D8; CuK radiation) and
a
Raman spectroscopy (Renishaw InVia : 514 nm laser excitation)
to be single phase (Fig. 1a).
Diglyme (diethylene glycol dimethyl ether) was obtained from
Aldrich (99.5%). As received it was determined to contain
ꢁ400 ppm H2O by Karl–Fischer titration. It was found to be
critically important to remove all traces of moisture to avoid side
reactions between GeCl4 and H2O. In our early experiments the
solvent was stirred for 12–48 h over Na/K alloys (Na: Aldrich,
99.95%; K: Fischer, 99.95%) followed by vacuum distillation that
was thought to provide sufficient drying. However, EXAFS experi-
ments and TEM results indicated GeO2 nanoparticles were pro-
duced along with the nano-Ge particles. Subsequent Karl–Fischer
titrations showed that even such ‘‘fully dried’’ solvents could
contain up to ꢁ250 ppm of H2O, that was either incompletely
removed initially or re-introduced during subsequent manipula-
tions. After further experimentation we found it was necessary to
maintain the H2O content below 30 ppm by refluxing the solvent
over Na for 7–14 days followed by distillation under N2 and
storage over a Na mirror. We also replaced the air-tight syringes
used in early experiments for liquid transfer into the reaction
chamber by cannula/septum techniques.
We followed the Ge nanoparticle synthesis procedures
reported earlier [35,36]. Mg2Ge (46 mg, 0.38 mmol) was placed
in a two-necked 500 ml round-bottom flask along with 100 ml of
dried diglyme. The mixture was stirred magnetically and heated
to reflux (ꢁ155–160 1C) overnight under dry N2 to give a grey
suspension. SEM examination showed the solid particles broke up
into 0.1–0.5 mm fragments and the X-ray diffraction peaks
became broadened, indicating structural disorder or nanoparticle
formation. We found no evidence for dissolution of the Zintl
phase in the liquid but in situ XAS experiments indicated the
presence of Mg2Ge particles suspended in the beam path.
GeCl4 (Aldrich 99.99%: 0.13 ml, 1.14 mmol) was then added to
the stirred diglyme–Mg2Ge suspension. This amount corresponds
to a 3-fold excess over the stoichiometric amount required for
reaction (1): that requirement was suggested previously to be an
important condition for successful synthesis [35,36]. After a few
minutes the grey Mg2Ge suspension became clear and the colour
changed to light golden-yellow, indicating the formation of Ge
nanoparticles. White wispy filaments identified as MgCl2 were
also formed during the reaction (1). The Ge nanoparticles were
then capped by adding excess butyllithium (BuLi: Aldrich, 2.5 M
solution in hexane) to the cooled reaction mixture. Excess BuLi
was removed by addition of deionised water. The contents were
transferred to a separating funnel and the capped Ge nanoparti-
cles extracted into hexane. The non-aqueous phase was washed
with deionised water to remove LiCl. After removal of hexane by
rotary evaporation, an aliquot of the capped nanoparticles was
resuspended in CDCl3 for 1H NMR spectroscopy using a Bruker
AMX-300 instrument. Peaks at
d¼1.35 and 1.52 ppm indicated
the presence of Ge-alkyl species in the liquid or perhaps attached
to the surface of the nanoparticles. Intense blue photolumines-
cence characteristic of Ge nanoparticles with a 2–4 nm size range
was observed first by UV lamp (365 nm) illumination and then
recorded spectroscopically using 325 nm laser excitation during
room temperature photoluminescence studies (Fig. 2).
Fig. 1. Characterisation of the Zintl phase compound Mg2Ge used as a precursor
for the nano-Ge synthesis experiments. (a) Raman spectrum showing the single
triply degenerate mode expected for the antifluorite structure (b) Background
subtracted EXAFS spectrum at the Ge K edge. (c) Fourier transformed EXAFS data.
In (b) and (c) the solid line represents a best fit to the data using a multiple
scattering model (r-value 7%).
2.2. Transmission electron microscopy
We examined the nanoparticles produced at various stages
during the experiments using transmission electron microscopy
(TEM) combined with selected area diffraction (SAD) and energy-
dispersive X-ray (EDX) or electron energy loss (EELS) analysis.
Samples were prepared by placing drops on an amorphous C film
supported by a 400 mesh Cu grid and allowing to dry at room
temperature. TEM imaging and SAD experiments were carried out
using a 100 kV JEOL JEM-100CX11 at UCL or a 300 kV JEOL

A.J. Pugsley et al. / Journal of Solid State Chemistry 184 (2011) 2345–2352
2347
detection with a solid state Ge detector because of the low
concentrations of reactants and products present. Edge positions
were calibrated using crystalline Ge and are reported as threshold
energies at the maxima in first derivative spectra of the X-ray
absorption vs. energy [50]. A ‘‘Quick EXAFS’’ mode was used to
achieve rapid collection times (400 s/scan) for in situ reaction
experiments at SRS 9.2. Similar rapid scan experiments at ESRF
BM26 were run using 840 s collection times for each spectrum.
Spectra of the solid-state materials were obtained in transmission
mode using finely ground materials mounted on tape in the beam
path. A special cell with 25 mm mica windows and 2 mm path
length was constructed to hold air-sensitive liquids including
GeCl4 in diglyme solution and GeBu4 in hexane for transmission
or fluorescence XAS/EXAFS measurements at both facilities.
In situ reaction studies were carried out using various generations
of a specially designed high-temperature reaction cell, that enabled
stirring the Mg2Ge precursor with diglyme under H2O-free condi-
tions, addition of GeCl4 to the reaction mixture, maintaining
controlled temperature under reflux and avoiding loss of volatile
materials from the reaction mixture [53]. These cells were generally
constructed from aluminium block: the sample chamber consisted of
a 2 cm diameter hole drilled to a depth of 18 cm along the long axis
of the cell. Two 1 cm openings drilled half-way up the cell and closed
with 25 mm Kapton or mica windows permitted XAS measurements.
In later versions Al tubing wrapped with a Cu cooling coil added
above the reaction cell provided a condenser for volatile components
under reflux conditions: this was connected to a dry N2 supply and
bubbler to exclude ingress of atmospheric O2/H2O. Electrical car-
tridge heaters were inserted into holes drilled in the cell body
parallel to the sample chamber and T was monitored using a K-type
thermocouple. Cells for in situ reaction studies were brought to the
synchrotron beamlines pre-sealed and loaded with solid Mg2Ge
(5.4 mg, 0.04 mmol), diglyme (ꢁ7 ml) and a 3 mm magnetic stirrer
follower. Liquid GeCl4 was added using a gas-tight syringe or cannula
and the nanoparticle formation reaction (1) was initiated by heating
the cell to 155 1C.
Normalisation and background procedures to treat the XAS/
EXAFS data were mainly carried out using programmes from
Daresbury SRS: Excalib, Exback, Exspline, Exbrook, etc. EXAFS
modelling was carried out using Excurve98 using theoretical
potentials and phase shifts to fit the first three shells with
coordination numbers N around the central atom, radial distances
(rij) and Debye–Waller factors. EXAFS studies of solid Mg2Ge
indicated the importance of including multiple scattering paths
modelled using FEFF and IFEFFIT (Fig. 1b and c) [54,55].
Fig. 2. Room temperature photoluminescence spectra of capped Ge nanoparticles
in suspension in diglyme following reaction (1) or after extraction into hexane
using 325 nm laser excitation show strong features in the 400–550 nm range.
(a) Products of an early synthesis containing nanocrystalline q-GeO2 in addition to
nano-Ge. (b) Later samples prepared in fully dried diglyme, compared with the PL
spectrum of bulk Ge obtained under similar conditions.
JEM-3010 at University of Southampton (Dr. B.A. Cressey). Energy
dispersive analyses were performed using a JEOL JEM-2200FS
operating at 200 kV in the Department of Materials, Imperial
College London (Drs. D. McComb and I. Khan): these measure-
ments revealed the presence of oxygen impurities (K edge
ꢁ0.5 keV) among samples prepared during the early part of our
study and this was confirmed by EELS at the University of Leeds.
Later samples showed no presence of oxygen.
EXAFS analyses of our first in situ data sets obtained during the
nanoparticle formation reaction showed a peak due to Ge–Ge
˚
first-neighbour distances that was observed to grow near 2.4 A, as
expected (Fig. 3). However, the EXAFS patterns were dominated
˚
2.3. XAS/EXAFS measurements
by a strong peak at 1.7–1.9 A that was assigned to Ge–O oscilla-
tions of quartz-structured GeO2 (q-GeO2) nanoparticles present in
the samples. These had been produced along with the nano-Ge
particles because of H2O present within the reaction mixture, at
concentrations of 200–400 ppm. In our later experiments we used
fully dried diglyme (with o30 ppm H2O) and improved reaction
cell design as well as modified transfer procedures, so that
moisture ingress was controlled or eliminated, and GeCl4 loss
was avoided during the nanoparticle formation reaction.
A first series of XAS/EXAFS studies were carried out at the Ge K
edge (ꢁ11.1 keV) to characterise various solid and liquid state
standards and precursor materials to help interpret the nanopar-
ticle formation data. We then examined aliquots removed from
the reaction mixtures after various times, and also carried out
in situ XAS/EXAFS studies while implementing the synthesis
reaction (1) using one of a series of specially designed chemical
reaction cells. XAS studies were carried out at stations 9.2, 9.3 and
16.5 at SRS Daresbury, UK [51] and also at BM26 (Dutch-Belgian
Beamline DUBBLE, BM26, ESRF Grenoble [52]). Double crystal
Si(2 2 0) or Si(1 1 1) monochromators were used, slightly detuned
to give 50% harmonic rejection in the incident beam. Transmis-
sion spectra were obtained using Ar/He-filled ion chambers
before and after the sample chamber for I0 and It measurements.
The in situ reaction studies were carried out using fluorescence
3. Results and discussion
3.1. TEM and photoluminescence studies
Nanoparticle size distributions were determined by direct
measurement of ꢁ500 particles from each of several TEM images

2348
A.J. Pugsley et al. / Journal of Solid State Chemistry 184 (2011) 2345–2352
Fig. 3. (a) Background subtracted EXAFS region of Ge K-edge XAS spectra in the
reaction cell as a function of time (SRS 9.2). The corresponding Fourier transforms
of the EXAFS data are shown below. These experiments were carried out
incompletely dried solvent. The main feature in the FT EXAFS is due to GeO2
nanoparticles (quartz structure) formed by reaction between GeCl4 and H2O
present within the solvent or introduced into the cell during transfer procedures.
Fig. 4. Nanoparticle size distributions determined by direct measurement from
TEM images containing ꢁ500 particles. Aliquots were examined for samples
removed from the reaction mixtures at 1, 3 and 6 h (above) and after 4 h (below).
The first set of samples contained GeO2 as well as nano-Ge particles: the second
run was carried out under moisture-free conditions.
˚
The growth of a peak near 2.5 A is apparent due to formation of nano-Ge particles.
obtained from samples during the runs (Fig. 4). In a first series of
samples, after 1 h the particle size range extended from 1.5 to
4.5 nm, with a mean value near 2.7 nm; after 3 h, the average size
had increased to 2.9 nm and to ꢁ3.7 nm by 6 h, with only a few
larger particles extending up to ꢁ6 nm in dimension. This direct
counting method can underestimate the numbers of the smallest
particles because they are often difficult to locate in the TEM
images. Particle size distributions for samples prepared under
fully anhydrous conditions were indistinguishable from those of
earlier samples that contained some nanocrystalline GeO2 along
with the nano-Ge particles.
Previous studies have indicated crystalline lattice fringes present
for Ge/Si nanoparticles produced by the same synthesis procedures
we have attempted to reproduce here [37]. However, in most of our
SAD-TEM studies of capped Ge nanoparticles formed by reaction (1)
we could detect only diffuse amorphous diffraction rings with a
spacing consistent with a-Ge (Fig. 5). Well defined diffraction spots
found to be arising from some nanoparticles were determined to be
due to q-GeO2, formed by reaction of GeCl4 with H2O in incomple-
tely dried solvent conditions. In later studies of nano-Ge particles
formed under fully anhydrous conditions we found the diffraction
patterns were not stable under electron beam irradiation. We could
not determine if (a) the nanoparticles were initially crystalline
but amorphised in the beam; (b) the particles were amorphous
but crystallised in the beam; or (c) a mixture of crystalline and
amorphous nanoparticles were present immediately following
synthesis by reaction (1). In certain cases, crystalline diffraction spots
were observed during initial exposure of the nanoparticle to the
electron beam, using accelerating voltages ranging 100–300 keV;
however, these quickly evolved to give amorphous rings within a few
seconds of irradiation. In a few cases crystalline spots or sharp rings
were also observed to appear during electron beam irradiation of
some nanoparticles. Our TEM results could not allow us to conclude
if the Ge nanoparticles formed by reaction (1) are amorphous,
crystalline or a mixture of both.
The room temperature PL spectra excited by 325 nm laser
radiation exhibits a characteristic double maximum at ꢁ430 and
ꢁ510 nm, similar to that observed for bulk Ge, but with the lower
energy peak displaced to shorter wavelength values (Fig. 2). That
corresponds to the blue PL observed in previous studies and
under broadband UV irradiation. We did not observe any sys-
tematic changes in peak position or the PL intensity as a function
of reaction time during our studies as the average particle size
evolved between 2 and 4 nm indicated by the TEM results. The PL
spectra from samples containing nanocrystalline GeO2 exhibited
an additional weak feature at ꢁ660 nm.
3.2. XAS/EXAFS results: model compounds
During our study it was first necessary to interpret the Ge K
edge position and XANES to understand the nature of the growing
nanoparticles. It is well known that the Ge K edge shifts in
response to the bandgap [47,49]. The growing semiconductor
nanoparticles are exposed to an O-rich environment within the

A.J. Pugsley et al. / Journal of Solid State Chemistry 184 (2011) 2345–2352
2349
Fig. 6. Ge K-edge spectra for GeCl4, Mg2Ge and bulk c-Ge.
Studi dell’ Aquila, Italy) [47] and G. Aquilenti (SISSA, Trieste). The
threshold energies for bulk c-Ge and a-Ge are nearly identical:
analysis of the EXAFS data for a-Ge indicates a slightly elongated
˚
average Ge–Ge bond length of 2.464(2) A [47,56–58]. For comparison
with the results below, we obtained an XAS spectrum for molecular
GeBu4 prepared prepared by reacting BuLi with GeCl4 in diglyme,
representing the Ge-alkyl species that might be formed in side
reactions with excess GeCl4 present in solution during the capping
reaction with BuLi, or existing at the capped nano-Ge particle surface.
The edge position for this molecular species occurs at 11103 eV, close
to the values observed for bulk c- and a-Ge.
The XAS spectrum of molecular liquid GeCl4 contains a sharp,
characteristic white line with a K-edge value of 11,108 eV (Fig. 6).
Sharp oscillations extending to ꢁ120 eV above the edge are due
to multiple scattering among Cl–Ge–Cl paths for ejected photo-
electrons within the highly symmetric tetrahedral molecules [59].
The solid state Zintl compound Mg2Ge has an antifluorite (CaF2)
structure with Mg2þ on tetrahedral sites and Ge4ꢀ anions on the
eight-coordinated sites. During an initial EXAFS analysis we found
that a single-scattering model could not reproduce the Fourier-
transformed data. We performed a calculation including multiple
scattering paths with associated phase and amplitude functions in
order to determine the structural parameters, which are in good
agreement with those determined by X-ray diffraction methods
(Fig. 1). The first neighbour Ge–Mg distance (8-fold coordination)
Fig. 5. TEM studies of Ge nanoparticles formed by reaction (1). (a) Typical image
obtained using the 300 kV JEOL instrument at University of Southampton. The
scale bar represents 10 nm. (b) SAD pattern showing the amorphous ring pattern
obtained for typical nanoparticles. The ring positions are consistent with diffrac-
tion from amorphous Ge. Occasionally, the a-Ge nanoparticles were observed to
recrystallise during electron irradiation; in a few other cases, crystalline spots
were observed initially but these disappeared within a few seconds to be replaced
by an amorphous pattern.
˚
was determined to be 2.7590(4) A with
a
Debye–Waller
˚
2
˚
factor 0.0076(2) A , and the GeyGe distance was 4.5427(3) A
(Debye–Waller factor 0.0129(5) A ). Data for phenacite and
2
˚
spinel-structured Ge3N4 polymorphs have been previously com-
pared with quartz- vs rutile-structured phases of GeO2 and
provide examples of Ge4þ in 4- and 6-coordination to O and N
with different degrees of covalent bonding [50,60]. Samples of
GeS and GeSe containing Ge–Ge and Ge–S, Se bonds within the
black phosphorus (LiSi) structure and GeSe2 with a tetrahedrally
bonded chain structure were also examined and these
‘‘standards’’ have been used in this study to look at the edge
effect of covalency and oxidation state.
The edge positions for Mg2Ge and elemental Ge occur at lower
energy than for materials with Ge in positive oxidation states.
It has already been proposed that X-ray absorption edge energies
are correlated with the effective charge on the absorbing atom or
ion [61]. In general, increasing oxidation state is associated with
an increase in the partial charge, so that an edge shift to lower
energy is expected to occur with lower (i.e., with less positive or
diglyme solvent, and Ge–C species are expected to appear at the
nanoparticle surface after the organic ’’capping’’ reaction has
taken place, so it was important to characterise changes that
could occur as a function of Ge coordination to O- vs C-bearing
species. We examined the edge shifts among several solid state
materials and molecular compounds containing Ge atoms in
different oxidation states and coordination environments to
establish correlations that would be useful in our studies of
nano-Ge particle formation.
The edge position for diamond-structured c-Ge occurs at 11103 eV
(Fig. 6). The edge crest is broad and it extends to ꢁ11118 eV. Analysis
˚
of the EXAFS yields a first-neighbour Ge–Ge distance of 2.451(2) A in
good agreement with X-ray diffraction. Reference spectra for bulk
a-Ge were supplied by Drs. A. Fillipponi (Dip. di Fisica, Univ. degli

2350
A.J. Pugsley et al. / Journal of Solid State Chemistry 184 (2011) 2345–2352
more negative) oxidation states [54]. We plotted the Ge K-edge
positions obtained in this study against the formal oxidation
states for various compounds relevant to this study (Fig. 7). We
find little variation in edge positions among materials with low
oxidation state values, but they evolve rapidly towards higher
energy values for compounds described by the þ4 state. The
highest energies occur for the oxides, with the maximum value
recorded for octahedrally coordinated rutile-structured r-GeO2.
These results have helped us interpret the XAS data obtained
during the nanoparticle formation reactions.
3.3. XAS/EXAFS results: nanoparticle synthesis studies
During our early synthesis experiments nanocrystalline GeO2
was produced along with Ge nanoparticles as a result of small
amounts of H2O (200–400 ppm) present within the diglyme
solvent, or introduced inadvertently into the reaction cell. We
first surmised the presence of GeO2 species from an unexpected
˚
peak dominating the EXAFS signal at r¼1.7–1.9 A (Fig. 4). This
could be correlated with the high edge value (11,107 eV) and a
sharp edge crest, that is typical for GeO2-containing materials
(Fig. 9) [50,60]. The edge position and shape were most char-
acteristic of tetrahedrally bonded q-GeO2, consistent with SAD-
TEM results for crystalline nanoparticles in samples quenched
from these initial runs.
Later synthesis studies were carried out using fully dried
solvent and under anhydrous reaction conditions. The recorded
edge position now occurred at a higher energy than that of bulk
c-Ge or a-Ge: the XANES was broadened and was featureless,
comparable with the signal for GeBu4 (Fig. 8). The EXAFS data
˚
were dominated by the first shell Ge–Ge peak at 2.45–2.46 A as
expected. No obvious second shell GeyGe correlation was
observed in the Fourier transformed EXAFS data supporting the
likelihood that the nanoparticles formed by reaction (1) are
mainly amorphous, in agreement with our TEM results (Fig. 5).
Fig. 8. Characteristic XAS edge spectra for various materials produced during the
nanoparticle formation reaction. (a) Comparison of crystalline and amorphous
bulk Ge with q-GeO2 and GeBu4. (b) Comparison of Ge, GeO2 and GeBu4 with a
typical aliquot removed from a nanoparticle synthesis run in diglyme (a157).
˚
A peak at low r (1.92–1.95 A) was also present (Fig. 9). This
feature could be assigned to Ge–C bonds from species capping the
nanoparticles. We determined from TEM studies that the nano-Ge
particles formed during typical reaction times of 1–6 h had
˚
diameters between 2 and 9 nm (20–90 A) (Fig. 4). The volume
3
2
˚
˚
(4000–380,000 A ) and surface area (1300–25,400 A ) determined
for these nanoparticles thus contain 180–16,000 Ge atoms, of
Fig. 9. EXAFS analysis of
a typical nano-Ge product produced in fully dried
diglyme.
which 80–2000 are present at the surface. The ratio of Ge–Ge to
Ge–C bonds should then lie between 2:1 and 8:1, within the range
observed from the EXAFS data (Fig. 9). However, we also carried
out mass spectrometry experiments on samples removed during
Fig. 7. XAS Ge K edge positions measured for various model compounds plotted
against the formal oxidation state.

A.J. Pugsley et al. / Journal of Solid State Chemistry 184 (2011) 2345–2352
2351
the nanoparticle formation reaction following capping by BuLi,
and observed characteristic peaks occurring for Ge-alkyl frag-
ments derived from monomeric and oligomeric Ge-butyl species.
These were formed by a side reaction occurring between BuLi and
excess GeCl4 that was present in solution from the initial
nanoparticle formation reaction. As noted earlier, previous synth-
esis studies had indicated that a large (3-fold) excess of GeCl4
should be used in the reaction mixture. However, in our present
synthesis, none of this reacting species was lost by evaporation,
and the solvent was thoroughly dried so no GeCl4 was used in
forming oxide, and thus the excess remained in solution and was
available to react with the capping agent. We then produced
GeBu4 by reacting GeCl4 with BuLi in diglyme solution and
obtained its XAS spectrum in which there was only evidence for
Ge-alkyl species (i.e., no Ge–Cl remaining). The edge position
occurred near 11,103 eV, close to that of bulk c- and a-Ge, and it
has a broad featureless edge crest that is only slightly broader
than the solid-state materials (Fig. 8). We conclude that the
˚
1.92–1.95 A feature observed in the EXAFS data for the Ge
Fig. 11. In situ XAS data obtained during the course of reaction (1) using fully
dried diglyme and the second generation reaction cell. The first spectrum was
taken after GeCl4 was added to the cell, before heating to 155 1C to initiate the
reaction. The strong white line signal for GeCl4 is off scale. Prior to this a signal for
Mg2Ge solid particles stirred into the beam path was observed, but this was
obscured by the strong GeCl4 spectrum. The sequence of spectra were taken as a
function of time after onset of the reaction. Each spectrum was recorded for 840 s
during 2.5 h. The strong GeCl4 signal diminishes rapidly in intensity as the
reaction proceeds. Growth of the feature at low energy (ꢁ1105 eV) is observed
due to the formation of Ge nanoparticles.
nanoparticle samples could contain contributions from both
Ge–C bonds formed at the surface of the nanoparticles, and from
GeBu4 or other Ge-alkyl species present in solution. We could
obtain reasonable fits to the observed XAS spectra of various
aliquots of our nanoparticle-containing suspensions by taking
ꢁ50:50 combinations of contributions from crystalline or amor-
phous Ge along with GeBu4 species (Fig. 10).
We carried out a series of in situ reaction studies at ESRF BM26
under anhydrous conditions (Fig. 11). A solution of diglyme and
Mg2Ge was initially stirred at room temperature before the
reaction was started and an edge jump was observed which can
be attributed to Mg2Ge particles suspended in the beam path.
GeCl4 was then added by cannula and the XAS spectrum became
dominated by the characteristic sharp edge peak of molecular
GeCl4. The heater was then switched on to initiate the nanopar-
ticle formation reaction at 155 1C. Each in situ run was carried out
for 2–3 h, with each data set accumulated for 840 s (Fig. 11). The
contribution of the strong GeCl4 signal to the overall XAS
decreased rapidly and a new feature grew at low energy signal-
ling the formation of germanium nanoparticles. However, some
GeCl4 component was still detected in the reaction mixture at up
to ꢁ2 h. These preliminary results indicate that in situ XAS
studies could be used to examine the kinetics and mechanism
of the nanoparticle formation reaction.
4. Conclusions
We studied formation of nano-Ge particles in an organic
solvent (diglyme) by a heterogeneous metathesis/comproportio-
nation reaction between dissolved GeCl4 (Ge4þ) and suspended
Mg2Ge (Ge4ꢀ). We have characterised the solid state starting
material Mg2Ge using XAS/EXAFS along with Raman spectro-
scopy. We noted the formation of nanocrystalline q-GeO2 along
with nano-Ge during early runs carried out using incompletely
dried diglyme solvent: we determined that o30 ppm H2O must
be maintained throughout the synthesis in order to avoid produc-
tion of oxide. However, the presence of nc-q-GeO2 does not affect
the photoluminescence properties. The Ge nanoparticles pro-
duced in our studies were mainly amorphous, with average size
2–4 nm for reaction times ranging between 0.5 and 6 h. Our in situ
XAS/EXAFS experiments offer a new way to monitor and deter-
mine the mechanism and kinetics of the nanoparticle formation
reaction.
Acknowledgments
We thank CCLRC (now STFC) for access to various stations at
the SRS (closed since August 2008): also Nick Chinnery, Bob
Bilsborrow, C. Corrigan and Mark Jackson for their help with
various XAS measurements and sample handling. M. Sheehy
(previously at the RI) is thanked for building the inital in situ cell
within the Davy-Faraday Research Laboratory. The experiments
were supported by a Wolfson Foundation-Royal Society Research
Merit Award and EPSRC Senior Research Fellowship to PFM
(EP/D07357X) and by EPSRC Portfolio award EP/D504782 to
PFM in collaboration with CRA Catlow and P Barnes.
References
[1] A.P. Alivisatos, A.L. Harris, N.J. Levinos, M.L. Steigerwald, L.E. Brus, Journal of
Chemical Physics 89 (1988) 4001–4011.
[2] L.E. Brus, Nature 353 (1991) 301–302.
Fig. 10. Result of fitting a typical nanoparticle XAS spectrum (‘‘fn’’) to a combina-
tion of GeþGeBu4 components.

2352
A.J. Pugsley et al. / Journal of Solid State Chemistry 184 (2011) 2345–2352
[3] D.J. Crouch, S. Norager, P. O’Brien, J.-H. Park, N. Pickett, Philosophical
Transactions of the Royal Society London A 361 (2003) 297–310.
[4] D.J. Crouch, P. O’Brien, M.A. Malik, P.J. Skabara, S.P. Wright, Chemical
Communications (2003) 1454–1455.
[34] H.P. Wu, J.F. Liu, Y.W. Wang, Y.W. Zeng, J.Z. Jiang, Materials Letters 60 (7)
(2006) 986–989.
[35] B.R. Taylor, S.M. Kauzlarich, G.R. Delgado, H.W.H. Lee, Chemistry of Materials
11 (1999) 2493–2500.
[5] A. Mews, A.V. Kadavanich, U. Banin, A.P. Alivisatos, Physical Review B 53
(1996) 13242–13245.
[36] B.R. Taylor, S.M. Kauzlarich, H.W.H. Lee, G.R. Delgado, Chemistry of Materials
10 (1998) 22–24.
[6] P. O’Brien, Current Opinion in Solid State and Materials Science 6 (2002) 335.
[7] R. Rossetti, J.L. Ellison, J.M. Gibson, L.E. Brus, Journal of Chemical Physics
(1984) 4464–4469.
[8] R. Rossetti, S. Nakahara, L.E. Brus, Journal of Chemical Physics (1983)
1086–1088.
[9] J.G. Brennan, T. Siegrist, P.J. Carroll, S.M. Stucynski, L.E. Brus, M.L. Steigerwald,
Joural of the American Chemical Society 111 (1989) 4141–4143.
[10] P. Lianos, J.K. Thomas, Chemical Physics Letters 125 (1986) 299–302.
[11] T. Dannhauser, M. O’Neil, K. Johansson, D. Whitten, G. McLendon, Journal of
Physical Chemistry 90 (1986) 6074–6076.
[37] C.-S. Yang, S.M. Kauzlarich, Y.C. Wang, Chemistry of Materials 11 (1999)
3666–3670.
[38] X. Zhang, D. Neiner, S. Wang, A.Y. Louie, S.M. Kauzlarich, Nanotechnology 18
(2007) 095601.
[39] R.A. Bley, S.M. Kauzlarich, Journal of the American Chemical Society 118
(1996) 12461–12462.
[40] R.A. Bley, S.M. Kauzlarich, H.W.H. Lee, J.E. Davis, in: Materials Research
Society Symposium Proceedings, vol. 351, 1994, pp. 275–280.
[41] S.M. Kauzlarich, Chemistry, Structure, and Bonding of Zintl Phases and Ions,
VCH, New York, 1996.
[12] A. Fojtik, H. Weller, U. Kock, A. Henglein, Berichte der Bunsengesellschaft fuer
Physikalische Chemie 88 (1984) 969–977.
[13] L. Spanhel, M. Haase, H. Weller, A. Henglein, Journal of the American
Chemical Society 109 (1987) 5649–5655.
[42] R. Nesper, Progress in Solid State Chemistry 20 (1990) 1.
[43] H. Scha¨fer, Annual Review of Materials Science 15 (1985) 1.
[44] N. Zaitseva, Z.R. Dai, C.D. Grant, J. Harper, C. Saw, Chemistry of Materials 19
(2007) 5174–5178.
[14] A.P. Alivisatos, Science 271 (1996) 933–937.
[45] B.K. Teo, EXAFS: Basic Principles and Data Analysis, Springer-Verlag
New York, 1986.
[46] G.E.J. Brown, G. Calas, G.A. Waychunas, J. Petiau, Reviews in Mineralogy 18
(1988) 431–512.
[47] A. Filipponi, A. Di Cicco, Physical Review B 51 (1995) 12322–12337.
[48] E.A. Stern, D.E. Sayers, F.W. Lytle, Physical Review B 11 (1975) 4836–4846.
[49] J. Bla¨sing, P. Kohlert, M. Zacharias, P. Veit, Journal of Applied Crystallography
31 (1998) 589–593.
[15] K. Brunner, Reports on Progress in Physics 65 (2002) 27–72.
[16] L.E. Brus, P.F. Szajowski, W.L. Wilson, T.D. Harris, S. Schuppler, P.H. Citrin,
Journal of the American Chemical Society 117 (1995) 2915–2922.
[17] R.T. Collins, P.M. Fauchet, M.A. Tischler, Physics Today 50 (1997) 24–31.
[18] M. Larsson, A. Elfving, P.O. Holtz, G.V. Hansson, W.A. Ni, Surface Science 532
(2003) 832–836.
[19] K.A. Littau, P.J. Szajowski, A.J. Muller, A.R. Kortan, L.E. Brus, Journal of Physical
Chemistry 97 (1993) 1224–1230.
[20] S. Miyazaki, K. Sakamoto, K. Shiba, M. Hirose, Thin Solid Films 255 (1995)
99–102.
[50] L. Bertini, P. Ghigna, M. Scavini, F. Cargnoni, Physical Chemistry Chemical
Physics (2003). doi:10.1039/b21202h.
[21] M. Nirmal, L.E. Brus, Accounts of Chemical Research 32 (1999) 407–414.
[22] T. Suemoto, K. Tanaka, A. Nakajima, T. Itakura, Physical Review Letters 70
(1993) 3659–3662.
[23] J.C. Vial, A. Bsiesy, F. Gaspard, R. Herino, M. Ligeon, R. Muller, R. Romestain,
Physical Review B 45 (1992) 14171–14176.
[24] L.T. Canham, Applied Physics Letters 57 (1990) 1046.
[25] H.W. Chiu, C.N. Chervin, S.M. Kauzlarich, Chemistry of Materials 17 (2005)
4858–4864.
[26] H.W. Chiu, S.M. Kauzlarich, E. Sutter, Langmuir 22 (12) (2006) 5455–5458.
[27] J. Cho, Electrochimica Acta 54 (2) (2008) 461–466.
[28] S.J. Kim, O.K. Quy, L.S. Chang, E.A. Stach, C.A. Handwerker, A. Wei, Journal of
Materials Chemistry 20 (2) (2010) 331–337.
[51] G.N. Greaves, C.C.R. A, G.E. Derbyshire, M.I. McMahon, R.J. Nelmes, G. van der
Laan, Nature Materials 7 (2008) 827–830.
[52] S. Nikitenko, A.M. Beale, M.J. van der Eerden, S.D.M. Jacques, O. Leynaud,
M.G. O’Brien, D. Detallenaere, R. Kaptein, B.M. Weckhyuysen, W. Bras,
Journal of Synchrotron Radiation 15 (2008) 632–640.
[53] A.J. Pugsley, Synthesis of Ge-Naoparticles in Organic Solution: Chemical and
Structural Characterisation and In Situ Studies of the Formation Reaction,
Ph.D. Thesis, University College London, 2008.
[54] M. Newville, B. Ravel, D. Haskel, J.J. Rehr, E.A. Stern, Y. Yacoby, Physica B 209
(1995) 154–156.
[55] J.J. Rehr, J.M. Deleon, S.I. Zabinsky, R.C. Albers, Journal of the American
Chemical Society 113 (1991) 5135–5140.
[29] K.T. Park, H.D. Um, S.W. Jee, J.Y. Jung, Y.C. Park, J.M. Yang, J.H. Lee, Chemical
Vapor Deposition 14 (11–12) (2008) 331.
[30] S. Prabakar, A. Shiohara, S. Hanada, K. Fujioka, K. Yamamoto, R.D. Tilley,
Chemistry of Materials 22 (2) (2010) 482–486.
[56] G. Bunker, Nuclear Instrumentation and Methods 207 (1983) 437.
[57] G. Dalba, P. Fornasini, M. Grazioli, F. Rocca, Journal of Non-Crystalline Solids
164&166 (1993) 159.
[58] D.E. Sayers, E.A. Stern, F.W. Lytle, Physical Review Letters 27 (1971) 1204.
[59] C.R. Natoli, D.K. Misemer, S. Doniach, F.W. Kutzler, Physical Review A 22
(1980) 1104–1108.
[31] A.Y. Shah, A. Wadawale, B.S. Naidu, V. Sudarsan, R.K. Vatsa, V.K. Jain,
V. Dhayal, M. Nagar, R. Bohra, Inorganica Chimica Acta 363 (14) (2010)
3680–3684.
[60] C.L. Bull, P.F. McMillan, J.-P. Itie´, A. Polian, Physica Status Solidi (a) 201 (2004)
909–916.
[32] H.J. Song, S.M. Yoon, H.J. Shin, H. Lim, C. Park, H.C. Choi, Chemical Commu-
nications 34 (2009) 5124–5126.
[61] A. Yoshiasa, T. Tamura, O. Kamishima, K.-I. Murai, Journal of Synchrotron
Radiation 6 (1999) 1051–1058.
[33] H.P. Wu, J.F. Liu, M.Y. Ge, L. Niu, Y.W. Zeng, Y.W. Wang, G.L. Lv, L.N. Wang,
G.Q. Zhang, J.Z. Jiang, Chemistry of Materials 18 (7) (2006) 1817–1820.