Inorganic Chemistry
Article
and the byproducts, NaF and unreacted NH4F, were removed by
filtration. Excess and coordinated NH3 was removed under vacuum at
T = −40 °C to recover NH4BH4. The metal borohydrides LiBH4 and
NaBH4 were purchased from Sigma-Aldrich, while the solvent- and
chloride-free metal borohydrides Mg(BH4)2, Ca(BH4)2, Sr(BH4)2,
Mn(BH4)2, Y(BH4)3, La(BH4)3, and Gd(BH4)3 were synthesized in-
house by combining solvent-based methods and mechanochemistry
according to previously described protocols.5,75−77
Cryo mechanochemistry (T = −196 °C) of NH4BH4−M(BH4)m
(M = Li, Na, Mg, Ca, Sr, Y, La, Gd) in appropriate ratios (see Table
1) was carried out using a 6770 Spex Freezer mill. The powders were
loaded in a polycarbonate cylinder (25 mL) or a stainless steel
cylinder (2 mL), in both cases with stainless steel end plugs and a
stainless steel rod. The canister was magnetically rotated back and
forth 15 times per second for 2 min intervened by a 2 min break, and
this sequence was repeated 15 times.
A Note on Safety. All reagents and starting materials are air- and
moisture-sensitive and may react violently with H2O. All sample
handling and preparation was performed under an inert argon
atmosphere using Schlenk techniques or using an argon-filled
glovebox with a circulation purifier, p(O2, H2O) < 1 ppm. The
materials are thermally sensitive and should be stored in a glovebox
freezer (below T = −34 °C), and samples should be kept cold during
handling.
Synchrotron Radiation Powder X-ray Diffraction. In situ
variable-temperature synchrotron radiation powder X-ray diffraction
data (SR PXD) were collected at the BM01 beamline at the European
Synchrotron Radiation Facility (ESRF, Grenoble, France) with
wavelengths λ = 0.69449 and λ = 0.8212 Å, at the beamline I11 at
the Diamond Light Source (Oxford, England) with wavelengths λ =
0.825775 and λ = 0.825873 Å, and at the beamline I711 at MAXII,
MAX-Lab (Lund, Sweden) with the wavelength λ = 0.9938 Å. The
samples were packed in 0.5 mm boron silicate capillaries, sealed under
an argon atmosphere, and heated at a heating rate of 2−5 °C/min in
the temperature range T = −25 to +500 °C. The samples were rotated
during data acquisition.
Density Functional Theory Calculations. The experimental
structures were optimized by DFT calculations. The calculations were
performed using the Vienna Ab initio Simulation Package (VASP)81
with van der Waals density functional (vdW-DF2) proposed by Lee et
al.82,83 The inclusion of van der Waals energy gives a better prediction
of the lattice parameters for systems with weak bonds as in this work.
A projector-augmented wave potential84 with a plane-wave cutoff
energy of 500 eV was used, and the unit cell parameters were
optimized using a higher plane-wave cutoff energy of 600 eV. As a
measure of agreement between the experimental structure solution
and the calculated structure, cell parameters and cell volumes are
Fourier Transform Infrared Spectroscopy (FTIR). The samples
were characterized by infrared absorption spectroscopy using a
NICOLET 380 FT-IR spectrometer from Thermo Electron
Corporation with a diamond attenuated total reflectance (ATR)
crystal. The samples were exposed to air for approximately 10 s when
they were transferred from the sample vial to the instrument. Data
were collected in the range 500−4000 cm−1, and 32 scans with a
spectral resolution of 4 cm−1 were collected per sample and averaged.
Thermal Analysis and Mass Spectroscopy. Thermogravimetric
analysis (TGA) and differential scanning calorimetry (DSC) were
measured using a PerkinElmer STA 6000 coupled with a mass
spectrometer (MS) (Hiden Analytical HPR-20 QMS sampling
system). Each sample (approximately 1−5 mg) was placed in an
Al2O3 crucible and heated with a rate of 2−5 °C/min and an argon
purge rate of 20 mL/min. The outlet gas was monitored for hydrogen
(m/z = 2), ammonia (m/z = 17), diborane (m/z = 26), and borazine
(m/z = 80) using mass spectrometry.
Temperature-Programmed Photographic Analysis (TPPA).
Samples were sealed under argon in a glass tube placed in a custom-
made oven as described in the literature.85 The samples were heated
from room temperature to 400 °C with a heating rate of 5 °C/min,
while photos of the sample were collected every 5 s.
RESULTS AND DISCUSSION
■
Synthesis and Initial Characterization. Extensive
systematic synthesis work has allowed the preparation of 10
novel ammonium metal borohydrides and 7 ammonium metal
borohydride derivatives by combining ammonium borohy-
Structural Solution and Refinement. The crystal structures
were solved and refined from SR PXD data. The general procedure
involved indexing of the unit cell and subsequently solving the
structure ab initio by global optimization in direct space, as
−
+
dride, NH4BH4, and metal borohydrides, M(BH4)m (Mm+
=
implemented in the program FOX, and treating BH4 , NH4 , NH3,
and NH3BH3 as rigid bodies during the structure solution process.78
The structural model was then refined by the Rietveld method using
the program Fullprof79 and was checked for higher symmetry using
the ADDSYM routine in Platon.80
Li+, Na+, Mg2+, Ca2+, Sr2+, Mn2+, Y3+, La3+, Gd3+, according to
the addition reaction in eq 1, using a cryo-mechanochemical
approach.86
Verification of the structural models was performed by evaluating
the structure after Rietveld refinements, where a poor fit to the PXD
data and unreasonable bond distances indicated a wrong structural
model. In these cases, a new structural model and in some cases
modified composition was attempted to obtain a better fit to the
observed PXD data.
xNH4BH4(s) + yM(BH4)m(s)
−196 °C
⎯⎯⎯⎯⎯⎯⎯→⎯ (NH4)xMy(BH4)x+my(s)
(1)
The precursors, ammonium borohydride and metal
borohydrides, were prepared as solvent- and halide-free
compounds.5,75−77 The synthesis conditions provide products
according to eq 1, but small amounts of remaining reactants
and/or decomposition products are often present. Further
optimization of the synthesis conditions may improve the
purity of the samples. In some cases the reaction partially
occurs during the cryo-mechanochemical treatment and
partially during the thermal treatment (i.e., observed during
an in situ SR PXD investigation). NH4BH4 gradually
decomposes during extended cryo-mechanical treatment to
[(NH3)2BH2][BH4]. However, if a stable ammonium metal
borohydride is not formed with a composition that matches
the initial ratio between M(BH4)m and NH4BH4 (with
NH4BH4 in excess), then an ammonium metal borohydride
derivative may be formed, containing neutral molecules such as
NH3 and/or NH3BH3 due to partial decomposition: e.g.,
NH4Y(BH4)4·NH3 and (NH4)2Y(BH4)5·NH3 are formed in
In the metal coordination sphere it is particularly challenging to
−
distinguish the ligands NH3 and BH4 , as they are isoelectronic and
often have similar bond distances to the metal. Thus, permutations of
the ligands coordinating the metal were considered, and the most
likely configuration was selected, considering both chemical proper-
ties and the results from DFT optimization. The coordinations of
−
NH3 and BH4 are different, since NH3 always coordinates to the
metal via the nitrogen lone pair as a terminal ligand, while BH4− may
act as a bridging ligand between two metal atoms. Furthermore, N−H
usually forms dihydrogen bonds to nearby BH4− complexes. All of the
−
structures containing NH3 also contain terminal BH4 ligands; thus
permutations other than those presented here may be possible.
DFT calculations were used for a final verification of the structural
models, and large deviations in the unit cell volume and atomic
positions indicated a wrong structural model. Crystallographic
information files (CIF) are reported for the DFT-optimized
structures. The atomic positions were not refined in subsequent
Rietveld refinements after DFT optimization.
C
Inorg. Chem. XXXX, XXX, XXX−XXX
Cerny, Radovan
