RETRACTED ARTICLE: Monodisperse Ni3Fe single-crystalline nanospheres as a highly efficient catalyst for the complete conversion of hydrous hydrazine to hydrogen at room temperature

Article abstract of DOI:10.1039/c3ta10321h

A facile, reliable, and robust synthetic method for the preparation of monodisperse Ni₃Fe single-crystalline nanospheres with 182.3 m ² g^-1 specific surface area is reported. The results of an investigation of the shape/size distribution, structural characteristics, and composition of these Ni₃Fe nanospheres are presented, and a possible mechanism for the formation of the nanospheres at high temperatures in a cetylamine-palmitic acid solution in the presence of tri-iso-butyl thallium and borane trimethylamine complexes is proposed. During the decomposition of hydrazine, when supported on carbon (Ni₃Fe/C), the monodisperse Ni₃Fe nanospheres exerted 100% H₂ selectivity, and showed exceedingly high activity at room temperature. The decomposition reaction is first-order in Ni₃Fe concentration, but zero-order in N ₂H₄ concentration; the activation energy was 48.1 ± 2.0 kJ mol^-1. Furthermore, the Ni₃Fe/C was a stable catalyst for N₂H₄ hydrolytic dehydrogenation; it provided 15840 total turnovers in 30 h. An Ni₃Fe nanosphere-based catalyst such as this that is efficient under mild reaction conditions represents a promising step toward the practical development of N₂H₄ as a feasible hydrogen storage medium for fuel cell applications. The Royal Society of Chemistry 2013.

Full text of DOI:10.1039/c3ta10321h

Journal of
Materials Chemistry A
View Article Online
View Journal | View Issue
PAPER
Monodisperse Ni3Fe single-crystalline nanospheres as a
highly efficient catalyst for the complete conversion of
hydrous hydrazine to hydrogen at room temperature†
Cite this: J. Mater. Chem. A, 2013, 1,
6425
a
a
b
Dong Ge Tong, Dong Mei Tang, Wei Chu, Gui Fang Gua and Ping Wua
*
*
A facile, reliable, and robust synthetic method for the preparation of monodisperse Ni3Fe single-crystalline
nanospheres with 182.3 m2 gꢀ1 specific surface area is reported. The results of an investigation of the shape/
size distribution, structural characteristics, and composition of these Ni3Fe nanospheres are presented, and a
possible mechanism for the formation of the nanospheres at high temperatures in a cetylamine–palmitic
acid solution in the presence of tri-iso-butyl thallium and borane trimethylamine complexes is proposed.
During the decomposition of hydrazine, when supported on carbon (Ni3Fe/C), the monodisperse Ni3Fe
nanospheres exerted 100% H2 selectivity, and showed exceedingly high activity at room temperature. The
decomposition reaction is first-order in Ni3Fe concentration, but zero-order in N2H4 concentration; the
activation energy was 48.1 ꢁ 2.0 kJ molꢀ1. Furthermore, the Ni3Fe/C was a stable catalyst for N2H4
hydrolytic dehydrogenation; it provided 15 840 total turnovers in 30 h. An Ni3Fe nanosphere-based
catalyst such as this that is efficient under mild reaction conditions represents a promising step toward
the practical development of N2H4 as a feasible hydrogen storage medium for fuel cell applications.
Received 21st January 2013
Accepted 25th March 2013
DOI: 10.1039/c3ta10321h
www.rsc.org/MaterialsA
To maximize the efficacy of hydrazine as a hydrogen storage
material, the undesired reaction pathway (2) should be avoided.
1 Introduction
Hydrogen is one of the most promising energy carriers to meet
the increasing demand for an efficient and clean energy
supply.1–3 The controlled storage and release of hydrogen is a
great technological challenge in the fuel cell-based hydrogen
economy.4–8 Various kinds of solid materials have been inves-
tigated and have been found to be promising for hydrogen
storage.1–11 Typical examples are metal/complex hydrides, metal
nitrides and imides, carbon materials, inorganic nano-
structures, and metal–organic frameworks. Recently, hydrazine
monohydrate (N2H4$H2O) was identied as one of the leading
candidates for hydrogen reservoirs, owing to its high hydrogen
content (8.0 wt%).12–19 Hydrazine can be decomposed in two
ways: (1) complete decomposition, as shown in eqn (1)
Although 100% H2 selectivity has already been successfully
achieved using Ni-based catalysts such as Ni–Ir, Ni–Rh, Ni–Pt,
and Ni–Fe,20–24 their kinetics are slow, which greatly hinders the
practical application of this system. Among these Ni catalysts,
Ni–Fe is one of the lowest-cost catalysts for the N2H4 hydrolysis
reaction. From the commercial viewpoint, the preparation of a
Ni–Fe catalyst with a high hydrolytic activity is highly desirable.
Generally, Ni–Fe nanocatalysts are prepared via the co-
reduction of their precursors with NaBH4 in aqueous solutions,
in the presence of hexadecyltrimethylammonium bromide
(CTAB).23 However, the nanoparticles (NPs) in the catalysts oen
undergo signicant aggregation aer the removal of CTAB and
form much larger particles. As a result, their catalytic activity is
poor because of the low surface-to-volume ratio available aer
agglomeration.23 For example, the Ni2Fe nanocatalyst was only
active at 343 K, and addition of NaOH was necessary for 100%
H2 selectivity.23 Meanwhile, only 89% H2 selectivity was
observed for the Ni3Fe nanocatalyst.23 In recent years, mono-
disperse metallic NPs with high surface-to-volume ratio avail-
able have received extensive interest, because of their high
activity under mild conditions in various catalytic reactions.25–30
For example, monodisperse Pt nanoparticles show enhanced
catalytic activity toward oxygen reduction, over twice as high as
that of commercial Pt nanoparticles,25 and monodisperse Pt–Cu
nanocubes exhibit highly active catalytic activity toward the
oxidation of formic acid.29 Therefore, the preparation of
monodisperse nickel–iron nanoparticles with high specic
N2H4 / N2 + 2H2,
and (2) incomplete decomposition, as shown in eqn (2)
3N2H4 / 4NH3 + N2.
(1)
(2)
aMineral Resources Chemistry Key Laboratory of Sichuan Higher Education
Institutions, College of Materials and Chemistry & Chemical Engineering, Chengdu
University of Technology, Chengdu 610059, China. E-mail: tongdongge@163.com;
Fax: +86-28-8407-9074
bCollege of Chemical Engineering, Sichuan University, Chengdu 610065, China.
E-mail: chuwei65@yahoo.com.cn; Fax: +86-28-8540-3397
† Electronic supplementary information (ESI) available. See DOI:
10.1039/c3ta10321h
This journal is ª The Royal Society of Chemistry 2013
J. Mater. Chem. A, 2013, 1, 6425–6432 | 6425

View Article Online
Journal of Materials Chemistry A
Paper
surface area is a way of enhancing their hydrolytic activity added. Aer 0.075 g of Tl(i-Bu)3 and 0.250 g of BTM were added
towards N2H4. It is well known that monodisperse metal NPs to the mixture, the temperature was raised to 200 ꢂC and
can be stabilized by surfactants such as long-chain thiols or maintained at this level for 30 min, under stirring. The solution
polymers that bind strongly to the surface of monodisperse was then cooled to room temperature. 30 mL of ethanol was
metal NPs;31–37 the presence of these molecules makes it diffi- added, and the product was separated by centrifugation and
cult to achieve high catalytic activity. Although there have been washed with anhydrous hexane for several cycles. The Ni3Fe
some studies reporting the removal of these types of surfactants nanospheres were eventually re-dispersed in hexane, forming a
from the surface of NPs,38 the aggregation of NPs and the stable colloidal suspension.
deterioration of the catalysis (as observed for Ni–Fe NPs
For the preparation of the Ni3Fe/C, 20 mg of Ni3Fe NPs was
prepared in the presence of CTAB) is unavoidable.23 In this dissolved in hexane in a 10 mL vial, and an equal amount of
regard, the preparation of monodisperse Ni–Fe NPs in the carbon (Vulcan XC-72R, Carbot, USA) support was carefully
presence of weakly binding surfactants that do not need to be added to it. This colloidal mixture was stirred for 2 h to ensure
removed is highly desirable. Recently, it has been reported that complete adherence of the Ni3Fe NPs on the carbon support.
the presence of weakly binding surfactants such as amines or Aer the hexane evaporated, 5 mL of oxygen-free distilled water
fatty acids is advantageous for the preparation of highly active was added to the Ni3Fe/C dispersion, and the mixture was
and stable monodisperse metal nanocatalysts.39–42 For example, sonicated for 30 min to ensure uniform distribution. The
monodisperse Ni NPs prepared in the presence of oleylamine aqueous Ni3Fe/C dispersion was protected in a sealed glass tube
and oleic acid exhibited much higher catalytic activity and with a helium atmosphere. All catalytic reactions were per-
stability for the hydrolytic dehydrogenation of ammonia borane formed using aqueous Ni3Fe/C catalyst dispersion; aer exper-
than commercial Ni catalysts, with no special treatment to iments, the dispersion was transferred into oxygen-free water
remove surfactants.42 Inspired by these results, and using and broken under water, to avoid the oxidation of Ni3Fe. The
cetylamine and palmitic acid as weakly binding surfactants for aggregated Ni3Fe on carbon (Vulcan XC-72R, Carbot, USA) was
the rst time, we obtained monodisperse nickel–iron (Ni3Fe) also prepared as a comparison, using the same procedure.
single-crystalline nanospheres with 182.3 m2 gꢀ1 specic
surface area; these nanoparticles exhibited high catalytic
activity, with 100% H2 selectivity for the decomposition of N2H4
2.2 Characterization of the catalyst
without the promotion of NaOH under ambient conditions. The X-ray diffraction patterns were measured using an X’Pert X-ray
success of the synthetic method developed here shows that powder diffractometer equipped with a Cu Ka radiation source
monodisperse spherical Ni3Fe nanocrystals can be prepared via (l ¼ 0.15406 nm). A JEOL-2100F TEM instrument operated at
the reduction of nickel(II) acetylacetonate ([Ni(acac)2]) and 200 kV was employed to take TEM images and collect selected-
anhydrous iron(II) chloride (FeCl2) with tri-iso-butyl thallium area electron diffraction (SAED) patterns for the Ni3Fe, and an
(Tl(i-Bu)3) and borane trimethylamine (BTM) complexes in the FEI Tecnai F20 FEG TEM operated at 200 kV was used to take
presence of cetylamine and palmitic acid. In this reaction, Tl(i- TEM images for the Ni3Fe/C samples. Samples for TEM analysis
Bu)3 and BTM served as reducing agents, and cetylamine and were prepared by depositing a single drop of diluted nano-
palmitic acid were added as weakly binding surfactants to particle dispersion in hexane on an amorphous carbon-coated
stabilize the Ni3Fe nanospheres. It should be noted that when copper grid. High-angle annular dark-eld scanning trans-
loaded on carbon as the catalyst for N2H4 decomposition, our mission electron microscopy (HAADF-STEM) images were taken
monodisperse Ni3Fe nanospheres did not require any special on a JEOL-2100F microscope using an electron probe (0.5 nm
treatment to remove the surfactants. Their turnover frequency diameter), at a diffraction camera length of 10 cm. The energy
(TOF) was 9.26 mol of H2 (mol of Ni3Fe)ꢀ1 minꢀ1, which was dispersive X-ray spectroscopy (EDS) data were acquired using an
better than any of the nickel-based catalysts with 100% H2 Oxford Instrument EDS detector. The measured compositions
selectivity previously reported for the decomposition of N2H4.
were typically averaged from 5–10 spots, with an estimated
systematic error of less than ꢁ2 at%. Infrared spectra were
acquired using a Bruker FTIR spectrometer (EQUINOX55). The
ICP-MS measurements were carried out on a JY2000 Ultrace ICP
Atomic Emission Spectrometer equipped with a JY AS 421 auto
2 Experimental methods
2.1 Preparation of the catalyst
FeCl2 (99.99%), Ni(acac)2 (95%), anhydrous hexane (98.5%), sampler and a 2400 g mmꢀ1 holographic grating. The surface
and BTM (99%) were purchased from Sigma Aldrich, and were electronic states were investigated using X-ray photoelectron
used as received. Tl(i-Bu)3 (95%), palmitic acid (98%), and spectroscopy (XPS; Perkin-Elmer PHI 5000C ESCA, using AlKa
cetylamine (98%) were provided by Changzhen High Tech- radiation). The binding energy (BE) values were calibrated using
nology Limited Company.
C1s ¼ 284.6 eV as a reference. For the in situ XPS measurements,
Ni3Fe nanospheres were synthesized via the co-reduction of 10 mg of the catalyst was xed in a homemade in situ XPS
Ni(acac)2 and FeCl2 in the presence of Tl(i-Bu)3, in a high- reactor cell. Aer drying in an argon atmosphere, the sample
temperature organic solution. In a typical experiment, 30.0 g of was transferred to an analysis chamber, and the XPS spectrum
cetylamine and 10.0 g of palmitic acid were loaded into a three- was recorded. UV-Vis spectra were collected on a UV5800 modal
ꢂ
neck ask and heated to 70 C under an argon stream, under spectrophotometer. For the UV-Vis measurements, 0.5 g of para-
stirring. Then, 0.10 g of Ni(acac)2 and 0.050 g of FeCl2 were dimethylaminobenzaldehyde (PDAB), 25 mL of ethanol, and
6426 | J. Mater. Chem. A, 2013, 1, 6425–6432
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Journal of Materials Chemistry A
2.5 mL of hydrochloric acid (1 mol Lꢀ1) were mixed together as
the chromogenic agent for N2H4. Then, 0.3 mL solutions of 0.5
mol Lꢀ1 hydrous hydrazine, 0.1 mL of chromogenic agent, and
3.4 mL of de-ionized water were added, and the sample solu-
tions were stored for 30 min before measurements. The room
temperature magnetic characterization of the samples was
performed using a vibrating sample magnetometer (MagLab-12,
Oxford). The specic surface area was measured at 77 K with a
Quantachrome Nova 4000 analyzer by using the BJH model.
2.3 Catalytic hydrolysis of hydrazine monohydrate
The catalytic activity of the Ni3Fe/C catalyst in generating
hydrogen from the hydrolysis of hydrazine monohydrate was
determined by measuring its rate of hydrogen generation,
using an on-line gas chromatography system (HP Series 6890),
as reported in our previous work.43–45 Under magnetic stirring,
the reaction was initiated via the introduction of an N2H4$H2O
solution into a three-necked round-bottom ask containing
the catalyst, which was prepared as described above. The gas
released during the reaction was passed through a trap con-
taining 15 mL of a standardized solution of 0.001 mol Lꢀ1 of
HCl at room temperature, to ensure the absorption of
ammonia if it was produced. Aer the gas generation ceased,
the resulting solution was titrated with a standard solution of
0.01 mol Lꢀ1 of NaOH, using phenolphthalein as an acid–base
indicator. The quantity of the liberated ammonia gas was
calculated from the difference of HCl concentration between
the two HCl solutions before and aer the reaction. For mass
spectral analysis, the acid trap was not used. The recycled
catalyst was collected using a magnetic separation technique,
and then washed thoroughly (at least three times) with water
and absolute ethanol.
Fig. 1 (a) TEM image; (b) high-resolution TEM image; (c) selected-area electron
diffraction pattern of the 8.1 nm Ni3Fe nanospheres prepared at the optimized
conditions in this work.
The XPS spectra (Fig. S3†) for the Ni3Fe nanospheres further
conrmed the coexistence of metallic Ni and Fe. As shown in
Fig. S4 and S5,† Fe and Ni were present in the metallic state,
with binding energies of 706.6 (Fe2p3/2), 720.0 (Fe2p1/2), 852.6
(Ni2p3/2), and 870.1 (Ni2p1/2) eV, respectively.47 No oxidation of
metallic Fe and Ni was observed. Compared to those standard
peaks of Fe (Fig. S4a†) or Ni (Fig. S5a†) metals,47 the Fe 2p3/2
peak of Ni3Fe shied to lower binding energy by ꢃ0.2 eV while
the Ni 2p3/2 feature shied to higher binding energy by ꢃ0.3 eV,
respectively. It further conrmed the formation of Ni3Fe alloy.
However, the binding energy shi was in contrast to that pre-
dicted by bulk electronegativities. The bulk electronegativity of
Fe (1.83) is lower than that of Ni (1.91) and therefore charge
transfer from Fe to Ni is anticipated.48 Accordingly the binding
energy of Fe should increase while that of Ni should decrease in
3 Results and discussion
Fig. 1a shows a typical low-magnication TEM image of Ni3Fe Ni3Fe. The different result obtained in our work is possibly
NPs prepared under optimized conditions from the hexane attributed to the formation of a Ni–Fe bond in the Ni3Fe alloy
dispersion; the particles had an average particle size of 8.1 nm. leading to a gain in electron density by Fe, which has the larger
These Ni3Fe nanospheres were highly monodisperse, with a fraction of empty states in the valence band (3d64s2) than Ni
homogeneous size distribution (Fig. S1a in the ESI†). The (3d84s2).48 Similar differences have been observed in other
HRTEM image of a selected single Ni3Fe nanosphere (Fig. 1b) bimetallic systems previously.48–50Using the depth prole curves
revealed a lattice spacing of 0.205 nm, which was consistent obtained from the XPS studies, we further analyzed the varia-
with the lattice constant for the (111) planes of the face-centered tion of Ni and Fe XPS signals as we sliced the nanoparticle layer
(fcc) Ni3Fe alloys.46 The SAED pattern for these nanospheres by layer using the Ar+ ion beam. The TEM data showed that the
(presented in Fig. 1c) conrmed their single-crystalline nature. average particle size was approximately 8 nm, and we used this
Fig. S1b (in the ESI†) presents an XRD pattern recorded for the knowledge to calibrate the ion dose in terms of particle size
Ni3Fe nanospheres. The peaks were attributed to well-dened along the beam axis. Using geometrical considerations and
fcc-Ni3Fe nanospheres with the space group Fm3m, conrming ignoring the preferential sputtering and inelastic mean free
that they possessed a Ni3Fe lattice structure in the Ni-displaced path of the core electrons, we qualitatively estimated the
fcc phase. No diffraction signals for pure Fe and/or pure Tl were composition ratio by assuming a radial dependence of the
detected in the XRD patterns. This suggested that only a single density51,52 of each element. The composition (relative%) of
phase of Ni3Fe existed in the nanospheres. The coexistence of Ni each element as calculated from the corresponding atomic
and Fe in the nanoparticles was conrmed by the cross- scattering factors is shown in Fig. S6.† It can be seen that the
sectional compositional line analyses in the HAADF-STEM nanoparticle was richer in Ni at 0 nm, which was possibly due to
experiments (Fig. S2a and b†), and the EDS measurements the fact that the reduction of Ni2+ was faster than that of Fe2+
performed at different points (Fig. S2c–e†).
,
because of the higher reduction potential of Ni2+/Ni (ꢀ0.26 V)
This journal is ª The Royal Society of Chemistry 2013
J. Mater. Chem. A, 2013, 1, 6425–6432 | 6427

View Article Online
Journal of Materials Chemistry A
Paper
compared with Fe2+/Fe (ꢀ0.44 V), which led to the preferential decomposition of Tl(i-Bu)3 alone, under reaction conditions
nucleation of Ni over Fe. The average stoichiometry of the identical to those used to prepare the Ni3Fe. However, we
nanoparticle was 3 : 1. This agreed with the previous atomic obtained only a transparent solution with no traces of solid
ratio that was recorded along the cross-sectional compositional material.
line of Ni and Fe in the HAADF-STEM image (Fig. S2†).
Palmitic acid and cetylamine were indispensable for the
It was revealed that the presence of Tl(i-Bu)3 was the key in preparation of the Ni3Fe nanospheres in our work (Fig. 2b and
controlling the shape of the Ni3Fe. Without Tl(i-Bu)3, Ni3Fe NPs c). To verify this, several tests were designed and carried out.
could still be generated, but their morphology appeared to be When cetylamine was used as the single reaction solvent, it did
much less controlled (Fig. 2). Generally, the formation of NPs not generate NPs with good shape control (Fig. 2b), and when
consists of a nucleation stage and a subsequent Ostwald palmitic acid was used alone as the solvent, it yielded a
ripening process.53,54 The rates of nucleation and the subse- precipitate that could not be re-dispersed in any solvent
quent NP growth are the keys to controlling the shape.53–55 Less (Fig. 2c). Ni3Fe nanospheres could not be prepared in the
shape-controlled products are oen obtained, due to the absence of either of the solvents. To distinguish the capping
nucleation burst that consumes most precursors and results in agent from the stabilizing agent—that is, to determine which
insufficient feedstock for the growth stage.54 To obtain shape- one chemically capped the surface of Ni3Fe, and which one
controlled Ni3Fe NPs, we chose to control the feedstock for their temporarily stabilized the Ni3Fe colloids—the Fourier-trans-
Ostwald ripening growth via the introduction of a foreign form infrared spectrum (FT-IR) of the as-prepared Ni3Fe nano-
element, thallium (Tl).56 Under our reaction conditions, Tl did spheres was measured. Fig. 3 shows an FT-IR spectrum for the
not alloy with Fe and Ni;57 this was conrmed in the EDS as-prepared Ni3Fe nanospheres coated with the organic species,
(Fig. S2c–e†) and XPS (Fig. S3c†) analyses. The Tl that resulted providing qualitative information on the molecules that capped
from the decomposition of Tl(i-Bu)3 was therefore able to help the surface of the nanospheres. For comparison, the spectra for
in reducing the Ni(acac)2 to Ni atoms (or seeds), due to its lower the as-used cetylamine and palmitic acid are also included;
redox-potential compared with Ni; this led to the rapid nucle- these were in good agreement with the reported data.58 The
ation of Ni in the early stages of the reaction. Thus, the resultant spectrum recorded for Ni3Fe was similar to that measured for
Tl cations accumulated in a relatively high concentration, which cetylamine, but not that measured for palmitic acid. For
decelerated the subsequent metallic particle growth in the example, a band at 1700 cmꢀ1 for palmitic acid that was
following reaction:
attributed to the carbonyl C]O stretching mode58 was not
observed in the spectrum of the Ni3Fe nanospheres, whereas
the bands at 1582 and 1660 cmꢀ1 that were assigned to the N–H
deformation vibration of a primary amine58 appeared in the
spectra for both the cetylamine and the nanospheres. These
observations suggested that the organic residues that capped
the Ni3Fe nanospheres consisted only of cetylamine, although
both cetylamine and palmitic acid were used in the preparation.
Another observation supporting this conclusion was that the
addition of cetylamine stopped the aggregation of newly
prepared Ni3Fe NPs, whereas the addition of palmitic acid did
not. We therefore believe that cetylamine may act as a capping
agent that coordinates to the Ni3Fe surface sites, slowing but
not stopping the NPs’ growth. During a nanoparticle’s growth in
Ni2+ + 2Tl0 / Ni0 + 2Tln+
.
(3)
In other words, the Ni precursor-Tl system acted as a
“buffer”, ensuring a steady growth of particles with a sufficient
feedstock. The “slowdown” of the nucleation of Ni favored the
evolution of the Ni3Fe NPs, because the Ni clusters could be
steadily and continuously provided for the Ostwald growth in
this case. As a result, the total surface energies for different
growth directions were tuned, and nanospheres with a unique
component distribution developed. An additional piece of
evidence supporting this hypothesis was the observation of
Tl4f5/2 and Tl4f7/2 peaks around 123.0 and 118.5 eV, respec-
tively, in the XPS spectra of the reaction residue, which was a
sign of the Tl+ chemical state47 (Fig. S7†). In another set of blank
experiments, we also attempted to prepare pure Tl NPs via the
Fig. 2 TEM images of the as-prepared Ni3Fe: (a) without Tl(Me)3; (b) without
Fig. 3 FT-IR spectra of (a) cetylamine; (b) palmitic acid; (c) the as-prepared Ni3Fe
palmitic acid; (c) without cetylamine.
nanospheres at the optimized conditions in this work.
6428 | J. Mater. Chem. A, 2013, 1, 6425–6432
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Journal of Materials Chemistry A
solution, the development of the shape of the particle depends Ni(acac)2 ¼ 2 led to only a small increase in the Fe molar ratio in
on the competition between the growth rates in the various the product, to Fe/Ni ¼ 0.35, when the synthesis was carried out
growth directions.52 The growth rate can be determined by the at 200 ꢂC. When the molar ratio of BTM/(Ni + Fe) was increased
total surface energy, for which the binding energy between a to 2 : 1, the Ni molar ratio only increased to Ni/Fe ¼ 3.17.
ligand and a given growth direction is also a contribution.
Furthermore, empirically, FeCl2 gave better quality Ni3Fe
Clearly, a surface-selective capping agent could alter the NPs than the other two anhydrous iron precursors [FeC2O4 and
morphology of NPs. We believe that cetylamine offers such a Fe(acac)2] under the current synthetic conditions (Fig. S8†). The
function, lowering the total surface energy of all growth direc- particle size of the Ni3Fe catalysts could be tuned by varying the
tions on Ni3Fe NPs. As indicated in Fig. 4, the presence of pal- molar ratio of Tl(i-Bu)3/(Ni + Fe) (Fig. S9 and Table S1†).
mitic acid with an optimized palmitic acid–cetylamine mass Surprisingly, Tl(i-Bu)3 had no effect on the composition of the
ratio was an essential condition to fabricate the nanospheres. nanospheres (Table S1†). It was therefore concluded that
To achieve well-dened Ni3Fe nanospheres, the best ratio the optimized conditions for producing the well-dened
between cetylamine and palmitic acid was found to be 4 : 1 8.1 nm Ni3Fe nanospheres were: keep the temperature at 200 ꢂC
(molar) (Fig. 4). Based on the above discussion, we believe that for 30 min, with a molar ratio of cetylamine–palmitic acid ¼ 4,
in our system, cetylamine is an excellent capping agent for Ni/Fe precursors ¼ 1, BTM/(Ni + Fe) ¼ 1.3, and Tl(i-Bu)3/(Ni +
developing Ni3Fe nanospheres, whereas palmitic acid acts as a Fe) ¼ 0.1.
stabilizing agent to tune the capping abilities of cetylamine
through the acid–base equilibrium.
The catalytic activity of the 8.1 nm Ni3Fe NPs prepared under
optimized conditions toward the hydrolytic dehydrogenation of
The effects of reaction parameters—including the reaction N2H4 was evaluated. They were supported on carbon (Carbot,
temperature, reaction time, and molar ratio of Ni/Fe precursors, USA), and were employed as a catalyst without any further
i.e., BTM/(Ni + Fe)—on the sphere percentage, average diam- treatment.42 A TEM image of the Ni3Fe/C catalyst (Fig. 5a and b)
eter, and chemical composition of the Ni3Fe NPs are presented shows that the Ni3Fe NPs preserved their size aer they were
in Table S1.† Elemental analyses revealed that none of the Ni3Fe dispersed on the carbon support. From Fig. 6, it can be seen that
NPs showed any boron contamination. From Table S1,† it can our Ni3Fe/C catalyst exhibited high activity, completing the
be seen that the reaction temperature had a strong inuence on hydrazine decomposition in only 27 min at room temperature,
the chemical composition of the resultant nanospheres. which was much faster (as much as 181%) than that of the most
Generally, increases in the reaction temperature resulted in active Ni-related catalysts with 100% H2 selectivity reported for
increases in the Fe content in the NPs. To achieve nanospheres this reaction.24 The H2/N2 ratio remained almost constant as the
with an Ni : Fe molar ratio of 3 : 1, the reaction temperature was reaction proceeded (Fig. 6), which indicated that the competing
set to 200 ꢂC, or slightly higher. Furthermore, the composition reactions (1) and (2) proceeded over Ni3Fe/C simultaneously.
of nanospheres could be changed slightly by varying the feed The decomposition of hydrazine over Ni3Fe/C can therefore be
ratio between the Ni and Fe precursors and the molar ratio of expressed by eqn (4)
BTM/(Ni + Fe). However, the inuence was not very signicant.
For example, increasing the FeCl2 input molar ratio to FeCl2/
3N2H4 / (1 + 2x)N2 + 6xH2 + 4(1 ꢀ x)NH3,
(4)
where x is the H2 selectivity.43–45 The 100% H2 selectivity and the
reaction completeness were further conrmed using mass
spectrometry (no NH3, and H2/N2 ¼ 2.0, Fig. S10†) and UV-Vis
spectroscopy (no N2H4 aer reaction, Fig. S11†), respectively.
Fig. S12† shows a plot of the hydrogen generated during the
decomposition hydrolysis of N2H4 as a function of time, for
different catalyst concentrations. A linear hydrogen evolution
was observed. The reaction time was dependent on the catalyst
concentration; the higher the catalyst or substrate concentra-
tion, the shorter the reaction time. Fig. 7 presents a plot of the
Fig. 4 TEM images of Ni3Fe synthesized by varying the molar ratio of cetyl-
amine–palmitic acid: (a) 1, (b) 2, (c) 3, (d) 4, and (e) 5, respectively.
Fig. 5 (a) TEM image and (b) enlarged TEM image of the 8.1 nm Ni3Fe nano-
spheres supported on carbon.
This journal is ª The Royal Society of Chemistry 2013
J. Mater. Chem. A, 2013, 1, 6425–6432 | 6429

View Article Online
Journal of Materials Chemistry A
Paper
Fig. 8 Plot of ln[N2H4] versus ln RH2 during the hydrazine decomposition over
Ni3Fe/C at different N2H4 concentrations ([Ni3Fe] ¼ 8 mmol Lꢀ1, T ¼ 20 ꢁ 1 ꢂC).
Fig. 6 n(H2 + N2)/n(N2H4) and H2/N2 molar ratio versus time for Ni3Fe/C during
the decomposition of hydrous hydrazine at room temperature.
performance of Ni3Fe/C (Fig. S15†). In other words, the addition
of NaOH was unnecessary for 100% H2 selectivity in our work.
The enhanced performance of our catalyst stemmed from the
small size and monodispersity of the particles, which led to a
high specic surface area (182.3 m2 gꢀ1) and a better dispersion
on the carbon support. Furthermore, the aggregated Ni3Fe with
40.7 m2 gꢀ1 specic surface area supported on C only exhibits a
hydrogen generation rate of 0.42 mL minꢀ1 with 80.3% H2
selectivity under the same conditions (Table S2†), because of its
poor dispersion. The catalyst lifetime of Ni3Fe/C was also
investigated. Over a period of 30 h and at 20 ꢂC, it provided
15 840 turnovers in the decomposition of N2H4. The average
TOF value (ATOF) was calculated as 8.80 minꢀ1, which was
95.0% of the initial TOF. The hydrogen generation slowed down
as the reaction proceeded, and ultimately stopped. ICP analysis
revealed that the leaching of Ni and Fe species was <1.0 ppm;
this ruled out the loss of active phases. Moreover, TEM
(Fig. S16a and b†) and XRD (Fig. S17a†) results indicated that no
aggregation and no structural changes occurred for Ni3Fe/C
aer repetitive use at room temperature. It appeared that the
oxidation of Ni and Fe species on the active surface of Ni3Fe/C
was responsible for the deactivation, as shown by XPS
(Fig. S18a†). The oxidized Ni3Fe lm displayed a decrease in the
catalytic activity of Ni3Fe/C, due to blocking of the active sites.44
hydrogen generation rate (RH2) versus the Ni3Fe concentration
([Ni3Fe]) on a logarithmic scale. A slope of 0.97 indicated that
the hydrolysis catalyzed by the Ni3Fe/C catalyst was rst-order in
the catalyst concentration, but zero-order in the N2H4 concen-
tration ([N2H4]) (Fig. S13† and 8).
Fig. S14† shows the volume of hydrogen generated during
the decomposition of N2H4 as a function of time, at various
temperatures in the range of 20–60 ꢂC. The rate constant k was
calculated at different temperatures from the slope of the linear
part of each plot shown in Fig. 9. The activation energy was
calculated to be 48.1 ꢁ 2.0 kJ molꢀ1 from the Arrhenius plot (ln
k vs. (1/T)) (Fig. 9). This compared favorably with the reported
values of 46.7 ꢁ 1.3 kJ molꢀ1, 49.3 ꢁ 3.2 kJ molꢀ1, and 54.3 ꢁ 2.7
kJ molꢀ1 on Co–B honeycomb, Fe–B/MWCNT, and Ni–Al2O3
structures, respectively.44,45,59
Activities in terms of TOF values (mol of H2 ꢄ (mol of catalyst
min)ꢀ1) are compared in Table S2† for previously reported
catalysts19–24,43–45,59–61 that have been tested in the decomposition
of N2H4. Our Ni3Fe/C catalyst had the highest activity among all
of the nickel-based catalysts with 100% H2 selectivity.19–24
Meanwhile, the activity and H2 selectivity of our Ni3Fe nano-
spheres are higher than those of Ni3Fe prepared by Singh et al.23
(Table S2†). NaOH has been reported as a promoter for the
reaction activity of N2H4 decomposition.23,24 In our case, the
presence of NaOH had no inuence on the catalytic
Fig. 9 Plot of ln k versus 1/T during the hydrazine decomposition over Ni3Fe/C
at different temperatures in the range 20–60 ꢂC ([Ni3Fe] ¼ 8 mmol Lꢀ1, [N2H4] ¼
0.5 mol Lꢀ1).
Fig. 7 Plot of ln[Ni3Fe] versus ln RH2 during the hydrazine decomposition over
Ni3Fe/C at different catalyst concentrations ([N2H4] ¼ 0.5 mol Lꢀ1, T ¼ 20 ꢁ 1 ꢂC).
6430 | J. Mater. Chem. A, 2013, 1, 6425–6432
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Journal of Materials Chemistry A
These results indicated that the oxidation of Ni and Fe on the membrane fuel cell (PEMFC) at 100% H2 use.62,63 Therefore, a
Ni3Fe active surfaces was the main reason for the deactivation of reactor lled with 1.0 g of our prepared Ni3Fe/C might produce
Ni3Fe during N2H4 hydrolysis. Because the peaks for oxidized Ni more than 1.79 L of hydrogen in 1 min during the hydrolysis of
and Fe species were not observed in the XRD (Fig. S17a†), we N2H4. In other words, it could supply H2 for a 290 W PEMFC at
inferred that their structure was amorphous, or they were 100% H2 use via the hydrolysis of N2H4.
present in amounts below the detection limit of the XRD
equipment. Acid washing is an effective approach to reactivate
catalysts deactivated by surface oxidation. However, acid
4 Conclusions
washing greatly shortened the catalyst lifetime of the Ni3Fe/C In summary, we have reported the facile synthesis of mono-
materials, due to the leaching of Ni and Fe species and the disperse Ni3Fe single-crystalline nanospheres with 182.3 m2 gꢀ1
destruction of the structure. Recently, solution plasma tech- specic surface area via the co-reduction of Ni(acac)2 and FeCl2
niques (SPP) have been shown to be powerful approaches for with Tl(i-Bu)3 and BTM complexes in the presence of cetylamine
the reactivation of catalysts.44 In this work, we attempted to and palmitic acid. When supported on carbon, the Ni3Fe
reactivate Ni3Fe/C using the solution plasma process in the nanospheres were highly active for the decomposition of N2H4,
presence of potassium chloride (KCl). Surprisingly, the reac- even at low catalyst and substrate concentrations and at room
tivated Ni3Fe/C exhibited the same activity as the fresh sample temperature. Such Ni3Fe NP-based catalysis under mild reac-
(Fig. S19†). As shown by the XPS results, the oxidized Ni and Fe tion conditions represents a promising step toward the practical
were reduced (Fig. S18b†). It is believed that the electrons yiel- development of N2H4 as a feasible hydrogen storage medium
ded during the SPP were the main factor inuencing the for fuel cell applications.
reduction of Ni and Fe oxides on the Ni3Fe surface. This process
can be expressed as follows:
Acknowledgements
(Ni3Fe)xOy + 2yeꢀ + yH2O / x(Ni3Fe) + 2yOHꢀ.
(5)
This work was nancially supported by the Cultivating pro-
gramme of Middle-aged backbone teachers (HG0092), the
Cultivating programme for Excellent Innovation Team of
Chengdu University of Technology (HY0084) and Innovative
Experimental Items for College Students of Sichuan Province
(SZH1106CX04).
However, the real mechanism responsible for the reac-
tivation of Ni3Fe by SPP requires further study; this work is
currently underway. There plasma treatment had no signicant
effects on the morphology (Fig. S16c and d†) or the composition
of the Ni3Fe/C aer reactivation (Fig. S20†). In other words, SPP
is a potential pathway to reactivate catalysts deactivated by
surface oxidation.
Notes and references
The easy magnetic separation of Ni3Fe/C is very important
for its lifetime in solution systems.43–45,59 The magnetic proper-
ties of the sample are shown in Fig. 10. For Ni3Fe/C, the typical
coercivity (Hc), saturation magnetization (Ms), and remnant
magnetization (Mr) at 298 K were 180.8 Oe, 45.3 emu gꢀ1, and
9.8 emu gꢀ1, respectively. These magnetic properties enabled
the effective separation of the Ni3Fe/C using a magnet (the inset
in Fig. 10) in a solution system.43–45
1 L. Schlapbach and A. Zuttel, Nature, 2001, 414, 353.
2 P. Chen, Z. Xiong, J. Luo, J. Lin and K. L. Tan, Nature, 2002,
420, 302.
3 W. Grochala and P. P. Edwards, Chem. Rev., 2004, 104, 1283.
4 N. L. Rosi, J. Eckert, M. Eddaoudi, D. T. Vodak, J. Kim,
M. O’Keeffe and O. M. Yaghi, Science, 2003, 300, 1127.
5 A. Gutowska, L. Li, Y. Shin, C. M. Wang, X. S. Li,
J. C. Linehan, R. S. Smith, B. D. Kay, B. Schmid, W. Shaw,
M. Gutowski and T. Autrey, Angew. Chem., Int. Ed., 2005,
44, 3578.
According to previous research, a hydrogen generation rate
of 1.0 L minꢀ1 can supply H2 for a 162 W proton exchange
6 C. W. Hamilton, R. T. Baker, A. Staubitz and I. Manners,
Chem. Soc. Rev., 2009, 38, 279.
7 H. L. Jiang, S. K. Singh, J. M. Yan, X. B. Zhang and Q. Xu,
ChemSusChem, 2010, 3, 541.
8 T. He, H. Wu, G. Wu, J. Wang, W. Zhou, Z. Xiong, J. Chen,
T. Zhang and P. Chen, Energy Environ. Sci., 2012, 5,
5686.
9 J. Graetz, Chem. Soc. Rev., 2009, 38, 73.
10 A. N. Rollinson, J. Jones, V. Dupont and M. V. Twigg, Energy
Environ. Sci., 2011, 4, 1216.
11 G. M. Wang, Y. C. Ling, X. H. Lu, H. Y. Wang, F. Qian,
Y. X. Tong and Y. Li, Energy Environ. Sci., 2012, 5, 8215.
12 E. W. Schmidt, Hydrazine and its Derivatives: Preparation,
Properties, Applications, John Wiley & Sons, New York, 2nd
edn, 1984.
Fig. 10 Hysteresis loop and magnetic properties of Ni3Fe/C. The inset is the
photograph of Ni3Fe/C after magnetic separation.
13 J. Prasad and J. L. Gland, Langmuir, 1991, 7, 722.
This journal is ª The Royal Society of Chemistry 2013
J. Mater. Chem. A, 2013, 1, 6425–6432 | 6431

View Article Online
Journal of Materials Chemistry A
Paper
14 R. R. Schrock, T. E. Glassman, M. G. Vale and M. Kol, J. Am. 38 Z. Peng, J. Wu and H. Yang, Chem. Mater., 2010, 22,
Chem. Soc., 1993, 115, 1760. 1098.
15 J. B. O. Santos, G. P. Valenc and J. A. J. Rodrigues, J. Catal., 39 S. Peng, Y. Lee, C. Wang, H. Yin, S. Dai and S. Sun, Nano Res.,
2002, 210, 1. 2008, 1, 229.
16 S. H. Wu and D. H. Chen, J. Colloid Interface Sci., 2003, 259, 40 V. Mazumder and S. Sun, J. Am. Chem. Soc., 2009, 131, 4588.
282; Y. Nakajima, A. Inagaki and H. Suzuki, Organometallics, 41 F. Zhao, M. Rutherford, S. Y. Grisham and X. Peng, J. Am.
2004, 23, 4040.
17 X. Chen, T. Zhang, M. Zheng, Z. Wu, W. Wu and C. Li, 42 O. Metin, V. Mazumder, S. Ozkar and S. Sun, J. Am. Chem.
J. Catal., 2004, 224, 473. Soc., 2010, 132, 1468.
18 H. Gu, R. Ran, W. Zhou, Z. Shao, W. Jin, N. Xu and J. Ahn, 43 D. G. Tong, X. L. Zeng, W. Chu, D. Wang and P. Wu, Mater.
J. Power Sources, 2008, 177, 323. Res. Bull., 2010, 45, 442.
19 S. K. Singh, X. B. Zhang and Q. Xu, J. Am. Chem. Soc., 2009, 44 D. G. Tong, W. Chu, P. Wu and L. Zhang, RSC Adv., 2012, 2,
Chem. Soc., 2009, 131, 5350.
¨
¨
131, 9894.
2369.
20 S. K. Singh and Q. Xu, J. Am. Chem. Soc., 2009, 131, 18032.
21 S. K. Singh and Q. Xu, Inorg. Chem., 2010, 49, 6148.
22 S. K. Singh and Q. Xu, Chem. Commun., 2010, 46, 6545.
45 D. G. Tong, W. Chu, P. Wu, G. F. Gu and L. Zhang, J. Mater.
Chem. A, 2013, 1, 358.
46 Willams, Am. Mineral., 1960, 45, 450.
23 S. K. Singh, A. K. Singh, K. Aranishi and Q. Xu, J. Am. Chem. 47 Reference: Manual for Operator for PHI PC Windows Soware
Soc., 2011, 133, 19638. Version 1.2b, Physical Electronic Division, Perkin-Elmer.
24 J. Wang, X. B. Zhang, Z. L. Wang, L. M. Wang and Y. Zhang, 48 J. A. Rodriguez, R. A. Campbell and D. W. Goodman, J. Phys.
Energy Environ. Sci., 2012, 5, 6885. Chem., 1991, 95, 5716.
25 C. Wang, H. Daimon, Y. Lee, J. Kim and S. Sun, J. Am. Chem. 49 J. A. Rodriguez, R. A. Campbell and D. W. Goodman, Surf.
Soc., 2007, 129, 6974. Sci., 1994, 307–309, 377.
26 Y. Zhang, M. E. Grass, J. N. Kuhn, F. Tao, S. E. Habas, 50 Z. J. Wang, F. Yang, S. Axnanda, C. J. Liu and
W. Huang, P. Yang and G. A. Somorjai, J. Am. Chem. Soc.,
2008, 130, 5868.
27 J. N. Kuhn, W. Huang, C.-K. Tsung, Y. Zhang and
G. A. Somorjai, J. Am. Chem. Soc., 2008, 130, 14026.
28 D. Xu, Z. Liu, H. Yang, Q. Liu, J. Zhang, J. Fang, S. Zou and
K. Sun, Angew. Chem., Int. Ed., 2009, 48, 4217.
D. W. Goodman, Appl. Catal., A, 2011, 391, 342.
51 I. Tunc, S. Suzer, M. A. Correa-Duarte and L. M. Marzan,
J. Phys. Chem. B, 2005, 109, 7597.
52 D. Q. Yang, J. N. Gillet, M. Meunier and E. Sacher, J. Appl.
Phys., 2005, 97(2), 024303.
53 Y. Xiong and Y. Xia, Adv. Mater., 2007, 19, 3385.
29 D. Xu, S. Bliznakov, Z. P. Liu, J. Y. Fang and N. Dimitrov, 54 C. B. Murray, C. R. Kagan and M. G. Bawendi, Annu. Rev.
Angew. Chem., Int. Ed., 2010, 49, 1282. Mater. Sci., 2000, 30, 545.
30 A. X. Yin, X. Q. Min, Y. W. Zhang and C. H. Yan, J. Am. Chem. 55 J. Zhang, K. Sun, A. Kumbhar and J. Fang, J. Phys. Chem. C,
Soc., 2011, 133, 3816. 2008, 112, 5454.
31 J. S. Bradley, B. Tesche, W. Busser, M. Maase and M. T. Reetz, 56 J. Zhang, H. Yang, J. Fang and S. Zou, Nano Lett., 2010, 10,
J. Am. Chem. Soc., 2000, 122, 4631. 638.
32 C. Pan, K. Pelzer, K. Philippot, B. Chaudret, F. Dassenoy, 57 Y. Y. Lu and W. Q. Zeng, Inorg. Chem.: Properties and
P. Lecante and M.-J. Casanove, J. Am. Chem. Soc., 2001, 123,
7584.
33 G. V. Chakroune, S. Ammar, L. Poul, D. Veautier,
M. M. Chemini, C. Mangeney, F. N. Villain and F. Fievet,
Langmuir, 2005, 21, 6788.
Applications of Thallium, Science Press, Beijing, 1990.
58 X. Y. Jin, S. L. Chen and E. N. Yao, Application Guide of
Infrared Spectra, Science and Technology, Press of Tian Jin,
Tian Jin, 1992.
59 L. He, Y. Huang, A. Wang, X. Wang, X. Chen, J. J. Delgado
and T. Zhang, Angew. Chem., Int. Ed., 2012, 51, 6191.
34 T. Tsukatani and H. Fujihara, Langmuir, 2005, 21, 12093.
35 J. Park, E. Kang, S. U. Son, H. M. Park, M. K. Lee, J. Kim, 60 S. K. Singh, Y. Iizuka and Q. Xu, Int. J. Hydrogen Energy, 2011,
K. W. Kim, H.-J. Noh, J.-H. Park, C. J. Bae, J.-G. Park and
T. Hyeon, Adv. Mater., 2005, 17, 429.
36 H. Wang, X. Jiao and D. Chen, J. Phys. Chem. C, 2008, 112,
36, 11794.
61 L. He, Y. Q. Huang, A. Q. Wang, Y. Liu, X. Y. Liu, X. W. Chen,
J. J. Delgado, X. D. Wang and T. Zhang, J. Catal., 2013, 298, 1.
62 Y. Bai, C. Wu, F. Wu and B. Yi, Mater. Lett., 2006, 60, 2236.
18793.
¨
¨
37 O. Metin, S. Sahin and S. Ozkar, Int. J. Hydrogen Energy, 2009, 63 N. Patel, R. Fernandes and A. Miotello, J. Power Sources,
34, 6304.
2009, 188, 411.
6432 | J. Mater. Chem. A, 2013, 1, 6425–6432
This journal is ª The Royal Society of Chemistry 2013