Towards optimizing cobalt based metal oxide nanocrystals for hydrogen generation via NaBH4 hydrolysis

Article abstract of DOI:10.1016/j.apcata.2019.117303

Here, we evaluate H₂ generation, via NaBH₄ hydrolysis, for a rationally designed library of cobalt based metal oxide nanocrystals. Specifically, we developed highly surface active wüsite cobalt oxide nanorods (wz-CoO-NRs) with varied dimensions through the thermal decomposition of cobalt oleate (Co-OL) and cobalt stearate (Co-ST) – yielding multiple ‘nail’ structures grown along (002) direction. The length of wz-CoO-NRs (from 30 to 200 nm) was controlled by the ratio of Co-OL to Co-ST, temperature ramp rate, and reaction time. Further, by adjusting oleic acid concentration (as a surfactant) and/or (other) metal co-precursors (e.g. iron oleate (Fe-OL) and nickel oleate (Ni-OL)), shape (e.g. nanorod to urchin-like, irregular cube, flower, or sphere), crystal structure(s), and composition (from wz-CoO to cubic type CoO (cu-CoO) or ferrite metal oxides (Fe₃O₄, CoFe₂O₄, and Co_0.5Ni_0.5Fe₂O₄)) were serially tuned. The library of cobalt based metal oxide nanocrystals was compared, as catalysts, for H₂ generation via NaBH₄ decomposition, and performance was observed to be a function of crystal length for wz-CoO-NRs. A maximum hydrogen production rate of 10,367 mL min^?1 g^?1 at 293 K with an activation energy of 27.4 kJ mol^-1 was observed for 200 nm wz-CoO-NRs, which is among the lowest reported activation energies for this reaction. Enhanced reactivity is likely due to increased favorable surface chemistries (multiple catalytically active (002) sites on the edge of multiple nails) and lowered charge transfer energy for longer (e.g. 200 nm) wz-CoO-NRs.

Full text of DOI:10.1016/j.apcata.2019.117303

AppliedCatalysisA,General589(2020)117303
Contents lists available at ScienceDirect
Applied Catalysis A, General
journal homepage: www.elsevier.com/locate/apcata
Towards optimizing cobalt based metal oxide nanocrystals for hydrogen
generation via NaBH4 hydrolysis
Changwoo Kima,1, Seung Soo Leeb,1, Wenlu Lia, John D. Fortnera,*
a Department of Chemical and Environmental Engineering, Yale University, New Haven, CT, 06520, United States
b Department of Energy, Environmental, and Chemical Engineering, Washington University in St. Louis, St. Louis, MO, 63130, United States
A R T I C L E I N F O
A B S T R A C T
Keywords:
Here, we evaluate H2 generation, via NaBH4 hydrolysis, for a rationally designed library of cobalt based metal
oxide nanocrystals. Specifically, we developed highly surface active wüsite cobalt oxide nanorods (wz-CoO-NRs)
with varied dimensions through the thermal decomposition of cobalt oleate (Co-OL) and cobalt stearate (Co-ST)
– yielding multiple ‘nail’ structures grown along (002) direction. The length of wz-CoO-NRs (from 30 to 200 nm)
was controlled by the ratio of Co-OL to Co-ST, temperature ramp rate, and reaction time. Further, by adjusting
oleic acid concentration (as a surfactant) and/or (other) metal co-precursors (e.g. iron oleate (Fe-OL) and nickel
oleate (Ni-OL)), shape (e.g. nanorod to urchin-like, irregular cube, flower, or sphere), crystal structure(s), and
composition (from wz-CoO to cubic type CoO (cu-CoO) or ferrite metal oxides (Fe3O4, CoFe2O4, and
Co0.5Ni0.5Fe2O4)) were serially tuned. The library of cobalt based metal oxide nanocrystals was compared, as
catalysts, for H2 generation via NaBH4 decomposition, and performance was observed to be a function of crystal
length for wz-CoO-NRs. A maximum hydrogen production rate of 10,367 mL min−1 g−1 at 293 K with an acti-
vation energy of 27.4 kJ mol-1 was observed for 200 nm wz-CoO-NRs, which is among the lowest reported
activation energies for this reaction. Enhanced reactivity is likely due to increased favorable surface chemistries
(multiple catalytically active (002) sites on the edge of multiple nails) and lowered charge transfer energy for
longer (e.g. 200 nm) wz-CoO-NRs.
H2 generation
NaBH4 hydrolysis
Cobalt oxide
Cobalt iron oxide
Cobalt ferrite
1. Introduction
occurring (e.g. without CO and CO2), resulting in high purity H2
[10,11]. Due to high thermodynamic stability and relative slow kinetics
Hydrogen (H2) is a zero-emission fuel stock and its generation has
been widely studied. One promising technology utilizes hydrides as
effective storage materials, including LiAlH4, LiBH4, NaBH4, MgH2,
NaAlH4, and CaH2, which can be rapidly decomposed to produce H2 in
the presence of appropriate catalysts [1–7]. Among these, NaBH4 is a
promising material due to its storage capacity of 10.8 wt% at 298 K,
which is higher than LiAlH4, MgH2, NaAlH4, and CaH2. NaBH4 is also
relatively thermally stable (H2 release occurs by thermolysis when he-
ated > 673 K), chemically stable under basic conditions (> pH 13), and
highly water soluble (solubility of NaBH4 is 55 g/100 g at 298 K), thus
allowing for safe handling [1,8,9]. NaBH4 decomposition via hydrolysis
is shown in Eq. (1).
(at higher, stabilizing pH values), for practical applications, NaBH4
requires suitable catalysts for controlled H2 production [12,13].
Transition metal and metal oxide nanocrystals have been previously
studied as effective heterogeneous catalysts to accelerate NaBH4 hy-
drolysis reactions under basic condition due to their unique redox
chemistries and catalytic properties [4,15]. Ruthenium (Ru), rhodium
(Rh), platinum (Pt), nickel (Ni) and cobalt (Co) based catalysts typically
exhibit superior H2 generation (via hydride decomposition reactions)
[4]. For these, platinum based (nano) catalysts have shown the best
performance with an observed rate of 23,000 mL min−1 g−1 when in-
corporated with activated carbon [4,16]. While Pt, Rh, and Ru catalysts
have demonstrated high performance under ambient conditions, their
cost, catalyst life time, and stability are still challenging aspects when
considering scale-up [4]. Compared to Pt, Rh, and Ru based catalysts,
Co based catalysts are attractive due to low cost and active redox
chemistries, with a wide range of oxidation states [4]. As described by
kJ
NaBH4 + 2H2 O→ 4H2 + NaBO2 + Heat (−210
)
(1)
mol
This reaction is exothermic, and has no major side reactions
⁎ Corresponding author.
E-mail address: john.fortner@yale.edu (J.D. Fortner).
1 Authors contributed equally to this work.
https://doi.org/10.1016/j.apcata.2019.117303
Received 2 July 2019; Received in revised form 10 October 2019; Accepted 11 October 2019
Availableonline19October2019
0926-860X/©2019PublishedbyElsevierB.V.

C. Kim, et al.
AppliedCatalysisA,General589(2020)117303
others [16–28], cobalt based hydrogels, cryogels, microgels, nano-, and
carbon materials have been developed for H2 production using NaBH4
(recent studies are summarized in Table A.1). In addition, cobalt oxide
catalysts have been demonstrated as promising catalysts for H2 gen-
eration (recent studies summarized in Table A.2, including this work)
[29–38]. A commercial available Co3O4 catalyst has shown catalytic
properties for H2 generation, via NaBH4 hydrolysis, with a maximum
rate of 860 mL min-1 g-1 at 298 K and 85.2 kJ mol-1 (Arrhenius activa-
tion energy) [29]. Other, highly crystalline Co3O4 catalysts were de-
monstrated with a maximum H2 generation rate of 2500 mL min-1 g-1
(313 K) [31]. Interestingly, faster rates with a lower activation energy
were found when 1-dimensional CoO nanorods (ranging from 50 to
70 nm length) were used, with a maximum rate of H2 generation of
5555 mL min−1 g−1 at 303 K and an activation energy below 53 kJ
mol−1 [39]. Zhang H. et al. also found a low activation energy of
45.94 kJ mol−1 with 6130 mL min−1 g−1 as a maximum H2 generation
rate (303 K) for porous CoO nanorods (length over 500 nm) [40].
Though cost efficient cobalt oxide nanorods have improved the effi-
ciency and reaction rates, rational design of the (nano)catalyst remains
outstanding with regard to surface area/availability, crystallinity, and
redox properties (by controlling the composition of the mixed metal
oxides), among other factors.
Recently, CoO nanorods have been synthesized via decomposition
of organometallic cobalt precursors such as cobalt oleate (Co-OL) or
cobalt stearate (Co-ST) in organic solvents at high temperatures (at
473 K for cobalt acetylacetonate and over 588 K for cobalt oleate and
cobalt stearate as cobalt precursors). Resulting CoO nanorods demon-
strate wüsite type hexagonal (ZnO like) structures, which were not
previously observed for bulk CoO [39,41–43]. However, these methods
are not consistently reproducible regarding product purity (i.e. % of
nanorod in a sample). Recently, Buck M. R. et al. reported the de-
composition of pure Co-OL in an organic solvent (1-octadecene) at
583 K did not yield crystals (Fig. A.1.A); however, when the hydrated
phase of Co-OL (Co(oleate)2-x(OH)x) was included as an additive (ori-
ginally discovered as an impurity), complete decomposition of Co-OL
occurred, resulting in high pure CoO nanorods [44]. Similarly, Co-ST
decomposition in 1-octadecene did not yield high quality, mono-
disprese CoO nanorods; instead, spherical and non-monodisperse cobalt
oxides, with a relatively wide size distribution, were observed in most
2. Materials and methods
2.1. Chemicals
Iron (III) chloride hexahydrate (FeCl3·6H2O, 97%), cobalt chloride
(CoCl2·6H2O, 98%), oleic acid (OA, 90%), and 1-octadecene (90%)
were purchased from Sigma-Aldrich. Cobalt stearate was purchased
from Strem. All nanocrystals were synthesized under nitrogen condition
(N2, 99.999%).
2.2. Synthesis of cobalt, iron, and nickel precursors
Metal oleate precursors were synthesized by reacting metal chloride
(30 mmol of cobalt (II) chloride hexahydrate (CoCl2·6H2O) for cobalt
oleate (Co-OL), 20 mmol of iron (III) chloride hexahydrate (FeCl3·6H2O)
for iron oleate (Fe-OL), and 30 mmol of nickel chloride hexahydrate
(NiCl2·6H2O) for nickel oleate (Ni-OL)) with sodium oleate in ethanol,
water, and hexane mixture (30:20:40 by volume ratio) at 333 K for 6 h
[41,43,44]. The prepared metal precursors were washed with ethanol/
water mixture in a separatory funnel and extracted using hexane. The
extracted metal precursors in hexane were further purified by cen-
trifugation at 11,000 rpm for 30 min to remove remaining metal salts
and sodium oleate. The purified metal precursors were then stored in
hexane and the concentration of each precursor was measured by ICP-
OES before uses.
2.3. Synthesis of cobalt oxide nanonails
One dimensional structured cobalt oxides (CoO-NR) were synthe-
sized via the decomposition of cobalt precursors (the mixture of Co-OL
(0.7 mmol) and Co-ST (0.3 mmol)) in the presence of oleic acid
(0.5 mmol) in 1-octadecene (5 g) at 593 K for 30 min using a controlled
ramp rate (from 1 to 20 K min−1) under inert (nitrogen) condition.
These two cobalt precursors were mixed with oleic acid in 1-octadecene
at room temperature and kept at 393 K for 1.5 h to evaporate hexane
and water. The reaction mixture was then heated to 593 K at a rate of 1
- 20 K min−1 and the reaction was kept at 593 K for 30 min. The length
of CoO-NR was tuned by varying reaction conditions including heating
rate, reaction time, and the ratio of cobalt precursors. Urchin shaped
CoO nanocrystals (CoO nanourchins) were synthesized using high
concentration of oleic acid (3 mmol) with the cobalt precursor mixture
(Co-OL (0.7 mmol) and Co-ST(0.3 mmol)) in 1-octadecene (5 g) at
593 K for 30 min using a slow ramp rate of 1 K min−1 under nitrogen.
The resulting colloidal cobalt oxide nanocrystals (typically 5 mL) were
mixed with 40 mL ethanol in a centrifuge tube and centrifuged at
4000 rpm for 10 min. The purification was repeated in triplicate with
the final step of cobalt oxide nanocrystals stored in hexane.
cases (17.6
4.1 nm in Fig. A.1).
In this research, we demonstrate the complete decomposition of
cobalt oleate (Co-OL), in the presence of cobalt stearate (Co-ST), re-
sulting in wüsite cobalt oxide nanorods (wz-CoO-NRs) with tunable
lengths (from 30 to 200 nm). In addition, size (length), shape and
composition were tuned by the addition of excess oleic acid, as a sur-
factant, or other metal co-precursors (e.g. iron oleate (Fe-OL) and nickel
oleate (Ni-OL)). High molar ratio of OA to cobalt precursor transformed
the shape from wz-CoO nanorods to cubic type CoO (cu-CoO) urchin-
like shaped nanostructures. The addition of Fe-OL to the wz-CoO-NR
reaction formed irregular cube (and flower) shaped cu-CoO type
Co0.9Fe0.1O nanocrystals and spherical ferrite type metal oxide nano-
crystals (CoFe2O4 and Co0.5Ni0.5Fe2O4). Cobalt based metal oxide na-
nocrystals were evaluated as catalysts for the generation of H2 via
NaBH4 under basic conditions (from 293 to 323 K). As a function of size
(length), shape, crystal structure, and composition, H2 generation was
measured and compared. Results indicate that NR length and compo-
sition are critical performance (structural) variables: the maximum rate
of H2 generation for 200 nm wz-CoO-NRs, 30 nm wz-CoO-NRs, cu-CoO
nanoflowers, cu-CoO nanourchins, and cu-CoO nano-irr-cubes was
10,366.7, 8220.0, 3426.0, 3180.0, and 2221.2 mL min−1 g−1 at 293 K,
respectively. Ferrite type metal oxide nanocrystals (Fe3O4, CoFe2O4,
and Co0.5Ni0.5Fe2O4) are poor catalysts for this reaction.
2.4. Synthesis of irregular cube- and flower shaped nanocrystalline cobalt
iron oxides
Decomposition of cobalt precursors (both Co-OL and Co-ST) in the
presence of iron precursor (Fe-OL) led to the formation flower or irre-
gular cube shaped cobalt iron oxide nanocrystals (CoFeO nanoflowers
and nano-irr-cubes). CoFeO nanoflowers were synthesized using iron
oleate (0.07 mmol), Co-OL (0.7 mmol), Co-ST (0.3 mmol), and oleic
acid (0.5 mmol) in 1-octadecene (5 g); the mixture was heated to 593 K
at a rate of 1 K min−1 and kept at 593 K for 30 min. For CoFeO nano-irr-
cubes, the condition was same as irregular shaped cobalt iron oxide
nanocrystals except a heating rate of 4 K min−1. The resulting nano-
crystals were purified using ethanol and hexane as described above. The
purified nanocrystals were well dispersed in nonpolar solvents such as,
toluene, ether, chloroform, and hexane.
2

C. Kim, et al.
AppliedCatalysisA,General589(2020)117303
2.5. Synthesis of spherical iron oxide and cobalt ferrite nanocrystals
method [1,49,50]. The reaction temperature was controlled by im-
mersing the reaction flask in a temperature controlled water bath. Ty-
pically, 10 mg of as-synthesized (dried) nanocrystals were mixed with
15 mg of water and 3 mL of NaBH4 (2.8 wt%) in alkaline solution
containing sodium hydroxide (NaOH; from 2.7 to 28 wt%) at a desired
temperature (from 293 to 323 K). As a function of temperature, acti-
E )
12 nm (as an average diameter) spherical iron oxides were synthe-
sized by Fe-OL decomposition at high temperature as described pre-
viously [41,45,46]. 1.8 mmol of Fe-OL was mixed with oleic acid
(1 mmol) in 1-octadecene (5 g) at room temperature and the reaction
mixture was heated to 593 K at a rate of 8 K min−1 and the reaction was
kept at 593 K for 1 h. For 12 nm spherical cobalt ferrite nanocrystals,
the reaction mixture of Co-OL (0.7 mmol), Co-ST (0.3 mmol), Fe-OL
(1.8 mmol), and oleic acid (1 mmol) in 1-octadecene (5 g) was heated to
593 K at a ramp rate of 15 K min−1 and the nanocrystals were grown at
593 K for 1 h. Cobalt nickel iron oxide nanocrystals were synthesized by
decomposition of Co-OL (3.5 mmol), Co-ST (1.5 mmol), Ni-OL
(0.4 mmol), and Fe-OL (2.1 mmol) were mixed with oleic acid (1 mmol)
in 1-octadecene (5 g) at room temperature and the reaction temperature
increased to 593 K at a ramp rate of 8 K min−1. The resulting black
colloids (typically 5 mL) were purified using ethanol, acetone, and
hexane as described above. Purified nanocrystals were dispersed in
various nonpolar solvents such as, chloroform, hexane, and toluene.
vation energy (
tion:
of the reaction was calculated by Arrhenius equa-
a
Ea
RT
⎛
⎞
r = k0 exp −
(3)
⎝
⎠
r
Here denotes the reaction rate,
presents the reaction constant, and T is the reaction temperature (K).
R
is the ideal gas constant, k0 re-
3. Results and discussion
3.1. Synthesis of wüsite-type hexagonal cobalt oxide nanorods
Thermal decomposition of cobalt precursor mixtures (both Co-OL
and Co-ST) formed cobalt oxide nanorods (CoO-NRs) in 1-octadecene at
593 K. As-synthesized, CoO-NRs are monodisperse and stable in organic
(nonpolar) media, such as hexane, toluene, and chloroform (Fig. 1).
Synthesis of CoO-NRs is highly reproducible when two cobalt pre-
cursors are used with specific ratios (see the Materials and Methods
above). Co-ST as a precursor additive allows for complete decomposi-
tion of Co-OL and accelerates nucleation and growth of CoO-NRs;
however, a single cobalt precursor, either Co-OL or Co-ST, does not
result in monodisperse CoO-NRs, rather spherical and non-mono-
disperse cobalt oxides with wide size distributions (Fig. A.1). As shown
in Fig. A.2, nucleation occurs by decomposition of Co-OL in the pre-
sence of Co-ST (molar ratio of Co-OL to Co-ST was 0.7/0.3 at 2 K min−1
of the ramp rate), forming triangle shaped nanocrystals (seed materials
for CoO-NRs) for reactions less than 10 min. Under these conditions
with longer reaction times, monodisperse, one-dimensional CoO-NRs
with two or three nails per structure, ca. 40 nm in length, were observed
(after 15 min). Growth continued, forming 80 nm length CoO-NRs, for a
30 min reaction time. Interestingly, CoO-NRs actually become shorter
for longer reaction times, > 60 min, and then lose monodispersivity
after 2 h.
2.6. Transmission Electron Microscope (TEM)
The size of the as-synthesized nanocrystals was measured through
transmission electron microscope (TEM). TEM specimens were pre-
pared using carbon support film on 300 mesh copper grids (electron
Microscopy Sciences) and ultrathin carbon type-A 400 mesh copper
grids (Ted Pella Inc.). The TEM micrographs were taken by a Tecnai G2
Spirit Twin microscope (FEI, Hillsboro, OR) operated at 120 kV.
HRTEM analysis was performed using a JEOL 2100 F microscope
(JEOL. Tokyo, Japan) operated at 200 kV.
2.7. X-Ray diffraction (XRD)
X-ray powder diffraction (XRD) patterns were obtained using a
Bruker d8 Advance X-ray Diffractometer. 2θ range was from 10 to 80 °
with a Cu Kα radiation (1.54 Å) and the X-ray was generated at 40 kV
and 40 mA.
2.8. Inductively coupled plasma optical emission spectroscopy (ICP-OES)
The morphology of CoO-NR is also sensitive to the reaction heating
rate (to the target temperature, 593 K). The morphology of CoO-NRs
was explored for a wide range of heating ramp rates from 1 to 20 K
min−1, keeping all other reaction conditions constant (Fig. 1). CoO-NRs
with a length of ca. 200 nm is observed with a heating rate of 1 K
min−1, while shorter CoO-NRs (below 50 nm length) are observed
under faster heating conditions (> 4 K min−1). Faster ramp rates likely
lead to fast(er) nucleation (by decomposition of cobalt precursors) and
growth for a shorter period time, resulting in short CoO-NRs, which has
also been observed by others [41,43]. Conversely, slower ramp rates
allow for relatively slower nucleation/growth, which is favorable for
the one-dimensional growth of longer CoO-NRs [41,43]. An, K. et al.
reported the effect of heating rates on the length control of CoO-NRs
whereby, short CoO-NRs (19 nm as length) were observed for faster
The particle concentrations were measured by an inductively cou-
pled plasma optical emission spectrometer (CP-OES, Perkin Elmer
Optima 7300DV).
2.9. Inductively coupled plasma mass spectroscopy (ICP-MS)
Metal concentrations (low level) of the nanocrystals were measured
by an inductively coupled plasma mass spectrometer (ICP-MS, Perkin
Elmer ELAN DRC II).
2.10. Optical bandgap evaluation
The indirect bandgap energy values of libraries of engineered cobalt
based metal oxide nanocrystals were acquired by the electron absorp-
tion spectra using Tauc equation as follows [47,48]:
heating rates (1.9 K min−1) and slow(er) heating rates (1 K min−1
resulted in longer NRs (26 nm) [43].
)
CoO-NRs (described here) are considerably longer than those syn-
thesized by An K. et al. under similar heating rate conditions.
Decomposition of 1 mmol of cobalt precursor mixture (Co-OL and Co-
ST) in 5 g of 1-octadecene for a 1 K min−1 heating (ramp) rate results in
nearly 10 times longer CoO-NR formation than those NR prepared by
An K. et al. (Co-OL (1 mmol) in 1-octadecene (50 g) at a heating rate of
1 K min−1; both reactions were kept at 593 K for 30 min) [43]. This is
likely due to the effect(s) of high monomer concentration(s) in the re-
action [51–53]. In addition to length control, the number of nails of a
CoO-NR is also sensitive to ramp rate. More nails (as part of a single
structure) are observed as the ramp rate decreased (Fig. 1). Longer
(αhʋ)n = B (hʋ − Eg)
(2)
α
Where is the absorption coefficient, hʋ is the photon energy, n is 1/2
for an indirect transition or 2 for a direct transition, B is a constant
value, and Eg is the optical bandgap. The estimated bandgap energy
values were determined by extrapolation through the curves of the plot
((αhʋ)n vs. (hʋ)).
2.11. Hydrogen generation experiments
Rate of H2 generation was monitored using water displacement
3

C. Kim, et al.
AppliedCatalysisA,General589(2020)117303
Fig. 1. Length control of wz-CoO-NRs depending on the ramp rate. The mixture of cobalt precursors (Co-OL and Co-ST) was heated to 593 K at different ramp rates
from 1 to 20 K min−1 (1 K min−1 (A), 2 K min−1 (B), 4 K min−1 (C), 8 K min−1 (D), 15 K min−1 (E), and 20 K min−1 (F)) and all NRs were grown at 593 K for 30 min.
All scale bars are 50 nm.
structures (80–200 nm) with 5–6 nails are observed with heating rates
of 1 - 2 K min−1, while shorter (below 50 nm), 2–3 nail structures were
molar ratios of Co-OL / Co-ST (0.7/0.1), few nanonails were observed
in addition to loss uniformity. For these, 0.1 mmol of Co-ST is likely not
adequate to fully nucleate/decompose 0.7 mmol of Co-OL, resulting in a
poor distribution of wz-CoO-NRs. Uniform wz-CoO-NRs with high
purity were found when the molar ratio of Co-OL / Co-ST was from 0.7/
0.3 to 0.7/0.7. Ca. 78 nm length wz-CoO-NRs were synthesized from the
mole ratio of 0.7/0.3 and the lengths of the nanonail structures de-
creased to ∼30 nm when the mole ratio of Co-OL to Co-ST was 0.7/0.7.
wz-CoO-NRs lost one-dimensional structures when excess Co-ST was
applied (with a mole ratio of 0.7/1.0, Fig. A.3).
formed when the heating rate was faster than 4 K min−1
.
Synthesized CoO-NR shows wüsite type hexagonal CoO (wz-CoO)
structures, which is consistent with previous reports [39,42–44,54,55].
As shown in Fig. 2, XRD patterns of CoO-NR (both short (30 nm) and
long (200 nm)) are identical to wüsite ZnO, employed here as a stable
hexagonal type reference material [39,43,44,55,56]. Unique wz-CoO
(100) and (002) diffraction were also identified by HR-TEM. As shown
in Fig. 3, CoO-NRs grow from base nanoplates (the bottom part of na-
norod) along the (002) direction and the crystalline lattice fringes of
0.266 nm and 0.278 nm, from both side and top views of wz-CoO-NR,
respectively. As shown in the HR TEM, micrographs match to the (002)
and (100) planes of wz-CoO. As observed by Hu et al., unstable polar
(002) planes, exposed to metal ions in a hexagonal wz-CoO structure,
grow faster in one direction, resulting in one-dimensional aspect(s) as
multiple ‘nails’ on a nanoplate (a head (i.e. bottom) aspect of a com-
bined structure) [39,57].
3.2. Shape control of cobalt based mixed metal oxide nanocrystals
While uniform nanorod shaped cobalt oxides with controllable sizes
(lengths) were achieved through careful control of reaction time,
heating ramp rate, and the ratio of cobalt precursors, the addition of
high concentration(s) of a surfactant (e.g. oleic acid) transforms one
dimensional wüsite structures into a submicron three-dimensional
cubic structure(s) of CoO. As shown in Fig. 4.A to C, morphology de-
viation was observed to occur in the form of spherical structures with
many spikes (termed as urchin like) with ca. 100 nm diameter, when
oleic acid (3 mmol) was applied to the reaction condition for the longest
(200 nm length) wz-CoO-NRs (1 mmol of cobalt precursor mixture (Co-
The molar ratio of cobalt precursors is another critical parameter for
optimizing morphology, uniformity, and composition of wz-CoO-NRs.
Increasing the ratio of Co-OL to Co-ST leads to longer wz-CoO-NRs
(from 0.7/0.1 to 0.7/1.0) in the presence of oleic acid (0.5 mmol) at
593 K for 30 min with a heating rate of 4 K min−1 (Fig. A.3). For higher
4

C. Kim, et al.
AppliedCatalysisA,General589(2020)117303
iron oxide nanocrystals were observed when 0.07 mmol of Fe-OL is
included for 200 nm wz-CoO-NR reaction conditions (1 mmol of cobalt
precursor mixture at 593 K for 30 min using a heating rate of 1 K
min−1) (Fig. 4G to I). Resulting nanocrystal composition was measured
to be Co0.93Fe0.07O via ICP-OES and ICP-MS. Faster ramp rates (at a rate
of 4 K min−1) lead to irregular, cube shaped cobalt-iron oxide nano-
crystals, as shown in TEM and STEM images in Fig. 4D to F, with a
composition ratio of Co0.9Fe0.1O. Both flower- and irregular cube
shaped cobalt-iron oxide nanocrystals (Co1-xFexO, x ≤0.1) showed
cubic phase CoO crystal structure (cu-CoO) (Fig. 5). In contrast to
flower-like and irregular cube shaped cobalt-iron oxide nanocrystals
(cu-Co0.9Fe0.1O nanoflowers and nano-irr-cubes), the crystal structure
of cu-CoO was observed to be similar to cubic phase magnetite (cu-
Fe3O4) when an excess amount of Fe-OL (over two times higher than
cobalt precursors) is used. As shown in Fig. 6, Fe-OL decomposition of
Co-OL and Co-ST, with a mole ratio of 2 to 1, yields CoFe2O4 (cobalt
ferrite) nanocrystals with an average diameter of 12.3
1.0 nm and a
structure consistent with cubic phase Fe3O4. Irregular cube shaped
nanocrystals were also observed for Co0.5Ni0.5Fe2O4 nanocrystals, syn-
thesized by addition/decomposition of nickel precursor (Ni-OL) in the
presence of the mixture of Fe-OL, Co-OL, and Co-ST at 593 K for 30 min
at a heating rate of 4 K min−1 (Figs. 5 and 6).
Fig. 2. XRD of wüsite- (wz-) CoO-NRs. The red vertical lines at the bottom of the
chart are wz-ZnO (JCPDS #36-1451) (For interpretation of the references to
colour in this figure legend, the reader is referred to the web version of this
article.).
3.3. Evaluation of H2 generation using libraries of cobalt based metal oxide
nanocrystals
OL and Co-ST) in 1-octadecene (5 g) at 593 K for 30 min using a heating
rate of 1 K min−1). Oleic acid concentrations from 0.5 to 2.0 mmol in
the reaction system did not affect the length and morphology control of
wz-CoO-NRs (maintaining 200 nm length with one dimensional struc-
ture). Excess oleic acid addition (> 5.0 mmol) yielded blue/purple
suspensions, resulting from incomplete decomposition of cobalt pre-
cursors and the blue/purple precipitates (after the reaction upon
cooling) did not dissolve in any non-polar solvent. Interestingly, the
crystal structure of urchin shaped CoO nanocrystals did not show wü-
site hexagonal structure, instead it was observed to be cubic CoO phase
(Fig. 5). HR TEM image of urchin shaped CoO nanocrystals also showed
the crystal lattice planes consistent with (111) of cu-CoO with lattice
fringe spacing of 0.244 nm (Fig. 4B).
The inclusion of other metal precursors, as additives (such as, iron
oleate (Fe-OL) and nickel oleate (Ni-OL)), also altered the composition
and shape of cobalt based metal oxide nanocrystals. Decomposition of
iron precursor (Fe-OL) with cobalt precursors (Co-OL and Co-ST) leads
to cobalt-iron alloyed nanostructures. Instead of one dimensional wz-
CoO nanostructures, flower-like and irregular cube shaped cobalt-iron
oxide nanocrystals are observed (Fig. 4D to I). Flower shaped cobalt-
From the synthesis routes described above, libraries of engineered
cobalt based metal oxide nanocrystals were evaluated as nanocatalysts
for H2 generation via hydrolysis of sodium borohydride (NaBH4) at pH
13 [9,13,58]. The reaction was explored by monitoring H2 production
as a function of time and compared with commercial metal oxide nano
powders (as control experiments). The rate of H2 production was ex-
amined in terms of nanocrystal size, shape, crystal structure, and
composition. Without an applied catalyst, no H2 was generated from
NaBH4 solution (2.8 wt%) at pH 13.8 [4,9]. The half-life time for NaBH4
is 430 days at pH 14 at 298 K [14].
As shown in Fig. 7, highest H2 generation rates correlate with the
length of wz-CoO-NRs, demonstrating a maximum H2 generation rate of
10,366.7 mL min−1 g−1 at 293 K for 200 nm wz-CoO-NRs. H2 genera-
tion for 200 nm NRs was enhanced by a factor of ∼1.3 compared to
30 nm NRs (a maximum rate for 30 nm wz-CoO-NRs was 8220 mL
min−1 g−1). Generation rates for 200 nm wz-CoO-NRs were also 180
times faster than commercial CoO nanopowders and nearly 3 fold faster
rates than other, as-synthesized cu-CoO type cobalt based metal oxide
nanocrystals (Fig. A.4). H2 Production rates for commercial CoO
Fig. 3. HR-TEM micrographs of wz-CoO-NRs. The side (A) and top (B) view images of the synthesized wz-CoO showed (002) and (100) lattice planes. All scale bars
are 5 nm.
5

C. Kim, et al.
AppliedCatalysisA,General589(2020)117303
Fig. 4. Urchin shaped- (from A to C) cobalt oxide, irregular cube shaped- (from D to F) cobalt iron oxide nanocrystals, and flower shaped cobalt iron oxide
nanocrystals (from G to I). Three different shaped nanocrystals were characterized by TEM (A, D, and G), HR-TEM (B, E, and H), and STEM (C, F, and I)).
Fig. 5. XRD of wüsite- (wz-) and cubic- (cu-)
CoO nanocrystals (A) and ferrite type mixed
metal oxide nanocrystals (B). The red, blue,
and black vertical lines at the bottom of each
charts are JCPDS #36-1451, JCPDS #43-1004,
and JCPDS #19-0629 for wz-CoO, cu-CoO, and
magnetite, respectively (For interpretation of
the references to colour in this figure legend,
the reader is referred to the web version of this
article.).
nanopowder, cu-CoO nanourchins, cu-Co0.9Fe0.1O nanoflowers, and cu-
Co0.9Fe0.1O nano-irr-cubes were 55.3, 3180.0, 3426.0, and 2221.2 mL
min−1 g−1 at 293 K, respectively. Ferrite type, mixed metal oxide na-
nocrystals (Fe3O4, CoFe2O4, and Co0.5Ni0.5Fe2O4) did not show any H2
evolution under these conditions, similar to other reports [59]. 200 nm
wz-CoO-NRs showed the lowest Arrhenius activation energy (Ea) of any
material evaluated, based on the temperature dependent H2 generation
reaction kinetics (Fig. 7B and Fig. A.5). According to Fig. 7B, the acti-
vation energy for 200 nm wz-CoO-NRs, 200 nm wz-CoO-NRs, 30 nm wz-
CoO-NRs, cu-CoO nanourchins, cu-Co0.9Fe0.1O nanoflowers, and cu-
6

C. Kim, et al.
AppliedCatalysisA,General589(2020)117303
Fig. 6. TEM micrographs of iron oxide (A), cobalt iron oxide (B and C), and cobalt nickel iron oxide (D) nanocrystals. The average diameters of iron oxide (A) and
cobalt iron oxide (C) nanocrystals were 12.0 1.4 and 12.3 1.0 nm, respectively. All scale bars are 50 nm.
Co0.9Fe0.1O nano-irr-cubes is 27.4, 35.1, 45.4, 57.5, and 63.5 kJ mol−1
respectively. The activation energy for 200 nm wz-CoO-NRs is con-
siderably lower than any previous values(s) reported (Table A.2) for
cobalt oxide based catalysts of NaBH4 hydrolysis [29–38].
,
length of NR became longer; 30 nm wz-CoO-NR typically have one or
two bold nanorods per structure, while 200 nm wz-CoO-NRs exists as
five or six thin nanorods per structure (Fig. 1). This geometry gives rise
to relatively more (002) planes, exposed with metal (Co2+) ions [39].
As a function of structural properties, we also observed lower optical
bandgap energies for longer wz-CoO-NRs, which may also contribute to
enhanced catalytic activity involving charge transition of lattice asso-
ciated ions. Lu et al. observed a good relationship between catalytic
activity and charge transfer energy gap for cobalt catalyst [37]. Others
have observed that some catalytic reactions proceed by charge transfer
from metal to ligand [60–67]. As shown in Fig. A.6, a very low charge
transfer band is observed for 200 nm wz-CoO-NRs with a bandgap en-
ergy of 1.74 eV; the first bandgap transitions from the spectra nearly
750 nm (1.54 eV for 200 nm wz-CoO-NRs and 1.58 eV for 30 nm wz-
CoO-NRs) are associated with the Co2+ d-d transition energy gap in the
crystal field and the second band gap transitions derived from the
spectra below 550 nm are associated with ligand to metal charge
transfer (LMCT) O2− to Co2+ (1.74 eV for 200 nm wz-CoO-NRs and
2.06 eV for 30 nm wz-CoO-NRs) [68–70]. The optical energy bandgap
for wz-CoO-NRs was highly sensitive to NR morphology (length),
especially the degree of (002), as reported by Lu A. L. et al. [39] In all
cases (Fig. A.7), wz-CoO-NRs showed lower band gap energies than cu-
CoO nanourchins (3.55 eV), cu-Co0.9Fe0.1O nanoflowers (3.38 eV), and
cu-Co0.9Fe0.1O nano-irr-cubes (3.47 eV) [48,71].
200 nm wz-CoO-NRs were also used to explore H2 generation rates
as
a function of NaOH concentration (2.7–28.0 wt%) (Fig. A.5).
Maximum rates of 14,138.8, 11,538.0, and 10,367 mL min−1 g−1 for
reactions with 2.8, 9.3, and 12 wt% of NaOH, respectively. Excess
NaOH concentration suppressed the catalytic properties of CoO mate-
rials significantly. Maximum H2 generation rates decreased from
5964.0 to 686.4 mL min−1 g−1 as NaOH concentration increased from
15 to 28 wt%. This is likely due to formation of precipitates during the
reaction such as, NaB(OH)4 or NaBO2, which are less soluble materials
in very high concentrations of NaOH (extremely high pH). Borate (or
borohydroxide) precipitates near the surface of the catalysts may also
interfere with NaBH4 hydrolysis as observed by others [4,15,60].
As observed, H2 generation rates and the remarkably low activation
energy from the longest wz-CoO-NRs (200 nm) could not be simply
explained by surface area analysis. BET surface area for 200 nm wz-
CoO-NRs, 30 nm wz-CoO-NRs, cu-CoO nanourchins, cu-Co0.9Fe0.1O na-
noflowers, and cu-Co0.9Fe0.1O nano-irr-cubes is 50.9, 33.2, 66.3, 23.2,
and 35.5 m2/g, respectively. We hypothesize that catalytically active
(002) planes of wz-CoO-NRs play a critical role in H2 generation in
contrast to relatively stable (111) or (200) plane(s), which are the
primarily exposed surface(s) for cu-CoO nanourchins, nanoflowers, and
nano-irr-cubes. As discussed above, wz-CoO-NRs consist multiple na-
nonails ((002) direction) and the number of nails increased as the
Lastly, optimized materials were assessed for repeated, recycled
hydrolysis reaction(s) with NaBH4. Here, 200 nm wz-CoO-NRs were
separated magnetically and reused over 5 times. As shown in Fig. A.8, a
7

C. Kim, et al.
AppliedCatalysisA,General589(2020)117303
Fig. 8. Oxidation states of cobalt ions on the surface of 200 nm length wz-CoO
NRs before and after H2 generation. The concentration of each oxidation state
of cobalt was calculated by the integration of the area under a curve (Table
A.3).
Fig. 7. H2 generation using engineered nanocatalysts. (A) Volume of H2 gen-
erated by NaBH4 hydrolysis in the presence of 200 nm wz-CoO NR (black tri-
angle), 30 nm wz-CoO NR (red triangle), cu- Co0.9Fe0.1O nanoflower (green
circle), cu-CoO nanourchin (blue star), cu- Co0.9Fe0.1O nano-irr-cube (purple
square), and ferrite type mixed metal oxide (brown circle) was monitored as a
function of time. (B) Arrhenius activation energy (Ea) of wz-CoO NRs and cu-
CoO (and Co0.9Fe0.1O) nanocrystal samples was calculated by the plots LN
(maximum rate) versus 1/T (For interpretation of the references to colour in
this figure legend, the reader is referred to the web version of this article.).
and composition (cu-CoO type Co0.9Fe0.1O or ferrite type metal oxides
(Fe3O4, CoFe2O4, and Co0.5Ni0.5Fe2O4)). With regard to H2 generation
via NaBH4 hydrolysis, 200 nm wz-CoO-NRs showed the fastest H2
generation rate of 10.366.7 mL min−1 g−1 at 293 K and the lowest
activation energy of 27.4 kJ mol−1. These materials demonstrated a
maximum H2 generation rate at least three times higher than cu-CoO
type metal oxide nanocrystals (cu-Co0.9Fe0.1O nanoflowers, cu-CoO
nanourchins, and cu-Co0.9Fe0.1O nano-irr-cubes). We hypothesize that
such performance is due to unique morphology of 200 nm wz-CoO-NR
(multiple thin nanonails along the direction of (002). Taken together,
research provides new synthetic routes, performance regimes, and cri-
tical understanding of cobalt based metal oxide nanocrystals (nanoca-
talysts) for H2 generation via NaBH4 hydrolysis.
maximum H2 generation rate for 200 nm wz-CoO-NRs decreased to
1318.4 mL min−1 g−1 after the 5th injection of NaBH4 solution. This
decrease is related to the stability of the materials over multiple reac-
tion cycles. 200 nm wz-CoO-NRs completely lost crystallinity after
multiple uses (5 times) and became amorphous based on XRD analysis
(Fig. A.9). Furthermore, an increase of reduced cobalt ions on the
surface of the NR was observed after H2 generation reaction(s) by XPS
analysis (Fig. 8); the ratio of Co(0)/Co(II)/Co(III) from the spectra of Co
2 P region for the sample before and after H2 generation reactions (5
times) was 0/42.7/11.0 and 6.0/56.2/13.5, respectively [72–77].
Acknowledgements
This work is supported by U.S. Army Corps of Engineers (W912HZ-
13-2-0009-P00001) and the US National Science Foundation (CBET
1437820). TEM, ICP-OES, and ICP-MS were provided by the Nano
Research Facility (NRF), HR-TEM and XRD were provided by the
Institute of Materials Science & Engineering (IMSE) at Washington
University in St. Louis.
4. Conclusions
Precise thermal decomposition of cobalt oleate (Co-OL) and cobalt
stearate (Co-ST) resulted in wüsite cobalt oxide nanorods (wz-CoO-NRs)
with tunable lengths (from 30 nm to 200 nm) and shapes (e.g. nanorod
and urchin). Inclusion of iron and nickel precursors (iron oleate (Fe-OL)
and nickel oleate (Ni-OL)) resulted in mixed metal oxide nanocrystals
with varied shape (irregular cube, flower, and sphere), crystal structure,
Appendix A. Supplementary data
Supplementary material related to this article can be found, in the
8

C. Kim, et al.
AppliedCatalysisA,General589(2020)117303
online version, at doi:https://doi.org/10.1016/j.apcata.2019.117303.
(2005) 6188–6189.
[43] K. An, N. Lee, J. Park, S.C. Kim, Y. Hwang, J.G. Park, J.Y. Kim, J.H. Park, M.J. Han,
J.J. Yu, T. Hyeon, J. Am. Chem. Soc. 128 (2006) 9753–9760.
[44] M.R. Buck, A.J. Biacchi, R.E. Schaak, Chem. Mater. 26 (2014) 1492–1499.
[45] C. Kim, S.S. Lee, B.J. Reinhart, M. Cho, B.J. Lafferty, W. Li, J.D. Fortner, Environ.
Sci. Nano 5 (2018) 2252–2256.
[46] S.S. Lee, W. Li, C. Kim, M. Cho, B.J. Lafferty, J.D. Fortner, J. Mater. Chem. A 3
(2015) 21930–21939.
[47] J. Tauc, Mater. Res. Bull. 3 (1968) 37-&.
References
[1] D.M.F. Santos, C.A.C. Sequeira, Renew Sust Energ Rev 15 (2011) 3980–4001.
[2] Z.T. Xia, S.H. Chan, J. Power Sources 152 (2005) 46–49.
[3] A.M.F.R. Pinto, D.S. Falcao, R.A. Silva, C.M. Rangel, Int J Hydrogen Energ 31
(2006) 1341–1347.
[4] P. Brack, S.E. Dann, K.G.U. Wijayantha, Energy Sci. Eng. 3 (2015) 174–188.
[5] D.D. Tuan, K.-Y.A. Lin, Chem. Eng. J. (2018).
[48] D. Barreca, C. Massignan, S. Daolio, M. Fabrizio, C. Piccirillo, L. Armelao,
E. Tondello, Chem. Mater. 13 (2001) 588–593.
[49] A. Levy, J.B. Brown, C.J. Lyons, Ind. Eng. Chem. 52 (1960) 211–214.
[50] A.C. Caetano de Souza, J.L. Silveira, Y. Kiros, R. Zanzi Vigouroux, A low cost & safe
system of hydrogen production utilizing NaBH4 and CoO catalysis, 2nd
International Congress University-Industry Cooperation (UNIDU07) (2007).
[51] S.S. Lee, H.G. Zhu, E.Q. Contreras, A. Prakash, H.L. Puppala, V.L. Colvin, Chem.
Mater. 24 (2012) 424–432.
[6] S. Duman, S. Özkar, Int. J. Hydrogen Energ. 43 (2018) 15262–15274.
[7] A.D. Chowdhury, N. Agnihotri, A. De, Chem. Eng. J. 264 (2015) 531–537.
[8] U.B. Demirci, O. Akdim, J. Andrieux, J. Hannauer, R. Chamoun, P. Miele, Fuel Cells
10 (2010) 335–350.
[9] T.P. Bartkus, J.S. T’ien, C.J. Sung, Int. J. Hydrogen Energ. 38 (2013) 4024–4033.
[10] D. Ross, Vacuum 80 (2006) 1084–1089.
[52] S.S. Lee, W.L. Li, C. Kim, M. Cho, J.G. Catalano, B.J. Lafferty, P. Decuzzi,
[11] J. Mao, D. Gregory, Energies 8 (2015) 430–453.
J.D. Fortner, Environ Sci-Nano 2 (2015) 500–508.
[53] S.S. Lee, C.G. Zhang, Z.A. Lewicka, M.J. Cho, J.T. Mayo, W.W. Yu, R.H. Hauge,
V.L. Colvin, J Phys Chem C 116 (2012) 10287–10295.
[54] K.M. Nam, J.H. Shim, D.W. Han, H.S. Kwon, Y.M. Kang, Y. Li, H. Song, W.S. Seo,
J.T. Park, Chem. Mater. 22 (2010) 4446–4454.
[55] M.J. Redman, E.G. Steward, Nature 193 (1962) 867-&.
[56] Z.W. Wang, L.L. Daemen, Y.S. Zhao, C.S. Zha, R.T. Downs, X.D. Wang, Z.L. Wang,
R.J. Hemley, Nat. Mater. 4 (2005) 922–927.
[57] J.Q. Hu, Q. Li, N.B. Wong, C.S. Lee, S.T. Lee, Chem. Mater. 14 (2002) 1216–1219.
[58] C.M. Kaufman, B. Sen, J. Chem. Soc. Dalton Trans. (1985) 307–313.
[59] H.I. Schlesinger, H.C. Brown, A.E. Finholt, J.R. Gilbreath, H.R. Hoekstra, E.K. Hyde,
J. Am. Chem. Soc. 75 (1953) 215–219.
[12] L. Schlapbach, A. Zuttel, Nature 414 (2001) 353–358.
[13] H. Dong, H.X. Yang, X.P. Ai, C.S. Cha, Int. J. Hydrogen Energ. 28 (2003)
1095–1100.
[14] S.C. Amendola, S.L. Sharp-Goldman, M.S. Janjua, N.C. Spencer, M.T. Kelly,
P.J. Petillo, M. Binder, Int. J. Hydrogen Energ. 25 (2000) 969–975.
[15] C.E. Bunker, M.J. Smith, J. Mater. Chem. 21 (2011) 12173–12180.
[16] Y. Bai, C.A. Wu, F. Wu, B.L. Yi, Mater. Lett. 60 (2006) 2236–2239.
[17] B.H. Liu, Z.P. Li, S. Suda, J. Alloys. Compd. 415 (2006) 288–293.
[18] N. Sahiner, S. Butun, T. Turhan, Chem. Eng. Sci. 82 (2012) 114–120.
[19] T. Turhan, Y.G. Avcıbası, N. Sahiner, J. Ind. Eng. Chem. 19 (2013) 1218–1225.
[20] T. Turhan, Y.A. Güvenilir, N. Sahiner, Energy 55 (2013) 511–518.
[21] F. Seven, N. Sahiner, J. Power Sources 272 (2014) 128–136.
[22] F. Seven, N. Sahiner, Int. J. Hydrogen Energy 39 (2014) 15455–15463.
[23] S. Yildiz, N. Aktas, N. Sahiner, Int. J. Hydrogen Energy 39 (2014) 14690–14700.
[24] N. Sahiner, S. Sagbas, J. Power Sources 246 (2014) 55–62.
[25] N. Sahiner, F. Seven, Fuel Process. Technol. 132 (2015) 1–8.
[26] B. Liu, A. Rose, N. Zhang, Y.-Y. Hu, M. Ma, J. Phys. Chem. C 121 (2017)
12610–12616.
[27] G. Bozkurt, A. Özer, A.B. Yurtcan, Int. J. Hydrogen Energy 43 (2018) 22205–22214.
[28] H. Zhang, X. Feng, L. Cheng, X. Hou, Y. Li, S. Han, Colloids Surf. A Physicochem.
Eng. Asp. 563 (2019) 112–119.
[29] D. Xu, X. Zhang, X. Zhao, P. Dai, C. Wang, J. Gao, X. Liu, Int. J. Energy Res. (2019).
[30] P. Krishnan, S.G. Advani, A.K. Prasad, Int. J. Hydrogen Energy 33 (2008)
7095–7102.
[31] A. Lu, Y. Chen, J. Jin, G.-H. Yue, D.-L. Peng, J. Power Sources 220 (2012) 391–398.
[32] N. Patel, A. Santini, V. Bello, G. Mattei, A. Miotello, Surf. Coat. Technol. 235 (2013)
784–791.
[60] R. Pena-Alonso, A. Sicurelli, E. Callone, G. Carturan, R. Raj, J. Power Sources 165
(2007) 315–323.
[61] M. Bonn, S. Funk, C. Hess, D.N. Denzler, C. Stampfl, M. Scheffler, M. Wolf, G. Ertl,
Science 285 (1999) 1042–1045.
[62] R.M. Lambert, R.L. Cropley, A. Husain, M.S. Tikhov, Chem. Commun. (2003)
1184–1185.
[63] Y. Zhang, A. Kolmakov, S. Chretien, H. Metiu, M. Moskovits, Nano Lett. 4 (2004)
403–407.
[64] J. Zhang, X. Liu, R. Blume, A. Zhang, R. Schlögl, D.S. Su, Science 322 (2008) 73–77.
[65] W.E. Kaden, T. Wu, W.A. Kunkel, S.L. Anderson, Science 326 (2009) 826–829.
[66] Z. Zhang, J.T. Yates Jr., J. Phys. Chem. C 114 (2010) 3098–3101.
[67] L.R. Baker, A. Hervier, H. Seo, G. Kennedy, K. Komvopoulos, G.A. Somorjai, J. Phys.
Chem. C 115 (2011) 16006–16011.
[68] H.A. Weakliem, J. Chem. Phys. 36 (1962) 2117–2140.
[69] M.A. White, S.T. Ochsenbein, D.R. Gamelin, Chem. Mater. 20 (2008) 7107–7116.
[70] Q. Xiaoqing, L. Guangshe, S. Xuefei, L. Liping, F. Xianzhi, Nanotechnology 19
(2008) 215703.
[71] K. Deori, S. Deka, Crystengcomm 15 (2013) 8465–8474.
[72] B. Cui, H. Lin, Y.Z. Li, J.B. Li, P. Sun, X.C. Zhao, C.J. Liu, J Phys Chem C 113 (2009)
14083–14087.
[73] H.C. Chen, J.J. Jiang, L. Zhang, H.Z. Wan, T. Qi, D.D. Xia, Nanoscale 5 (2013)
8879–8883.
[74] H.C. Chen, J.J. Jiang, Y.D. Zhao, L. Zhang, D.Q. Guo, D.D. Xia, J Mater Chem A 3
(2015) 428–437.
[75] R.B. Ba, Y.H. Zhao, L.J. Yu, J.J. Song, S.S. Huang, L.S. Zhong, Y.H. Sun, Y. Zhu,
Nanoscale 7 (2015) 12365–12371.
[76] W.H. Ryu, J. Shin, J.W. Jung, I.D. Kim, J Mater Chem A 1 (2013) 3239–3243.
[77] J. Yang, H.W. Liu, W.N. Martens, R.L. Frost, J Phys Chem C 114 (2010) 111–119.
[33] A. Lu, Y. Chen, D. Zeng, M. Li, Q. Xie, X. Zhang, D.-L. Peng, Nanotechnology 25
(2013) 035707.
[34] T. Pfeil, T. Pourpoint, L. Groven, Int. J. Hydrogen Energy 39 (2014) 2149–2159.
[35] H. Zhang, T. Ling, X.-W. Du, Chem. Mater. 27 (2014) 352–357.
[36] L. Xie, K. Wang, G. Du, A.M. Asiri, X. Sun, Int. J. Hydrogen Energy 42 (2017)
30639–30645.
[37] G.R.M. Tomboc, A.H. Tamboli, H. Kim, Energy 121 (2017) 238–245.
[38] B.H. Liu, Z.P. Li, S. Suda, J. Alloys. Compd. 415 (2006) 288–293.
[39] A.L. Lu, Y.Z. Chen, D.Q. Zeng, M. Li, Q.S. Xie, X.X. Zhang, D.L. Peng,
Nanotechnology 25 (2014).
[40] H. Zhang, T. Ling, X.W. Du, Chem. Mater. 27 (2015) 352–357.
[41] J. Park, K.J. An, Y.S. Hwang, J.G. Park, H.J. Noh, J.Y. Kim, J.H. Park, N.M. Hwang,
T. Hyeon, Nat. Mater. 3 (2004) 891–895.
[42] W.S. Seo, J.H. Shim, S.J. Oh, E.K. Lee, N.H. Hur, J.T. Park, J. Am. Chem. Soc. 127
9