Lanthanum hydroxide nanorods were employed as both a template and catalyst for carbon synthesis by chemical vapor deposition. The resulting carbon possesses hollow nanorod shapes with graphitic walls. The hollow carbon nanorods were interconnected at some junctions forming a mazelike network, and the broken ends of the tubular carbon provide accessibility to the inner surface of the carbon, resulting in a surface area of 771 m2/g. The hollow carbon was tested as an electrode material for supercapacitors. A specific capacitance of 128 F/g, an energy density of 55 Wh/kg, and a power density of 1700 W/kg at 1 A/g were obtained using the ionic liquid, 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide, as the electrolyte.
Lanthanum hydroxide nanorods were employed as both a template and catalyst for carbon synthesis by chemical vapor deposition. The resulting carbon possesses hollow nanorod shapes with graphitic walls. The hollow carbon nanorods were interconnected at some junctions forming a mazelike network, and the broken ends of the tubular carbon provide accessibility to the inner surface of the carbon, resulting in a surface area of 771 m2/g. The hollow carbon was tested as an electrode material for supercapacitors. A specific capacitance of 128 F/g, an energy density of 55 Wh/kg, and a power density of 1700 W/kg at 1 A/g were obtained using the ionic liquid, 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide, as the electrolyte.
Graphene-based materials
with designed nanoarchitectures have been
studied for a wide range of applications.[1,2] Many
methodologies have been established to prepare three-dimensional or
porous carbons with graphitic walls to achieve high surface area and
high conductivity, which are of great importance for electrical double-layer
capacitors.[3−7] The hydrophobic nature and high surface area of carbon-based materials
also provide advantages when employed as catalyst supports, which
allows better adsorption and diffusion of reactions.[8,9] Among these carbon materials, templated carbon has been prepared
using hard templates, such as zeolites and mesoporoussilica.[10−16] Unfortunately, most of them have amorphous carbon framework, which
limited their use in energy storage as well as catalysis.[17] Graphene- and carbon nanotube (CNT)-based carbon
electrodes with porous or three-dimensional structures have shown
excellent performance in supercapacitors.[18−22] Therefore, it is of great interest to make templated
carbon with graphene-like walls for energy storage and other applications.
Mesoporoussilica has been used with various organic precursors to
produce carbons with improved graphitic content.[17] Recently, a lanthanum-exchanged zeolite was reported to
be an excellent template for carbon synthesis.[23] The resulting carbon replicated the zeolite pore structure
with graphene-like walls and conductivity comparable to gold. It is
assumed that the carbon source reacts with La first to form LaC2, which is then turned to La(OH)3 by reacting with
steam leading to the growth of graphitic carbon (Figure ). This work demonstrated the
application of La as a catalyst for the synthesis of graphitic carbon
with controlled morphology and structure. In the present study, pure
lanthanum hydroxide nanorods (LaNR) were employed as templates to
produce hollow carbon nanorods. Compared to the La-zeolite, the pure
lanthanum hydroxide templates are removed by HCl, and the resulting
solution can be used to make new nanorods.
Figure 1
Schematic illustration
of the carbon formation mechanism on La(OH)3 surface.
Schematic illustration
of the carbon formation mechanism on La(OH)3 surface.
Results and Discussion
The LaNRs
were hydrothermally synthesized using a solution of La(NO3)3·6H2O and NaOH at 110 °C
for 24 h followed by centrifugation, drying, and annealing at 200
°C for 1 h. The lanthanum hydroxide nanorod-templated carbon
(LaNRTC) was synthesized by chemical vapor deposition (CVD) using
acetylene and steam at 600 °C. Washing with concentrated HCl
removes the La(OH)3 template leaving hollow replicas of
the nanorods.The LaNR shown in Figure a,d possesses a rod shape with an average
diameter of 16 nm
and an average length of 140 nm. The lattice fringes of LaNR (Figure b) have a d-spacing of 0.285 nm, corresponding to the (200) face of
La(OH)3 (JCPDS 36-1481). Figure b,e shows the morphology of the templated
carbon before removal of LaNR. The transmission electron microscopy
(TEM) image shows that carbon covers the surface of the LaNR. The
scanning electron microscopy (SEM) image (Figure b) shows better contrast compared to that
of the pure LaNR since the carbon covered on the surface of LaNR is
conductive. These images indicate that the carbon was grown only on
the surface of LaNR without formation of amorphous carbon elsewhere.
This demonstrates that the La(OH)3 serves as both a template
and catalyst for carbon growth, which is consistent to our previous
work that no carbon was formed under the same conditions if only mesoporoussilica was placed in the reactor.[24] And
after La(OH)3 was deposited on mesoporoussilica, carbon
was formed along the surface of the templates. After removal of the
LaNR by using acid, the resulting carbon retains the rod shape morphology
with similar particle size (Figure c), and the inner space is empty as shown in the TEM
image (Figure f).
The acid may access LaNR through defects in the carbon coating, leading
to openings in the tubular carbon network. The edge of the resulting
carbon exhibits lattice fringes with a d-spacing
of 0.40 nm attributed to the (002) planes of graphitic carbon, indicating
the graphitic nature of the lanthanum hydroxide nanorod-templated
carbon. The d-spacing is bigger than graphite (0.34
nm) due to less than five graphene layers on the edge of the carbon,
where the d-spacing has been reported to increase
as the number of the graphene layers decreases.[25]
Figure 2
SEM and TEM images of (a, d) lanthanum hydroxide nanorods (LaNR);
lanthanum hydroxide nanorod-templated carbon (b, e) before (C@LaNR)
and (c, f) after (LaNRTC) template removal.
SEM and TEM images of (a, d) lanthanum hydroxide nanorods (LaNR);
lanthanum hydroxide nanorod-templated carbon (b, e) before (C@LaNR)
and (c, f) after (LaNRTC) template removal.The LaNRTC after template removal shows a broad peak around
22°
in the X-ray diffraction (XRD) and the corresponding d-spacing is calculated using Bragg equation, which is consistent
to the d-spacing measured in the TEM image. This
is attributed to the (002) planes of the carbon, whereas the LaNR
and C@LaNR mainly exhibit XRD patterns for nanocrystalline La(OH)3 (JCPDS 36-1481) (Figure a). This is also consistent to the TEM images (Figure ), where the LaNRTC
shows a few layers of carbon. The Raman spectra exhibit a strong G
band (∼1600 cm–1) with a IG/ID ratio of 1.03. Previously,
a La-zeolite-templated carbon was reported to possess five- or seven-member
rings, which contributed to the peaks for disorderedcarbon (D band),
but they are still sp2 hybridized.[23] In that case, the overall carbon material has graphitic framework
with good conductivity. Combining the results from TEM and Raman,
the LaNRTC possesses graphitic walls.
Figure 3
(a) Powder X-ray diffraction patterns
of LaNR, C@LaNR, and LaNRTC;
(b) Raman spectrum of LaNRTC; (c) N2 adsorption–desorption
isotherms of LaNR and LaNRTC; (d) density functional theory (DFT)
pore size distributions of LaNR and LaNRTC.
(a) Powder X-ray diffraction patterns
of LaNR, C@LaNR, and LaNRTC;
(b) Raman spectrum of LaNRTC; (c) N2 adsorption–desorption
isotherms of LaNR and LaNRTC; (d) density functional theory (DFT)
pore size distributions of LaNR and LaNRTC.Figure c
shows
the N2 adsorption–desorption isotherms for the LaNR
template and the resulting carbon, LaNRTC. The isotherm of the pure
LaNR shows a type II isotherm with a low Brunauer–Emmett–Teller
(BET) surface area of 54 m2/g, indicating no porosity but
space between nanoparticles. In contrast, the LaNRTC shows a type
IV isotherm corresponding to mesoporous features with a clear hysteresis
between 0.4 and 1.0. The adsorption below 0.1 also indicates some
microporosity. The BET surface area of the LaNRTC is much higher than
that of the pure LaNR (771 vs 54 m2/g). The micropore volume
and total pore volume of LaNRTC is much higher than LaNR, as shown
in Table S1. Both LaNRTC and LaNR show
similar DFT pore size distribution, which might due to the absence
of uniform pore structure (Figure d). The large increase in surface area is due to the
transformation of the morphology from rods to tubular network with
accessibility to inner surface of the tubes. The TEM images in Figure show both closed
ends and open ends. The open ends result in pore openings as large
as 30 nm (Figure ).
Micropores resulting from defects are not apparent in these images.
Additionally, some of the tube junctions were fused together as interconnected
L-shaped channels. It is assumed that the carbon grows from one rod
to the other across the tightly packed junctions rather than just
covering an individual rod. This arises because during the CVD process
the carbon cannot completely coat the randomly stacked La(OH)3 nanorods. Thus, there are hollow nanorods and a complicated
tubular carbon network. Overall, the material is composed of carbon
tubes, tube stacks, and fused mazelike nano carbon tube networks.
The large increase in the surface area from pure LaNR to LaNRTC is
much higher than that of most CNTs (<500 m2/g).[7] The TEM images in Figure also represent the relationship between
the nanorods template and the pore structure of resulting carbon.
Since the synthesis of La(OH)3 nanorods with various sizes
and shapes has not been well established, the pore structure is not
controlled by different La(OH)3 templates here, but it
is of great interest for future work. The hollow carbon tubes with
different length and diameters are presented in Figure , and they are formed on La(OH)3 nanorods with various sizes. The size of La(OH)3 nanorods
controls the resulting carbon’s dimensions and inner diameters,
which affect the porosity of each carbon tube in the same batch. By
using different La(OH)3 templates, it is possible to control
the pore structure of the resulting carbon in bulk.
Figure 4
Schematic illustration
of the formation and TEM images of broken
ends and connected junctions in the LaNRTC.
Schematic illustration
of the formation and TEM images of broken
ends and connected junctions in the LaNRTC.The LaNRTC was tested as an electrode material for supercapacitors
using 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide
(EMIM-TFSI) as the electrolyte. Figure a shows the cyclic voltammograms (CVs) of the LaNRTC
at different scan rates (10, 25, 50, 75, and 100 mV/s) in the range
of −2.0–2.0 V.[26−29] The shape of the CV is nearly rectangular indicating
ideal double-layer capacitive behavior. The capacitances of the electrodes
were calculated from CV using the following equationwhere I is the current as
a function of time, υ is the scan rate. The specific capacitance
in different scan rates was calculated using following equationwhere m is mass of both electrodes.
The highest specific capacitance obtained for the LaNRTC was 128 ±
2 F/g at 10 mV/s. Table summarizes the obtained specific capacitance values at different
scan rates. To compare the LaNRTC with a mesoporouscarbon templated
by mesoporoussilica, wrinkled mesoporouscarbon was prepared according
to reported procedure.[24] The capacitance
of LaNRTC is higher than the mesoporouscarbon templated by La2O3 supported on wrinkled mesoporoussilica (Figure S1), which has a capacitance of 70 F/g.
The capacitance of LaNRTC is among the best results for mesoporousgraphenes,[20,30−36] and aligned CNTs[18,37] as well as activated carbons[38−41] using ionic liquids as the electrolyte. Unlike these carbons, the
LaNRTC does not require complicated porous silica templates, lengthy
preparation process, or additional activation steps. And the ionic
liquids allow a higher operating voltage window, which is more practical
for various applications.
Figure 5
Electrochemical characterizations of LaNRTC:
(a) cyclic voltammograms;
(b) charge–discharge curves from 1 to 6 A/g; (c) Nyquist plots;
(d) Ragone plot; (e) specific capacitance retention as a function
of cycle number.
Table 1
Specific
Capacitance (F/g) in Different
Scanned Rates
scan rates (mV/s)
specific
capacitance (F/g)
10
128.1
25
126.2
50
125.0
75
122.6
100
120.2
Electrochemical characterizations of LaNRTC:
(a) cyclic voltammograms;
(b) charge–discharge curves from 1 to 6 A/g; (c) Nyquist plots;
(d) Ragone plot; (e) specific capacitance retention as a function
of cycle number.Figure b shows
the charge–discharge curves for the LaNRTC in different current
densities (1–10 A/g). The JME cells were charged to 3.5 V and
discharged completely at different current densities. Linear discharge
curves show the ideal capacitive behavior. After charging up to 3.5
V, it is important to stabilize the coin cell before it starts to
discharge. A plateau at 3.5 V is purposely introduced to investigate
the stability of the coin cell. At 1 A/g, the IR drop
is ∼0.2 V, which shows the good conductive behavior of the
JME cell. Energy densities (Ed) and power
densities (Pd) were calculated using following
equationswhere t is the time taken
to discharge to 0 V from the initial voltage (ΔV), subtracting IR drop at the beginning of discharge.
The energy density and power density of the LaNRTC were found to be
55 Wh/kg and 1700 W/kg, respectively, at 1 A/g. In contrast, the mesoporouscarbon templated by La2O3 supported on wrinkled
mesoporoussilica (Figure S1) has an energy
density of 35.8 Wh/kg and a power density of 1499 W/kg at 1 A/g.Figure c shows
the Nyquist plots for the LaNRTC in EMITFSI electrolyte. Electrochemical
impedance spectroscopy (EIS) can provide information such kinetics
in the double-layer charging–discharging processes. The almost
vertical increase in the lower frequency region is indicative of an
ideal capacitance behavior. In the Nyquist plot, the x axis intercepts at the highest frequency region indicating bulk
electrolyte resistance, which mainly depends on the electrolyte solution.Figure d shows
the Ragone plot of the LaNRTC, polyacrylonitrile (PAN) nanofibers,[42] and carbon nanotubes (CNTs) at different current
densities 1–10 A/g by comparing energy densities and power
densities. Energy densities were determined by discharging from 3.5
to 0 V. When the current density increases, the energy density decreased,
but retained 90% of its original energy density at discharge rate
of 10 A/g, whereas PAN and CNTs retained only 75 and 80%, respectively.
The energy density of the LaNRTC is much higher than PAN nanofibers
and the pure CNTs, as shown in Figure d.Figure e shows
the plot of specific capacitance retention against cycle number. After
the 100th cycle, 98% of the specific capacitance was retained, whereas
97% of the capacitance was retained after 1000th cycle. Capacitance
(89%) was retained after 5000th cycles, indicating the good cycling
ability. This combined with the high energy density, and power density
of the LaNRTC makes this carbon a good electrode material for supercapacitors.
Conclusions
In conclusion, the lanthanum hydroxide nanorods function as both
catalyst and template for carbon growth, resulting in hollow carbon
nanorods. The hollow carbon nanorods possess connected junctions and
open ends, which result in a mazelike structure with high surface
area and porosity. The LaNRTC shows excellent performance as an electrode
material for supercapacitors taking advantage of its unique structure
as well as low resistance. The templating method using lanthanum hydroxide
nanoparticles also creates possibilities for the preparation of other
graphitic carbons with controlled morphology.
Experimental Section
Synthesis
of La(OH)3 Nanorods
In a typical
synthesis, 5.77 g of La(NO3)3·6H2O in 10 mL of deionized water was mixed with a solution of 14.7 g
of NaOH in 20 mL of water and heated to 110 °C in a 45 mL Teflon-lined
autoclave for 24 h. The solid product was collected by centrifugation
and washed with water several times then dried overnight at 100 °C.
The product was then annealed at 200 °C for 1 h.
Synthesis of
La(OH)3 Nanorod-Templated Carbon
The La(OH)3 nanorods (∼1 g) were placed in a ceramic
boat centered in a horizontal quartz tube reactor. The reactor was
heated to 600 °C under N2 (200 mL/min). Then, acetylene
(30 mL/min) and steam (water is pumped at 5 mL/h) were fed into the
reactor for 1.5 h followed by carbonization under N2 at
850 °C for 2 h. The resulting carbon/La(OH)3 composite
was washed with concentrated HCl and water to remove the La(OH)3 template. For comparison, a mesoporouscarbon was prepared
with the same method, but by using La(OH)3 supported on
wrinkled mesoporoussilica as the template following reported procedure.[24]
Fabrication of Coin Cell-Type Supercapacitors
Symmetric
JME cells were fabricated according to a previously reported procedure
for JME packaging (Figure S2).[43] Electrodes were prepared by mixing the La(OH)3 nanorod-templated carbon with 2% poly(tetrafluoroethylene)
binder. The carbon electrodes were immersed in the electrolyte, 1-ethyl-3-methylimidazolium
bis(trifluoromethylsulfonyl)imide (EMIM-TFSI), for 1 h at room temperature
under vacuum prior for assembling the coin cell. The two electrodes
were separated by a Teflon film (0.02 mm) and constructed under inert
environment.
Characterization
Scanning electron
microscope (SEM)
images were recorded on a Zeiss-LEO model 1530 scanning electron microscope.
Transmission electron microscope (TEM) images were recorded using
a JEOL 2100 transmission electron microscope. Powder X-ray diffraction
(XRD) patterns were acquired on a Rigaku Ultima IV diffractometer
using Cu Kα radiation. Raman spectra was obtained on a Jobin
Yvon Horiba high-resolution LabRam Raman microscope. The N2 adsorption–desorption isotherms and BET surface areas were
measured using a Quantachrome AS1 Autosorb. Electrochemical measurements
were made using a BT2000 Arbin battery testing system. Electrochemical
impedance spectroscopy (EIS) was measured on a EG&G Princeton
Applied Research potentiostat/galvanostat.