PCCP

View Article Online

PAPER View Journal | View Issue

Energy storage properties of a two-dimensional

TiB4 monolayer
Published on 25 May 2019. Downloaded by University of Frankfurt on 7/26/2019 4:59:43 AM.

Cite this: Phys. Chem. Chem. Phys.,

2019, 21, 13151 ab
Zhiyang Liu, Erdong Wu,a Jiangxu Liab and Shi Liu *a

Received 3rd April 2019,
Accepted 24th May 2019
DOI: 10.1039/c9cp01864f
rsc.li/pccp
Herein, the energy storage properties of TiB4 monolayers were studied within the density functional
theory framework. Both CH4 and H2 were chosen as adsorption molecules, and their interactions with a
TiB4 sheet were investigated. TiB4 attracted gas molecules via open Ti sites, and each Ti atom could
adsorb a maximum of two molecules. Via the electronic density of the states and atomic charge
analysis, we found that the mechanism for gas adsorption was mainly electrostatic. For H2 adsorption
cases, orbital interactions also made contributions. As the combustion energy of one CH4 molecule is
three times that of one H2, the TiB4–2CH4 compound can achieve the best equivalent gravimetric
hydrogen density of 10.14 wt% with the average adsorption energy of 0.38 eV. Ab initio molecular
dynamics calculations on this compound showed that there was no kinetic barrier during CH4
desorption. Moreover, the stacking of the TiB4 monolayers could weaken the energy storage capacity.
Therefore, it should be avoided in practial usage. Based on the abovementioned results, the TiB4 mono-
layer was suggested to be a promising candidate for onboard energy storage.

1. Introduction an obvious limitation: re-hydrogenation occurs only under high

Energy is a main issue facing humanity today. Compared with
traditional fossil fuels, hydrogen is a clean, efficient and renew-
able energy source; thus, it can be an efficient source for the
next generation of fuels.1 The challenge in hydrogen utilization
is the storage of hydrogen with high gravimetric and volumetric
density under near ambient thermodynamic conditions. Hydro-
gen can be stored in two forms: atomic and molecular. As a
common and typical atomic hydrogen storage material,
magnesium can store up to 7.7 wt% hydrogen; however, the
hydrogen desorption temperature of MgH2 is high, and the
(de)hydrogenation kinetics is poor; thus, additives are needed
to improve these properties.2,3 Bulk alloys, such as vanadium
alloys,4–6 store hydrogen by forming metal hydrides, and the
stored hydrogen can be used at around room temperature and
standard atmospheric pressure. However, they usually undergo
chemical reactions during storage, and bonds are broken or
formed during hydrogen usage; moreover, the hydrogen per-
cent is too far to reach the target of 5.5 wt% (ultimate 6.5 wt%)
hydrogen by 2025 according to the US Department of Energy
(DOE);7 in this regard, a covalently bonded compound, i.e.
LiBH4, can be a promising hydrogen storage material because
of its high hydrogen storage capacity (18.4 wt%); however, it has
a
Institute of Metal Research, Chinese Academy of Sciences, 72 Wenhua Road,
Shenyang, 110016, China. E-mail: sliu@imr.ac.cn
b
School of Materials Science and Engineering, University of Science and Technology
of China, 72 Wenhua Road, Shenyang, 110016, China
temperature and hydrogen pressure conditions.8 The storage of
hydrogen in molecular form could be an effective way to deal
with the abovementioned problems due to no bond formation
or breaking during the process of usage. Many calculations
have been conducted on nanomaterials9–20 and framework
materials21–26 for molecular hydrogen storage due to their high
surface/volume ratios. For instance, Duan et al.17 investigated
Li-decorated CN via simulation and achieved the theoretical
gravimetric hydrogen density of 10.81 wt%. Yang et al.14 simu-
lated Ti-decorated g-C3N4 and achieved the hydrogen capacity
of 9.70 wt%. However, considering the experimental results of
these materials obtained to date, most storage capacities
obtained at ambient temperature and pressure still failed to
reach the DOE target.27–30 As a result, molecular hydrogen
storage materials need to be further explored.
Recently, a new kind of 2D material, i.e. TiB4, has been
predicted.31 It has much lower weight than TM-TCNQ,32,33 and
a similar open metal site that may attract gas molecules.
Considering the valences of TM, one Sc (3d14s2) atom can
coordinate with three B atoms to form a carbon-like electronic
shell—ScB3 nanotube, which is expected to be a stable hydro-
gen storage material.34 Then, the TiB4 nanostructure contain-
ing one Ti (3d24s2) atom and four B atoms could also be stable.
Via thermodynamic calculations, Qu and coworkers found that
the TiB4 sheet still remained stable at 2500 K,31 much higher
than the thermal stability of isoreticular MOFs (673 K),35 which
could be conducive to the actual usage. Therefore, in this study,
we chose TiB4 as a substrate material. For energy storage, CH4

This journal is © the Owner Societies 2019 Phys. Chem. Chem. Phys., 2019, 21, 13151--13156 | 13151

Paper
could be another choice due to its low cost and limited air
pollution upon combustion. Although the CH4 molecules have
larger repulsion force than H2, the energy of one CH4 is
3.1 times that of one H2 molecule. Then, the addition of CH4
could also increase the energy capacity of the storage system.
For example, Wu and coworkers found that the mixture of CH4
and H2 significantly enhanced the energy storage capacity of
Li-decorated graphene and Sc-decorated armchair graphene
nanoribbons.36 Therefore, herein, via first principles calcula-
tions, not only the hydrogen storage properties of TiB4 were
Published on 25 May 2019. Downloaded by University of Frankfurt on 7/26/2019 4:59:43 AM.
studied, but also CH4 adsorption on the TiB4 monolayer was
considered.
2. Computational methods
Density functional theory (DFT) calculations were performed
using the Vienna ab initio simulation package (VASP)37–39 and
the projector augmented-wave (PAW)40,41 method. The Perdew–
Burke–Ernzerhof (PBE)42 functional within the generalized
gradient approximation (GGA) was used to determine the
exchange–correlation potential. The DFT-D3 scheme of
Grimme for the vdW correction43–45 was applied to investigate
the gas adsorption of H2 and CH4 on the TiB4 sheet. The
original structure parameters of TiB4 chosen herein were
obtained from the literature.31 A vacuum region of more than
20 Å was used to avoid the interaction between adjacent layers.
Fully optimized structures were obtained when the energy and
maximum force were less than 10 5 eV and 0.01 eV Å 1,
respectively. The kinetic energy cut-off for the electron wave
functions was 500 eV. We used 12  12  1 G-centered
Monkhorst–Pack k-point samplings in the Brillouin zone for
geometry optimization and property calculation.
The average gas adsorption energy (Egave) and the continuous
adsorption energy (Egcon) were defined as follows:
Egave = (nEg + ETiB4 ETiB4
ng)/n
Egcon = Eg + ETiB4
(n 1)g ETiB4
ng
where g represents the gas molecule, Eg is the total energy of one
H2 or CH4 molecule in the gas phase, and ETiB4, ETiB4
(n 1)g, and
Fig. 1 (a) The top view of a 2  2  1 TiB4 super cell and (b) the corresponding ELF map sliced perpendicular to the (001) direction. (c) Phonon dispersion
of the TiB4 monolayer. G(0.0 0.0 0.0), X(0.0 0.5 0.0), and M(0.5 0.5 0.0) refer to special q-points in the first Brillouin zone of the reciprocal space. Ti and B
atoms are denoted by blue and green spheres, respectively. The black dashed square marks a unit cell.
13152 | Phys. Chem. Chem. Phys., 2019, 21, 13151--13156
View Article Online
PCCP
ETiB4
ng are the total energy of the unit TiB4 sheet and its counter-
parts with (n 1) and n adsorbed gas molecules, respectively.
3. Results and discussion
The optimized structure of the TiB4 monolayer is shown in
Fig. 1(a). The calculated lattice constants of the unit cell were
a = b = 4.10 Å with the Ti–B bond distance of 2.22 Å, the B–B
bond length in the B squares of 1.71 Å, and the B–B bond
length in the sharing wheel edges of 1.68 Å. These values were
in accordance with the previously reported values of 4.12 Å,
2.22 Å, 1.72 Å and 1.68 Å.31 The electron localization function
(ELF)40 was used to analyze the electron distribution and is
shown in Fig. 1(b). The regions where the ELF values were 0 and
1 corresponded to no and fully localized electrons, respectively.
It can be seen from Fig. 1(b) that a few electrons are distributed
around the Ti atoms; this indicates the ionic interaction
between Ti and the B network. Electrons accumulated in the
middle of the B–B bonds; this confirmed the presence of strong
covalent states between B atoms. Phonon dispersion calculations46
were performed to evaluate our optimized TiB4 monolayer, and no
imaginary frequency existed in the first Brillouin zone (Fig. 1(c));
this confirmed the stability of this monolayer.
For gas adsorption on the TiB4 monolayer, at first, we
considered the adsorption of a single gas molecule (H2 or
CH4) and varied its position above the surface of the TiB4 unit.
Via optimization, we found that the preferred adsorption site
was above the Ti atom, and the gas structures changed after
adsorption. The optimized equilibrium configurations of the
gas-adsorbed TiB4 monolayers are shown in Fig. 2(a)–(f), corres-
ponding to the side and top views. When one H2 adsorbed at
the Ti site (Fig. 2(b)), its H–H bond length increased from 0.75 Å
to 0.79 Å. For the adsorption of one CH4 (Fig. 2(d)), two H atoms
of the CH4 molecule stayed close to the Ti site with two
(1) corresponding C–H bond lengths of 1.11 Å. Moreover, the other
two H atoms were far away from the TiB4 monolayer, and their
(2)
C–H bond lengths were 1.09 Å. As more gas molecules were
added to the TiB4 surface, we found that at each side of the
open Ti site, only one gas molecule stably adsorbed. When two
This journal is © the Owner Societies 2019
 View Article Online

PCCP Paper

Table 1 The adsorption energy (Egave and Egcon, eV per g), Ti–g distance
(dTi–g, Å) and Bader charges (e) of the mH2–nCH4–TiB4 compounds. g
stands for H2 or CH4

dTi–g
Host Egave Egcon 1st g 2nd g QTi QB QH2 QCH4
TiB4 +1.51 0.38
TiB4
1H2 0.54 1.94 +1.48 0.36 0.035
TiB4
2H2 0.38 0.22 2.02 2.02 +1.44 0.34 0.030
TiB4
1CH4 0.41 2.71 +1.53 0.38 0.029
TiB4
2CH4 0.32 0.23 2.74 3.61 +1.53 0.37 0.022
Published on 25 May 2019. Downloaded by University of Frankfurt on 7/26/2019 4:59:43 AM.

TiB4
1H2
1CH4 0.38 0.22 1.94 3.48 +1.48 0.36 0.030 0.019

between C and Ti was 2.71 Å. When the second CH4 molecule

Fig. 2 The side and top views of the optimized TiB4 monolayers with (a)
zero gas molecule; (b) one H2; (c) two H2; (d) one CH4; (e) two CH4; (f) one
H2 and one CH4. H and C atoms are denoted by white and grey spheres,
respectively.
H2 were adsorbed, as shown in Fig. 2(c), each H2 molecule had
the bond length of 0.78 Å. Fig. 2(f) shows that when two CH4
molecules adsorbed, the first adsorbed CH4 molecule had the
same structure as that of the TiB4–CH4 system, whereas the
C–H bond lengths of the second CH4 molecule were consistent
with that of a free CH4 molecule (1.10 Å). Fig. 2(f) shows
the optimized TiB4–CH4–H2 system, in which the bond length
of the adsorbed H2 molecule is 0.79 Å. Compared to the
second adsorbed CH4 molecule in the TiB4–2CH4 system,
the C–H bond lengths in Fig. 2(f) were the same (1.10 Å), but
the H atoms of the CH4 molecule were located in different
areas: three H atoms were near the TiB4 surface, and the
remaining H atom was off the surface. Using eqn (1) and (2), it
was found that all the adsorption binding energies calculated
herein lied in the previously reported ideal range of 0.2–0.6 eV
per molecule.47,48 We found that the binding energy of one
gas molecule was 0.54 eV per H2 (0.41 eV per CH4), and it
decreased to 0.38 eV per H2 (0.32 eV per CH4) due to the
relatively weak interaction between the second gas molecule
and the TiB4 sheet, as shown in Table 1. It could be seen
from Fig. 2 and Table 1 that the distance between H2 and the
open Ti site was B2.02 Å, much shorter than that for the
adsorbed CH4. For one CH4 adsorption case, the distance
was added, it was a little far away from the Ti site (dTi–g =
CH4
3.61 Å), whereas the Econ was 0.23 eV per CH4, still showing
the possibility of adsorption. For the mixture adsorption, the
CH4
Econ value of 0.22 eV per CH4 was close to that of pure CH4.
However, considering the energy storage capacity, TiB4–2CH4
was the better choice.
To obtain further information on the adsorption mecha-
nism for the gas molecules on the surface of the TiB4 mono-
layer, we performed a Bader charge analysis.49 It could be seen
from Table 1 that, in the pure TiB4 sheet, each Ti transferred
B1.51 e to the B network; this was consistent with the former
ELF result and confirmed the ionic interactions between Ti and
B. In the gas-adsorbed TiB4 systems, each gas molecule had a
small and negative charge, indicating Coulomb interactions
between the gas molecules and the host materials. When H2
adsorbed onto the TiB4 sheet, it obtained a charge of 0.030 to
0.035 e and changed the charges of both Ti and B; this
indicated that the adsorbed H2 interacted with these atoms.
For the CH4 adsorption cases, CH4 had a relatively small charge
( 0.019 to 0.029 e). Moreover, compared to the TiB4 sheet,
only the charge of Ti changed after the addition of CH4; this
indicated that the electrostatic interaction in the CH4–TiB4
system resulted from CH4 and Ti. In addition, considering
the adsorption energy, we found that it was positively correlated
with the absolute charge value of the corresponding gas
molecule.
The density of states (DOS) calculation is another way to
investigate the bond interactions of the TiB4 sheet. As shown in
Fig. 3(a), the DOS peaks for the TiB4 sheet occurred at approxi-
mately 3 eV, which originated from the interactions between
the B p electrons and the Ti d electrons and were in good
agreement with previously reported results.31 After H2 adsorbed
on the TiB4 sheet (Fig. 3(b)), it was found that the H 1s orbital
hybridized with the Ti 3d and B 2p orbital at 7 to 8 eV below
the Fermi level. Therefore, the orbital interactions also con-
tributed to the H2 adsorption. As presented in Fig. 3(c), the
adsorption of the CH4 molecules onto the TiB4 sheet did not
influence the binding interactions between the Ti and the B
atoms. Moreover, no orbital interactions existed between the
CH4 molecules and the host sheet.
To satisfy the DOE target for practical use, it is necessary to
investigate the adsorption and desorption kinetics of the gas
molecules on the TiB4 monolayer. To illustrate this, we carried

This journal is © the Owner Societies 2019 Phys. Chem. Chem. Phys., 2019, 21, 13151--13156 | 13153
 View Article Online

Paper PCCP
Published on 25 May 2019. Downloaded by University of Frankfurt on 7/26/2019 4:59:43 AM.

Fig. 3 Density of states for the compounds (a) TiB4, (b) TiB4–2H2 and (c) TiB4–2CH4. For each compound, total DOS are shown in black at the top of the
figures. The projected DOS on the atoms with s, p and d contributions are shown in red, green and blue, respectively. The origin of energy was set at the
highest occupied state and is indicated with a vertical dashed line.

out ab initio molecular dynamics (AIMD) calculations50 on the
optimized 2  2  1 TiB4–2H2 and TiB4–2CH4 super cells at
different temperatures (50 K, 100 K, 150 K, 200 K, 300 K, 400 K
and 500 K) followed by equilibration for 10 000 steps at the 1 fs
time scale. During the MD simulation, the gas molecules that
were relatively far away from the surface were deleted. Fig. 4
shows the structures after 10 ps running. It could be seen
that the desorption behaviors were the same for TiB4–2H2 and
TiB4–2CH4 at 50 K, 100 K, and 150 K, with zero, one, and three
corresponding gas molecules evaporated, respectively. Five and
four H2 molecules remained adsorbed at 200 K and 300 K.
When the temperature reached 500 K, three H2 molecules were
still stably adsorbed, leading to only a 2.65 wt% of reversible
hydrogen storage capacity in the TiB4–2H2 system. As for
TiB4–2CH4, four and two CH4 molecules adsorbed at 200 K
and 300 K, and all were desorbed at temperatures Z400 K.
Considering that the combustion heat for one CH4 molecule
was 3.1 times that of one H2 molecule, the reversible stored energy
of TiB4–2CH4 was equivalent to the gravimetric hydrogen density
of 10.14 wt%, which was comparable to that of previously
reported Li-decorated g-CN (10.81 wt%)17 and Ti-decorated
g-C3N4 (9.70 wt%).14 The abovementioned behaviors were consistent
with the adsorption energies and showed that as energy storage
materials, TiB4 had great potential for practical applications.
With regard to the energy storage properties of the TiB4
monolayer system, the distance between adjacent layers was
large enough to avoid an interaction. For practical applications,
the adjacent TiB4 monolayers may be close enough to each
other to get stacked. Therefore, the multilayer structure of TiB4
and its corresponding energy storage properties should also be
considered. To illustrate this, we investigated the bilayer TiB4
system. Several different TiB4 bilayers were optimized, and the
most stable structure is shown in Fig. 5(a). It could be seen that
each Ti atom in one layer was above the center of the four
membered B ring in the other layer. Moreover, the B–B layer
distance and interlayer interaction energy were 2.80 Å and
329 meV per atom, similar to the previously reported values
of 2.78 Å and 351 meV per atom, respectively.31 Via the Bader

Fig. 4 Snapshots after 10 ps MD simulations of TiB4–2H2 and TiB4–2CH4 systems at different temperatures.

13154 | Phys. Chem. Chem. Phys., 2019, 21, 13151--13156 This journal is © the Owner Societies 2019

PCCP
Published on 25 May 2019. Downloaded by University of Frankfurt on 7/26/2019 4:59:43 AM.
Fig. 5 The side and top views of the optimized TiB4 bilayers with (a) zero
gas molecules, (b) two H2 molecules, and (c) two CH4 molecules.
charge analysis, it was found that the charges of the Ti and B atoms
in each layer of the stacked TiB4 were similar to those in the
individual TiB4 monolayers; this indicated that the moderate
interlayer interactions were mainly caused by ionic interactions.
For the energy storage of the TiB4 bilayer, due to the interlayer
interactions and space cases, there was no inside site for gas
adsorption. After optimization, we found that gas molecules tended
to adsorb at the Ti sites and above the bilayer surface. For H2
absorption, as shown in Fig. 5(b), each Ti site could only adsorb
one H2 molecule. The adsorbed H2 molecule had the charge of
0.0061 e, whose absolute value was much smaller than that in the
TiB4 monolayer systems; this resulted in the weak adsorption
energy of 0.08 eV per H2 that was unsuitable for mobile applica-
tions. Fig. 5(c) shows the optimized configuration of the CH4-
adsorbed TiB4 bilayer. Every Ti site of the TiB4 bilayer also attracted
a maximum of one CH4 molecule, and the absolute charge of the
adsorbed CH4 molecule ( 0.012 e) was still smaller than that in the
corresponding monolayer cases, whereas it maintained the mod-
erate binding energy of 0.22 eV per CH4. This binding energy could
be caused by the charge of the C atom. Via further charge analysis,
we found that the C atom of CH4 in the TiB4 bilayer systems
obtained a negative charge ( 0.13 e), which was in the range of the
C charges for the TiB4 monolayer systems ( 0.11 to 0.17 e). This
negatively charged C would interact with a Ti atom (+1.51 e); this
would lead to effective CH4 adsorption. As a result, the maximum
stored energy in the TiB4 bilayer was equivalent to the gravimetric
hydrogen density of 5.07 wt%, only half that of the TiB4 monolayer.
4. Conclusions
In summary, we studied the 2D TiB4 monolayer for high energy
density storage by first principles calculations. For the TiB4
This journal is © the Owner Societies 2019
View Article Online
Paper
sheet, the interactions were mainly ionic between Ti and B and
covalent in the B network. The stability of TiB4 was demon-
strated by the phonon spectrum calculation. When the H2 and
CH4 gas molecules were adsorbed, they preferred to locate
above the Ti sties, and each side of Ti could adsorb only
one molecule. The first adsorbed molecule had much higher
binding energy than the second adsorbed molecule; this was
consistent with the AIMD results. Via comparison, it was found
that TiB4–2CH4 had better reversibility than TiB4–2H2 and
obtained the effective energy storage of 10.14 wt% under
ambient conditions. Moreover, when the TiB4 monolayers were
stacked, their energy capacity dramatically decreased. Thus, in
practical application, the interlayer spacing should be noted.
We expect that this study will provide new insights into
molecular hydrogen storage and accelerate the experimental
studies on energy economy.
Conflicts of interest
There are no conflicts to declare.
Acknowledgements
The authors gratefully acknowledge the assistance of Dr Süleyman
Er, Dutch Institute for Fundamental Energy Research in perform-
ing the simulation.
Notes and references
1 R. Coontz and B. Hanson, Science, 2004, 305, 957.
2 S. Er, G. A. de Wijs and G. Brocks, J. Phys. Chem. Lett., 2010,
1, 1982–1986.
3 C. Zhou, Z. Z. Fang, C. Ren, J. Li and J. Lu, J. Phys. Chem. C,
2013, 117, 12973–12980.
4 H. Yukawa, M. Takagi, A. Teshima and M. Morinaga,
J. Alloys Compd., 2002, 330, 105–109.
5 M. V. Lototsky, V. A. Yartys and I. Y. Zavaliy, J. Alloys Compd.,
2005, 404, 421–426.
6 Z. Liu, L. Xiong, J. Li, S. Liu and S. Er, Int. J. Hydrogen Energy,
2018, 43, 21441–21450.
7 DOE Technical Targets for Onboard Hydrogen Storage for
LightDuty Vehicles, http://www.energy.gov/eere/fuelcells/
doe-technical-targets-onboard-hydrogen-storage-light-duty-
vehicles, accessed October, 2018.
8 A. Zuttel, P. Wenger, S. Rentsch, P. Sudan, P. Mauron and
C. Emmenegger, J. Power Sources, 2003, 118, 1–7.
9 S. Er, G. A. de Wijs and G. Brocks, J. Phys. Chem. C, 2009,
113, 18962–18967.
10 S. S. Mao, S. H. Shen and L. J. Guo, Prog. Nat. Sci., 2012, 22,
522–534.
11 S. Banerjee, C. G. S. Pillai and C. Majumder, Appl. Phys. Lett.,
2013, 102, 073901.
12 J. Wang, H. Y. Zhao and Y. Liu, ChemPhysChem, 2014, 15,
3453–3459.
Phys. Chem. Chem. Phys., 2019, 21, 13151--13156 | 13155

Paper
13 S. Er, G. A. de Wijs and G. Brocks, J. Mater. Chem. A, 2015, 3,
7710–7714.
14 W. Zhang, Z. Zhang, F. Zhang and W. Yang, Appl. Surf. Sci.,
2016, 386, 247–254.
15 A. Hashmi, M. U. Farooq, I. Khan, J. Son and J. Hong,
J. Mater. Chem. A, 2017, 5, 2821–2828.
16 X. Tan, H. A. Tahini and S. C. Smith, Energy Storage
Materials, 2017, 8, 169–183.
17 Y.-D. Chen, S. Yu, W.-H. Zhao, S.-F. Li and X.-M. Duan, Phys.
Chem. Chem. Phys., 2018, 20, 13473–13477.
Published on 25 May 2019. Downloaded by University of Frankfurt on 7/26/2019 4:59:43 AM.
18 Z. Sheng, S. Wu, X. Dai, T. Zhao and Y. Hao, Phys. Chem.
Chem. Phys., 2018, 20, 13903–13908.
19 L. Wang, X. Chen, H. Du, Y. Yuan, H. Qu and M. Zou, Appl.
Surf. Sci., 2018, 427, 1030–1037.
20 X. Liang, S.-P. Ng, N. Ding and C.-M. L. Wu, Appl. Surf. Sci.,
2019, 473, 174–181.
21 S. S. Han, W.-Q. Deng and W. A. Goddard III, Angew. Chem.,
Int. Ed., 2007, 46, 6289–6292.
22 Y. H. Hu and L. Zhang, Adv. Mater., 2010, 22, E117–E130.
23 E. Tylianakis, E. Klontzas and G. E. Froudakis, Nanoscale,
2011, 3, 856–869.
24 M. Samolia and T. J. D. Kumar, J. Phys. Chem. C, 2014, 118,
10859–10866.
25 Y. Pramudya and J. L. Mendoza-Cortes, J. Am. Chem. Soc.,
2016, 138, 15204–15213.
26 K. Koizumi, K. Nobusada and M. Boero, Phys. Chem. Chem.
Phys., 2019, 21, 7756–7764.
27 Y. W. Li and R. T. Yang, J. Am. Chem. Soc., 2006, 128,
8136–8137.
28 A. G. Wong-Foy, A. J. Matzger and O. M. Yaghi, J. Am. Chem.
Soc., 2006, 128, 3494–3495.
29 L. J. Murray, M. Dinca and J. R. Long, Chem. Soc. Rev., 2009,
38, 1294–1314.
30 S.-Y. Liu, P. Kundu, T.-W. Huang, Y.-J. Chuang, F.-G. Tseng,
Y. Lu, M.-L. Sui and F.-R. Chen, Nano Energy, 2017, 31,
218–224.
31 X. Qu, J. Yang, Y. Wang, J. Lv, Z. Chen and Y. Ma, Nanoscale,
2017, 9, 17983–17990.
13156 | Phys. Chem. Chem. Phys., 2019, 21, 13151--13156
View Article Online
PCCP
32 Y. D. Ma, Y. Dai, W. Wei, L. Yu and B. B. Huang, J. Phys.
Chem. A, 2013, 117, 5171–5177.
33 Q. Deng, T. Wu, G. Chen, H. A. Hansen and T. Vegge, Phys.
Chem. Chem. Phys., 2018, 20, 5173–5179.
34 Y. F. Zhao, M. T. Lusk, A. C. Dillon, M. J. Heben and
S. B. Zhang, Nano Lett., 2008, 8, 157–161.
35 M. Eddaoudi, J. Kim, N. Rosi, D. Vodak, J. Wachter,
M. Keeffe and O. M. Yaghi, Science, 2002, 295, 469.
36 Q. Xue, M. Wu, X. C. Zeng and P. Jena, J. Mater. Chem. A,
2018, 6, 8916–8922.
37 G. Kresse and J. Hafner, Phys. Rev. B: Condens. Matter Mater.
Phys., 1993, 48, 13115–13118.
38 G. Kresse and J. Furthmuller, Comput. Mater. Sci., 1996, 6,
15–50.
39 G. Kresse and J. Furthmuller, Phys. Rev. B: Condens. Matter
Mater. Phys., 1996, 54, 11169–11186.
40 P. E. Blochl, Phys. Rev. B: Condens. Matter Mater. Phys., 1994,
50, 17953–17979.
41 G. Kresse and D. Joubert, Phys. Rev. B: Condens. Matter
Mater. Phys., 1999, 59, 1758–1775.
42 J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett.,
1996, 77, 3865–3868.
43 S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem. Phys.,
2010, 132, 154104.
44 C. D. Zeinalipour-Yazdi, J. S. J. Hargreaves, S. Laassiri and
C. R. A. Catlow, Phys. Chem. Chem. Phys., 2017, 19,
11968–11974.
45 F. Bakhshi and N. Farhadian, Int. J. Hydrogen Energy, 2018,
43, 8355–8364.
46 A. Togo and I. Tanaka, Scr. Mater., 2015, 108, 1–5.
47 S. K. Bhatia and A. L. Myers, Langmuir, 2006, 22, 1688–1700.
48 Y.-H. Kim, Y. Zhao, A. Williamson, M. J. Heben and
S. B. Zhang, Phys. Rev. Lett., 2006, 96, 016102.
49 G. Henkelman, A. Arnaldsson and H. Jonsson, Comput.
Mater. Sci., 2006, 36, 354–360.
50 D. Marc and J. Hutter, Ab Initio Molecular Dynamics: Theory
and Implementation, John von Neumann Institute for Com-
puting, Julich, 2000.
This journal is © the Owner Societies 2019