Journal of

Materials Chemistry C
View Article Online
PAPER View Journal
Published on 09 May 2017. Downloaded by University of California - San Diego on 31/05/2017 01:39:38.

A computational study of hydrogen detection

by borophene
Cite this: DOI: 10.1039/c7tc00976c
Michal Novotný, ab Francisco Javier Domı́nguez-Gutiérrez b
and
Predrag Krstić *b

In this work we present a quantum-classical molecular dynamics study of hydrogen irradiation of a

single corrugated boron sheet in the incident energy range of 0.25–100 eV and report the resulting
reflection, adsorption, and transmission probability. For this we use self-consistent charge tight binding
density functional theory. A comparison of the irradiation results with our previous study on hydrogen
detection using graphene shows a significant increase in the adsorption probability and a lower reflection
rate for borophene. We benchmark our method of finding the interaction potentials of hydrogen and the
Received 7th March 2017, boron sheet with plane wave density functional theory for validating our approach. We utilized the open
Accepted 9th May 2017 boundary non-equilibrium Green’s function method to obtain conductivity of borophene as a function of
DOI: 10.1039/c7tc00976c hydrogen coverage. For voltages of up to 300 mV the systems have a linear volt–ampere characteristic
and the resulting conductance decreases exponentially as a function of the hydrogen coverage. Our
rsc.li/materials-c results suggest that borophene has favorable properties for its use as a hydrogen detector.

1 Introduction development.27–29 In these applications accurate measurements

of particle fluxes of hydrogen are essential. There are several
Since the discovery of the first truly 2D material, graphene,1 a works which study thermal atom adsorption on graphene and
number of other 2D materials, such as germanene and silicane, the change of its properties by hydrogenation.30 However no
have been theoretically predicted2 and later experimentally such studies have yet been done for borophene, apart from
confirmed.3,4 Most p-block elements of the first 3 periods can calculations of the electron transport properties of a fully
form stable homonucleic 2D crystals. Even though the boron’s hydrogenated borophene termed borophane. Borophene has
2D structure, borophene, was theoretically predicted by several been observed so far on a Ag substrate8 rather than in a free-
authors,5–7 it was only recently reported by Mannix et al.8 and standing form. This is consistent with a recent theoretical
Feng et al.,9 who managed to synthesize it for the first time. The study31 that free-standing borophene has imaginary phonon
challenge in a synthetic realization of borophene was mainly modes and is therefore dynamically unstable. However it has
due to the required costly and often toxic precursors. What been recently argued,31 based on recent studies,32 that intrinsic
finally justified an experimental investigation was the calculated defects or interactions with non-reactive surfaces can remove
properties of borophene which suggested8,10–18 that it would be structural instabilities of borophene. In addition, it was also
metallic10 and have an anisotropic electronic structure11–16 with inferred19,33 that free standing borophane gains structural stability.
mechanical stiffness.8,17,18 Most of these theoretical predictions The idea of using a 2D material for hydrogen detection has
were confirmed in recent experiments.19–24 been previously explored by us using graphene.34–36 We showed
A 2D material composed of light atoms with metallic properties that a single layer of graphene is not a good adsorber of hydrogen
can find a wide variety of applications in the field of electronic, atoms at impact energies of above 2.5 eV and therefore is not
magnetic, photonic, and mechanical devices.25,26 Our work is suited as a detector of hydrogen at higher kinetic energies. We
focused on the study of the properties of borophene for its use studied the adsorption, transmission, and reflection properties of
as a hydrogen nano-sensor. Such an application is of particular borophene irradiated by hydrogen in the range of kinetic energies
interest in battery development and fusion energy research and 0.25–100 eV and we calculated its electron transport properties as
a function of hydrogen coverage and voltage bias. By conducting
this study we hope to provide a more workable understanding of
a
Institute for Advanced Computational Science, Stony Brook University,
the use of borophene as an effective hydrogen detector.
Stony Brook, NY 11749, USA
b
Department of Physical and Theoretical Chemistry, Faculty of Natural Sciences,
Borophene has two experimentally confirmed structures, a
Comenius University in Bratislava, 84215 Bratislava, Slovakia. body centered orthorhombic lattice corrugated in one direction8
E-mail: [email protected] and a hexagonal, fully planar one.9 We have chosen the former

This journal is © The Royal Society of Chemistry 2017 J. Mater. Chem. C

 View Article Online

Paper Journal of Materials Chemistry C

structure reported to be stable by several authors.19,37,38 Even

though the bond distances in corrugated borophene (0.16 nm
and 0.18 nm) are longer than in graphene (0.14 nm) borophene
has a larger planar atomic density (103 806 at per nm2) than
graphene (79 459 at per nm2). This is a result of the boron’s
higher coordination number (6) in this lattice than that of
carbon (3) in graphene. Therefore we expect that borophene
will have a lower transmission and a higher reflection probability
Published on 09 May 2017. Downloaded by University of California - San Diego on 31/05/2017 01:39:38.

of hydrogen. We simulate the irradiation process with quantum-

classical molecular dynamics (QCMD) simulations,39 using self
consistent charge density functional tight binding theory (SCC-
DFTB).40–44 A large number of independent trajectories, impacting
randomly chosen locations on borophene, were generated to
obtain the desired adsorption and transmission probabilities
with a small statistical error. To ensure the validity of our results
we compare the DFTB pair potentials between borophene and a
hydrogen atom with respective potential calculated using full
plane wave DFT.
What adds to the possibility of borophene’s application as a
hydrogen detector are its outstanding thermal stability and electronic
properties.31,38,45 Previous studies of the electronic properties of
borophene and borophane show a significantly lower conductance
for the latter,46–48 which motivates us to quantify the sensitivity
of borophene to hydrogen coverage. We calculate the electron
transmission probability by applying the non-equilibrium
Green’s function (NEGF) method.49,50
In Section 2 we describe the computational methods used in
this paper. The results of our hydrogen irradiation simulations of
borophene as well as of electron transport calculations are presented
in Section 3. Finally, in Section 4 we give our concluding remarks.

2 Computational method Fig. 1 Top and side view of borophene (a–c) and borophane (d–f).
Because of corrugation the borophene lattice has two layers. Atoms in
2.1 Structures the lower plane (denoted by 1) are colored in gray. The structure is formed
of linear chains of boron atoms along the x direction. These chains are
We have chosen the corrugated boron structure reported by shifted from each other to form the corrugations depicted in (c) and (f). In
Mannix et al.8 for our model 2D system. This structure has a body the case of borophane the hydrogen atoms bond to boron parallel to the
centered orthorhombic unit cell with the lattice parameters 0.66  surface normal always in the convex-top positions (f).
0.29  0.16 nm with a basis of two boron atoms. We used a
supercell composed of 3  6 unit cells for the calculations of the simulation package (VASP).52,53 The borophene structure was first
interaction potential energy curves of borophene and a hydrogen optimized using both PWDFT and SCC-DFTB. Then using both
atom, while in the case of graphene we use a 4  4 supercell. This methods we carried out a series of adiabatic calculations of the
way we created similar in size supercells that were large enough to potential energy curves of a hydrogen atom interacting with the
eliminate intercell H–H interactions. For the QCMD simulations relaxed fully periodic, borophene sheet for various distances and
we used a 6  12 supercell with 144 atoms with a size of 1.96  adsorption sites of the hydrogen atom. The results are shown in
1.74  5.0 nm. The size of the hydrogen irradiated target is Fig. 2. For borophene the distances were counted from the ‘‘lower’’
approximately 1 nm2 in the middle of the supercell to avoid inter plane of the borophene structure marked by gray coloring in
cell hydrogen interactions. All of the borophene and graphene Fig. 1(a)–(c).
structures were energy optimized before the calculations and The SCC-DFTB40 approach is a tight binding approximation
the resulting lattice parameters and atomic positions were in to DFT which also includes the charge corrections up to second
agreement with the theoretical and experimental structures.1,8 order as implemented in the DFTB+ code.43,54 The main idea
behind the DFTB method, originating from Slater and Koster,55 is to
2.2 Interaction potentials describe the Hamiltonian eigenstates with an atomic-like basis set
To establish the quality of our calculations we compare hydrogen– and replace the Hamiltonian with a parameterized Hamiltonian
borophene DFTB pair-potentials51 with those calculated using the matrix whose elements depend only on the internuclear distances
plane wave (PW) DFT method implemented in the Vienna ab initio (this requires the integrals of more than two centers to be neglected)

J. Mater. Chem. C This journal is © The Royal Society of Chemistry 2017

 View Article Online

Journal of Materials Chemistry C Paper

and orbital symmetries. The method includes ab initio concepts in for graphene. The pair potentials of SCC-DFTB for borophene
relating the Kohn–Sham (KS) orbitals of the atomic configu- are generally slightly stronger in bonding and repulsion compared
ration to a minimal basis of the localized atomic valence to the DFT results. These deficiencies of DFTB are on average only
orbitals of the atoms. The energy may be approximated as a 0.05 eV per atom which will impose only insignificant inaccuracies
sum of the occupied KS single-particle energy and a repulsive in our MD simulations, given that the kinetic energy spectrum of
energy, which can be obtained from DFT calculations in the projectile is in the range of 0.25–100 eV. DFTB potential energy
properly chosen reference systems, and presented through curves agree well in Fig. 2 with the DFT ones for the sites that have
the so called Slater–Koster (SK) parameter files.55 We use the a lower electronic density. These differences are most visible in
Published on 09 May 2017. Downloaded by University of California - San Diego on 31/05/2017 01:39:38.

semi-relativistic SK parameter set for materials science (matsci) Fig. 2(a) and (b) where the depicted adsorption sites are chemically
simulations for B–B and B–H interactions.51 Because the KS very different. We see that the bridge 1 and bridge 3 positions
orbitals are well defined, one can derive expressions within the located along the linear boron chains show a greater error than the
DFTB method for the properties in the same way as within a full bridge 2 and hollow positions situated between the chains. On the
DFT scheme. Thus, in order to have a better description of other hand DFTB performs poorly for graphene (Fig. 2d), especially
electronic systems and better transferability of DFTB in the in the hollow site where it predicts a bonded state which strongly
cases where long-range Coulomb interactions are significant, disagrees with the result obtained by DFT. In addition, DFTB
the method has been improved, giving rise to the so called self- potentials in the bridge and top positions for graphene are more
consistent charge correction DFTB (SCC-DFTB).40 diffuse and shifted toward longer bond lengths. All typical
The DFT calculations were performed with a plane wave basis adsorption sites of borophene have bonded states with higher
set using the projector-augmented-wave (PAW) pseudopotentials binding energies (0.042–0.085 eV per at.) than graphene does
of Blochl,56 as adapted by Kresse and Joubert57 and with the (0.017–0.026 eV per at.). These results verify our computational
correlation–exchange GGA functional of Perdew, Burke, and approach using SCC-DFTB. They also give first insights into how
Ernzerhof (PBE).58,59 A Monkhorst–Pack k-point mesh of 16  hydrogen will perform in impacts with borophene. Because of
8  1 was used in both DFT and DFTB calculations.60 The kinetic the higher atomic density it is more probable for the impinging
energy cutoff for the VASP plane wave basis was set to 400 eV. atom at borophene to encounter a top position than at graphene.
The adsorption sites and their respective calculated potential Also as a result of the boron’s larger coordination number there
energies are shown in Fig. 2(a–c) for borophene and in Fig. 2(d) are more bridge positions per unit cell with higher barriers for

Fig. 2 Potential energy curves of different hydrogen adsorption sites for (a–c) borophene and (d) graphene.

This journal is © The Royal Society of Chemistry 2017 J. Mater. Chem. C

 View Article Online

Paper Journal of Materials Chemistry C

the hydrogen to overcome. The hollow positions in borophene

have much smaller surface areas than in graphene. The deeper
potentials result in a stronger bonding of hydrogen to the surface
which decreases the probability of thermal desorption thereby
ensuring a higher detection accuracy. The obtained results indicate
the favorable properties of borophene for hydrogen capture.

2.3 Hydrogen irradiation on borophene

Published on 09 May 2017. Downloaded by University of California - San Diego on 31/05/2017 01:39:38.

The study of the transmission, adsorption, and reflection processes

of hydrogen irradiation of a quasi-planar borophene sheet is done
by QCMD simulations, using the SCC-DFTB40 approach described
in Section 2.1, while the collision dynamics is performed using the
velocity Verlet algorithm. We applied periodic boundary conditions
to the borophene surface cell, irradiating only the central 1 nm2 of Fig. 3 Structures used in electron transport calculations: (a) the x-direction,
the surface. The borophene structure is energy optimized and (b) the y-direction. The blue and red areas indicate leads with the dashed line
separating the two principal layers. The middle region is the investigated device.
thermalized to 300 K using a Nose–Hoover thermostat prior to
bombardment with hydrogen atoms.
In order to calculate adsorption, transmission, and reflection We calculate the electron transport through pristine and the
probabilities, we repeat the impact of a hydrogen atom on the partially covered hydrogen detection surface using the non-
same, optimized, and thermalized (at 300 K) borophene sheet equilibrium Green’s function method implemented in the SCC-
using 2000 independent trajectories. For each trajectory the DFTB+NEGF code.49,50 We have performed the calculation with
target borophene structure was identical. The initial position 26 voltages ranging from 0–300 mV and 11 hydrogen coverages
of the hydrogen projectiles is uniformly distributed on the x–y in range 0–100%.
plane covering the target borophene with a separation distance SCC-DFTB-NEGF is the tight-binding DFT code adapted for
of 0.22 Å, which enables us to obtain acceptable statistics of the calculation of the quantum transport using non-equilibrium
collision events. Each hydrogen atom is emitted vertically Green’s function (NEGF) method.49 In general, this method is a
toward the borophene sheet (Fig. 1) from a height of 0.6 nm generalization of the Landauer approach,62 which uses Green’s
above the pink borophene layer in Fig. 1. Considering that an functions to calculate the tunneling current between two conducting
approaching H+ ion would be neutralized by the electron Fermi leads, which are subject to the open boundary conditions. With a
sea of the conducting surface, the projectile trajectory is calculated molecular device between the biased leads (source and drain) the
using QCMD,39 the total electronic energy of the adiabatic system NEGF approach49,50 calculates charge densities consistently, solving
is solved quantum-mechanically at each time step (0.25 fs), the Poisson equation under these non-equilibrium conditions. This
maintaining fixed atom positions after that the forces on each enables self-consistent solution of the Kohn–Sham equations of
atom are updated, and the atoms are moved classically for the DFT, for either coherent or incoherent (phonon scattering) transport.
time step.34,61 An electronic temperature of 1000 K, equivalent
to Fermi–Dirac smearing, is used in the QCMD simulations.
This procedure34 is repeated for the following impact energies: 3 Results and discussion
0.25, 1, 2, 5, 8, 10, 20, 50, and 100 eV. A typical collision lasts for
300–500 fs. The calculation is done using the embarrassing 3.1 Hydrogen bombarding of a borophene sheet
parallelization at 160–240 cores of the computer cluster. We have calculated the probability of transmission, reflection,
and adsorption of hydrogen irradiation on the corrugated boron
2.4 Electron transport calculations sheet for various impact energies using the approach described in
In the electron transport simulations,49,50 we used a 5  5 super- Section 2.3. The probabilities in Fig. 4 are calculated as the NH/NTH
cell of borophene for the detection surface on which we varied the ratio, where NTH is the total number of the incident hydrogen
amount of adsorbed hydrogen from 0–100% at convex top 1 atoms and NH is the number of atoms which are reflected,
adsorption positions. In our study we consider two cases in which transmitted, or bonded to the target. We compare our results for
we cover either one or both sides of the borophene sheet with borophene with those reported for a graphene layer in ref. 34.
hydrogen atoms. A full two-sided coverage corresponds to the Both graphene and borophene behave similarly below 0.5 eV
borophane structures shown in Fig. 1(d) and (e). Electron transport impact energy. However significant differences arise at higher
is calculated along the x-direction of the linear boron chains, (2e) energies.34 The probability of adsorption for graphene rapidly
as well as along the y-direction orthogonal to the corrugation (2f). decreases in favor of reflection and ceases completely for
The electrodes used in these calculations were 10  5 supercells energies above 2.5 eV. Instead the probability of adsorptions
of pristine borophene layers. The studied system as well as its in borophene stays above 50% for energies below 8 eV in which
division into the respective parts is shown in Fig. 3. The whole transmission takes over reaching close to 100% at 100 eV. In
device was structurally optimized before the electron transport graphene transmission starts increasing after the material
calculations. stops absorbing hydrogen and becomes a dominant process

J. Mater. Chem. C This journal is © The Royal Society of Chemistry 2017

 View Article Online

Journal of Materials Chemistry C Paper

Published on 09 May 2017. Downloaded by University of California - San Diego on 31/05/2017 01:39:38.

Fig. 5 Angular distribution of reflected and transmitted hydrogen atoms

Fig. 4 Comparison of the probability of transmission, adsorption, and at different impact energies on borophene. We show the angular distributions
retention as a function of the impact energy for borophene (solid lines with full only for those energies for which a significant probability of the given incident
points) and graphene (dashed lines). There are significant differences between was observed (410%).
borophene and graphene profiles for adsorption and reflection probability. On
the other hand the transmission probability follows a very similar trend in both
materials. This leads us to conclude that the processes governing adsorption The angular distribution in Fig. 5 shows that in the case of reflection
and reflection are considerably different for borophene and graphene whereas we observed a wide range of reflection angles for the whole impact
the process of transmission is similar if not the same for both materials.
energy range where reflection was non-negligible (1–20 eV). This
indicates that the trajectories of the atoms were strongly affected by
at around 20 eV. It is interesting that the transmission curves of the highly inhomogeneous potential of the surface before being
borophene and graphene are quite similar in the whole range backscattered. For transmission we can see that at the highest
of incident energies. This result is somewhat intuitive for boron studied energy (100 eV) the atoms passed trough the sheet with their
since after 2 eV (probability of transmission 44%) the projectile trajectories unaffected by the encountered potential with all of them
would have enough energy to pass through the structure without having a 1801 transmission angle. For 50 eV we see that some
getting caught in the binding potential. We did not notice that any trajectories were affected only to a small extent. Only at energies
boron atoms were kicked out of the structure or that the structure below 20 eV one can see noticeably diverse angular distributions but
would relax into a different one, even at 100 eV. A very important none would exceed 1201. This is another argument in favor of
feature is the low reflection probability of borophene in comparison multilayer detector concept, since atoms passing trough the first
with graphene. This indicates the possibility of a multilayer boro- layers would be scattered only to a minor degree.
phene detection system which would efficiently detect hydrogen In Fig. 6, we illustrate the position of adsorption events
at impact energies larger than 1 eV. Such a device could suffer a obtained at various impact energies. At 1 eV the majority of the
smaller loss of detection due to small reflection and high adsorption. adsorbed hydrogen atoms are bonded to the upper layer atoms
The angular distribution of reflection and transmission of of the borophene sheet, primarily along the lines of the top 1
hydrogen atoms is independent of the azimuthal angle. In Fig. 5 and bridge 1 positions established in Fig. 2. This is indicated by
we define the polar angle y with respect to the surface normal. The a clear lack of adsorption events along the lines of gray boron
direction of the velocity of the impact particle is in the negative atoms in the xy-plane view of the system. With the increase in
z-axis direction. We have calculated the angular distribution of the impact energy the hydrogen atoms transmit to the other
the reflected and transmitted hydrogen atoms according to the side where they might be adsorbed. When transmitted they mostly
following equation:34 remain bonded in the concave bridge 3 and top 2 positions. This is
  most probably caused by the loss of kinetic energy when passing
Dy the upper layer after which the hydrogen atom is not mobile
N yþ
dN 2 enough to pass further to the more stable top 1 and bridge 1
ffiC ; (1)
dO 2p sin yDy positions. Hydrogen atoms with kinetic energies above 10 eV
where C is a normalization constant for the angular distribution, remain almost entirely bonded to the most stable top 1 position.
  This is in agreement with the findings of several authors that
Dy
such that its maximum value is at the unit circle, N y þ is the borophane structure is very stable and therefore the hydro-
2
the number of reflected or transmitted atoms into a Dy range genated system will ultimately converge toward it.37,46–48
with the center at the polar angle. We present the calculated
angular distribution of reflected (0 o y o 901) and transmitted 3.2 Conductance of hydrogenated borophene
(901 o y o 1801) hydrogen atoms for different impact energies. Borophene has an atypical electronic structure since it behaves as
We notice that reflected atoms follow a cos4(y) function and an anisotropic metal8 as shown in its band structure in Fig. 7(a).
transmitted hydrogen atoms tend to follow a cos6(y) function. The gray areas are paths through the k-space parallel to the x-axis

This journal is © The Royal Society of Chemistry 2017 J. Mater. Chem. C

 View Article Online

Paper Journal of Materials Chemistry C

Published on 09 May 2017. Downloaded by University of California - San Diego on 31/05/2017 01:39:38.

Fig. 6 Position of absorption events for the corrugated boron sheet. Top
(z) and side (x) (y) view for impact energies of (a and b) 1 eV, (c and d) 5 eV
and (e and f) 10 eV.

Fig. 8 The total transmission probabilities summed over all channels in

the x-direction (a) and the y-direction (b) for pristine borophene and
borophene hydrogenated on one and both sides.

consistent with the finding of other authors.8,34,44 Electronic bands

also cross the Fermi level at other paths, in white areas, which are
T–U and H–G. These two paths have a significant x component
which is in the conducting direction. Borophane exhibits a similar
anisotropic electronic structure but instead of bands crossing the
Fermi level it contains only Dirac cones, Fig. 7(b). The band
structures suggest that both of these materials are conductors
although borophane is expected to have a significantly lower
conductance, as shown by our electron transport calculations.
We calculated the band structure of the unit cells of both
borophene and borophane using VASP with a kinetic energy
cutoff of 800 eV and using a Monkhorst–Pack k-point mesh of
50  25  1 in both VASP and DFTB+. The bands obtained by
both codes, VASP and DFTB+, below their respective Fermi levels
are almost identical albeit for a slight shift. The difference in
bands above the Fermi level comes from the different treatment
of high energy bands in the SCC-DFTB method compared to DFT.
The borophene’s unique band structure indicates higher
conductance in the direction of corrugation, the x-direction,
then in the y-direction.46 We have calculated the transmission
probabilities depicted in Fig. 8. There is a significant difference
between x and y directions with significance lower transmission
in the y direction. The coverage of borophene with hydrogen
Fig. 7 The band structure of (a) borophene and (b) borophane. The gray panels
significantly reduces the transmission.
indicate paths through the k-space perpendicular to the direction of corrugation.
The band structures were shifted so that their Fermi levels are at 0 eV. We calculate the conductance of our system using the
Landauer formula62

in Fig. 1(b) which are perpendicular to the direction of structural e2 X

GðEÞ ¼ Tn ðV; Ef ; T Þ; (2)
corrugation. In these paths the bands cross the Fermi level. This is p
h n

J. Mater. Chem. C This journal is © The Royal Society of Chemistry 2017

 View Article Online

Journal of Materials Chemistry C Paper

single and double sided coverages in the range 0–100%. For

voltages below 300 mV, the obtained I–U characteristics are
linear. The conductance in the x-direction decreases about
8 times when the hydrogen atom’s two sided coverage increases
from virgin borophene to full coverage, illustrating the high
potential of borophene as an effective hydrogen detector.
Published on 09 May 2017. Downloaded by University of California - San Diego on 31/05/2017 01:39:38.

Acknowledgements
We acknowledge the support from the Fulbright Commission
(Grant 1516093) to MN and CONACyT for postdoctoral scholarship
no. 267898 to FJDG. MN appreciates also the support from Slovak
Grant Agencies (VEGA Grant 1/0092/14, SRDA Grant APVV-15-
0105). PSK acknowledge support of the Department of Energy,
Laboratory Directed Research and Development funding, under
contract DE-AC02-09CH11466 with Princeton University for
operating Princeton Plasma Physics Laboratory. Results in this
paper were obtained using the LI-red institutional cluster at the
Institute for Advanced Computational Science in Stony Brook
University and the Comet cluster of the San Diego super
computer center which is part of (XSEDE). The DFT benchmark
calculations were conducted in the Computing Center of the
Fig. 9 Conductance of borophene as a function of hydrogen coverages
Slovak Academy of Sciences using the supercomputing infra-
in both metallic (x), hollow symbols, and nonmetallic (y), solid symbols, structure acquired in project ITMS 26230120002 and 26210120002
directions. Single and two sided hydrogen coverages are shown by squares (Slovak infrastructure for high-performance computing) supported
and circles, respectively. The inset depicts the linear I–U characteristics in by the Research & Development Operational Program funded by
the range of 0–300 mV.
the ERDF.

where G(E) is the electrical conductance, n is the transport

channel, and Tn is the transmission probabilities of the channels
References
which depend on voltage V, Fermi energy Ef, and temperature T. 1 K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang,
The inset of Fig. 9 shows the tunneling currents in the x-direction S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004,
as a function of the voltage for various coverages of hydrogen. I–U 306, 666.
characteristics stay linear for voltages below 300 mV, at all 2 S. Cahangirov, M. Topsakal and E. Akturk, et al., Phys. Rev.
coverages. From the slopes of the I–U lines we obtain conductances Lett., 2009, 102, 236804.
shown by hollow circles in Fig. 9. Conductance as a function of 3 M. E. Davila, L. Xian and S. Cahangirov, et al., New J. Phys.,
percentage for one side coverage of hydrogen is shown by hollow 2014, 16, 095002.
squares. Obviously, the conductance is a strong function of 4 J. Sone, T. Yamagami and Y. Aoki, et al., New J. Phys., 2014,
hydrogen coverage. Full circles and square symbols depict that 16, 095004.
the conductance is in the corrugated y-direction. 5 I. Boustani, Phys. Rev. B: Condens. Matter Mater. Phys., 1997,
55, 16426.
4 Conclusions 6 Z. A. Piazza, H. S. Hu and W. L. Li, Nat. Commun., 2014,
5, 3113.
In this paper we perform computer simulations of hydrogen 7 H. Liu, J. Gao and J. Zhao, Sci. Rep., 2013, 3, 3238.
irradiation of a corrugated borophene sheet for collision energies 8 A. J. Mannix, X. F. Zhou and B. B. Kiraly, et al., Science, 2015,
in the range of 0.25–100 eV. We use the quantum-classical 350, 1513.
molecular dynamics with the SCC-DFTB approach, whose quality 9 B. Feng, J. Zhang and Q. Zhong, et al., Nat. Chem., 2016, 8, 563.
we verify using the PW-DFT method of the VASP package. We 10 A. Garcı́a-Fuente, J. Carrete, A. Vega and L. J. Gallego, Phys.
calculate and analyze the transmission, reflection, and adsorption Chem. Chem. Phys., 2017, 19, 1054.
probabilities. The corrugated borophene sheet is a better hydrogen 11 Z. Wang, T.-Y. Lü and H.-Q. Wang, et al., Phys. Chem. Chem.
adsorber than graphene in the whole studied impact energy range. Phys., 2016, 18, 31424.
Upon relaxation most of the hydrogen atoms are bonded to the 12 B. Mortazavi, O. Obaidur Rahaman and A. Dianat, et al.,
convex top positions at both sides of the surface. The quantum Phys. Chem. Chem. Phys., 2016, 18, 27405.
conductance of the borophene sensor as function of the hydrogen 13 H. Tang and S. Ismail-Beigi, Phys. Rev. Lett., 2007, 99, 115501.
coverage (in percents) is calculated using the NEGF method at 11 14 X. F. Zhou, et al., Phys. Rev. Lett., 2014, 112, 085502.

This journal is © The Royal Society of Chemistry 2017 J. Mater. Chem. C

 View Article Online

Paper Journal of Materials Chemistry C

15 K. C. Lau and R. Pandey, J. Phys. Chem. C, 2007, 111, 2906. 40 M. Elstner, D. Porezag and G. Jungnickel, et al., Phys. Rev. B:
16 E. S. Penev, S. Bhowmick and A. Sadrzadeh, et al., Nano Lett., Condens. Matter Mater. Phys., 1998, 58, 7260.
2012, 12, 2441. 41 Q. Cui, M. Elstner and E. Kaxiras, et al., J. Phys. Chem. B,
17 C. Ozdogan, et al., J. Phys. Chem. C, 2010, 114, 4362. 2001, 105, 569.
18 X. Wu, et al., ACS Nano, 2012, 6, 7443. 42 W. Han, M. Elstner and K. J. K. J. Jalkanen, et al., Int.
19 H. Sachdev, Science, 2015, 350, 1468. J. Quantum Chem., 2000, 78, 459.
20 H. Sun, Q. Li and X. G. Wan, Phys. Chem. Chem. Phys., 2016, 43 B. Aradi, B. Hourahine and T. J. Frauenheim, J. Phys. Chem.
18, 14927. A, 2007, 111, 5678.
Published on 09 May 2017. Downloaded by University of California - San Diego on 31/05/2017 01:39:38.

21 Z. Zhang, E. S. Penev and B. I. Yakobson, Nat. Chem., 2016, 44 D. Porezag, T. Frauenheim and T. Köhler, et al., Phys. Rev. B:
8, 525. Condens. Matter Mater. Phys., 1995, 51, 12947.
22 T. Tsafack and B. I. Yakobson, Phys. Rev. B: Condens. Matter 45 H. Sun, Q. Li and X. G. Wan, Phys. Chem. Chem. Phys., 2016,
Mater. Phys., 2016, 93, 165434. 18, 14927.
23 H. Shu, F. Li and P. Liang, et al., Nanoscale, 2016, 8, 16284. 46 J. E. Padilha, R. H. Miwa and A. Fazzio, Phys. Chem. Chem.
24 R. D. Dewhurst, R. Claessen and H. Braunschweig, Angew. Phys., 2016, 18, 25491.
Chem., Int. Ed., 2016, 55, 4866. 47 Y. Liu, Y. Y. Dong and Z. Tang, et al., Phys. Chem. Chem.
25 J. B. Levine, et al., Adv. Funct. Mater., 2009, 19, 3519. Phys., 2016, 18, 31424.
26 M. Herrmann, et al., Int. J. Refract. Met. Hard Mater., 2013, 39, 53. 48 L. C. Xu, A. Du and L. Kou, Phys. Chem. Chem. Phys., 2016,
27 B. P. Lavrov, M. Osiac and A. V. Pipa, et al., Plasma Sources 18, 27284.
Sci. Technol., 2003, 12, 576. 49 A. Pecchia and A. Di Carlo, Rep. Prog. Phys., 2004, 67, 1497.
28 P. J. Kervalishvili, Amer. J. Condens. Matt. Phys., 2015, 5, 10. 50 A. Pecchia, G. Penazzi and L. Salvucci, et al., New J. Phys.,
29 T. Hübert, L. Boon-Brett and G. Black, et al., Sens. Actuators, 2008, 10, 065022.
B, 2011, 157, 329. 51 B. Lukose, A. Kuc, J. Frenzel and T. Heine, Beilstein
30 A. Krasheninnikov and K. Nordlund, J. Appl. Phys., 2010, J. Nanotechnol., 2010, 1, 60.
107, 071301. 52 G. Kresse and J. Hafner, Phys. Rev. B: Condens. Matter Mater.
31 B. Peng, H. Zhang and H. Shao, et al., J. Mater. Chem. C, Phys., 1993, 47, 558.
2016, 4, 3592. 53 G. Kresse and J. Hafner, Phys. Rev. B: Condens. Matter Mater.
32 Y. Liu, E. S. Penev and B. I. Yakobson, Angew. Chem., Int. Ed., Phys., 1994, 49, 14251.
2013, 52, 3156. 54 http://www.dftb-plus.info/.
33 C.-H. Hu, A. R. Oganov, Q. Zhu, G.-R. Qian, G. Frapper, 55 P. C. Slater and G. F. Koster, Phys. Rev., 1954, 94, 1498.
A. O. Lyakhov and H.-Y. Zhou, Phys. Rev. Lett., 2013, 110, 165504. 56 P. E. Blochl, Phys. Rev. B: Condens. Matter Mater. Phys., 1994,
34 R. C. Ehemann, P. S. Krstić and J. Dadras, et al., Nanoscale 50, 17953.
Res. Lett., 2012, 7, 198. 57 G. Kresse and D. Joubert, Phys. Rev. B: Condens. Matter
35 I. Deretzis, G. Fiori and G. Iannaccone, et al., Physica E, Mater. Phys., 1999, 59, 1758.
2012, 44, 981. 58 J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett.,
36 F. Schedin, A. K. Geim and S. V. Morozov, et al., Nat. Mater., 1996, 77, 3865.
2007, 6, 652. 59 J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett.,
37 L. C. Xu, A. Du and L. Kou, Phys. Chem. Chem. Phys., 2016, 1997, 78, 1396.
18, 27284. 60 H. J. Monkhorst and J. D. Pack, Phys. Rev. B: Solid State,
38 Y. Liu, Y.-J. Dong and Z. Tang, et al., J. Mater. Chem. C, 2016, 1976, 13, 5188.
4, 6380. 61 A. F. Oliviera, G. Seifert and T. Heine, et al., J. Braz. Chem.
39 P. S. Krstic, R. J. Harrison and B. G. Sumpter, Phys. Scr., Soc., 2009, 20, 1193.
2006, T124, 101. 62 R. Landauer, IBM J. Res. Dev., 1957, 1, 223.

J. Mater. Chem. C This journal is © The Royal Society of Chemistry 2017