Using first-principles calculations, we computationally designed a new two-dimensional (2D) inorganic material, Be3N2 monolayer with a flat hexagonal structure similar to graphene. Good stability of the Be3N2 monolayer is demonstrated by its moderate cohesive energy, the absence of imaginary modes in its phonon spectrum, and the high melting point predicted by molecular dynamics (MD) simulations. The Be3N2 monolayer is a direct band gap semiconductor with a band gap of 3.831 eV that can be effectively tuned by employing an external strain. The wide band gap and outstanding strain-engineered properties make Be3N2 monolayer a highly versatile and promising 2D material for innovative applications in microelectromechanical and nanoelectronic devices. Additionally, the one-dimensional Be3N2 nanoribbons which divided by Be3N2monolayer, are computed to have quite rich characteristics such as direct band gaps with various values, depending on the direction of the division and the width of nanoribbons.

The discovery of graphene has ignited universal interest in two-dimensional nano materials. Graphene is overwhelming because it possesses extraordinary physical properties and meets the strong demands of nano electronics. However, graphene is not the single member in the family of 2D nanomaterials. Other graphene-like nanomaterials, such as silicene,1 germanene,2 metal dichalcogenides,3 borophene4,5 and Black Phosphorus6 etc, also joined the 2D family. As research on graphene has increased, methods for finding materials of similar nature and structure to graphene have been created. Recently, 2D metal-nitride sheets received great attention due to their novel versatile electronic, superconducting, optical, plasmonic photocatalytic properties, as well as many potential applications, such as semiconductor devices, high-performance catalyst for energy conversion and storage, Li-ion batteries, and ultracapacitors.7–11 The painstaking research has uncovered a large number of nitride single-layer structures with special properties, for example the BeN2 monolayer12 which has a direct band gap and a high carrier mobility.

In this paper, we report a first principles prediction of a new 2D inorganic material, namely, the Be3N2 monolayer. Using material prediction software, the lowest-energy structure can be found by providing a given stoichiometry of Be atoms and N atoms. Applying the same method, a novel binary layer of beryllium nitride with similar graphene structure is predicted. We examined various properties and structures of the Be3N2 monolayer in the same manner as graphene was studied. Based on the calculated band structure, Be3N2 was found to be a direct wide bandgap semiconductor which has its potential application in light-emitting devices such as light-emitting diodes (LEDs) and semiconductor lasers as well as RF signal processing devices.

We used the Vienna ab initio simulation package (VASP13) to performed the DFT computations, and used the projector-augmented plane ware (PAW14)approach to represent the ion-electron interaction. The electron exchange-correlation functional was treated using generalized gradient approximation (GGA15) in the form proposed by Perdew, Burke and Ernzerhof (PBE). The 500 eV energy cutoff was adopted and Brillouin zone was with a 4x4x2 Γ-centered Monkhorst-Pack k-points grid for geometry optimization and self-consistent calculation. The energy precision was set to 10-6eV and atomic position was fully relaxed until the maximum force on each atom was less than 10-4eV/Å.16 

The band gaps was underestimated by the standard GGA, but we prefer to use the Heyd-Scuseria-Ernzerhof (HSE0617,18) hybrid function, which was proven to be a reliable method for the calculation of calculating the band structure of Be3N2 monolayers. The band structure of Be3N2 was computed along the special lines of Γ(0, 0, 0)→M(1/2, 0, 0)→K (-1/3, 2/3, 0)→Γ(0, 0, 0) in the k-space.

To determine the stability of this Be3N2 monolayer, we measured its phonon spectra and molecular dynamics data. Among them, phonon dispersion19 analysis was performed using the Phonopy code interfaced with the density functional perturbation theory (DFPT) as implemented in VASP. Ab initio Molecular Dynamics20 (AIMD) simulations were performed using the PAW method and the PBE function to assess all the thermal predictions of the Be3N2 monolayer. The vibrational and thermodynamic properties are all obtained with PBE. And in the MD simulations, the initial configuration of Be3N2 monolayers with 4×4 supercell (48 Be atoms and 32 N atoms) was annealed at different temperatures, each MD simulation in NVT ensemble lasted for 10 ps with a time step of 2.0 fs, and the temperature was controlled by using the Nose-Hoover method.

The lowest-energy structure of Be3N2 with a flat hexagonal structure predicted by CALYPSO21 code is shown in Fig. 1a, The Be3N2 monolayer is placed in the x-y plane with the z-direction perpendicular to the layer plane and a vacuum space of 19.8075Å in the z-direction is used to avoid the interaction between adjacent layers. The optimized Be3N2 flake crystallizes in the hexagonal space group P6/mmm (No.191) with a lattice constant of a = b = 5.24212 Å. There is a unique Be atom (site symmetry 3g) and An independent N atom (site symmetry 3g). Be atoms at the Wyckoff22 3g site act as a “bridge”, forming a strong band with N atoms, each of which is a triangular coordination without any distortion characteristics. The length of the sp2-hybridized, Be-N band is 1.51327Å with an angle of 120 °as shown in Fig. 1b. Structural evolution can be seen as the process of inserting a “bridge” into the network by expanding the pore size in flat nitrogen.

FIG. 1.

(a) Top view of the optimized geometry of the Be3N2 monolayer with a unit cell labeled by the black solid line. (b) Zoom-in structure of the Be3N2 monolayer with the bond distances and bond angles marked.

FIG. 1.

(a) Top view of the optimized geometry of the Be3N2 monolayer with a unit cell labeled by the black solid line. (b) Zoom-in structure of the Be3N2 monolayer with the bond distances and bond angles marked.

Close modal

We calculated the charge density of the Be3N2 monolayer to further understand the bond formation between the Be and N atoms (see Fig. 2). It can be seen from the figure that the electron cloud around the Be atom is relatively sparse, the electrons are all surrounded by N atoms, and the N atom is electronegative. Two electrons in the outermost 2s orbit of the Be atom are transferred to the N atom and participate in the N atom 2s and 2p orbital hybrids go in. The N-Be-N connected rings form a graphene-like benzene six-membered ring structure.23 

FIG. 2.

Charge density map of Be3N2 nanosheet, isosurface of charge density plotted with the value of 0.2 au and the charge density map sliced perpendicular to (0 0 1) direction for Be3N2 monlayer.

FIG. 2.

Charge density map of Be3N2 nanosheet, isosurface of charge density plotted with the value of 0.2 au and the charge density map sliced perpendicular to (0 0 1) direction for Be3N2 monlayer.

Close modal

We further calculated the charge density difference and electron localization function (ELF) of the Be3N2 monolayer to understand its unique structure and chemical bond formation mechanism. Charge density difference is a useful tool for determining the bond. The negative position which is usually shown in blue is the electron divergence area, where the electron is lost, while the positive position is the electron convergence area which is shown in yellow, where the electron is gained. The covalent bond electrons converge toward the middle of the two atoms, and the ionic bonds converge the electrons that are biased toward one side of them. As fig. 3(a) shows, electrons converge between Be and N atoms to form a Be-N covalent bond. The blue area on both sides of the Be atom and the yellow area in the middle of the six-ring indicate that the Be atom contributes to the construction of N-Be-N six-ring frameworks. Electron localization function (ELF) can effectively describe the chemical bonding of molecules and solids in the form of a contour pot in real space with values ranging from 0 to 1. In areas surrounded by isosurfaces with high numerical values, the electrons are more strongly localized in their regions. And in areas with low ELF values, the electronic localization is weak. we plotted the the ELF map of Be3N2 monolayer with the iso-value of 0.5 au. shown in fig. 3(b). As a whole, the Be and N atoms are surrounded by a red region, connected into a six-ring. Moreover, the ELF value of yellow color between the adjacent Be atoms is lower than the red between Be and N atoms, Showing that the possibility of forming a covalent bond between Be and N atoms is higher than that between Be atoms. Accordingly, this electron distribution analysis suggests that Be3N2 monolayer is a honeycomb structure consisting of a six-ring structure consisting of N-Be-N covalent bonds.

FIG. 3.

(a) Deformation charge density map of Be3N2 monolayer with the iso-value of 0.2 au. (b) Electron localization function map of Be3N2 monolayer with the ELF-value of 0.5 au.

FIG. 3.

(a) Deformation charge density map of Be3N2 monolayer with the iso-value of 0.2 au. (b) Electron localization function map of Be3N2 monolayer with the ELF-value of 0.5 au.

Close modal

By calculating the energy band structure of the Be3N2 monolayer, We studied its electrical properties. Fig. 4 shows the band structure and PDOS of Be and N. The PBE calculations give a band gap of 2.773 eV, while HSE06 brings a band gap of 3.831 eV (Fig. 4). Obviously, the Be3N2 sheet is a direct band gap semiconductor with the valence band maximum (VBM) and conduction band minimum (CBM) at the same point Γ. Considering the bulk structure of Be3N2, the beryllium nitride alpha phase has the same direct band gap in the central Brillouin zone Γ point of 4.43 eV. The α-Be3N2 is a cubic bcc structure like bixbyte, in which the metal ion has four tetrahedral neighbors and N has two types of distorted octahedral coordination groups.24 And the phase β-Be3N2, converted from α-Be3N2 crystals to 1400 degrees Celsius, is a wide band gap semiconductor with an indirect gap of 12.50 eV.25 The structure and properties of the two body structures of Be3N2 are very different from those of Be3N2 monolayer.

FIG. 4.

Band structure and DOS of Be3N2 nanosheet based on HSE06.

FIG. 4.

Band structure and DOS of Be3N2 nanosheet based on HSE06.

Close modal

It can be seen from the energy band structure diagram of Be3N2 monolayer that the band closest to the Fermi surface is the valence band. And, the N atom contributes most to the valence band according to the density of states. Of course, the electrons on the orbit of the Be atom are also involved in the formation of the electron orbits near the Fermi surface. Based on the PDOS of Be3N2 nanosheet (shown in Fig. 5), the Be atoms and N atoms have the similar shape in the valence band near the Fermi surface. The electrons of the Be atom s orbital and the N atom p orbital participate in the formation of electron orbitals. In order to further understand the bonding situation between Be3N2 monolayer atoms, we calculated the electronic charge distribution near the Fermi energy (see Fig. 6). It can be seen that some of the electrons surround the N atom participate in the electron orbital formation of the conduction band near the Fermi surface, and the other part forms the electron orbit of the valence band near the Fermi surface. Seen from the electron cloud profile, the VB is mainly filled with the electrons of the N-atom π orbital, and the CB is mainly filled with the electrons of the σ orbit of the N atom.

FIG. 5.

The PDOS of a nitrogen atom in Be3N2.

FIG. 5.

The PDOS of a nitrogen atom in Be3N2.

Close modal
FIG. 6.

Electronic cloud orbit profile near Fermi surface of Be3N2 nanosheet. (a) Electronic cloud distribution near the Fermi valence band. (b) Electronic cloud distribution near the Fermi surface conduction band.

FIG. 6.

Electronic cloud orbit profile near Fermi surface of Be3N2 nanosheet. (a) Electronic cloud distribution near the Fermi valence band. (b) Electronic cloud distribution near the Fermi surface conduction band.

Close modal

In order to evaluate the thermodynamic stability, we calculated the cohesive energy Ecoh=3xEBe+2xENEBe3N2/5x, with EBe, EN and EBe3N2 being the total energies of a single Be atom, N atom and one unit cell of the Be3N2 monolayer. The Be3N2 monolayer has a cohesive energy26 of 5.82 eV/atom. For comparison, using the same calculation method, the cohesive energies of graphene27 and silicide28 were 7.85, 3.98 eV/atom, respectively. Thus, while the cohesion can be significantly lower than the Be and N 3D bulk phases, this is sufficient when the material is limited to 2D, because of its stability between the graphene and silicence.

We evaluated the kinetic stability of the Be3N2 monolayer by calculating the phonon dispersion along the high line of symmetry in the first Brillouin zone, shown in Fig. 7. There is no apparent hypocentric frequency in the phonon spectrum of the Be3N2 monolayer, indicating the kinetic stability of the Be3N2 monolayer. The highest frequencies of Be3N2 sheet were 1265cm-1 (= 37.95THz) higher than the highest frequency of t-TiC sheet (810cm-1),29 Cu2Si (420cm-1)30 and MoS2 monolayers (473cm)31 These high frequencies in the phonon spectrum also indicate a strong Be-N interaction in the newly predicted Be3N2 monolayer.

FIG. 7.

Phonon dispersion of the Be3N2 monolayer.

FIG. 7.

Phonon dispersion of the Be3N2 monolayer.

Close modal

The stability of the Be3N2 sheet can be further tested by AIMD simulations at different temperatures. Simulations were performed using relatively large 4 × 4 supercell at temperatures of 300K, 600K and 900K. The geometry at the end of the 10 ps simulation (Fig. 8) shows that the Be3N2 sheet can retain its original structure at 300K and collapse at very high temperatures of 600 and 900K. Its high temperature stability is not very good, but it can be stable at room temperature.

FIG. 8.

Snapshots for equilibrium structures of Be3N2 monolayer at the temperatures of (a) 300K, (b) 600K and (c) 900K, at the end of 10ps AIMD simulations.

FIG. 8.

Snapshots for equilibrium structures of Be3N2 monolayer at the temperatures of (a) 300K, (b) 600K and (c) 900K, at the end of 10ps AIMD simulations.

Close modal

To effectively use the band gap of the Be3N2 monolayer, the relevant application might be achieved by changing the size of the band gap by altering the strain. From the Fig. 9, we can find in the process of changing uniaxial strain32 of Be3N2 monolayer, The sizes of the band gap decrease linearly with the strain within a certain range without disrupting the cell structure of Be3N2.

FIG. 9.

Band gaps of the Be3N2 monolayer as a function of the biaxial strain.

FIG. 9.

Band gaps of the Be3N2 monolayer as a function of the biaxial strain.

Close modal

We further examined the Be3N2 nanoribbons33 by dividing the Be3N2 monolayer into armchair and zigzag nanoribbons along the x and y directions, calculating the band along the direction of the nanoribbon repetition (shown in Fig. 10). As examples, the band structures of H-terminated Na-Armchair with Na = 3–9 and Nz-Zigzag with Nz = 3–9 are shown in Fig. 10(a) and (b), respectively. The results obtained show that the bottom of the conduction band and the top of the valance band appeared at Γ point are always separated, representing a direct semiconductor character, indicating that the properties of electron transport are similar in both directions. From Fig. 10(a) and (b) we also find the number of both conduction band and valance band increases with increasing the ribbon width, additionally, as demonstrated in Fig. 11, the size of the band gaps both decrease with the increase of the width of nanoribbon from 0.5nm to 3.5nm while the band gap of armchair one starts to increase when zigzag one converges to 3.5ev.

FIG. 10.

The band structures of the H-terminated Be3N2 nanoribbons (a) Armchair with Na = 3–9 and (b) Zigzag with Nz = 3–9,respectively. The Fermi level is set to zero.

FIG. 10.

The band structures of the H-terminated Be3N2 nanoribbons (a) Armchair with Na = 3–9 and (b) Zigzag with Nz = 3–9,respectively. The Fermi level is set to zero.

Close modal
FIG. 11.

Variation of energy band gap as a function of the width for Be3N2 nanoribbons.

FIG. 11.

Variation of energy band gap as a function of the width for Be3N2 nanoribbons.

Close modal

We performed comprehensive DFT computations to check the possibility to obtain graphene-like materials containing Be atoms and N atoms which process planar hexagonal structure. Be3N2 monolayer has sound thermodynamic, kinetic and thermal stabilities and presents a direct band gap semiconductor characteristics which bandgap can be effectively tuned by employing an external strain. Also, the Be3N2 nanoribbons are proved to have direct band gaps with various values, depending on the direction of the division and the width of nanoribbons. Those quite rich characteristics might render Be3N2 monolayer an outstanding 2D nanomaterial with great promise for applications in novel nanodevices.

1.
B.
Feng
,
Z.
Ding
,
S.
Meng
 et al,
Nano Letters
12
(
7
),
3507
3511
(
2012
).
2.
L.
Li
 et al,
Advanced Materials
26
(
28
),
4820
4824
(
2014
).
3.
H. Q.
Wang
 et al,
Nature Nanotechnology
7
(
11
),
699
(
2012
).
4.
W.-L.
Li
 et al,
Journal of the American Chemical Society
136
(
35
),
12257
12260
(
2014
).
5.
J.
Liu
 et al,
International Journal of Modern Physics B
1850033
(
2017
).
6.
L.
Li
 et al,
Nature Nanotechnology
9
(
5
),
372
(
2014
).
7.
X.-F.
Zhou
,
X.
Dong
,
A. R.
Oganov
,
Q.
Zhu
,
Y.
Tian
, and
H.-T.
Wang
,
Phys. Rev. Lett.
112
,
085502
(
2014
).
8.
T.
Cui
,
R.
Lv
,
Z.-H.
Huang
,
H.
Zhu
,
J.
Zhang
,
Z.
Li
,
Y.
Jia
,
F.
Kang
,
K.
Wang
, and
D.
Wu
,
Carbon
49
,
5022
5028
(
2011
).
9.
L.
Qu
,
Y.
Liu
,
J. B.
Baek
, and
L.
Dai
,
ACS Nano
4
,
1321
1326
(
2010
).
10.
A. L. M.
Reddy
,
A.
Srivastava
,
S. R.
Gowda
,
H.
Gullapalli
,
M.
Dubey
, and
P. M.
Ajayan
,
ACS Nano
4
,
6337
6342
(
2010
).
11.
H. M.
Jeong
,
J. W.
Lee
,
W. H.
Shin
,
Y. J.
Choi
,
H. J.
Shin
,
J. K.
Kang
, and
J. W.
Choi
,
Nano Lett.
11
,
2472
2477
(
2011
).
12.
C.
Zhang
and
Q.
Sun
,
The Journal of Physical Chemistry Letters
7
(
14
),
2664
2670
(
2016
).
13.
G.
Sun
,
J.
Kürti
,
P.
Rajczy
,
M.
Kertesz
,
J.
Hafner
, and
G.
Kresse
,
Journal of Molecular Structure: THEOCHEM
624
(
1
) (
2003
).
14.
T.
Rangel
,
D.
Caliste
,
L.
Genovese
, and
M.
Torrent
,
Computer Physics Communications
208
(
2016
).
15.
R.
Ashraf
,
T.
Mahmood
,
S.
Riaz
, and
S.
Naseem
,
Chinese Journal of Physics
55
(
4
) (
2017
).
16.
W.
Mahmood
,
Advanced Materials Research
3913
(
1101
) (
2015
).
17.
B.
Ali
,
T.
Dahame
, and
B.
Bentria
,
Optical Materials
54
(
2016
).
18.
J.
Yang
and
Q.
Fan
,
IOP Conference Series: Materials Science and Engineering
167
(
1
) (
2017
).
19.
X.-P.
Luo
and
H.-L.
Yi
,
International Journal of Heat and Mass Transfer
114
(
2017
).
20.
B.
Curchod
,
T.
Penfold
,
U.
Rothlisberger
, and
I.
Tavernelli
,
Open Physics
11
(
9
) (
2013
).
21.
Y. C.
Wang
,
J.
Lv
,
L.
Zhu
, and
Y. M.
Ma
,
Phys. Rev. B: Condens. Matter Mater. Phys.
82
,
094116
(
2010
).
22.
P.
Avery
and
E.
Zurek
. RandSpg: Computer Physics Communications,
2016
.
23.
C.
Li
,
R.
Wang
,
H.
Xu
,
X.
Lu
,
H.
Konishi
, and
K.
He
,
American Mineralogist
101
(
7
) (
2016
).
24.
M. G. M.
Armanta
and
A.
Reyes-Serrato
,
Computational Materials Science
21
(
1
),
95
100
(
2001
).
25.
A.
Mokhtari
and
H.
Akbarzadeh
,
Physica B: Condensed Matter
324
(
1-4
),
305
311
(
2002
).
26.
A.
Safaei
,
Philosophical Magazine
91
(
10
) (
2011
).
27.
K.
Lyon
and
Z. L.
Miškovic
,
Nanoscale Systems: Mathematical Modeling, Theory and Applications
1
(
2014
).
28.
J.
Roger
,
V.
Babizhetskyy
,
R.
Jardin
,
R.
Guérin
,
C.
Moinet
,
U.
Burkhardt
, and
J.-F.
Halet
,
Zeitschrift für Kristallographie
221
(
5–7
) (
2006
).
29.
G.
Saeid
,
H.
Mohammad
,
F. J.
de Serres
,
S. N.
Bajestani
,
M.
Hamid
, and
L.
Andrea
,
Investigative Ophthalmology & Visual Science
48
(
1
) (
2007
).
30.
L.-M.
Yang
 et al,
Journal of the American Chemical Society
137
(
7
),
2757
2762
(
2015
).
31.
I.
Popov
,
S.
Gotthard
, and
T.
David
,
Physical Review Letters
108
(
15
) (
2012
).
32.
F.
Khoeini
,
Superlattices and Microstructures
81
(
2015
).
33.
F.
Ma
,
Z.
Guo
,
K.
Xu
, and
P. K.
Chu
,
Solid State Communications
152
(
13
) (
2012
).
All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).