types: zigzag
a
-graphyne nanoribbons (Z
a
GyNRs) and armchair
a
-graphyne nanoribbons (A
a
GyNRs). All of A
a
GyNRs are
nonmagnetic, and the band gap oscillates with a period of three
when the ribbon width increases [30], while the Z
a
GyNRs have a
stable antiferromagnetic configuration [31]. Yue et al. demon-
strated the bipolar spin-filtering effect in the symmetric Z
a
GyNRs,
and found the 100% spin-polarized current can be produced and
modulated by the direction of bi as voltage and/or magnetization
configuration of the electrodes [32].
To expand its practical applicati on in spintronics, doping
graphyne nanoribbons with 3d transition-metal atoms has been
considered as an effective way [33e35]. He et al. found tha t the
adsorption of 3d tran sition-metal atom not only efficiently mod-
ulates the elect ronic structures of graphyne, but also introduces
excellent magnetic properti es [36]. Since previous calculations
have indicated that the edge states have a significant effect on the
nanoribbons' electronic and magnetic properties [16] ,inthispa-
per we mainly focus on the effects of 3d transition terminated
atoms doping on the s ingle-edge-passivated Z
a
GyNRs. From first
principal method, the sp in-dependent transport properties of Fe-,
Co- and Ni-passivated Z
a
GyNRs are calculated, and we suggest a
route to manipulate spin current and design the spintronics
devices.
2. Model and method
The structural optimizations and spin-dependent t ransport
properties are performed by the Atomistix ToolKit (ATK) package
[37,38], which is based on the spin density functional theory (DFT)
and the fully self-cons istent non-equilibrium Green's function
(NEGF) method. The local spin density approximation (LSDA) is
employed for the exchange-correlation potential. The wave
functions of all the atoms are expanded by double-zeta pola rized
basis set. The real space grid technique is used with the en ergy
cutoff of 200 Ry as the required cutoff energy in numerical in-
tegrations. The k-point sampling 1, 1, and 100 are used in t he x, y, z
direction, respectively, where z is the direction of electron trans-
port. A vacuum layer of 16 Å is added in x and y di rections to avoid
the interaction between adja cent ribbons . The electronic states of
the electrodes are obtained from a st andard bulk ca lculation. The
basis set is a dopted for elements of systems and the convergence
criterion for the total energy is 10
5
Ry to achieve a balance be-
tween calcul ation efficiency and accuracy. The electron temper-
ature is set to 300 K. The geometrical structures are fully
optimized in the presence of the two-leads until the residual force
on each atom is smaller than 0.01 eV/Å. The spin-polarized cur-
rent through the s ystem is calculated using the Land auer-Büttiker
formula [3 9],
I
s
ðV
b
Þ¼
e
h
Z
m
R
ðV
b
Þ
m
L
ðV
b
Þ
f
T
s
ðE; V
b
Þ½f
L
ðE; V
b
Þf
R
ðE; V
b
ÞgdE (1)
where e is the electron charge, h is the Planck's constant, and T
s
is
the transmission of an electron with spin
s
. f
LðRÞ
ðE; V
b
Þ is the
Fermi-Dirac distribution function of the l eft (right) electrode, and
m
LðRÞ
ðV
b
Þ is the chemical potential of the left (right) electrode,
which depends on th e exte rnal bias voltage, V
b
. Under the
external bias V
b
, the electrochemical potential of the left and right
electrodes will be moved rigidly relative to each other, that
is,
m
L
ðV
b
Þ¼
m
L
ð0ÞþeV
b
=2and
m
R
ðV
b
Þ¼
m
R
ð0ÞþeV
b
=2.T
s
is obtained
from the equation,
T
s
ðE; V
b
Þ¼Tr
"
Im
(
X
r
L
s
ðE; V
b
Þ
)
G
r
s
ðE; V
b
Þ
Im
(
X
r
R
s
ðE; V
b
Þ
)
G
r
s
ðE; V
b
Þ
#
(2)
where G
r
ðG
a
Þ is the retarded (advanced) Green's function matrix,
and
P
r
L
ð
P
r
R
Þ is the retarded self-energy matrix for the left (right)
electrode.
3. Results and discussions
The Z
a
GNRs are originally obtained by cutting a 2D
a
-graphyne
along zigzag direction, and the edge carbon atoms are passivated by
single-H atoms to remove the dangling bonds. The studied models
are shown in Fig.1. The blue shade areas are presented the electrode
repeat units. Considering the electronic properties of Z
a
GyNRs [30]
and the calculated amount, the ribbon widths of 4 (about 2.5 nm)
have been chosen, and both edges are passivated by single-H atoms,
denoted as H-4Z
a
GyNRs-H, short for M1, as shown in Fig. 1(a). In
order to realize the spin-polarization in H-4Z
a
GyNRs-H, we also
adopt 3d transition-metal atoms (Fe, Co and Ni) to substitute the H
atoms on one edge as Fe-4Z
a
GyNRs-H, Co-4Z
a
GyNRs-H and
Ni-4Z
a
GyNRs-H as shown in Fig. 1(bed), for convenience, we
labeled them as M2, M3, and M4, respectively. Take M1 for example,
the corresponding device structure is shown in Fig. 1(e), which
consists of three parts: the left electrode, the center scattering re-
gion, and the right electrode. It is noted that the electrodes and the
contact between graphyne and the electrodes have crucial in-
fluences on the spin-polarized electronic transport properties of the
Z
a
GyNRs. Nevertheless, the contact between graphyne and the
electrodes is very complex in real practices, we thus just only
considered the internal electronic transport properties in Z
a
GyNRs
and a supercell of Z
a
GyNRs has been taken as electrode in our
calculations.
Fig. 1. (aed) The structure unit cells of M1-M4, respectively. The numbers denote
different C atoms along one hexagonal side. (e) Two-probe model of M1, and the Left,
Center, and Right represent the lef t electrode, the center scattering region and the right
electrode, respectively. The geometry in red dashed rectangle is unit cell. (For inter-
pretation of the references to colour in this figure legend, the reader is referred to the
web version of this article.)
M. Li et al. / Organic Electronics 44 (2017) 168e175 169