Full Paper

www.advelectronicmat.de

Strain Engineering for 2D Ferroelectricity in Lead

Chalcogenides
Tao Xu,* Xiaoyuan Wang, Jiawei Mai, Jingtong Zhang, Jie Wang, and Tong-Yi Zhang*

random-access memory (FeRAM), and

The quest for new 2D ferroelectric materials continues to arouse interest. electromechanical devices.[2,3] Achieving
Based on first-principles calculations, here, 2D ferroelectric properties in ferroelectric properties in low-dimensional
materials is of great importance for the
lead chalcogenides PbXs (X = S, Se, and Te) with a thickness of two atomic
development of miniaturized, integrated
layers via strain engineering is demonstrated. Although these materials electronic devices as well as for the fun-
are stable in a rocksalt-type cubic structure and are intrinsically non­ damental science attributed to the exotic
ferroelectric materials, an appropriate mechanical strain can readily activate physical properties and the emergence of
a paraelectric to ferroelectric phase transition and induce in-plane electric unusual ferroic patterns.[4–7] However, the
polarization in these atomic layers. The induced polarization magnitude ferroelectric order is inevitably destroyed
below the critical thickness in conven-
can be as large as 1.90 × 10−10 C m−1 at a biaxial strain of ε = 4.0%, which is
tional ferroelectrics, for example, 3–5 nm
comparable in magnitude to that of ultrathin ferroelectric SnTe. The Curie for BaTiO3 nanostructures,[8,9] due to the
temperature Tc of the materials is also estimated using effective Hamilto- significantly enhanced depolarization
nian simulations. The origin of the emerged ferroelectric phase is attributed field and finite size effects at nanoscale
to softening of the polar mode with applied strain. In addition, the band dimensions.
The rise of 2D materials opens a new
gaps of the crystals are found to be tunable with applied strain, which can
avenue for the discovery of atomic-scale
be adjusted to the ideal value of 1.3 eV for photovoltaic applications. The ferroelectrics.[10,11] From theoretical calcu-
results not only provide a new route to explore ferroelectricity in 2D mate- lations, several 2D ferroelectric materials,
rials but also suggest promising semiconducting ferroelectrics for solar including compounds[12–14] and elemental
applications. materials,[15,16] have been recently pro-
posed. In experiments, 2D ferroelectrics
were first established in two atomic layer
(AL)-thick SnTe,[17] which subsequently
1. Introduction stimulated research on the ferroelectricity (or ferroelasticity)
of other 2D SnTe analogues MXs (M = Si, Ge, Sn; X = S,
Ferroelectrics, such as the perovskite PbTiO3 and its solid solu- Se, Te).[12,13,18,19] Nevertheless, the lead chalcogenide PbX (X = S,
tions, are a typical class of functional materials that possess Se, and Te) atomic layers are intrinsically non­ferroelectric mate-
not only switchable electric polarization but also superior func- rials, although their corresponding bulk materials share the
tionalities, including a large piezoelectric response and a high same space group (Fm3m) as bulk SnTe at room temperature.
dielectric permittivity.[1] These fascinating functionalities are The discovery of new 2D ferroelectric materials and the under-
widely exploited for applications, including sensors, nonvolatile lying mechanism will not only provide a better understanding
of the physical mechanism behind the low-dimensional ferro-
electricity but also may facilitate the application of nanoscale
Dr. T. Xu, J. Mai, Prof. T.-Y. Zhang
Materials Genome Institute ferroelectric devices.
Shanghai University On the other hand, 2D materials are normally fabricated on
Shanghai 200444, China a substrate in practice, in which lattice mismatch and resulting
E-mail: xutao6313@shu.edu.cn; zhangty@shu.edu.cn misfit strain are inevitably present. Strain engineering has been
Dr. X. Wang proven to be a powerful tool in the control and modulation of
Key Laboratory of Pressure Systems and Safety
Ministry of Education, East China University of Science and Technology
the structural, electronic, optical, and magnetic properties of 2D
Shanghai 200237, China materials.[20–23] In perovskite oxides, elastic strain has been rec-
J. Zhang, Prof. J. Wang ognized as a viable means to modify the structural instabilities
Department of Engineering Mechanics and induce ferroelectricity.[24] In this paper, we demonstrate, by
School of Aeronautics and Astronautics using first-principles calculations, that appropriate mechanical
Zhejiang University strain can activate a paraelectric to ferroelectric phase transi-
Hangzhou 310027, China
tion in 2 AL-thick PbX films and modulate the resulting fer-
The ORCID identification number(s) for the author(s) of this article
can be found under https://doi.org/10.1002/aelm.201900932.
roelectric properties. The present results provide a prospective
means of achieving 2D ferroelectric materials in otherwise non­
DOI: 10.1002/aelm.201900932 ferroelectric materials and designing nanoscale devices.

Adv. Electron. Mater. 2020, 6, 1900932 1900932  (1 of 7) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
www.advelectronicmat.de

2. Results and Discussion Table 1.  Calculated results of the lattice parameters, elastic constants,
and cohesive energies of 2 AL PbX films.
The initial configurations of the 2 AL PbXs were cleaved along
the (001) plane from the bulk PbXs. Due to the breaking of the a0 = b0 [Å] C11 = C22 [GPa] C12 [GPa] C44 [GPa] Ec [eV]
mirror symmetry along the [001] direction, the 2 AL PbXs are PbS 4.23 25.8 13.7 17.1 −7.17
reduced to P4/nmm space group from the Fm3m space group PbSe 4.40 19.3 6.1 9.8 −6.58
of bulk materials. As illustrated in Figure  1a, the crystals con-
PbTe 4.64 18.8 10.8 17.1 −5.88
sist of two flat atomic layers with two Pb atoms and two X
atoms in a primitive unit cell in a square lattice. The primitive
cells of 2 AL PbXs were adopted to investigate their structural a constant value. All these results ensure the stability and plau-
properties. The optimized lattice constants, as illustrated by the sibility of the 2 AL PbX films, and these films are expected to
dashed black line, are a0  =  b0  = 4.23 Å, a0  =  b0  = 4.40 Å, and be experimentally fabricated considering that few-atomic-layer
a0  =  b0  = 4.64 Å for PbS, PbSe, and PbTe, in good agreement PbXs have been successfully grown.[17,28,29]
with previous reports.[25,26] Furthermore, the PbX bond length Having confirmed the stability of atomic-thick PbXs, we turn
increases from PbS through PbSe to PbTe, consistent with the to the investigation of the ferroelectricity of these 2D systems.
growing atomic radius. These structures are different from Recent experiments have demonstrated strong in-plane ferroe-
the closely related bi-layer SnTe (space group Pmn21), which lectricity in 2D SnTe,[17] attributed to the relative displacements
contains a relative in-plane displacement between Sn and Te between Sn and Te atoms with different electronegativities.
atoms.[18] Generally, a larger relative displacement implies a larger value
We then investigated the stability of the 2 AL PbX films. The of the ferroelectric spontaneous polarization. Unlike in the 2D
cohesive energies of the crystals, defined as Ec  = EPbX – EPb – SnTe material, the centres of the positive and negative charges
EX, were calculated, where EPbX is the total energy of 2 AL PbXs, in the 2 AL PbXs coincide due to the absence of a relative
and EPb and EX are the energies of free Pb and X atoms, respec- atomic displacement between Pb and X atoms. Consequently,
tively. The calculated cohesive energies are collected in Table 1. the PbXs preserve the inversion center and are intrinsically
Negative values indicate that formation of the atomic layers nonferroelectric materials.
is energetically favorable. The independent elastic constants Mechanical strain is an effective tool to modulate the struc-
of the studied materials are also summarized in Table 1, in tural properties of 2D materials. We further investigated the
which all three crystals meet the Born criteria for 2D materials: structural properties of 2 AL PbXs with respect to biaxial ten-
C11C22 –C122 > 0 and C11, C22, C44 > 0.[27] This result indicates sile strain up to 8.0%. With their structures analogous to 2 AL
that they can be mechanically stable. Furthermore, we also SnTe, we calculated the strain dependence of the total energy of
verified the dynamic stabilities of these structures by calcu- 2 AL PbS in the centrosymmetric P4/nmm structure (denoted
lating their phonon dispersions, and the results are provided in as phase C) and the distorted Pmn21 structure (denoted as
Figure 1b and Figure S1, Supporting Information. It is notice- phase A). The obtained results are depicted in Figure  2 and
able that the studied crystals are free of any imaginary phonon Figure S3, Supporting Information. We can see that the A phase
modes within the Brillouin zone, indicating that all of them are is unstable and spontaneously relaxes to the C phase in all
kinetically stable. Finally, AIMD simulations were performed three crystals without strain. However, for tensile strains larger
at 300 K to examine the thermal stability of the crystals. As than 1.5%, the non-centrosymmetric A phase becomes energet-
illustrated in Figure S2, Supporting Information, a snapshot of ically favorable in PbS, indicating the transformation from the
the equilibrium structure indicates that the PbS atomic layer C phase into the A phase. The A phase of PbS becomes notably
remains nearly unchanged after a 5 ps simulation, and the total more stable with increasing strain. An analogous trend is also
energy of the simulation system only slightly fluctuates around observed in 2 AL PbSe and PbTe with the application of biaxial

Figure 1.  a) Top view and side view of 2 AL PbX film. The dashed square denotes the primitive cell. The two basis vectors are denoted as a0 and b0.
b) Phonon dispersion curves of 2 AL PbS film.

Adv. Electron. Mater. 2020, 6, 1900932 1900932  (2 of 7) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
www.advelectronicmat.de

no appreciable imaginary-frequency modes are observed over

the entire Brillouin zone in the polar phase, which confirms
that the distorted structure is indeed the ground state. More-
over, our AIMD simulations also demonstrate that the strained
films remain intact at finite temperature, as illustrated in Figure
S5, Supporting Information. Therefore, the biaxial strain leads
to breaking of the structural centrosymmetry, which is the nec-
essary condition for spontaneous polarization.
The lattice distortion of crystals was quantified by the vector
d, defined as the off-center displacement of the Pb atom, that
is, the relative in-plane displacement between neighboring Pb
and S atoms, as illustrated in Figure 3a. The centrosymmetric
C phase corresponds to d  = 0 with four equivalent in-plane
PbS bonds, which forbids the formation of ferroelectricity,
while the A phase has a nonzero value of d, and the two cova-
lent PbS bonds denoted by white lines are strengthened,
Figure 2.  Total energy of C and A phases versus biaxial strain in 2 AL
leading to breaking of the inversion symmetry. Upon reversion,
PbS film. The insets show the atomic structures of the C and A phases.
the A phase can transform to the energy degenerate ground-
state A′ phase with opposite d. As a result, the A and A′ phases
strain, as demonstrated in Figure S3, Supporting Informa- are expected to have opposite polarizations. The evolution of
tion, although the critical strains for this transition are larger the atomic distortion of 2 AL PbS with applied strain is shown
(i.e., ≈2.0% and 3.0% for PbSe and PbTe, respectively). We fur- in Figure 3b. It is evident that the distortion displacement
ther explicitly computed the full phonon dispersion curves of increases as a function of strain. Similar behaviors can also be
the strained PbX atomic layers. As shown in Figure S4, Sup- found in PbSe and PbTe films (see Figure S6, Supporting Infor-
porting Information, for PbS under 4.0% strain as an example, mation), in which the induced polar distortions are slightly

Figure 3.  a) Top view of the atomic structures and charge density distributions of two energy-degenerated distorted phases (A and A´) and the cen-
trosymmetric C phase. The arrows and white lines indicate the directions of electric polarization and the covalent Pb–S bonds, respectively. b) Polar dis-
tortions of 2 AL PbS as a function of biaxial strain. c) Double-well potential of 4.0% strained 2 AL PbS film along the normalized distortion displacement.

Adv. Electron. Mater. 2020, 6, 1900932 1900932  (3 of 7) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
www.advelectronicmat.de

Figure 4.  a) Polarization of 2 AL PbS film versus biaxial strain. b) Minimum energy pathway and activation barrier for the transformation between two
ferroelectric phases (A and A´) in the 2 AL PbS film with 4.0% strain.

smaller than the corresponding values in PbS under an equiva- tric polarization vanishes. Following the approach in previous
lent strain amplitude. The variation in the energy of strained studies,[18,19,36] we constructed the free energy of the PbS super-
2 AL PbS (ε = 4.0%) with respect to the normalized distortion cell expressed in the Landau–Ginzburg expansion in terms of
displacement was also calculated, and the results are plotted in the polarization:
Figure 3c, which clearly shows double well potential character-
istics. This result further suggests the existence of ferroelectric E = ∑ i ( Api2 + Bpi4 + Cpix2 piy2 ) + ∑ ijαβ J ijαβ piα p jβ (1)
polarization.
The spontaneous polarization of the strained films was Here, pi is the polar distortion of the local unit cell i including
confirmed and evaluated using the modern Berry phase components of pix and piy, and A, B and C are constant param-
method,[30,31] which takes into account the electronic and ionic eters. The first three energy terms represent the anharmonic
contributions. We find that the distorted (A and A′) phases double-well potential, as shown in Figure 3c. The last term
have opposite spontaneous polarizations along the in-plane describes the coupling effect between the nearest-neighbor
(100) direction. The evolution of the calculated polarization of local modes i and j, in which interaction parameter Jijαβ (i and
2 AL PbS with biaxial strain is depicted in Figure  4a, which j represent the local modes, and α and β represent directions)
increases with increasing applied strain. Specifically, 2 AL PbS can be derived by fitting the DFT results using mean-field
has a large spontaneous polarization of 1.90 × 10−10 C m−1 theory.[37] Taking PbS at strains of 4.0% and 8.0% as examples,
at a strain of ε  = 4.0%. This value is comparable to those in we fitted all the related coefficients and list them in Table S1,
ultrathin ferroelectric SnTe (≈10−10 C m−1[18]) and GeSe (1.6 × Supporting Information. With the established effective Hamil-
10−10 C m−1[32]). PbSe and PbTe films under this strain also tonian model, we then performed MC simulations to investi-
exhibit polarization with the same order of magnitude. The gate the ferroelectric phase transition. As illustrated in Figure
activated ferroelectric polarization is expected to be detected in S7, Supporting Information, the Tc of 2 AL PbS with 4% strain
experiments, as has been done for the 2D ferroelectric SnTe, is estimated to be ≈130 K and is increased to about 300 K with
due to their comparable polarizations. Furthermore, to evaluate 8% strain. These values are larger than the result predicted in
the feasibility of polarization conversion, the nudged elastic defect-free 2 AL SnTe film (i.e., 30 K),[38] which suggests the fer-
band (NEB) method[33] was used to compute the minimum roelectric state is more stable in the 2D PbX structures.
energy pathway and the activation barrier in the ferroelectric We now discuss the origin of the ferroelectric polarization
transition of PbS film, in which the initial and final structures in the 2 AL PbXs with biaxial strain. The ferroelectricity in con-
were set as the A and A′ phases, respectively (see Figure 4b). ventional perovskite ferroelectrics with the displacive transition
The calculated energy barrier is approximately 23 meV per is usually associated with softening of the polar optical modes
unit cell, which is close to that in SnTe[18] and comparable to in the undistorted paraelectric structure.[39] The polar optical
that of the conventional ferroelectric BaTiO3 (18 meV/f.u.).[34] modes become unstable in the high-symmetry paraelectric
Similarly, we obtain energy barriers of 22 and 16 meV for PbSe phase and are condensed due to their imaginary frequencies,
and PbTe films at a strain of 4.0%. These small values indi- which causes ferroelectric distortion. Taking the 2 AL PbS at
cate that polarization reversal can be easily achieved under an a strain of ε = 3.0% as an example, we plot the phonon disper-
external electric field, which is beneficial for energy efficient sion curve of its centrosymmetric structure in Figure  5a. One
ferroelectric devices. It is worth mentioning that ferroelectric can see that this undistorted structure exhibits pronounced
grain boundaries can further decrease the transition barrier structural instability around the Γ point, as denoted by the
and make the polarization switching easier due to the disconti- imaginary (negative) frequency. This unstable mode has a
nuity in the polarization and the depolarizing field induced by doubly degenerate Eu symmetry. Its real-space eigendisplace-
the boundaries.[35] ment manifests as a relative displacement between adjacent
To assess the stability of the emerging ferroelectricity, we fur- Pb and S atoms along the [100] or [010] directions, thus corre-
ther employed the Monte Carlo (MC) method to estimate the sponding to the ferroelectric soft mode. This soft mode tends to
Curie transition temperature Tc of 2 AL PbS, above which elec- break the structural centrosymmetry and yields the ferroelectric

Adv. Electron. Mater. 2020, 6, 1900932 1900932  (4 of 7) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
www.advelectronicmat.de

Figure 5.  a) Phonon dispersion curves of centrosymmetric 2 AL PbS film at 3.0% strain. The inset shows the real-space eigendisplacements of the
doubly degenerate unstable mode in the zone centre. b) Frequency of the polar distortion of 2 AL PbS film versus biaxial strain.

ground state in the strained PbS film. Note that there is a pro- 2 AL PbS without strain is shown in Figure S10, Supporting
nounced imaginary frequency at the R point, which is associ- Information, which shows semiconductor characters with a
ated with antiferroelectric (AFE) distortion. However, the AFE band gap of about 1.22 eV. This value is consistent with pre-
state is metastable and higher in energy (e.g., 20.5 meV at a vious reports[45,46] and is larger than that of its bulk counter-
strain of 4.0%) than the ferroelectric ground state. The frequen- part (i.e., about 0.5 eV[47]). The variation of the band gaps of
cies of the polar mode in 2 AL PbS versus strain are depicted PbS, PbSe, and PbTe versus biaxial strain are also shown in
in Figure 5b. We clearly see that the frequencies are strongly Figure  6. It is well known that the traditional ferroelectric
strain dependent and that the onset of softening occurs at a photovoltaic materials suffer from large band gap and thus
critical strain of ≈ε = 1.5%, which produces the structural phase have low energy conversion efficiency. Surprisingly, the band
transition from the paraelectric C phase to the ferroelectric A gaps of PbX films in the current study are generally close to
(or A′) phase. This phonon mode is increasingly softened with 1.0 eV and are increased by the application of biaxial strain.
increasing strain such that the polar distortion is increasingly In particular, the PbS film with 4.0% strain has a band gap of
preferred. A similar strain-induced soft polar mode is also 1.35 eV, which is the ideal value of band gap for photovoltaic
observed in PbSe and PbTe films. applications. Since ferroelectric polarization and suitable band
The above results thus demonstrate that biaxial strain gap are two key properties in photovoltaic devices, our strategy
can trigger the paraelectric to ferroelectric transition due of strain engineering in lead chalcogenides further provides
to the frozen of the polar mode and effectively modulate the a means for excellent 2D ferroelectric photovoltaic materials.
ferroelectric properties in 2 AL PbXs. Note that 2D materials Despite the challenge in the synthesis of 2D group IV chal-
have superior mechanical flexibility and can sustain much cogenides at this moment, 2 AL SnTe film and few-atomic-
higher strain than their bulk counterparts, for example, ≥25% layer PbXs have been successfully grown in experiments.[17,29]
in some 2D materials.[40] The biaxial strain in the present study We hope that our prediction can stimulate further experi-
can be introduced by choosing suitable substrates, such as Si mental work on this class of crystals to design atomic scale
and Ni,[41] in practice. It can also be intentionally imposed and ferroelectrics and high-efficiency ferroelectric photovoltaic
controlled by defects,[42] functional wrapping,[43] or even direct nanodevices.
mechanical application.[44] In addition to under biaxial strain,
the paraelectric to ferroelectric phase transition in 2 AL PbX
films can also occur in different strain states. The [100] uniaxial
strain above a certain level, for instance, also breaks the original
P4/nmm symmetry of PbX films and leads to the ferroelectric
Pmn21 phase with polarization along the [100] direction. As
illustrated in Figure S8, Supporting Information, the critical
uniaxial strains for the ferroelectric phase transitions are ≈2.5%,
3.0%, and 4.0% for PbS, PbSe, and PbTe, respectively. The polar-
ization of the films versus uniaxial strain was also calculated,
and the results are shown in Figure S9, Supporting Informa-
tion. For instance, the ferroelectric polarization of 2 AL PbS at
a uniaxial strain of 4.0% is 1.42 × 10−10 C m−1, which is slightly
smaller than the corresponding value of 1.90 × 10−10 C m−1
at the same biaxial strain. All these results demonstrate the
effective tuning of ferroelectricity by the application of strain
engineering.
To further improve the understanding of the electronic
properties of the thin films, we calculated the electronic den-
sity of states (DOS) of all the investigated crystals. The DOS of Figure 6.  Band gaps of 2 AL PbX films versus biaxial strain.

Adv. Electron. Mater. 2020, 6, 1900932 1900932  (5 of 7) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
www.advelectronicmat.de

3. Conclusion Author Contributions

T.X. and X.Y.W. contributed equally to this work. T.X. conceived the
In summary, we have proposed a strain engineering approach
project, designed and directed computational experiments, and wrote
for realizing 2D ferroelectricity in nonferroelectric 2 AL PbX the entire manuscript. X.Y.W. and J.W.M. supported the calculations and
films based on DFT calculations. The atomic structures and discussed the results. J.T.Z. and J.W. conducted the MC calculations and
physical properties of the crystals were explored, and their sta- discussed the results. T.Y.Z. supervised the work and provided critical
bilities were confirmed by cohesive energy calculations, ab initio feedback on the manuscript. All authors read and commented on the
molecular dynamics simulations, and phonon-spectrum calcula- manuscript.
tions. The strain-free atomic layers are characterized by a par-
aelectric P4/nmm structure. As the applied strain is increased,
PbS, PbSe, and PbTe in turn transform into a ferroelectric phase Keywords
with in-plane electric polarization. The emerging ferroelectricity
2D materials, ferroelectric photovoltaic materials, ferroelectricity, strain
is enhanced with increasing strain. The strain dependence of
engineering
phonon modes in 2 AL PbXs was also studied to clarify the origin
of the electric polarization. Moreover, our work also reveals that Received: August 30, 2019
the strained PbX films are promising ferroelectric photovoltaic Revised: September 30, 2019
(PV) materials because of their notable ferroelectric polarization Published online: November 18, 2019
and suitable band gap. The present results provide great oppor-
tunities for realizing a ferroelectric phase in low-dimensional
materials and designing multifunctional nanoscale devices.
[1] M. E. Lines, A. M. Glass, Principles and Applications of Ferroelectrics
and Related Materials, Clarendon Press, Oxford, UK 1979.
[2] R. Ramesh, Thin Film Ferroelectric Materials and Devices, Kluwer Aca-
4. Experimental Section demic, Boston, MA 1997.
The first-principles calculations were conducted based on density [3] J. F. Scott, Ferroelectric Memories, Springer, Berlin, Germany 2000.
functional theory (DFT), as implemented in the Vienna Ab initio [4] T. Shimada, T. Xu, Y. Uratani, J. Wang, K. Kitamura, Nano Lett. 2016,
Simulation Package.[48,49] The generalized gradient approximation 16, 6774.
(GGA)[50] of the Perdew–Burke–Ernzerhof (PBE)[51] form was employed [5] Y. Nahas, S. Prokhorenko, L. Louis, Z. Gui, I. Kornev, L. Bellaiche,
for the exchange-correlation functional. The kinetic energy cutoff was set Nat. Commun. 2015, 6, 8542.
to 400 eV, and the Monkhorst–Pack k-point scheme[52] with a size of 12 [6] R. Mazumder, D. P. Sujatha, D. Bhattacharya, P. Choudhury, A. Sen,
× 12 × 1 was used for the Brillouin zone integrations. To avoid spurious M. Raja, Appl. Phys. Lett. 2007, 91, 062510.
interactions due to the periodic boundary condition, a vacuum space of [7] J. Liu, Y. Ji, S. Yuan, L. Ding, W. Chen, Y. Zheng, npj Comput. Mater.
20 Å normal to the layer plane was applied. The atomic positions were 2018, 4, 78.
fully optimized until the Hellmann–Feynman force on each atom was [8] M. J. Polking, M.-G. Han, A. Yourdkhani, V. Petkov, C. F. Kisielowski,
less than 0.001 eV Å−1. The phonon dispersion was studied using the V. V. Volkov, Y. Zhu, G. Caruntu, A. P. Alivisatos, R. Ramesh, Nat.
finite displacement method implemented in the Phonopy package,[53] Mater. 2012, 11, 700.
and the ferroelectric polarization was calculated by the Berry phase [9] I. I. Naumov, L. Bellaiche, H. Fu, Nature 2004, 432, 737.
approach.[30,31] A 7 × 7 × 1 supercell containing 196 atoms and k-mesh [10] C. Cui, F. Xue, W.-J. Hu, L.-J. Li, npj 2D Mater. Appl. 2018, 2, 1.
grids of 1 × 1 × 1 were used in the phonon spectra calculations. Ab [11] M. Wu, P. Jena, Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2018, 8,
initio molecular dynamics (AIMD) simulations with the NVT ensemble
e1365.
and Nosé–Hoover thermostat were carried out to assess the thermal
[12] M. Wu, X. C. Zeng, Nano Lett. 2016, 16, 3236.
stability of the crystals. For the modeling of epitaxial 2 AL PbX films, a
[13] H. Wang, X. Qian, 2D Mater. 2017, 4, 015042.
biaxial strain, defined as ε11 =  ε22 = (a–a0)/a0 = (b–b0)/b0, was imposed
by constraining the in-plane lattice parameters a and b, where a0 and b0 [14] S. Shen, C. Liu, Y. Ma, B. Huang, Y. Dai, Nanoscale 2019, 11,
are the equilibrium lattice constants. 11864.
[15] C. Xiao, Adv. Funct. Mater. 2018, 28, 1707383.
[16] Y. Wang, C. Xiao, M. Chen, C. Hua, J. Zou, C. Wu, J. Jiang,
S. A. Yang, Y. Lu, W. Ji, Mater. Horiz. 2018, 5, 521.
Supporting Information [17] K. Chang, J. Liu, H. Lin, Na. Wang, K. Zhao, A. Zhang, F. Jin,
Supporting Information is available from the Wiley Online Library or Y. Zhong, X. Hu, W. Duan, Q. Zhang, L. Fu, Q.-K. Xue, X. Chen,
from the author. S.-H. Ji, Science 2016, 353, 274.
[18] W. Wan, C. Liu, W. Xiao, Y. Yao, Appl. Phys. Lett. 2017, 111, 132904.
[19] R. Fei, W. Kang, L. Yang, Phys. Rev. Lett. 2016, 117, 097601.
[20] M. Wu, H. Fu, L. Zhou, K. Yao, X. C. Zeng, Nano Lett. 2015, 15,
Acknowledgements 3557.
[21] D.-B. Zhang, S.-H. Wei, npj Comput. Mater. 2017, 3, 32.
The authors acknowledge the financial support from the National
Natural Science Foundation of China (Grant No. 11802169 and No. [22] I. Niehues, R. Schmidt, M. Drüppel, P. Marauhn, D. Christiansen,
11602252) and Fundamental Research Funds for the Central Universities M. Selig, G. Berghäuser, D. Wigger, R. Schneider, L. Braasch,
(Grant No. 50321071915017). R. Koch, A. Castellanos-Gomez, T. Kuhn, A. Knorr, E. Malic,
M. Rohlfing, S. M. de Vasconcellos, R. Bratschitsch, Nano Lett.
2018, 18, 1751.
[23] C. Long, Y. Dai, Z.-R. Gong, H. Jin, Phys. Rev. B 2019, 99, 115316.
Conflict of Interest [24] D. G. Schlom, L.-Q. Chen, C. J. Fennie, V. Gopalan, D. A. Muller,
The authors declare no conflict of interest. X. Pan, R. Ramesh, R. Uecker, MRS Bull. 2014, 39, 118.

Adv. Electron. Mater. 2020, 6, 1900932 1900932  (6 of 7) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
www.advelectronicmat.de

[25] X.-B. Li, P. Guo, Y.-N. Zhang, R.-F. Peng, H. Zhang, L.-M. Liu, [39] H. Uwe, T. Sakudo, Phys. Rev. B 1976, 13, 271.
J. Mater. Chem. C 2015, 3, 6284. [40] K. S. Kim, Y. Zhao, H. Jang, S. Y. Lee, J. M. Kim, K. S. Kim, J.-H. Ahn,
[26] X. Zhang, Z. Yang, Y. Chen, J. Appl. Phys. 2017, 122, 064101. P. Kim, J.-Y. Choi, B. H. Hong, Nature 2009, 457, 706.
[27] M. Born, K. Huang, Dynamical Theory of Crystal Lattices, Clarendon [41] Z. Fu, M. Liu, Z. Yang, Phys. Rev. B 2019, 99, 205425.
Press, Oxford, UK 1954. [42] L. Seravalli, M. Minelli, P. Frigeri, P. Allegri, V. Avanzini, S. Franchi,
[28] G. B. Bhandari, K. Subedi, Y. He, Z. Jiang, M. Leopold, N. Reilly, Appl. Phys. Lett. 2003, 82, 2341.
H. P. Lu, A. T. Zayak, L. Sun, Chem. Mater. 2014, 26, 5433. [43] X. L. Wu, F. S. Xue, Appl. Phys. Lett. 2004, 84, 2808.
[29] S. Khan, Z. Jiang, S. M. Premathilka, A. Antu, J. Hu, A. A. Voevodin, [44] Y. G. Wang, Q. L. Zhang, T. H. Wang, W. Han, S. X. Zhou, J. Phys. D:
P. J. Roland, R. J. Ellingson, L. Sun, Chem. Mater. 2016, 28, 5342. Appl. Phys. 2011, 44, 125301.
[30] R. D. King-Smith, D. Vanderbilt, Phys. Rev. B 1993, 47, 1651. [45] W. Wan, Y. Yao, L. Sun, C.C. Liu, F. Zhang, Adv. Mater. 2017, 29,
[31] R. Resta, Rev. Mod. Phys. 1994, 66, 899. 1604788.
[32] S. Guan, C. Liu, Y. Lu, Y. Yao, S. A. Yang, Phys. Rev. B 2018, 97, 144104. [46] X. B. Li, P. Guo, Y. N. Zhang, R. F. Peng, H. Zhang, L. M. Liu,
[33] G. Henkelman, B. P. Uberuaga, H. Jónsson, J. Chem. Phys. 2000, J. Mater. Chem. C 2015, 3, 6284.
113, 9901. [47] W. Li, Q. Y. He, Y. Z. Wang, T. Wang, Chem. Phys. Lett. 2017, 687,
[34] R. E. Cohen, Nature 1992, 358, 136. 101.
[35] W. Shu, J. Wang, TY. Zhang, J. Appl. Phys. 2012, 112, 064108. [48] G. Kresse, J. Hafner, Phys. Rev. B 1993, 47, 558.
[36] W. Zhong, D. Vanderbilt, K. M. Rabe, Phys. Rev. Lett. 1994, 73, [49] G. Kresse, J. Furthmüller, Phys. Rev. B 1996, 54, 11169.
1861. [50] S. Grimme, J. Comput. Chem. 2006, 27, 1787.
[37] W. Zhong, D. Vanderbilt, K. M. Rabe, Phys. Rev. B 1995, 52, 6301. [51] M. Ernzerhof, G. E. Scuseria, J. Chem. Phys. 1999, 110, 5029.
[38] K. Liu, J. L. Lu, S. Picozzi, L. Bellaiche, H. J. Xiang, Phys. Rev. Lett. [52] H. J. Monkhorst, J. D. Pack, Phys. Rev. B 1976, 13, 5188.
2018, 121, 027601. [53] A. Togo, F. Oba, I. Tanaka, Phys. Rev. B 2008, 78, 134106.

Adv. Electron. Mater. 2020, 6, 1900932 1900932  (7 of 7) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim