The boom of graphene, ^{[1, 2]} which exhibits outstanding physical and chemical properties, has given rise to the exploration of other two-dimensional (2D) materials. Silicene, which is the silicon counterpart of graphene, holds great promise for future applications in microelectronic devices due to its excellent electronic properties and easy integration into current Si-based semiconductor technology. However, unlike graphene, silicene is not absolutely planar but has a buckled height of about 0.44  Å .^{[3, 4]} In spite of the low buckled geometry, silicene possesses outstanding electronic properties similar to those of graphene.^{[1– 11]} Most importantly, silicene is a zero-gap semiconductor with linear dispersion around Fermi level, forming the famous “ Dirac cone” at the symmetric point K in the reciprocal space. The charge carriers in silicene behave like massless Dirac fermions in a small energy range around the Fermi level, with a Fermi velocity of 10⁵  m/s– 10⁶  m/s, ^{[3, 8, 11]} comparable to that of graphene.^{[12– 14]} Moreover, the spin– orbit coupling (SOC) effect in silicene is much stronger than that in graphene.^{[9, 15]} Therefore, a more prominent quantum spin Hall effect (QSHE) in an experimentally accessible temperature regime is anticipated in silicene.

Substantial fabrication of high quality silicene is the precondition for its future applications. Since the bulk phase of silicon is a cubic diamond lattice, the mechanical stripping method of fabricating graphene^[1] and many other 2D materials^[16] does not work for silicene. Up to now, the main method of synthesizing silicene has been epitaxial growth. Before large-scale silicene sheets were produced, silicene nanoribbons were successfully fabricated on Ag (001)^[17] and Ag (110)^{[18– 20]} surfaces. Afterwards, several experiments were conducted to fabricate silicene by epitaxial growth on Ag (111) surface and a variety of superstructures were observed, including (4× 4), , , with respect to Ag (111)^{[21– 26]} and with respect to silicene 1× 1 lattice.^{[7, 21, 27]} A recent experiment has been combined with density functional theory (DFT) calculations to reveal that two energy-degenerate and mirror-symmetric rhombic phases of silicene epitaxially grown on Ag (111) surface can form.^[27] Owing to the low transition barrier between these two phases, they can interchange through dynamic flip– flop motion, resulting in the superstructure observed in experiments. The superiority of Ag as a substrate for the growth of silicene originates from the moderate and homogeneous interaction between silicene and Ag substrate, which results in rather small local tension in silicene in the growth process.^[28]

In addition to the Ag (111) substrate, a silicene sheet has been successfully synthesized on zirconium diboride (ZrB₂₎ thin films grown on Si wafers through epitaxial growth^[29] and Ir (111) surface by directly depositing silicon atoms and annealing at 670  K.^[30] Very recently, a highly buckled silicene sheet was fabricated on a semiconducting MoS₂ substrate, which paves the way to obtaining an isolated silicene sheet while avoiding the influence of metal substrates on the electronic structures.^[31]

Beyond the monolayer silicene, multilayer silicene sheets have been fabricated on a Ag (111) surface by the epitaxial technique.^{[25, 32– 36]} Unlike monolayer silicene, bilayer and multilayer silicene only shows one structural phase on Ag (111) surface, namely superstructure with respect to silicene. Dirac fermions were reported to exist in multilayer silicene by angle-resolved photoemission spectroscopy measurements.^{[32, 33]} Multilayer silicene films showed only a very small top surface oxidation after 24-h exposure to ambient air while keeping the rest of the silicene film underneath intact.^[36] In addition, germanene, which possesses similar electronic properties to those of silicene, ^{[3, 9, 11]} was recently synthesized successfully on Pt (111) and Au (111), respectively.^{[37, 38]}

The nature of the zero band gap of silicene severely inhibits its application in microelectronic devices. Fortunately, a tunable band gap can be achieved in silicene by applying a perpendicular electric field, ^{[39, 40]} which originates from symmetry breaking of the two sublattices in silicene. Benefiting from its chemically active surface, the electronic properties of silicene can be easily tailored by adsorbates. Upon exposure to hydrogen, H atoms will form covalent bonds with silicene through sp³ hybridization; therefore, opening an appreciable band gap in silicene.^{[41– 46]} Fully hydrogenated silicene (namely, silicane) has the most stable chair configuration and is an indirect semiconductor with a band gap of 4.0  eV, as obtained from HSE calculation. Moreover, one-sided semihydrogenated silicene has been predicted to be a ferromagnetic semiconductor with a direct gap of 1.74  eV.^[44] In the same manner, halogen functionalized X-silicene (X = F, Cl, Br, and I) can also convert silicene from a half metal into a direct semiconductor with band gaps of 1.469  eV, 1.979  eV, 1.950  eV, and 1.194  eV, respectively.^{[47, 48]}

A recent experiment combined with density functional theory (DFT) calculations suggests that an electronic band gap can be induced in monolayer silicene on an Ag (111) surface by oxidation.^[49] The bridge site is the most energetically favored configuration for adsorption of oxygen adatoms on the surfaces of silicene. In fully oxidized silicene, the buckled silicene structure vanishes with subsequent crumpling of the sample and exposure of bare Ag (111) surface areas.

The electronic properties of silicene can also be easily tuned by the adsorption of various metal or nonmetal atoms.^{[50– 56]} It is noteworthy that the adsorption of alkali metals (Li, Na, and K), which favor the hollow site of silicene, does not induce any significant distortion or stress in the silicene lattice. Meanwhile, alkali metal adsorption opens a small band gap at the Dirac point submerged under Fermi level, nearly preserving the linear dispersion of the pristine silicene.^[50]

Very recently, Tao et al. successfully fabricated a silicene field-effect transistor (FET) with a measured room-temperature mobility of ∼ 100  cm²· V^{− 1}· s^{− 1}.^[57] This experiment was conducted by a growth– transfer– fabrication process, in which the devised-silicene was formed from an encapsulated delamination with native electrodes. This is no doubt a breakthrough for the development of silicene devices and it opens a way to the future application of silicene in microelectronics.

All of this exciting progress indicates that silicene is a fascinating 2D material with versatile electronic properties, depending on the substrate, structural defects, functionalization, metal adsorption, etc. In this review, we intend to present an overview of the modification of silicene by point defect, substrates, alkali metal intercalation, and transition metal alloying from recent first-principles calculations, along with some relevant experiments. The atomistic mechanism for the growth of silicene on an Ag surface will also be addressed. General information about the atomic structures, electronic properties, and synthesis of silicene can be found in the previous review articles.^{[58– 66]}

In a vacuum, silicon clusters prefer cage-like structures to planar structures.^{[67– 69]} Since silicon atoms do not prefer sp² hybridization, small planar Si_N clusters (N = 6, 10, 13, 16, 19, 22, 24) as aggregates of six-membered rings are unstable and would transform into severely buckled three-dimensional (3D) configurations upon relaxation.^[28] Therefore, to obtain silicene through epitaxial growth, a complete wetting substrate, such as an Ag (111) surface, is needed to keep the silicon clusters planar.

Here we adopt the Ag (111) that is most frequently used as representative to explain the growth behavior of silicene. Clarifying of the stablest configurations of small silicon clusters on Ag (111) surface is a crucial step towards understanding the growth of silicene patch. The favorite adsorption site for single Si atom on Ag (111) surface is the fcc hollow site^{[28, 70]} (see Fig.  1). Small 2D honeycomb silicon clusters composed of one-six hexagonal rings on an Ag (111) surface are then constructed. After geometry optimization, these small silicene-like structures are well preserved, except for Si₆ and Si₁₃, implying that an Ag (111) surface is a good substrate for silicene growth.^[28]

On the other hand, 3D configurations (ground state of Si_N clusters in vacuum) are also placed on an Ag (111) surface and their energetic stability is compared with those of 2D clusters.^[27] As shown in Fig.  2, the formation energies of 3D clusters decrease after they have been deposited on a Ag (111) surface due to the metal passivation effect; however, 2D clusters have lower formation energies than 3D clusters. Therefore, no 3D clusters will occur at the early stage of silicene growth; i.e., the wetting of Ag  (111) surface is complete. Moreover, the formation energy for silicene-like 2D cluster decreases as the number of hexagonal ring increases, which implies the continuous growth of the silicene sheet. Crystal growth is the reverse process of dissolution. During the growth of silicene, silicon atoms aggregate and dissolve at the same time. When the aggregation rate prevails that of dissolution, the cluster grows bigger. Therefore, surface diffusion of silicon atoms on Ag  (111) surface is crucial to the epitaxial growth of silicene on an Ag  (111) surface. Indeed, a very small diffusion energy barrier (0.031  eV) for silicon monomer on an Ag  (111) surface is given by DFT calculation.^[28] Consequently, small 2D silicon clusters will easily form after depositing silicon atoms on an Ag  (111) surface. After the clusters reach some critical size, they will grow into large silicene sheets by continuously uptaking silicon atoms on the Ag  (111) surface.

Fig.  1.

	Figure Option ViewDownloadNew Window
	Fig.  1. Structures and formation energies (eV per Si atom) of 2D SiN clusters on the Ag (111) surface.[28]

Fig.  2.

	Figure Option ViewDownloadNew Window
	Fig.  2. Formation energies of SiN clusters in vacuum and on Ag (111) surface: (a) freestanding SiN clusters with ground state configurations in vacuum (GS– SiN); (b) direct deposition of GS– SiN clusters on Ag (111) surface (GS– SiN– Ag); (c) 2D silicene-like clusters on Ag (111) surface (2D-SiN– Ag).[28]

A recent study has demonstrated that the nucleation rate of silicene is very sensitive to both the growth temperature and the chemical potential Δ μ of the silicon atoms.^[70] At a typical growth temperature (500  K), the nucleation rate showed a ∼ 23-orders of magnitude increase (from ∼ 10^{− 10}  cm^{− 2}· s^{− 1} to ∼ 10¹³  cm^{− 2}· s^{− 1}) as the Δ μ varies from 0.05  eV to 0.15  eV. For a specified Δ μ (e.g. Δ μ = 0.1  eV), reduction of the temperature from 500  K to 300  K would lead to an ∼ 11-orders of magnitude drop in the nucleation rate (from 10⁷  cm^{− 2}· s^{− 1} to 10^{− 4}  cm^{− 2}· s^{− 1}). Usually, a high nucleation rate will result in a large number of defects. Thus, the growth conditions should be carefully chosen in order to reduce the defects in silicene.

As discussed above, defects are nearly inevitable in the growth of silicene. Hence, how defects affect the electronic properties of silicene becomes an important issue. Here we briefly discuss the geometries and stabilities of some point defects in a silicene monolayer, as well as the defect effect on the electronic properties of silicene, which have been comprehensively investigated by our group through using first-principles calculations.^[56]

The geometries of various point defects in silicene, including Stone– Wales rotation, single and double vacancies (abbreviated as SW, SV, and DV, respectively, hereafter), and Si adatom are displayed in Fig.  3. Here, the formation energy is defined as E_F = E_T− N × E_Si, where E_T is the total energy of defective silicene, N is the number of silicon atoms in the supercell of defective silicene, and E_Si is the energy per silicon atom in a perfect silicene sheet. For Stone– Wales rotation, two five-membered rings and two seven-membered rings are generated with a formation energy of 2.09  eV. The barrier for creating an SW defect is 2.64  eV, while the reverse barrier is as low as 0.5  eV, indicating that annealing at an appropriate temperature may eliminate the existing SW defects. Two kinds of SV defects are considered: SV-2 with three dangling atoms is less stable than SV-1 with (55| 66) rings. In the case of DV defects, two five-membered rings and one eight-membered ring are introduced. A reconstruction of DV defect with an energy barrier of 1.20  eV transforms (5| 8| 5) rings into (555| 777) and lowers the formation energy by 0.86  eV. Note that the formation energy for a DV is much smaller than that for two isolated SVs, which implies that two SVs may coalesce into a DV by diffusion. The DFT calculations showed that an SV-1 defect can migrate within the silicene sheet with a small energy barrier of 0.12  eV. Thus, the combination of two SVs into a DV is probable. In contrast, a large energy barrier (2.06  eV) would prevent the diffusion of DV-2. The Si adatom prefers the top site of silicene with the original lattice Si atom pressed down (Fig.  3(f)). To our surprise, the formation energy for Si adatom may even be negative.

Defects can effectively modify the electronic properties of silicene. SW and DV-1  (5| 8| 5) defects would introduce small band gaps of 33  meV and 161  meV in silicene, respectively. The SV-1  (55| 66) defect would transform semi-metallic silicene into metallic silicene. DV-2  (555| 777) would destroy the Dirac cone, but keep the semimetal characteristic of silicene. Most interestingly, an Si adatom not only opens a noticeable band gap of 91  meV but also induces a local magnetic moment of 2  μ _B in silicene.

Fig.  3.

	Figure Option ViewDownloadNew Window
	Fig.  3. Geometries and formation energies of various point defects in silicene: (a) SW; (b) SV-1 by (55| 66) rings; (c) SV-2 with three dangling atoms; (d) DV-1  (5| 8| 5); (e) reconstructed DV-2  (555| 777); (f) Si adatom.[56]

On the other hand, porous silicene with DV-1 and DV-2 defects are chemically inert to noble gases (He, Ne, and Ar), H₂, H₂O, CO, CO₂, N₂, and CH₄ without additional hydrogen or nitrogen atoms to passivate the edge silicon atoms, showing a great prospect for gas purification.^{[71, 72]} The diffusion energy barrier for H₂ through the DV-1 of silicene is only about 0.34  eV, suggesting that an H₂ molecule can permeate the porous silicene at moderate temperature and pressure.^[71] The calculated diffusion energy barriers are 1.03, 0.99, 1.01, and 1.66  eV for N₂, CO, CO₂, and CH₄, respectively. Therefore, a porous silicene sheet with DV-1 would serve as a good hydrogen purification membrane.^[71] Moreover, a porous silicene sheet with DV-1 or DV-2 also exhibits high selectivities to He/Ne and He/Ar.^[72]

Without a substrate, free-standing monolayer silicene is unstable due to its high formation energy and active surface.^[56] Moreover, monolayer silicene is a zero gap semimetal, which certainly hinders its applications in microelectronics and optoelectronics. It is generally known that stacking is a universal method to modulate properties of 2D materials and may offer new opportunities to open a band gap.^{[103, 104]} Multilayer silicene has been successfully synthesized on an Ag (111) surface by the epitaxial technique in recent experiments.^{[25, 32– 35]} For the as-prepared multilayer silicene, the bottom layer (bonded with Ag surface) usually exhibits a (3× 3) superstructure with respect to free-standing silicene. superstructure with respect to free-standing silicene is only occasionally observed in the bottom layer. All the upper layers show only one superstructure; i.e., superstructure with respect to free-standing silicene.

On the theoretical side, significant efforts have been devoted to bilayer silicene and in total six kinds of bilayer silicene have been brought forward.^{[105– 114]} Five possible structures of bilayer silicene, namely, planar bilayer silicene (P-BS), AB-stacked bilayer silicene (AB-BS), AA-stacked bilayer silicene (AA-BS), AA′ -stacked bilayer silicene (AA′ -BS), and slide-AA bilayer silicene (slide-AA-BS), are shown in Fig.  7. Indeed, the search for the most stable structure of bilayer silicene can be traced back to 2008.^[105] By quenching liquid silicon confined in slit nanopores through using molecular-dynamics (MD) calculations with the Tersoff potential, Morishita et al. predicted the formation of AA′ -BS.^[105] P-BS was first proposed by Bai et al. in 2010 via MD simulations.^[107] Subsequently, these five kinds of bilayer silicene are all constructed by stacking two silicene layers, ^{[108– 112]} and their stability and electronic properties have been systematically investigated.^{[109, 113]} Recent DFT calculations combined with the particle swarm optimization algorithm have predicted a new silicene bilayer structure, called Si-Cmme, with lower energy than the widely accepted lowest energy configuration (P-BS).^[114]

Fig.  7.

	Figure Option ViewDownloadNew Window
	Fig.  7. Atomic structures (upper for top view, lower for side view) of various silicene bilayers along with formation energies per Si atom and lattice constants. The black lines highlight the primitive cell of bilayer silicene.[113]

For P-BS, AA-BS, and AA′ -BS, the two silicene layers coincide from the top view. The two silicene layers are planar in P-BS, but buckled in AA-BS and AA′ -BS. In AA-BS, the lower atoms in the first layer are located right above the lower atoms in the second layer, while in AA′ -BS, the lower atoms in the first layer are located right above the upper atoms in the second layer. In slide-AA-BS, silicon atoms in the upper layer are not directly located on the top sites of lower atoms but slightly off the exact AA stacking.^[109] AB-BS is indeed a slab model of an Si  (111) surface. Two silicene layers bond together covalently for all the six stacking patterns, which is different from the bilayer graphene with a vdW interlayer interaction only.^[115]

According to the formation energy, P-BS is stablest in all bilayer systems, except for Si-Cmme. Interestingly, P-BS is 0.025  eV/atom, which is even stabler than the slab model of Si  (111) surface; i.e., AB-BS, suggesting that a phase transition will occur when the silicon film becomes as thin as only two-layers. Indeed, the strain-induced phase transition from AB-BS to P-BS was recently reported.^[108]

Bilayer silicene systems distinctly show different electronic properties depending on the stacking style.^{[106, 107, 109– 112]} Unlike bilayer graphene, the Dirac cone in pristine silicene is badly destroyed in bilayer silicene due to the strong interaction between two silicene layers. The band structures calculated with PBE functional for P-BS, AB-BS, AA-BS, and AA′ -BS are shown in Fig.  8. The P-BS exhibits metallic character with an indirect overlap between the CBM and VBM (Fig.  8(a)). For AB-BS, two parabolic bands cross the Fermi level (Fig.  8(b)) AA-BS exhibits apparent metallic feature with several bands crossing the Fermi level (Fig.  8(c)). For AA′ -BS, two crossings of nearly linear dispersions emerge along two directions, namely G– K and G– M (Fig.  8(d)). One crossing point is slightly above the Fermi level (by 70  meV), and another one is right below the Fermi level (by 65  meV). The newly found Si-Cmme is also metallic.^[114] In particular, the band structure calculated with Heyd– Scuseria– Ernzerhof (HSE) hybrid functional suggests that slide-AA-BS is an indirect semiconductor with a band gap of 1.16  eV.^[109]

Fig.  8.

	Figure Option ViewDownloadNew Window
	Fig.  8. Band structures of four bilayer silicene sheets: (a) P-BS, (b) AB-BS, (c) AA-BS, and (d) AA′ -BS.[113]

The absence of a Dirac cone in bilayer silicene no doubt hinders their utilization in future microelectronic devices. Fortunately, an effective method, potassium intercalation, has been proposed to decouple the strong interactions between two silicene layers, thus recovering the superior electronic properties of silicene.^[113] As displayed in Fig.  9, bilayer silicene sheets intercalated with alkali metal atom (including Li, Na, K, and Rb), namely, LiSi₄, NaSi₄, KSi₄, and RbSi₄, are systematically investigated by first-principles calculations.^[113] After intercalation, two silicene layers are separated with an interlayer distance larger than 3.8  Å and the low buckled structure of pristine silicene is basically recovered. Moreover, the formation energies of LiSi₄, NaSi₄, KSi₄, and RbSi₄ are 0.275  eV, 0.302  eV, 0.309  eV, and 0.507  eV per atom, respectively, which are lower than that of P-BS (0.561  eV per atom) and much lower than that of free-standing silicene (0.779  eV per atom). The ab initio molecular dynamics (AIMD) simulation at 500  K confirms the thermal stability of KSi₄. Most importantly, the feature of Dirac cone of silicene is restored after intercalation. For KSi₄, a double-degenerated Dirac cone with a band gap of 0.27  eV appears at the K point in the first Brillouin zone (Fig.  10(b)). However, the band gap is submerged under the Fermi level due to the charge transfer from K atoms to silicene. The Fermi level can be lifted up by replacing the neutral potassium atom with its cation (K⁺ ). As displayed in Fig.  10(c), the band structure of K⁺ Si₄ exhibits a semiconducting character with a moderate gap of 0.43  eV. In an FET device, the charge doping of this type can be realized by applying a suitable gate voltage.^[116]

Fig.  9.

	Figure Option ViewDownloadNew Window
	Fig.  9. Atomic structure of alkali metal intercalated bilayer silicene (upper for top view, lower for side view). The yellow balls represent silicon atoms and the purple balls represent alkali metal atoms. The black lines define the primitive cell with four silicon atoms and one metal atom.[113]

Fig.  10.

	Figure Option ViewDownloadNew Window
	Fig.  10. Band structures of (a) monolayer silicene, (b) KSi4, and (c) K+ Si4.[113]

The 2D magnetic materials are important for developing spintronics and magnetic storage. Ferromagnetism can be induced in silicene by half hydrogenation^{[44, 45]} or direct adsorption of transition metal adatoms.^{[50, 53]} However, ordered phase is hard to form in these systems since the transition metal atoms tend to aggregate. Recently, seven monolayers of transition metal silicides (TMSi₂), namely., TiSi₂, VSi₂, CrSi₂, MnSi₂, FeSi₂, NbSi₂, and MoSi₂, with magnetic orders were reported by our group.^[117] These monolayer silicides can be constructed by placing a transition metal atom in the center of each hexagonal hole of honeycomb lattice of silicene, as shown in Fig.  11. Among them, TiSi₂, CrSi₂, MnSi₂, and FeSi₂, are planar (Fig.  11(a)), while VSi₂, NbSi₂, and MoSi₂ are buckled (Fig.  11(b)).

Fig.  11.

Figure Option ViewDownloadNew Window

Fig.  11. Top views and side views of (a) planar monolayers of TiSi2, CrSi2, MnSi2, and FeSi2; (b) buckled monolayers of VSi2, NbSi2, and MoSi2. The blue or red balls represent transition metal atoms and the yellow balls represent silicon atoms. The black rectangle represents the unit cell. The values of buckled height (d) for VSi2, NbSi2, and MoSi2 are 0.289, 0.756, and 0.647  Å respectively.[117]

Fig.  12.

	Figure Option ViewDownloadNew Window
	Fig.  12. Magnetic moments per 3d or 4d atom of TMSi2 silicide monolayer. The red line represents 3d silicide and the blue line represents 4d one.[117]

Compared with the atomic spins of 2, 3, 6, 5, 4, 1, and 2  μ _B for individual Ti, V, Cr, Mn, Fe, Nb, and Mo atoms, respectively, the magnetic moment per TM atom in the TMSi₂ monolayers is reduced by about half. As shown in Fig.  12, the on-site moments per TM atom in the TMSi₂ monolayers are 0.563, 2.148, 3.008, 2.512, 1.508, 0.618, and 0.207  μ _B for Ti, V, Cr, Mn, Fe, Nb, and Mo, respectively, while the remaining 3d and 4d TMSi₂ monolayers are all non-magnetic. Among all of the magnetic systems, TiSi₂, VSi₂, CrSi₂, NbSi₂, and MoSi₂ exhibit ferromagnetic ordering, while MnSi₂ and FeSi₂ monolayers favor antiferromagnetism. Among them, the formation energies of ferromagnetic TiSi₂ and antiferromagnetic FeSi₂ monolayers are lower than that of silicene, and the formation energy of antiferromagnetic MnSi₂ with large on-site moment is comparable to that of silicene. As a result, these three TMSi₂ monolayers are expected to possibly be synthesized in experiments due to their low formation energies. An outstanding advantage of these novel TMSi₂ monolayers is the tunable magnetic ordering from ferromagnetic to antiferromagnetic, implying their great potential in future spintronic devices.

In this review, the growth behavior, effects of defects and substrates, and bilayer silicene are discussed. The wetting of the Ag (111) surface is crucial for silicene growth, since small silicon clusters prefer 2D structures due to passivation of the Ag atoms. Moderate and homogeneous interaction between silicene and the Ag (111) surface accounts for the superiority of Ag substrate for silicene growth. Defects, which are nearly inevitable in the fabrication, would greatly affect the electronic properties of silicene, either opening a small band gap or transforming it into metal. Moreover, porous silicene sheets with double vacancies are promising for gas purification.

After being synthesized on a metal surface, silicene is very likely to lose its Dirac cone due to strong hybridization between silicene and substrate, although there is still some debate about this point. Excitingly, many inert substrates (semiconductor or insulator) can afford effective mechanical support without destroying the geometric and electronic properties of silicene. Furthermore, a small band gap can be opened in silicene supported on these inert substrates due to the weak vdW interactions which break the symmetry of the two sublattices in silicene.

Bilayer silicene has a lower formation energy than monolayer silicene but loses the Dirac cone due to the strong covalent interaction between two silicene layers. Alkali metal intercalation can effectively decouple the two silicene layers, thus restoring the Dirac cone in pristine silicene with a small band gap at the vertex of the Dirac cone. Besides, alkali metal intercalation can further lower the formation energy compared with bilayer silicene, making intercalated bilayer silicene a promising prospective in future microelectronic devices, especially FET. As an extension of silicene, some monolayers of transition metal silicides are ferromagnetic or antiferromagnetic, which is useful in future spintronic devices.

Even though the use of silicene is promising in many applications, the research of silicene still has a long way to go. Various superstructures of silicene coexist during the epitaxial growth and thus a large number of structural defects and grain boundary emerge in the epitaxial silicene sheets. Fabricating large-scale defect-free silicene with a uniform superstructure is still a considerable challenge. Transferring silicene from metal substrates onto inert substrates without destroying the structural integrity of the 2D sheet is another difficult issue. Alternatively, one may try to grow silicene directly on some inert substrates. Recent success in building a silicene field-effect transistor has introduced a new and effective method of developing silicene-based microelectronic devices. Optimistically, all of these problems might be overcome in the near future, which will in turn open a bright future for silicene in future microelectronics and other nanoscale devices.