Novel Donor Materials in Excitonic Solar Cells - ACS Publications

Oct 17, 2013 - Computational Dissection of Two-Dimensional Rectangular Titanium Mononitride TiN: Auxetics and Promises for Photocatalysis. Liujiang Zh...
0 downloads 0 Views 3MB Size
Letter pubs.acs.org/NanoLett

SiC2 Siligraphene and Nanotubes: Novel Donor Materials in Excitonic Solar Cells Liu-Jiang Zhou,†,‡ Yong-Fan Zhang,§ and Li-Ming Wu*,† †

State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Fuzhou, Fujian 350002, People’s Republic of China ‡ University of Chinese Academy of Sciences, Beijing 100039, People’s Republic of China § Department of Chemistry, Fuzhou University, Fujian 350108, People’s Republic of China S Supporting Information *

ABSTRACT: In excitonic solar cells (XSC), power conversion efficiency (PCE) depends critically on the interface band alignment between donor and acceptor materials. Graphene or silicene is not suitable for donor materials due to their semimetallic features (zero band gaps); it is therefore highly desired to open an energy gap in graphene or silicene to extend their application in optoelectronic devices, especially in photovoltaics. In this paper, based on the global particle-swarm optimization algorithm and the density functional theory methods, we predict a novel SiC2 siligraphene (g-SiC2) with a direct band gap of 1.09 eV showing infinite planar geometry, in which Si and C atoms adopt sp2 hybridization and C atoms form delocalized 4 C-domains that are periodically separated by Si atoms. Such a g-SiC2 siligraphene (with a global minimum of energy) is 0.41 eV/atom lower and thermally stabler than the isomeric pt-SiC2 silagraphene containing planar 4-fold coordinated silicon (3000 K vs 1000 K). Interestingly, the derivative (n, 0), (n, n) nanotubes (with diameters greater than 8.0 Å) have band gaps about 1.09 eV, which are independent of the chirality and diameter. Besides, a series of g-SiC2/GaN bilayer and g-SiC2 nanotube/ZnO monolayer XSCs have been proposed, which exhibit considerably high PCEs in the range of 12−20%. KEYWORDS: g-SiC2 siligraphene, graphene, silicene, particle-swarm optimization (PSO) algorithm, excitonic solar cell

G

eV) band gaps or gaps depending on the chirality and width of the nanoribbons. To date, graphene or silicene with mediate band gaps (i.e., 1.0−2.0 eV) that are highly desired by the field effect transistor and the solar cell7 are still extremely difficult to achieve. Intensive studies of 2D nanomaterials lead to not only successful syntheses of silicene,7,8 SiC,17 CN,18 C3N4,19 BN,20 ZnO,21 and MS2,20,22 but also theoretical predictions of some novel compounds, such as B with low-buckled configurations, 23,24 graphitic GaN−ZnO, 25 boron−carbon compounds,26,27 carbon nitride,11 germanene,12,28 pt-SiC2,29 GeC,30 SnC,30 tetragonal TiC,31 group III−VI compounds,30,32 and MX2 chalcogenides.33,34 Regarding these 2D nanomaterials, several points are worth noting: (i) C- and Si-based 2D nanomaterials are extremely rare, and only four examples are known, that is, graphene, SiC sheet,17 silicene, and pt-SiC2;29 (ii) most of them are either metal26,27,29 or semimental;12 (iii) band gaps of their derivative nanotubes are usually dependent on the chirality or/and the tube diameter.35−37 No example

raphene, a two-dimensional one-atom-thick honeycomblike layered material, shows intriguing electronic and mechanical properties, such as the chiral quantum Hall effect, unique electronic conductivity, high intrinsic mechanical strength, and large surface area, which are highly desired in the field of next generation of faster and small electronic devices.1−6 On the other hand, silicene adopting a graphenelike lattice with a low-buckled geometry has also been intensively studied,7,8 and it is particularly desired for device applications because of its compatibility.8 However, graphene and silicene both exhibit zero band gaps, making them unsuitable for the controlled and reliable transistor operation, consequently limiting their widespread applications in optoelectronic devices, such as light-emitting diodes, field effect transistors, and solar cells. It is therefore highly desired to open an energy gap in graphene and silicene. But this is a big challenge, because their unique electronic structures originate from the massless Dirac fermion-like behavior of the charge carriers. Several ways have been proposed, such as a substrateinduced gap,9 cutting into nanoribbons though the confinement,10 metal adsorption,9,11 applying an external electrostatic gate,12 chemical functionalization,13 hydrogenation,13,14 doping,11 or creating defects.15,16 Unfortunately, these methods provide either too small (about tenths of eV)6,9 or too large (>5 © XXXX American Chemical Society

Received: August 11, 2013 Revised: October 4, 2013

A

dx.doi.org/10.1021/nl403010s | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

Figure 1. (a) Optimized structure of g-SiC2 siligraphene with unit cell outlined as red lines. Red vectors: a1, a2: graphene lattice vectors; b1, b2: siligraphene lattice vectors. 1 and 2 represent the two crystallographic independent C atoms. (b) The unit cell of g-SiC2 siligraphene with bond distance, angle, and atom number. Blue: Si; gray: C. (c) The band structure of g-SiC2 monolayer calculated based on PBE and HSE06. The π bands are marked in red. (d) The total DOS and projected DOS (PDOS) of g-SiC2 siligraphene. The Fermi energy is set as zero. (e) The total electronic charge density projected on the g-SiC2 siligraphene. (f) The difference charge density (Δρ) of g-SiC2 siligraphene. Isosurface value: 0.014 e/A3; pink: Δρ > 0, green: Δρ < 0.

derivative (n, n), (n, 0) g-SiC2 nanotubes with diameters greater than 8.0 Å have uniform band gaps independent of the chirality and diameter. More interestingly, a series of g-SiC2/GaN bilayer and g-SiC2 nanotube/ZnO monolayer XSCs have been proposed, which show power conversion efficiency (PCE) in a tunable range of 12−20%, higher than the highest certified efficiency for a polymer solar cell to date (about 8.62%).46 The structure, stability, elastic, and electronic properties are also discussed. All possible 2D structures having a Si:C ratio of 1:2 were simulated by PSO and optimized by DFT. Two stable phases, g-SiC2 siligraphene and pt-SiC2 silagraphene,29 were obtained at the fourth generation with z = 2 and the 13th generation with z = 1, respectively. The new g-SiC2 siligraphene is 0.41 eV/atom lower in energy than the pt-SiC2 silagraphene.29 Their binding energies are 6.46 and 6.05 eV/atom, respectively, lower than that of graphene (8.66 eV), but higher than those of Si solid and the low-buckled Si sheet (4.57 and 3.93 eV/atom).29 Therefore g-SiC2 siligraphene shows strong covalent bonds and is relatively stable. The optimized g-SiC2 siligraphene crystallizes in the hexagonal space group, P6̅2m (no. 189), with a = 5.019 Å (Figure 1a and b). As similar as in the pure carbon graphene, Si adopts an undistorted trigonal coordination with a Si−C distance of 1.798 Å and bond angle of 120°. The C1 atom at the Wyckoff 1b site is also trigonally coordinated without any distortion. The C2 atom at the 3g site is in a distorted trigonal coordination with bond angles of 126.3 and 107.4° and bond distances of 1.798 and 1.445 Å. Such a small distortion comes from the radius and electronegativity differences between Si and C atoms, which also results in the electron polarization along the Si−C bond. Nevertheless, this distortion is limited within the g-SiC2 plane and does not push the C2 atom out of the plane. As a resemblance of the pure carbon graphene, the gSiC2 siligraphene features sp2-hybridization of C and Si atoms. The C−C bond (1.445 Å) shows the characteristic of a weak

with moderate band gap (1.0−2.0 eV) that is independent of the diameter or chirality is reported to date. 1D or 2D nanomaterials are often used as donors in an excitonic solar cell (XSC).38 A few C- or Si-based examples are recently reported, such as a graphene-based heterojunction,39 graphene−BN heterojunction,40 quantum-dot-based photovoltaics,41 and nanocarbon-based photovoltaics.11,41−43 Therefore, new layered materials containing both Si and C atoms with moderate band gaps will be of great scientific importance and are highly desired by the industry, and nanotubes with band gaps that are independent of the chirality and diameter will be equally interesting. Recently, several theoretical predictions have guided successful syntheses of the SiC sheet17 and SiC nanotubes.44 We therefore further explore the possible 2D “SixCy” materials starting with two Si:C stoichiometries of 1:2 and 2:1. Note that a “Si2C” sheet containing more Si than C atoms tends to have high formation energy, to adopt a silicene-like low-buckled geometry, and it is thus unstable. In this paper, we only explore all of the possible monolayer geometries with a Si:C ratio of 1:2. Among these possibilities, a SiC2 sheet containing planar 4-fold coordinated silicon (denoted as pt-SiC2) has been already predicted recently,29 which is metallic and shows a relative low melting temperature (below 1000 K) that limits their application in the microelectronic semiconductor industry. By utilizing the global particle swarm optimization (PSO)45 methodology and density functional theory (DFT) methods (Supporting Information), we first confirm the existence of the reported pt-SiC2 2D planar structure and then systematically search all other possible 2D planar structures of the same stoichiometry (Si:C = 1:2). Fortunately, a new 2D planar SiC2, a resemblance of the pure carbon graphene honeycomb lattice, siligraphene (denoted as g-SiC2) is found to be more stable than the isomeric pt-SiC2 silagraphene structure.29 Interestingly, such a g-SiC2 siligraphene has a direct band gap of 1.09 eV and shows excellent thermostability up to 3000 K. The B

dx.doi.org/10.1021/nl403010s | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

mainly from Si to C2 atom via Si−C bonding interaction, and the centering C1 atom in the 4 C-domain is almost neutral. Usually, Si prefers sp3 hybridization, and C, sp2 hybridization. This has been verified in silicene (mixture of sp2 and sp3 hybridization leading to the low-buckled geometry) and in graphene (sp2 hybridization). In g-SiC2 siligraphene, the Si and C atoms adopt sp2 hybridization, and the Si−C bond is a twocenter sp2 covalent bond. The sp2 hybridization and the electronegativity difference between C and Si atoms make the π electrons mainly delocalizing over the 4 C-domain, which has relatively weak π interaction with the pz orbitals of Si atoms. Such π distribution leads to the semiconducting feature of gSiC2 siligraphene. Differently, in graphene and silicene, π electrons are distributed evenly within the entire layer. Nanotubes have been constructed by rolling the g-SiC2 siligraphene along specific directions according to a wellestablished method.54 The (n, 0) g-SiC2 tubes are armchair type, and the (n, n) g-SiC2 nanotubes are zigzag type. Two representative types, (n, n) (3 ≤ n ≤ 10) and (n, 0) (2 ≤ n ≤ 6) g-SiC2 nanotubes, are presented in Figure S3. Along the axial direction, the unit cell parameters for (4, 4) and (8, 0) nanotubes are 5.017 and 8.699 Å. Other (n, n) and (n, 0) gSiC2 nanotubes with diameters ranging from 4.20 to 14.50 Å have also been studied. As shown in Figure S4, the Forcite bond distribution analyses of g-SiC2 show the Si−C and C−C bond lengths do not differ significantly no matter in a tube or in a sheet, except in the thin (2, 2) and (3, 0) tubes. The bond lengths in larger tubes, such as (4, 4) and (8, 0) tubes, are already close to those of a monolayer. In the thin (2, 2) and (3, 0) tubes, the C−C bonds elongate approximately parallel to tube axis, while the Si−C bonds shorten perpendicular to the tube axis. The stability of g-SiC2 nanotubes is evaluated by the strain energy, which is defined as the cohesive energy difference between a tube and a sheet. The average strain energies of (n, n) (3 ≤ n ≤ 10) and (n, 0) (2 ≤ n ≤ 6) g-SiC2 tubes have been shown in Figure 2a. For both types, the strain energy inversely changes with the increase of the diameter; this is consistent with the fact that the thin tubes (with small diameters) are expected to possess larger strain energies because of the stronger curvature effect. Nevertheless, all tubes wider than 8 Å in diameter have small stain energies (