Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2024 Sep 5;18(37):25478–25488. doi: 10.1021/acsnano.4c04789

Identification and Manipulation of Atomic Defects in Monolayer SnSe

Chengguang Yue †, Zhenqiao Huang ‡, Wen-Lin Wang †, Zi’Ang Gao †, Haicheng Lin †,*, Junwei Liu ‡,*, Kai Chang †,*
PMCID: PMC11411721  PMID: 39236319

Abstract

graphic file with name nn4c04789_0008.jpg

SnSe, an environmental-friendly group-IV monochalcogenide semiconductor, demonstrates outstanding performance in various applications ranging from thermoelectric devices to solar energy harvesting. Its ultrathin films show promise in the fabrication of ferroelectric nonvolatile devices. However, the microscopic identification and manipulation of point defects in ultrathin SnSe single crystalline films, which significantly impact their electronic structure, have been inadequately studied. This study presents a comprehensive investigation of point defects in monolayer SnSe films grown via molecular beam epitaxy. By combining scanning tunneling microscopy (STM) characterization with first-principles calculations, we identified four types of atomic/molecular vacancies, four types of atomic substitutions, and three types of extrinsic defects. Notably, we have demonstrated the ability to convert a substitutional defect into a vacancy and to reposition an adsorbate by manipulating a single atom or molecule using an STM tip. We have also analyzed the local atomic displacement induced by the vacancies. This work provides a solid foundation for engineering the electronic structure of future SnSe-based nanodevices.

Keywords: atomic defects, monolayer SnSe, two-dimensional ferroelectricity, local density of states, molecular beam epitaxy, scanning tunneling microscopy

Group-IV monochalcogenides are a family of semiconductors with orthorhombic lattices that resemble staggered black phosphorus. Their relatively low crystalline symmetry makes their physical properties highly tunable. Among them, SnSe, with a moderate bandgap,1−3 is utilized in various applications including photodetectors,4 solar cells,5 photocatalysis,6 supercapacitors,7 gas sensors,8 memristors,9 thermoelectric materials,10,11 and anode materials for batteries.12 Recent studies have identified topological crystalline insulating phases in rocksalt Pb1–xSnxSe single crystals and SnSe thin films.13−16 SnSe is also renowned for its thermoelectricity, attributed to robust anharmonicity and outstanding in-plane electrical transport,10,11 especially in the β-SnSe phase. Notably, α-SnSe demonstrates two-dimensional (2D) ferroelectricity,17−20 which allows for the switchable in-plane spontaneous polarization at the monolayer level and at room temperature.17 Since the properties and applications above are highly sensitive to the electronic structure and chemical potential of SnSe, understanding the impact of point defects holds significant interest.

Extensive research on SnSe’s point defects is primarily driven by their significant impact on the material’s exceptional thermoelectric properties. Density functional theory (DFT) modeling has extensively addressed intrinsic defects such as vacancies (VSn, VSe), antisites (SnSe, SeSn), and interstitials (Sni, Sei), as well as extrinsic defects, in numerous studies.21−27 Notably, charge defects such as VSn and VSe, which increase carrier concentration, are key to achieving a high power factor.21 These calculations identify VSn as the main factor in SnSe’s intrinsic p-type conductivity, attributed to its relatively low formation energy and shallow defect energy level.22−25 Experimental observations utilizing scanning tunneling microscopy (STM) have revealed the vacancies of Sn and Se atoms, as well as bunched vacancies involving multiple atoms (referred to as “multivacancies”), through atom-resolved images and differential conductance (dI/dV) spectra, from which the origin of p-type doping from Sn vacancies was affirmed.25,28,29 Furthermore, the structures of VSn, Sei, and multivacancies have been resolved using scanning transmission electron microscopy.29−32 Studies on the formation of VSn and VSe at different annealing temperatures were conducted using positron annihilation spectroscopy combined with transport measurements.33 Nevertheless, there remains a lack of systematic research focusing on the local density of states (LDOSs) of all types of intrinsic defects in SnSe. Moreover, previous studies on SnSe defects were predominantly static and lacked exploration into the transition between different types of defects, particularly in regards to controlled conversion.

Here, by combining low-temperature STM with DFT calculations, we have conducted a comprehensive analysis of the atomic and electronic structures of point defects in a single van der Waals monolayer (two atomic layers) of SnSe. This includes 8 types of intrinsic defects and 3 types of extrinsic defects. Furthermore, we demonstrated the capability to convert a substitution defect into a vacancy through the manipulation of the electric field between the STM tip and the sample surface. This study provides essential insights for future explorations in SnSe-based devices.

Results and Discussion

As previously reported, the lattice structure of SnSe varies with temperature and epitaxial conditions, alternating between an orthorhombic α phase (space group Pnma) or β phase (space group Cmcm), or a rocksalt phase (space group Fm3̅m).3,16 In this study, all monolayer SnSe samples crystallize in the α phase, which has a spontaneous in-plane polarization along its a1 direction. This polarization reduces the crystalline symmetry to Pnm21, as illustrated in Figure 1a.

Figure 1.

Figure 1

Open in a new tab

Point defects in monolayer SnSe. (a) Side view (upper panel) and top view (lower panel) of the lattice structure of monolayer SnSe. (b,c) Typical STM topography images of monolayer SnSe grown with the substrate kept at room temperature (b) and 100 °C (c), and subsequently annealed at 100 °C for 30 min. (d) The apparent height profile was extracted along the dashed line in (c). (e) dI/dV spectrum obtained from a defect-free area on the surface of monolayer SnSe. The arrows indicate the energies of the CBM and VBM. (f–i) Typical point defects in monolayer SnSe. Dashed circles in different colors are used to mark the type of defects. (f)/(g) and (h)/(i) are pairs of images acquired at the same positions but with different Vs. Tunneling parameters: (b) Vs = – 1.0 V, It = 5 pA; (c) Vs = + 1.7 V, It = 10 pA; (e) Vs = + 2.0 V, It = 50 pA, the sinusoidal modulation voltage VOSC = 20 mV for positive Vs, and Vs = – 0.8 V, It = 50 pA, VOSC = 8 mV for negative Vs; (f) Vs = – 0.7 V, It = 5 pA; (g) Vs = + 1.7 V, It = 5 pA; (h) Vs = – 0.9 V, It = 100 pA; (i) Vs = + 1.7 V, It = 200 pA. The scale bars in (b,c) are 80 nm, while those in (f–i) are 5 nm.

Our investigation commenced with synthesizing monolayer SnSe films on graphene by directly depositing SnSe molecules on a graphitized 4H–SiC(0001) substrate at either room temperature (Figure 1b) or 100 °C (Figure 1c). We deliberately keep the substrate temperature low to increase the concentration of point defects. The apparent height of the monolayer, consisting of two atomic layers, ranges between 0.72 and 0.78 nm, depending on the applied sample bias voltage Vs. Annealing at 250 °C leads to the formation of square-shaped nanoplates with a lower defect concentration, as reported in a previous study.17 The dI/dV spectrum, acquired at defect-free areas, reveals a band gap of 2.12 eV, with the conduction band minimum (CBM) at 1.59 eV and the valence band maximum (VBM) at – 0.53 eV (Figure 1e), in agreement with previous measurements.17 The band gap of monolayer SnSe is significantly larger than those reported in bulk SnSe, which range between 0.86 and 1.35 eV.3 This difference is attributed to the quantum size effect in the atomically thin nanoplates, where the bulk electronic bands further quantize into 2D quantum well states with parabolic dispersion. As the material’s thickness decreases, the energy separation between the apexes of these parabolic bands increases, leading to a corresponding rise in the band gap. Our measurement displays a quick drop of band gap size from 2.12 eV to approximately 1.4 eV as the thickness increases from one monolayer to six monolayers.

Figure 1f–i exhibit atom-resolved STM topography images of the point defects in monolayer SnSe films. The crystalline orientation can be easily identified from the moiré stripes (Figure S1), with the a1 and a2 axis being perpendicular and parallel to the stripes, respectively.17Figure 1f,g show the same group of point defects, resolved at Vs below the VBM and above the CBM, corresponding to the filled and empty states, similar as in Figure 1h,i. Only the Sn sublattice is resolved at both positive and negative Vs because the Sn atoms are lifted compared to the Se atoms on the surface of this staggered black phosphorus lattice (Figures 1a and S2).24 The atoms are usually more clearly resolved at a negative Vs, probably due to differences in local atomic orbitals near the CBM and VBM. The point defects within a single image act as references for each other, enabling precise positioning of each point defect, irrespective of the resolution of atomic lattices.

The comprehensive studies involving STM topography, dI/dV spectra, and DFT calculations enable us to identify the atomistic configurations of various point defects, especially those with similar structures. For instance, we can distinguish the same type of vacancy defects occurring in different atomic layers, as well as the same type of substitution defects with minor variations in atomic structure. Consequently, we have cataloged 8 types of intrinsic point defects and 3 types of extrinsic point defects, as listed in Table 1. These include 4 types of atomic/molecular vacancies (V1–V4), 4 types of Se antisite substitution of Sn (S1–S4), 1 type of Pb-substitution of Sn (S5), and 2 types of adsorbates (A1, A2). The justification for these defects is detailed below.

Table 1. Types of Point Defects and Their Notations.

defect type notation
Sn vacancy in SAL V1
Sn vacancy in BAL V2
combined defect a Sn vacancy in SAL and a Se vacancy in BAL V3
combined defect a Se vacancy in SAL and a Sn vacancy in BAL V4
Sn-substituted by Se in SAL (configuration 1) S1
Sn-substituted by Se in BAL (configuration 1) S2
Sn-substituted by Se in SAL (configuration 2) S3
Sn-substituted by Se in BAL (configuration 2) S4
Sn-substituted by Pb in SAL S5
Open in a new tab

For the vacancy defects, we attribute V1/V2 to atomic Sn vacancies in the surface/bottom atomic layer (SAL/BAL), and V3/V4 to vertical molecular Sn–Se vacancies at the Sn/Se site in SAL, respectively (Figure 2a–d). These assignments correspond with the features in STM topography images: V1 and V3 show a suppression in LDOSs at a Sn site in SAL during tunneling into both filled and empty states (Figures 2e,g; S5), while V2 and V4 are located at the center of the four nearest Sn atoms in the SAL. Noncentrosymmetric features, induced by the in-plane polarization along a1, are observed in these defects, especially in V1 and V2, near the VBM. Bright branches indicating higher LDOSs extend from V1 along the [11̅0] and [110] directions, similar to features reported in early STM studies on surficial Sn vacancies in bulk SnSe.25,28,29 The four nearest Sn atoms on top of V2 are highlighted, with the two atoms in the direction antiparallel to in-plane polarization appearing brighter. In comparison, V3 and V4 introduce less significant changes to the LDOSs in SAL, likely due to their charge-neutral nature. According to previous studies, Sn vacancies are a primary source of intrinsic p doping in SnSe,22,23,25,28,29,31 while SnSe molecule vacancies do not introduce additional charge carriers. The STM topography features, including noncentrosymmetric appearances, are well reproduced by the DFT calculations (Figure 2i–l). Noticeably, the simulated topography of V2 shows enhanced LDOS at the site of a Se atom in SAL; this feature is also experimentally observed in the dI/dV mapping images when Vs is set around −1.0 V (Figure S3).

Figure 2.

Figure 2

Open in a new tab

Atomic structures and the topographic appearances of vacancy defects V1–V4. (a–d) Side view (upper panels) and top view (lower panels) of the atomic structures of defects V1–V4, respectively. (e–h) Atom-resolved topography images of defects V1–V4. All scale bars correspond to 1 nm. The tunneling parameters for each image are specified as follows: (e) Vs = – 0.9 V, It = 10 pA; (f) Vs = – 0.9 V, It = 10 pA; (g) Vs = – 0.8 V, It = 30 pA; (h) Vs = – 1.0 V, It = 30 pA. (i–l) Simulated atom-resolved STM topography images of all vacancies. The sample bias voltages used in simulation are −0.3 V for all. The simulations of the topography images have set the VBM equal to the Fermi level. Distribution of height (m–p) and interatomic distances (q–t) of Sn sublattice within SAL around defects V1–V4. The red dashed circles indicate the positions of Sn and Sn–Se vacancies. The smaller dots are Se atoms with a height higher than Sn atoms in the (m) and (n).

Based on the relaxed atomic structures obtained from DFT calculations, we have extracted the lattice distortion map at each atomic site (Figure 2m–t). For all the four types of vacancy defects, the induced noncentrosymmetric distortion extends over 2–3 Sn atom sites from the defect center, aligning with the experimental results. Interestingly, the enhanced LDOSs around V1 and V2 do not derive from an increase in the corresponding Sn atoms’ height in SAL. In contrast, many of these Sn atoms shift downward. Therefore, the primary cause of LDOS enhancement appears to be changes in electronic states, which is further characterized by the dI/dV spectra obtained at these defects (Figure 3). None of these defects introduce observable states inside the semiconducting gap of monolayer SnSe. Directly at the defect site, V1 and V3 show a suppression in LDOSs at both their CBM and VBM (Figure 3a,c,e–g,k–m), while the CBM and VBM of V2 and V4 are hardly affected (Figure 3b,d,h–j,n–p). This can be understood as both V1 and V3 involve the absence of a Sn atom in SAL, while only Sn atoms can be resolved in STM topography images. Notably, V2 shows significant LDOS enhancement when Vs is set below −0.8 V (Figure 3b,j), consistent with the DFT simulation (Figure 2j) and dI/dV mapping (Figure S3). Interestingly, as a molecular vacancy, V4 hardly affects the LDOS of the CBM (Figures 3o and S5), which is consistent with its charge neutrality, and its position sitting in between four Sn atoms.

Figure 3.

Figure 3

Open in a new tab

dI/dV spectra of defects V1–V4. (a–d) Comparison of the dI/dV spectra acquired right at the defect position and those from defect-free areas. (f,g,i,j,l,m,o,p) Spatially resolved dI/dV spectra of V1–V4, obtained along the dash arrows in (e,h,k,n), respectively. Because the dI/dV intensity of conduction and valence bands are significantly different, the spectra above and below the Fermi level were measured under different tunneling parameters for clarity: (f,l) Vs = + 2.4 V, It = 200 pA, VOSC = 24 mV; (i,o) Vs = + 2.4 V, It = 100 pA, VOSC = 24 mV; (g) Vs = – 1.5 V, It = 200 pA, VOSC = 15 mV; (j,p) Vs = – 1.5 V, It = 100 pA, VOSC = 15 mV; (m) Vs = – 1.2 V, It = 200 pA, VOSC = 12 mV.

The spatial oscillations in Figure 3g–p correspond to the atomic corrugation of the SnSe lattice. When acquiring dI/dV spectra along the a1 direction of SnSe, the period of oscillation is about 4.4 Å; while along the [11] direction, it extends to about 6.1 Å. These measurement values are consistent with the lattice parameters of SnSe. Significantly, these spatial oscillations can only be observed under negative bias voltage. This phenomenon is also evident in the topography images, where atoms appear more distinctly at negative Vs due to the less localized electronic states around the CBM. It is also worth noting that, although the VBM mainly consists of the orbitals of Se atoms, the atoms resolved at negative bias voltage are still Sn. This is attributed to the Sn sublattice being slightly higher than the Se at the surface, as reported in previous studies.17,24

We noted that Sn atom vacancies are broadly reported in STM studies of bulk SnSe crystals,25,28,29 however, reports of Sn–Se molecular vacancies are rare. This is probably because our SnSe films were grown from the deposition of SnSe molecules, rather than synthesizing them from single Sn and Se elements. Different from the growth of bulk crystals, the molecular beam epitaxy (MBE) growth of thin films can be a process far away from thermal equilibrium, because the latter usually happens at a much lower temperature and in a much shorter time than the former. For instance, our growth of monolayer SnSe nanoplates happened at a substrate temperature of 300–370 K, much lower than SnSe’s melting point, and the growth only took several minutes. Therefore, growth kinetics largely affect the crystalline structure and the defect types in the MBE grown films.34−37 When directly depositing SnSe molecules, the formation of atomic defects involves breaking the bond between Sn and Se atoms in a molecule (our calculation yields a bonding energy of 5.04 eV), and forming new bonds with the existing film. Therefore, only the atomic defects with low enough formation energy can appear, such as Sn vacancies and the Se-substitution of Sn. The latter even has negative formation energy in a monolayer SnSe film (Table 2), implying that it would automatically appear as long as excess Se exists. However, the formation energy of Se vacancies in monolayer SnSe is 0.778 eV according to our calculations, much higher than that of Sn vacancies (0.382 eV). On the other hand, this growth kinetics increases the possibility of forming SnSe molecular vacancies, because this process only involves the deposition of whole molecules, while does not need to break the bond inside a molecule. In fact, similar molecular vacancies have also been observed in other materials that were grown from the evaporation of a single compound, such as those in CdTe crystals.38

Table 2. Formation Energies and Densities of the Intrinsic Point Defects.

    defect density (1010 cm–2)
defect type formation energy (eV) deposit at RT deposit at 100 °C
V1 0.382 ± 0.005 2.89 2.87
V2 19.86 10.75
V3 4.407 ± 0.002 2.89 3.58
V4 0.00 0.72
S1 –2.680 ± 0.028 17.33 (total number of S1∼S4) 19.35 (total number of S1∼S4)
S2
S3 –2.358 ± 0.005
S4
Open in a new tab

Besides the vacancy defects, our findings regarding the antisite substitution defects (S1–S4) are more intriguing. In these defects, two distinct atomistic configurations emerge when a Se atom substitutes a Sn atom, as illustrated in Figure 4a–d. At a positive Vs close to the CBM, all antisite substitution defects appear as bright spots in STM topography images. However, at higher Vs, their apparent heights are suppressed, resulting in a dip (Figures 4j,k, S6). Conversely, at a negative Vs close to the VBM, the two types of defects in SAL, S1 and S3, show a dip at the original Sn sites that were substituted (Figure 4e,g). However, the other two types in BAL, S2 and S4, hardly show any features at negative Vs (Figure 4f,h). A comparison between experimental topography images and the DFT simulated images shows a strong correlation across different ranges of Vs (Figure 4i–n).

Figure 4.

Figure 4

Open in a new tab

Atomic structures and the topographic appearances of substitution defects S1–S4. (a–d) Side view (upper panels) and top view (lower panels) of the atomic structures of defects S1–S4, respectively. (e–h) Atom-resolved topography images. The features of defects S2 and S4 are clearer at positive Vs (see Supporting Information). (i–k) STM images of defects S1 and S3 at different sample bias voltages. The defects S1, S3 and V3 are indicated by gray, red and purple dash circles respectively. (l–n) Corresponding simulated STM images of defect S3. The tunneling parameters for each image are specified as follows: (e) Vs = – 1.0 V, It = 30 pA; (f) Vs = – 0.9 V, It = 50 pA; (g) Vs = – 1.0 V, It = 30 pA; (h) Vs = – 0.6 V, It = 10 pA; (i) Vs = – 1.0 V, It = 30 pA; (j) Vs = + 1.7 V, It = 10 pA; (k) Vs = + 2.2 V, It = 10 pA. All scale bars correspond to 1 nm.

Although the two types of antisite substitution are challenging to distinguish using STM topography images alone, the extra electronic states they introduce within the band gap are very different, making them easy to identify through dI/dV spectra. For the first type of atomistic configuration (S1 and S2, where S1 occurs in SAL and S2 in BAL), although their topography appearances are completely different, dI/dV spectra reveal almost identical in-gap states located right at the CBM energy of defect-free areas (Figure 5a,b,f,i), which introduce shallow n-type doping levels. These in-gap states are highly localized, expanding no further than the nearest unit cells. On the other hand, the second type of atomistic configurations (S3 and S4, with S3 in SAL and S4 in BAL), feature in-gap states approximately 0.6 eV lower than the CBM (Figure 5c,d,l,o), indicating deeper n-doping levels compared to those of S1 and S2. Furthermore, when measured under identical Vs and tunneling current It, the dI/dV spectrum weight of the in-gap states of S3/S4 is much lower than that of S1/S2.

Figure 5.

Figure 5

Open in a new tab

dI/dV spectra of defects S1–S4 and the corresponding DFT simulations. (a–d) Comparison of the dI/dV spectra acquired right at the defect position and those from defect-free areas. (f,g,i,j,l,m,o,p) Spatially resolved dI/dV spectra of defects S1–S4, obtained along the dash arrows in (e,h,k,n), respectively. Tunneling parameters: (f,o) Vs = + 2.4 V, It = 300 pA, VOSC = 24 mV; (i) Vs = + 2.4 V, It = 100 pA, VOSC = 20 mV; (l) Vs = + 2.4 V, It = 100 pA, VOSC = 24 mV; (g) Vs = – 1.5 V, It = 200 pA, VOSC = 15 mV; (j) Vs = – 1.0 V, It = 100 pA, VOSC = 10 mV; (m) Vs = – 1.4 V, It = 100 pA, VOSC = 14 mV; (p) Vs = – 1.5 V, It = 100 pA, VOSC = 15 mV. (q,r) DFT simulated topography images of S1 (q) and S3 (r) at a negative sample bias voltage. (s,t) DFT simulated LDOSs distribution across the defects S1 and S3, respectively. In-gap states that agree well with the experiments can be resolved.

The main difference between the spectra of S1 and S2 is found in their behavior under negative sample bias voltages. At the VBM, the LDOS at the center of S1 is suppressed, while S2 displays almost no apparent features. This behavior can be attributed to the atomistic configuration of the antisite defect. Specifically, when the substitution happens at the SAL, the substituted Se atom is positioned significantly lower than the surrounding Sn atoms in this layer, resulting in a reduced LDOS at S1. In contrast, when the substitution happens in the BAL, the height of the atoms in the SAL is just slightly affected, which explains the indistinct feature in the LDOS at S2. Similarly, a comparable feature is observed in the spectra of S3 and S4.

It should be noted that the whole spectra of S1 is shifted downward by 0.1 eV compared to the other defects. This shift occurs because the spectra of S1 was acquired from a SnSe nanoplate grown on monolayer graphene, while those of S2∼S4 were derived from nanoplates on bilayer graphene. This energy shift indicates the differences in the work functions between monolayer and bilayer graphene surfaces, which has been consistently observed in our experiments.

Our DFT calculations also support the interpretation of the dual configurations of antisite substitution defects, indicating that both configurations are at local energy minima. Notably, the formation energy of S1/S2 is 0.32 eV lower than that of S3/S4 (Table 2), implying that the S1/S2 configurations are more stable. Furthermore, the spatially and energetically resolved LDOS distribution of S1 and S3 reproduces the in-gap states, with those of S1/S2 being closer to the CBM, aligning well with the experimental results.

Having characterized all 8 types of intrinsic point defects in monolayer SnSe, we calculated their formation energies and compared these theoretical results with the experimental data in Table 2. The formation energy of V1/V2 is 1 order of magnitude lower than that of V3/V4, which is consistent with the experimentally observed lower defect density of V1/V2. However, in practical samples, the density of V2 is 4 to 7 times higher than that of V1, despite being the same type of defect with identical formation energy in the calculations. This discrepancy is most likely from the influence of the grapitized SiC substrate, which modifies the chemical environment of the SnSe film from the bottom side and breaks the equivalence between V1 and V2. Using the Gundlach oscillation depicted in the dz/dV spectra,39 we can readily extract the difference in work functions between monolayer SnSe and the substrate (Figure S9). The comparatively higher work function of SnSe prompts electron transfer from graphene, leading to an accumulation of negative charge at the SnSe side of the interface. Theoretical studies have shown that the formation energy of Sn vacancies decreases as the VBM of SnSe moves farther from the Fermi energy,23 hence favoring their formation in BAL. Surprisingly, though the calculated formation energies of all the antisite substitution defects are negative, their densities are not significantly higher than those of the vacancy defects in experiments. (The antisite substitution defects were counted altogether in Table 2 because it was difficult to distinguish them merely from topography images.) This is probably due to our growth method of directly depositing SnSe molecules, during which most of the SnSe molecules do not break into atoms, limiting extra Se atoms. The negative formation energy of these defects can be explained by the phase diagram of Sn and Se elements. Given that a stable Se-richer phase, SnSe2, exists between these two elements, extra Se flux tends to form patches of SnSe2 inside SnSe, with antisite substitution defects serving as nucleation centers. In fact, people have observed the transition from SnSe films to SnSe2 during postannealing in a Se-rich environment, implying a lower formation energy of SnSe2.40,41

In addition, we have identified three types of extrinsic point defects: a Pb-substitution of a Sn atom in SAL (S5), originating from impurities in the evaporation material, and two types of surface adsorbates, derived from the residual gases in the vacuum chamber (A1 and A2).

The topographic features of S5 are similar to those of S1 and S3, except that the dip at negative Vs is shallower and the brightness at positive Vs is lower (Figure 6a–c), aligning with DFT calculations (Figure 6d–f). Considering that Pb-substitution for Sn in SnSe is isovalent, it naturally follows that S5 does not introduce extra in-gap states but merely modifies the structure of the band edges at the CBM and the VBM slightly (Figure 6g). Although the specific doping atom cannot be directly identified from the spectra, Pb doping is the most probable cause because (i) no impurities listed on the datasheet of the SnSe granules used for evaporation can lead to isovalent substitution; (ii) the MBE chamber was used for the growth of PbSe at the same time, potentially leading to slight cross-contamination. No sign of Pb-substitution defects were found in BAL, probably because it hardly affects the neighboring atoms and are thus not detectable.

Figure 6.

Figure 6

Open in a new tab

Three types of extrinsic defects. (a–c) Typical atomic-resolved STM topographic images of S5. The defects S1 and S5 are indicated by gray and magenta dash circles respectively. The tunneling parameters used are Vs = – 0.9 V, It = 10 pA; Vs = + 1.5 V, It = 10 pA; Vs = + 2.3 V, It = 10 pA, respectively. (d–f) Simulated atomic-resolved STM topographic images for S5. (g) Atomic-resolved STM topographic image of S5. (h,i) Spatially resolved dI/dV spectra obtained along the black dash arrow indicating in (g). Tunneling parameters: (h) Vs = + 2.4 V, It = 200 pA, VOSC = 24 mV; (i) Vs = – 1.5 V, It = 200 pA, VOSC = 15 mV. (j–m) STM images of typical adsorbates A1 and A2 which are indicated by purple and green dash circles respectively. The tunneling parameters used are Vs = – 0.9 V, It = 30 pA; Vs = + 1.6 V, It = 10 pA; Vs = – 0.9 V, It = 30 pA; Vs = + 1.6 V, It = 10 pA, respectively. All scale bars correspond to 1 nm.

Besides S5, two types of adsorbates positioned directly above a Sn atom were observed: one activating three atoms in a row (A1, Figure 6j) and the other activating only one atom (A2, Figure 6l) at negative Vs. Interestingly, the brightness of the three atoms in A1 is not equal. The vector from the brightest to the dimmest atom is always parallel to the in-plane polarization of monolayer SnSe, making A1 a useful local indicator of polarization within SnSe. At positive Vs, the appearances of A1 and A2 are similar. It is not yet known the exact types of molecules that are responsible for the adsorbates, but it can be assumed that A1 is from polar molecules like H2O or CO, while A2 is from nonpolar molecules like H2, N2, and O2. Further details about the extrinsic point defects can be found in the Supporting Information.

We have not only identified the point defects in monolayer SnSe but also developed techniques to manipulate them. Figure 7a–e illustrate the process of converting a substitution defect S3 into a vacancy defect V1 through the interaction between the STM tip and the defect. Specifically, we position the tip above a defect S3 with tunneling parameters set at Vs = 1.6 V and It = 1.6 nA, then turn off the feedback loop and move the tip horizontally around the defect, and subsequently retract the tip from the surface. Following this operation, S3 is almost 100% converted into V1 by extracting the antisite Se atom out from the defect. Since the formation energy of S3 is lower than that of V1, this operation is a process of energy injection. Such manipulations can be applied to rationally adjust the local electronic states in monolayer SnSe. However, even when increasing Vs to over 4.0 V, we have not observed the conversion of S1 to any type of vacancy defect, implying that S1 maintains a stable configuration, while S3 is metastable, which consistent with their formation energies. Meanwhile, adsorbates A1 and A2 can also be relocated on the surface of monolayer SnSe following similar procedures (Figure 7f–j). Additionally, we have demonstrated the capability to remove a single Sn atom from a defect-free area of monolayer SnSe using the STM tip (Figure S11). Although the success rate of this operation currently stands at approximately 10%, it implies the potential to deliberately design patterns of vacancy defects through a series of manipulations with the STM tip.

Figure 7.

Figure 7

Open in a new tab

Manipulating point defects S3, A1 and A2 with a STM tip. (a) A schematic illustrating the process of converting S3 into V1. (b,c) Filled (b) and empty (c) state STM topography images before manipulating S3. The inset shows a typical dI/dV spectrum at S3. The S5 defect beside S3 serves as a local marker. The dash lines are guide for eyes to show the topography features around the S3. (d,e) Filled (d) and empty (e) state STM topography images after converting S3 into V1. The features marked by dash lines keep unchanged after the defect manipulation. (f) A schematic illustrating the process of relocating an adsorbate. (g,h) STM topography images before and after relocating an A1 defect. The dash line marks the distance from another V2 defect to the target A1 before moving, and the red solid arrow indicates the path of movement of A1. (i,j) STM topography images before and after relocating an A2 defect. Tunneling parameters: (b,d) Vs = – 0.9 V, It = 100 pA; (c,e) Vs = + 1.7 V, It = 10 pA; (g–j) Vs = – 0.9 V, It = 100 pA. All scale bars correspond to 2 nm.

Conclusions

In conclusion, we have extensively investigated the point defects in monolayer SnSe grown by MBE, combining both STM studies and DFT calculations. Eight types of intrinsic defects were identified, including 4 types of vacancies and 4 types of antisite substitutions. The vacancy defects consist of the loss of either a single Sn atom or a vertically oriented SnSe molecule. Most of the vacancy defects exhibit noncentrosymmetric appearances that are consistent with the in-plane polarization in monolayer SnSe. Moreover, the density of atomic Sn vacancies in BAL is significantly higher than in SAL, likely due to the influence of the graphitized SiC substrate. Surprisingly, the antisite substitution defects, involving a Se atom replacing a Sn atom, exhibit in two distinct atomistic configurations. Despite their similar topographic appearances, the energies of the extra electronic states they introduce within the band gap of monolayer SnSe show significant differences. All substitution defects exhibit negative formation energies, yet their densities are limited by the growth method involving directly deposition SnSe molecules. Most interestingly, we have achieved nearly 100% success in converting an antisite substitution defect into a Sn vacancy using STM tip manipulation. Furthermore, we identified 3 types of extrinsic point defects, including a Pb-substitution of a Sn atom and 2 types of adsorbates. Our study has unambiguously revealed all observable point defects as well as their atomic and electronic structures, establishing methods for their manipulation, hence clarifying the influence of the point defects on the electronic structure of SnSe. The results of this study can be applied in the rational band engineering of both ultrathin and bulk SnSe for applications in thermoelectric, photovoltaic and nonvolatile logical devices.

Materials and Methods

Sample Growth

A monolayer of SnSe was grown on a graphitized 4H–SiC(0001) substrate using MBE under a base pressure of 1 × 10–10 mbar. The substrate preparation process, involving ultrahigh vacuum annealing, has been described in previous reports.17 SnSe molecules were evaporated from high-purity SnSe granules (99.999%, Alfa Aesar) contained in Knudsen cell, which was kept at 420 °C. The substrates were held at either room temperature or 100 °C during the deposition, and were subsequently annealed at 100 °C for 30 min to improve film quality.

Low Temperature STM Characterization

The STM data were acquired with a Unisoku USM 1300 system directly linked to the MBE chamber. The as-prepared samples were characterized without exposure to the air. The measurements were performed at 4.2 K using mechanically sheared Pt/Ir alloy tips. Prior to measurements, both the topography and electronic states of the tip were calibrated on the surface of Ag(111) islands grown on a Si(111) substrate. The dI/dV spectra were obtained through lock-in technique, by applying a sinusoidal modulation at a frequency of 913 Hz.

Manipulating the Defects

To convert a substitution defect S3 into a vacancy V1, the STM tip was first suspended above the target defect at Vs = 1.6 V and It = 1.6 nA. The feedback loop was then deactivated to fix the sample-tip distance. Subsequently, the tip was laterally moved (speed 1 nm/s) away from the defect along the ±a1 and ±a2 directions, before being retracted. To ensure a higher success rate, the tip movement were carried out repeatedly in all the four directions. Using this method, the probability of successfully converting S3 into V1 is nearly 100%. We have also attempted to manipulate S1 at a Vs up to 4.0 V, but no conversion was observed. At higher Vs, the SnSe film could breakdown. The technique for relocating adsorbates was similar as above.

DFT Calculations

We performed the calculations using the Vienna ab initio simulation package (VASP) code,42,43 with the projector augmented wave method44 employing the Perdew–Burke–Ernzerhof functional45 within the generalized gradient approximation to describe exchange correlation interactions. Defect structures were based on a 14 × 14 large supercell, with a vacuum space 12 Å to avoid interlayer interaction. All structures were relaxed until forces on each atom were smaller than 0.01 eV/Å, and the convergence criteria for electronic iteration was set to 10–6 eV. STM images were simulated based on partial charge densities from the VASP code, while LDOS for simulating dI/dV curves were calculated using the GPAW package.46 The formation energy is calculated by subtracting the total energy of the pristine bulk material from the total energy of the system containing the single defect.

Acknowledgments

This work is funded by the National Natural Science Foundation of China (12074038, 92165104, 12204048, 12304206, 12022416), National Key R&D Program of China (2021YFA1401500), the Hong Kong Research Grants Council (26302118, 16305019), Beijing Municipal Science & Technology Commission (Z221100002722013), Innovation Program for Quantum Science and Technology (2023ZD0300500), Beijing Natural Science Foundation (1242037) and Open Research Fund Program of the State Key Laboratory of Low-Dimensional Quantum Physics (KF202208).

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsnano.4c04789.

  • Ferroelectricity in monolayer SnSe, buckled lattice in SnSe, V2 and S2 lattice site assignment, DFT simulations of the defects, work function difference between SnSe and graphene, extrinsic defects—S5, A1 and A2, and creating V1 defect from a defect-free area (PDF)

Author Contributions

K.C. designed and led the experiments. C.Y. prepared the samples and conducted the STM experiments together with W.-L.W. and Z.G. J.L. and Z.H. performed the DFT calculation. C.Y., Z.H., H.L., J.L. and K.C. analyzed and interpreted the data. C.Y., Z.H., H.L., J.L. and K.C. drafted the manuscript. All the authors contributed on the revision of the manuscript.

The authors declare no competing financial interest.

Supplementary Material

nn4c04789_si_001.pdf (1.8MB, pdf)

References

  1. Wu M. Two-Dimensional van der Waals Ferroelectrics: Scientific and Technological Opportunities. ACS Nano 2021, 15 (6), 9229–9237. 10.1021/acsnano.0c08483. [DOI] [PubMed] [Google Scholar]
  2. Barraza-Lopez S.; Fregoso B. M.; Villanova J. W.; Parkin S. S. P.; Chang K. Colloquium: Physical Properties of Group-IV Monochalcogenide Monolayers. Rev. Mod. Phys. 2021, 93 (1), 011001. 10.1103/RevModPhys.93.011001. [DOI] [Google Scholar]
  3. Li F.; Wang H.; Huang R.; Chen W.; Zhang H. Recent Advances in SnSe Nanostructures beyond Thermoelectricity. Adv. Funct. Mater. 2022, 32 (26), 2200516. 10.1002/adfm.202200516. [DOI] [Google Scholar]
  4. Butt F. K.; Mirza M.; Cao C.; Idrees F.; Tahir M.; Safdar M.; Ali Z.; Tanveer M.; Aslam I. Synthesis of Mid-Infrared SnSe Nanowires and Their Optoelectronic Properties. Cryst. Eng. Comm. 2014, 16 (17), 3470–3473. 10.1039/c4ce00267a. [DOI] [Google Scholar]
  5. Nair P. K.; Barrios-Salgado E.; Nair M. T. S. Cubic-Structured Tin Selenide Thin Film as a Novel Solar Cell Absorber. Phys. Status Solidi A 2016, 213 (8), 2229–2236. 10.1002/pssa.201533040. [DOI] [Google Scholar]
  6. Chen P.; Dai X.; Xing P.; Zhao X.; Zhang Q.; Ge S.; Si J.; Zhao L.; He Y. Microwave Heating Assisted Synthesis of Novel SnSe/g-C3N4 Composites for Effective Photocatalytic H2 Production. J. Ind. Eng. Chem. 2019, 80, 74–82. 10.1016/j.jiec.2019.07.033. [DOI] [Google Scholar]
  7. Zhang C.; Yin H.; Han M.; Dai Z.; Pang H.; Zheng Y.; Lan Y.-Q.; Bao J.; Zhu J. Two-Dimensional Tin Selenide Nanostructures for Flexible All-Solid-State Supercapacitors. ACS Nano 2014, 8 (4), 3761–3770. 10.1021/nn5004315. [DOI] [PubMed] [Google Scholar]
  8. Pawbake A. S.; Jadkar S. R.; Late D. J. High Performance Humidity Sensor and Photodetector Based on SnSe Nanorods. Mater. Res. Express 2016, 3 (10), 105038. 10.1088/2053-1591/3/10/105038. [DOI] [Google Scholar]
  9. Wang H.; Lu W.; Hou S.; Yu B.; Zhou Z.; Xue Y.; Guo R.; Wang S.; Zeng K.; Yan X. A 2D-SnSe Film with Ferroelectricity and Its Bio-Realistic Synapse Application. Nanoscale 2020, 12 (42), 21913–21922. 10.1039/D0NR03724A. [DOI] [PubMed] [Google Scholar]
  10. Zhao L.-D.; Lo S.-H.; Zhang Y.; Sun H.; Tan G.; Uher C.; Wolverton C.; Dravid V. P.; Kanatzidis M. G. Ultralow Thermal Conductivity and High Thermoelectric Figure of Merit in SnSe Crystals. Nature 2014, 508 (7496), 373–377. 10.1038/nature13184. [DOI] [PubMed] [Google Scholar]
  11. Chang C.; Wu M.; He D.; Pei Y.; Wu C.-F.; Wu X.; Yu H.; Zhu F.; Wang K.; Chen Y.; Huang L.; Li J.-F.; He J.; Zhao L.-D. 3D Charge and 2D Phonon Transports Leading to High Out-of-Plane ZT in n-Type SnSe Crystals. Science 2018, 360 (6390), 778–783. 10.1126/science.aaq1479. [DOI] [PubMed] [Google Scholar]
  12. Chen K.; Wang X.; Wang G.; Wang B.; Liu X.; Bai J.; Wang H. A New Generation of High Performance Anode Materials with Semiconductor Heterojunction Structure of SnSe/SnO2@Gr in Lithium-Ion Batteries. Chem. Eng. J. 2018, 347, 552–562. 10.1016/j.cej.2018.04.125. [DOI] [Google Scholar]
  13. Hsieh T. H.; Lin H.; Liu J.; Duan W.; Bansil A.; Fu L. Topological Crystalline Insulators in the SnTe Material Class. Nat. Commun. 2012, 3 (1), 982. 10.1038/ncomms1969. [DOI] [PubMed] [Google Scholar]
  14. Dziawa P.; Kowalski B. J.; Dybko K.; Buczko R.; Szczerbakow A.; Szot M.; Łusakowska E.; Balasubramanian T.; Wojek B. M.; Berntsen M. H.; Tjernberg O.; Story T. Topological Crystalline Insulator States in Pb1–xSnxSe. Nat. Mater. 2012, 11 (12), 1023–1027. 10.1038/nmat3449. [DOI] [PubMed] [Google Scholar]
  15. Okada Y.; Serbyn M.; Lin H.; Walkup D.; Zhou W.; Dhital C.; Neupane M.; Xu S.; Wang Y. J.; Sankar R.; Chou F.; Bansil A.; Hasan M. Z.; Wilson S. D.; Fu L.; Madhavan V. Observation of Dirac Node Formation and Mass Acquisition in a Topological Crystalline Insulator. Science 2013, 341 (6153), 1496–1499. 10.1126/science.1239451. [DOI] [PubMed] [Google Scholar]
  16. Wang Z.; Wang J.; Zang Y.; Zhang Q.; Shi J.-A.; Jiang T.; Gong Y.; Song C.-L.; Ji S.-H.; Wang L.-L.; Gu L.; He K.; Duan W.; Ma X.; Chen X.; Xue Q.-K. Molecular Beam Epitaxy-Grown SnSe in the Rock-Salt Structure: An Artificial Topological Crystalline Insulator Material. Adv. Mater. 2015, 27 (28), 4150–4154. 10.1002/adma.201501676. [DOI] [PubMed] [Google Scholar]
  17. Chang K.; Küster F.; Miller B. J.; Ji J.-R.; Zhang J.-L.; Sessi P.; Barraza-Lopez S.; Parkin S. S. P. Microscopic Manipulation of Ferroelectric Domains in SnSe Monolayers at Room Temperature. Nano Lett. 2020, 20 (9), 6590–6597. 10.1021/acs.nanolett.0c02357. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Du R.; Wang Y.; Cheng M.; Wang P.; Li H.; Feng W.; Song L.; Shi J.; He J. Two-Dimensional Multiferroic Material of Metallic p-Doped SnSe. Nat. Commun. 2022, 13 (1), 6130. 10.1038/s41467-022-33917-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Orlova N. N.; Timonina A. V.; Kolesnikov N. N.; Deviatov E. V. Switching Ferroelectricity in SnSe across Phase Transition. Europhys. Lett. 2021, 135 (3), 37002. 10.1209/0295-5075/ac2247. [DOI] [Google Scholar]
  20. Chiu M.-H.; Ji X.; Zhang T.; Mao N.; Luo Y.; Shi C.; Zheng X.; Liu H.; Han Y.; Wilson W. L.; Luo Z.; Tung V.; Kong J. Growth of Large-Sized 2D Ultrathin SnSe Crystals with In-Plane Ferroelectricity. Adv. Electron. Mater. 2023, 9 (4), 2201031. 10.1002/aelm.202201031. [DOI] [Google Scholar]
  21. Dewandre A.; Hellman O.; Bhattacharya S.; Romero A. H.; Madsen G. K. H.; Verstraete M. J. Two-Step Phase Transition in SnSe and the Origins of its High Power Factor from First Principles. Phys. Rev. Lett. 2016, 117 (27), 276601. 10.1103/PhysRevLett.117.276601. [DOI] [PubMed] [Google Scholar]
  22. Huang Y.; Wang C.; Chen X.; Zhou D.; Du J.; Wang S.; Ning L. First-Principles Study on Intrinsic Defects of SnSe. RSC Adv. 2017, 7 (44), 27612–27618. 10.1039/C7RA03367B. [DOI] [Google Scholar]
  23. Zhou Y.; Li W.; Wu M.; Zhao L.-D.; He J.; Wei S.-H.; Huang L. Influence of Defects on the Thermoelectricity in SnSe: A Comprehensive Theoretical Study. Phys. Rev. B 2018, 97 (24), 245202. 10.1103/PhysRevB.97.245202. [DOI] [Google Scholar]
  24. Kim S.; Duong A.-T.; Cho S.; Rhim S. H.; Kim J. A Microscopic Study Investigating the Structure of SnSe Surfaces. Surf. Sci. 2016, 651, 5–9. 10.1016/j.susc.2016.03.013. [DOI] [Google Scholar]
  25. Duvjir G.; Min T.; Thi Ly T.; Kim T.; Duong A.-T.; Cho S.; Rhim S. H.; Lee J.; Kim J. Origin of p-Type Characteristics in a SnSe Single Crystal. Appl. Phys. Lett. 2017, 110 (26), 262106. 10.1063/1.4991003. [DOI] [Google Scholar]
  26. Sirikumara H. I.; Morshed M.; Jameson C.; Jayasekera T. Dopant-Induced Indirect-Direct Transition and Semiconductor-Semimetal Transition of Bilayer SnSe. J. Appl. Phys. 2019, 126, 224301. 10.1063/1.5128337. [DOI] [Google Scholar]
  27. Zhou T.; Du J.; Wang C.; Huang Y. Chemical Doping of the SnSe Monolayer: A First-Principle Calculation. Phys. Chem. Chem. Phys. 2019, 21 (27), 14629–14637. 10.1039/C9CP02414J. [DOI] [PubMed] [Google Scholar]
  28. Asakawa K.; Oguro F.; Yoshida Y.; Sakai H.; Hanasaki N.; Hasegawa Y. Defect-Induced Electronic Structures on SnSe Surfaces. Jpn. J. Appl. Phys. 2019, 58 (SI), SIIA06. 10.7567/1347-4065/ab147d. [DOI] [Google Scholar]
  29. Nguyen V. Q.; Trinh T. L.; Chang C.; Zhao L.-D.; Nguyen T. H.; Duong V. T.; Duong A. T.; Park J. H.; Park S.; Kim J.; Cho S. Unidentified Major p-Type Source in SnSe: Multivacancies. NPG Asia Mater. 2022, 14 (1), 42–47. 10.1038/s41427-022-00393-5. [DOI] [Google Scholar]
  30. Wu D.; Wu L.; He D.; Zhao L.-D.; Li W.; Wu M.; Jin M.; Xu J.; Jiang J.; Huang L.; Zhu Y.; Kanatzidis M. G.; He J. Direct Observation of Vast Off-Stoichiometric Defects in Single Crystalline SnSe. Nano Energy 2017, 35, 321–330. 10.1016/j.nanoen.2017.04.004. [DOI] [Google Scholar]
  31. Wei W.; Chang C.; Yang T.; Liu J.; Tang H.; Zhang J.; Li Y.; Xu F.; Zhang Z.; Li J.-F.; Tang G. Achieving High Thermoelectric Figure of Merit in Polycrystalline SnSe via Introducing Sn Vacancies. J. Am. Chem. Soc. 2018, 140 (1), 499–505. 10.1021/jacs.7b11875. [DOI] [PubMed] [Google Scholar]
  32. Qin B.; Zhang Y.; Wang D.; Zhao Q.; Gu B.; Wu H.; Zhang H.; Ye B.; Pennycook S. J.; Zhao L.-D. Ultrahigh Average ZT Realized in p-Type SnSe Crystalline Thermoelectrics through Producing Extrinsic Vacancies. J. Am. Chem. Soc. 2020, 142 (12), 5901–5909. 10.1021/jacs.0c01726. [DOI] [PubMed] [Google Scholar]
  33. Sraitrova K.; Cizek J.; Holy V.; Plechacek T.; Benes L.; Jarosova M.; Kucek V.; Drasar C. Vacancies in SnSe Single Crystals in a Near-Equilibrium State. Phys. Rev. B 2019, 99 (3), 035306. 10.1103/PhysRevB.99.035306. [DOI] [Google Scholar]
  34. Neugebauer J.; Zywietz T. K.; Scheffler M.; Northrup J. E.; Chen H.; Feenstra R. M. Adatom Kinetics on and below the Surface: the Existence of a New Diffusion Channel. Phys. Rev. Lett. 2003, 90 (5), 056101. 10.1103/PhysRevLett.90.056101. [DOI] [PubMed] [Google Scholar]
  35. Vogt P.; Hensling F. V. E.; Azizie K.; McCandless J. P.; Park J.; DeLello K.; Muller D. A.; Xing H. G.; Jena D.; Schlom D. G. Extending the Kinetic and Thermodynamic Limits of Molecular-Beam Epitaxy Utilizing Suboxide Sources or Metal-Oxide-Catalyzed Epitaxy. Phys. Rev. Appl. 2022, 17 (3), 034021. 10.1103/PhysRevApplied.17.034021. [DOI] [Google Scholar]
  36. Heckingbottom R.; Todd C. J.; Davies G. J. The Interplay of Thermodynamics and Kinetics in Molecular Beam Epitaxy (MBE) of Doped Gallium Arsenide. J. Electrochem. Soc. 1980, 127 (2), 444–450. 10.1149/1.2129685. [DOI] [Google Scholar]
  37. Adkison K. M.; Shang S.-L.; Bocklund B. J.; Klimm D.; Schlom D. G.; Liu Z.-K. Suitability of Binary Oxides for Molecular-Beam Epitaxy Source Materials: A Comprehensive Thermodynamic Analysis. APL Mater. 2020, 8 (8), 081110. 10.1063/5.0013159. [DOI] [Google Scholar]
  38. Liszkay L.; Corbel C.; Baroux L.; Hautojärvi P.; Bayhan M.; Brinkman A. W.; Tatarenko S. Positron trapping at divacancies in thin polycrystalline CdTe films deposited on glass. Appl. Phys. Lett. 1994, 64 (11), 1380–1382. 10.1063/1.111994. [DOI] [Google Scholar]
  39. Lin C. L.; Lu S. M.; Su W. B.; Shih H. T.; Wu B. F.; Yao Y. D.; Chang C. S.; Tsong T. T. Manifestation of Work Function Difference in High Order Gundlach Oscillation. Phys. Rev. Lett. 2007, 99 (21), 216103. 10.1103/PhysRevLett.99.216103. [DOI] [PubMed] [Google Scholar]
  40. Barrios-Salgado E.; Nair M. T. S.; Nair P. K. Thin Films of n-Type SnSe2 Produced from Chemically Deposited p-Type SnSe. Thin Solid Films 2016, 598, 149–155. 10.1016/j.tsf.2015.11.075. [DOI] [Google Scholar]
  41. Tian Z.; Zhao M.; Xue X.; Xia W.; Guo C.; Guo Y.; Feng Y.; Xue J. Lateral Heterostructures Formed by Thermally Converting n-Type SnSe2 to p-Type SnSe. ACS Appl. Mater. Interfaces 2018, 10 (15), 12831–12838. 10.1021/acsami.8b01235. [DOI] [PubMed] [Google Scholar]
  42. Kresse G.; Furthmüller J. Efficiency of Ab-Initio Total Energy Calculations for Metals and Semiconductors Using a Plane-Wave Basis Set. Comput. Mater. Sci. 1996, 6 (1), 15–50. 10.1016/0927-0256(96)00008-0. [DOI] [Google Scholar]
  43. Kresse G.; Joubert D. From Ultrasoft Pseudopotentials to the Projector Augmented-Wave Method. Phys. Rev. B 1999, 59 (3), 1758–1775. 10.1103/PhysRevB.59.1758. [DOI] [Google Scholar]
  44. Blöchl P. E. Projector Augmented-Wave Method. Phys. Rev. B 1994, 50 (24), 17953–17979. 10.1103/PhysRevB.50.17953. [DOI] [PubMed] [Google Scholar]
  45. Perdew J. P.; Burke K.; Ernzerhof M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77 (18), 3865–3868. 10.1103/PhysRevLett.77.3865. [DOI] [PubMed] [Google Scholar]
  46. Enkovaara J.; Rostgaard C.; Mortensen J. J.; Chen J.; Dułak M.; Ferrighi L.; Gavnholt J.; Glinsvad C.; Haikola V.; Hansen H. A.; Kristoffersen H. H.; Kuisma M.; Larsen A. H.; Lehtovaara L.; Ljungberg M.; Lopez-Acevedo O.; Moses P. G.; Ojanen J.; Olsen T.; Petzold V.; Romero N. A.; Stausholm-Møller J.; Strange M.; Tritsaris G. A.; Vanin M.; Walter M.; Hammer B.; Häkkinen H.; Madsen G. K. H.; Nieminen R. M.; Nørskov J. K.; Puska M.; Rantala T. T.; Schiøtz J.; Thygesen K. S.; Jacobsen K. W. Electronic Structure Calculations with GPAW: A Real-Space Implementation of the Projector Augmented-Wave Method. J. Phys.: Condens. Matter 2010, 22 (25), 253202. 10.1088/0953-8984/22/25/253202. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

nn4c04789_si_001.pdf (1.8MB, pdf)

Articles from ACS Nano are provided here courtesy of American Chemical Society

RESOURCES