Self-Limiting Layer-by-Layer Oxidation of Atomically Thin WSe2

Subscriber access provided by Columbia Univ Libraries

Communication 2

Self-Limiting Layer-by-Layer Oxidation of Atomically Thin WSe Mahito Yamamoto, Sudipta Dutta, Shinya Aikawa, Shu Nakaharai, Katsunori Wakabayashi, Michael S Fuhrer, Keiji Ueno, and Kazuhito Tsukagoshi Nano Lett., Just Accepted Manuscript • Publication Date (Web): 03 Feb 2015 Downloaded from http://pubs.acs.org on February 5, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Nano Letters is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Self-Limiting Layer-by-Layer Oxidation of Atomically Thin WSe2 Mahito Yamamoto,*, † Sudipta Dutta,†, ‡ Shinya Aikawa,†, § Shu Nakaharai,† Katsunori Wakabayashi,† Michael S. Fuhrerǁ, ┴ Keiji Ueno,# and Kazuhito Tsukagoshi*, † †

WPI Center for Materials Nanoarchitechtonics (WPI-MANA), National Institute for Materials

Science, Tsukuba, Ibaraki 305-0044, Japan, ‡International Center for Young Scientists, National Institute for Materials Science, Tsukuba, Ibaraki 305-0044, Japan, §Research Institute for Science and Technology, Kogakuin University, Tokyo 192-0015, Japan, ǁCenter for Nanophysics and Advanced Materials, University of Maryland, College Park, Maryland 20742-4111, United States, ┴School of Physics, Monash University, Melbourne VIC 3800, Australia, #Department of Chemistry, Graduate School of Science and Engineering, Saitama University, Saitama 338-8570, Japan, KEYWORDS: Layered transition metal dichalcogenides, tungsten diselenide, oxidation, Raman spectroscopy, photoluminescence, x-ray photoelectron spectroscopy, ab-initio calculations

ABSTRACT: Growth of a uniform oxide film with a tunable thickness on two-dimensional transition metal dichalcogenides is of great importance for the electronic and optoelectronic applications. Here we demonstrate homogeneous surface oxidation of atomically thin WSe2 with a self-limiting thickness from single- to tri-layers. Exposure to ozone (O3) below 100 ˚C leads to

ACS Paragon Plus Environment

1

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 24

the lateral growth of tungsten oxide selectively along selenium zigzag-edge orientations on WSe2. With further O3 exposure, the oxide regions coalesce and oxidation terminates leaving a uniform thickness oxide film on top of unoxidized WSe2. At higher temperatures, oxidation evolves in the layer-by-layer regime up to trilayers. The oxide films formed on WSe2 are nearly atomically flat. Using photoluminescence and Raman spectroscopy, we find that the underlying single-layer WSe2 is decoupled from the top oxide but hole-doped. Our findings offer a new strategy for creating atomically thin heterostructures of semiconductors and insulating oxides with potential for applications in electronic devices.


2

Page 3 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Oxide-semiconductor heterostructures are at the heart of electronics and optoelectronics such as field effect transistors (FETs); growing a high-quality oxide film on a semiconductor is, hence, critical for the development of the electronic devices. The simplest approach to growing an oxide on a semiconducting material is thermal oxidation of the surface. On a silicon surface, for example, thermal treatment in oxygen (O2) gas or water vapor can lead readily to the growth of a uniform oxide film with a controllable thickness.1 This simple oxide fabrication process, together with superior properties of the oxide, was a major factor in the advancement in technologies based on silicon over other semiconductors such as gallium arsenide and germanium, where thermal treatments result in inhomogeneous oxidation of the surfaces.2, 3 Layered group VI transition metal dichalcogenides with a trigonal prismatic geometry are semiconductors with band gaps ranging from 1 to 2 eV, known for decades.4 In recent years, however, atomically thin layers of semiconducting dichalcogenides have attracted renewed interest for electronic and optoelectronic applications, owing to their sizable band gaps and direct band gap properties at the single-layer limit.5, 6 Atomically thin semiconducting dichalcogenides have been applied for a variety of devices such as FETs,7- 9 photodetectors,10, 11 and lightemitting diodes,12-14 promising performances comparable to silicon devices. However, in contrast to silicon, the pristine surface of a layered dichalcogenide is resistant to oxidation because of the absence of dangling bonds, whereas the edges and the surface defective sites show larger oxidative reactivity.15-18 Therefore, growth of thermal oxides that could be used for devices has yet to be demonstrated on atomically thin dichalcogenides. Alternatively, dichalcogenide-based devices have been fabricated mostly with substrates of silicon dioxide (SiO2) or atomic-layerdeposited (ALD) oxide films as insulating and dielectric environments, in common with


3

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 24

graphene. Controllable and uniform surface oxidation of atomically thin dichalcogenides is a challenging endeavor for more realistic applications. Here we demonstrate that ozone (O3) exposure leads to homogenous surface oxidation of atomically thin WSe2 with a self-limiting thickness from single- to tri-layers, depending on temperature. Below 100 ˚C, the O3 treatment of WSe2 results in the formation of tungsten oxide with a thickness of about 2 nm from edges and defects on the surface. With further exposure to O3, the oxides grow laterally along selenium zigzag edge (1¯ 010) orientations, until the top layer is fully oxidized. However, no underlying layers are oxidized at those temperatures. With increasing temperature, oxidation progresses layer-by-layer up to tri-layers. The oxide films formed on WSe2 are nearly atomically flat. Single-layer WSe2 produced by surface oxidation of a few-layer counterpart shows well-defined Raman and photoluminescence peaks consistent with single-layer thickness, but with shifts in peak positions due to hole-doping. Our findings are a first step toward controllable growth of oxides on layered transition metal dichalcogenides for applications ranging from tunnel barriers to gate dielectrics.


4

Page 5 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Figure 1. AFM images of thin WSe2 flakes on SiO2 after O3 exposure (a) at RT for 1.5 hours, (b) at RT for 1 hour, (c) at 50 ˚C for 1 hour, and at 70 ˚C for (d) 0.5, (e) 1, and (f) 1.5 hours. The number of layers (NL with N = 1 to 6) of each flake is indicated. The scale bars are 2 µm. Atomically thin films of WSe2 were mechanically exfoliated from bulk crystals onto SiO2. The samples were exposed to O3 at various temperatures (see Methods). Figure 1a is an atomic force microscope (AFM) image of single- and bi-layer WSe2 after O3 exposure at room temperature (RT) for 1.5 hours. The O3 treatment results in the formation of serrated oxides along the edges and dendritic islands on the surface, whose thicknesses are about 2 nm higher than the surface of the WSe2 layers (Fig. S1 in Supporting Information). Previous studies have reported local growth of similar triangular islands on WSe2 by either irradiating high-power laser19 or applying high voltages20 to the surfaces in air. The orientations of the islands are identical on each layer surface, but are 180˚-inverted between the surfaces of single- and bi-layer WSe2, indicating that WSe2 is oxidized preferentially along either selenium zigzag edge (1¯ 010) or tungsten zigzag edge (101¯ 0) orientations, as observed in MoS2 (see Fig. S2 in Supporting Information for the crystal structure of 2H-WSe2).21, 22 To determine the orientations of the oxidized edges, we examine the structural stability of WSe2 nanoribbons with either selenium- or tungsten-zigzag edges oxidized, using ab-initio calculations (see Methods). We find that the stability is reduced when tungsten atoms at the zigzag edge are substituted with oxygen atoms, while the nanoribbon structure becomes more stable when the selenium edge is oxidized (Fig. S3 in Supporting Information). These results suggest that WSe2 is oxidized preferentially along selenium zigzag edge (1¯ 010) orientations. The density of oxide islands varies from sample to sample, with no clear dependence on the number of layers of WSe2, exposure time, and temperature. Figure 1b is an AFM image of


5

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 24

tri- and four-layer WSe2 oxidized at RT for 1 hour, where quasi-triangular islands are formed over the surface with the density more than twenty times higher than that in Fig. 1a. In contrast, Fig. 1c shows a low density of triangular islands on the surface of four- to six-layer WSe2 after the O3 treatment at 50 ˚C for 1 hour. As shown in Fig. 1d, the O3 treatment at 70 ˚C for 0.5 hours leads to the formation of triangular islands, whose density is comparable to that in Fig. 1a. Additional 0.5 hours exposure results in the lateral growth of the triangular islands on each layer surface, but the number of islands rarely increases (Fig. 1e). With further exposure, the oxides coalesce to each other and the surface of WSe2 is almost fully oxidized in 1.5 hours (Fig. 1f). In Fig. S4 in Supporting Information, we show additional AFM images of oxidized WSe2 samples, where the density of oxide islands shows some sample-to-sample variation, independent of oxidation temperature and exposure time. The observations suggest that the oxide islands nucleate mostly from preexisting defects on WSe2, which are most likely selenium vacancies, rather than randomly on the surface, as observed in MoS2.21

Figure 2. AFM images of bilayer (2L), trilayer (3L), and four-layer (4L) WSe2 on SiO2 (a) before and (b) after the O3 treatment at 100 ˚C for 1 hour. The insets are optical images of the WSe2 flake. The regions surrounded by the black dashed lines correspond to the AFM scanned


6

Page 7 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

areas. The scale bars are 2 µm. (c) Height profiles of pristine (black line) and oxidized (red line) WSe2 along the white dashed lines in (a and b). Figure 2a and b shows AFM images of bi- to four-layer WSe2 before and after O3 exposure at 100 ˚C for 1 hour. The O3 treatment results in the formation of a homogeneous oxide film with a higher thickness than the pristine WSe2 layers by about 1.8 nm (Fig. 2c), which is consistent with the thicknesses of triangular oxide islands on WSe2 (Fig. S1 in Supporting Information). Remarkably, the oxide is nearly atomically flat with rms roughness of oxidized bilayer WSe2 of ~ 0.24 nm, which is only slightly larger than that of pristine bilayer WSe2 on SiO2 (~ 0.14 nm; see Fig. S6 in Supporting Information). As shown in the optical images, the optical color contrast between layers changes upon oxidation. Since the contrast between oxidized single-layer WSe2 and SiO2 are almost equivalent (Fig. S7 in Supporting Information), the optical contrast could be used to identify the number of layers of unoxidized WSe2 on SiO2 after the O3 treatment.23 By comparing the contrast of atomically thin WSe2 before and after oxidation, we estimate that only the top single-layer is oxidized by the O3 treatment at 100 ˚C for 1 hour (see Fig. S8 in Supporting Information for the detailed optical contrast analysis). We find that further exposure to O3 at 100 ˚C does not lead to a change in the optical contrast of WSe2 with the oxide, implying that O3 oxidation of WSe2 is self-limited at the single-layer thickness (see Fig. S9 in Supporting Information for optical images of few-layer WSe2 before and after oxidation at 100 ˚C for 1 hour and 2 hours). The self-limiting oxidation is likely because the top oxide film blocks vertical diffusion of oxygen into the underlying WSe2 layers.24, 25 However, oxidation progresses to the second and third layers with increasing temperature, as discussed below.


7

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 24

Figure 3. (a) Raman spectra of single-layer (1L) and bilayer (2L) WSe2 before (black lines) and after (red lines) the O3 treatment at 100 ˚C for 1 hour. The Raman peak intensities of pristine and oxidized 2L WSe2 are normalized with the E12g peak intensity of pristine 1L WSe2. (b) PL spectra of 1L and 2L WSe2 before (black lines) and after (red lines) the O3 treatment at 100 ˚C for 1 hour. The PL intensities of pristine and oxidized 2L WSe2 are normalized with the A peak intensity of pristine 1L WSe2. The inset shows normalized PL intensities of the A peaks of 1L and 2L WSe2 without (black squares) and with (red circle) the top oxide. We use Raman spectroscopy to further confirm the number of WSe2 layers oxidized by O3 exposure and to characterize the vibrational properties of atomically thin WSe2 with the top oxide. Figure 3a shows Raman spectra of single- and bi-layer WSe2 after the O3 treatment at 100


8

Page 9 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

˚C for 1 hour. After oxidation of single-layer WSe2, a prominent peak of the E12g or A1g mode at 250.5 cm-1 disappears (these modes are almost degenerate for atomically thin WSe2 26-28 and indistinguishable experimentally in our measurement, but we denote the mode by E12g in the following for simplicity), indicating no residual crystalline WSe2 structures. A previous study has shown that high-power laser irradiation to single-layer WSe2 in air results in the formation of WO3 with Raman peaks at 302 and 822 cm-1.19 We observe no Raman peaks of WO3 in oxidized single-layer WSe2 even after additional annealing at 300 ˚C in air (Fig. S10 in Supporting Information), which may imply that the oxide film is amorphous, or that the Raman intensity is simply too small to be detected. Atomically thin layers of 2H-WSe2 exhibit a Raman active B12g mode that is inactive for bulk.26, 27, 29 The B12g peak is also absent in single-layer WSe2 and, thus, can be a hallmark to distinguish between single- and few-layers. Figure 3a shows that the B12g peak of bilayer WSe2 at 310 cm-1 disappears after the O3 treatment at 100 ˚C for 1 hour, while the E12g peak remains with a frequency downshifting from 252 to 251 cm-1. This strongly suggests only the top layer of bilayer WSe2 is oxidized (consistent with the change in the optical contrast as discussed above), and the remaining single-layer WSe2 is mechanically decoupled from the overlying oxide. Raman frequencies and relative intensities of the E12g and B12g modes also suggest that atomically thin WSe2 is oxidized only at the top layers (Fig. S11 in Supporting Information). The E12g peak of single-layer WSe2 with the top oxide blueshifts by 0.5 cm-1 from that of pristine single-layer WSe2, likely due to doping30 or strain31 introduced in WSe2 through the O3 treatment. A previous study has reported that electron-doping of single-layer WSe2 leads to a redshift in the E12g peak.30 Alternatively, tensile strain has been observed to split the E12g mode, with the higher (lower) frequency peak blueshifting (redshifting) with increasing the magnitude


9

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 24

of strain.31 Since no splitting of the E12g peak is observed, we conclude that the blueshift in single-layer WSe2 with the top oxide is due to hole-doping. The hole-doping is likely caused by local oxidation of the underlying single-layer WSe221 and/or electron transfer from WSe2 to the top oxide that is tungsten oxide WOx with x ≤ 3 (see Fig. 4), known to be a strong hole-dopant for organic semiconductors.32 We investigate the electronic structure of atomically thin WSe2 with the top oxide using photoluminescence (PL) spectroscopy. In Fig. 3b, we show PL spectra of single- and bi-layer WSe2 before and after the O3 treatment at 100 ˚C for 1 hour. Pristine single-layer WSe2 shows a prominent PL peak of a direct band gap emission at 1.65 eV (A peak).26, 33, 34 The A peak disappears after the O3 treatment, indicating WSe2 is fully oxidized. Pristine bilayer WSe2 shows two PL peaks from the A emission at 1.61 eV and an indirect gap emission at 1.55 eV (I peak),33, 34

whose intensities are one order of magnitude smaller than that of the A peak of single-layer

WSe2 (inset of Fig. 3b). After the O3 treatment of bilayer WSe2, only a single peak from the A emission is observed, again indicating that the top layer is oxidized and the underlying singlelayer remains, as confirmed with Raman spectroscopy. However, the PL energy of single-layer WSe2 with the top oxide redshifts from that of pristine single-layer WSe2 by 40 meV (Fig. S12 in Supporting Information). The full width at half maximum (fwhm) of the PL peak increases from 49 meV for pristine single-layer WSe2 to 77 meV for single-layer WSe2 with the oxide. Furthermore, the PL intensity of single-layer WSe2 with the oxide is more than one order of magnitude smaller than that of pristine single-layer WSe2 (inset of Fig. 3b). Previous studies have shown that the signatures of PL of single-layer dichalcogenides are largely modified with surrounding environments such as deposited films35, 36 and substrates,37, 38 with dielectric screening of Coulomb potentials39, 40 and carrier doping40-43 in the crystals as the causes. With a


10

Page 11 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

higher dielectric environment, Coulomb potentials in a single-layer dichalcogenide are screened, leading to a blueshift in the PL energy, an enhancement of the intensity, and a narrowing of the peak,39 which are opposite trends to our observations. Alternatively, carrier doping changes the contribution of the PL emission from neutral excitons to negatively or positively charged trions that show lower PL energies and intensities (or vice versa), depending on the carrier types of the crystal and the dopant. For example, upon hole-doping of single-layer MoS2, which is commonly n-type, the PL shows a blueshift in the energy and an enhancement of the intensity because neutral excitons become more contributory to the PL emission than negatively charged trions.40, 41, 43

In contrast, electron-doping of single-layer MoS2 leads to a redshift in the PL energy, a

reduction of the intensity, and a broadening of the peak.41, 43 Similarly, the PL intensity of p-type single-layer WSe2 has been observed to be significantly suppressed upon hole-doping.40 Raman spectroscopy measurements imply that single-layer WSe2 with the top oxide is hole-doped. Hence, although we cannot identify the initial carrier type of our single-layer WSe2, we suspect that the observed changes in the PL of single-layer WSe2 with the top oxide are due to holedoping which induces the PL emission predominantly from positively charged trions rather than neutral excitons. Oxidative defects produced in underlying single-layer WSe2 could be another explanation for the reduction of the PL intensity, as observed in single-layer MoS2 upon O2 plasma treatments.44 However, we rule out this possibility because, in contrast to the previous study, the linewidth of the Raman E12g peak of single-layer WSe2 with the oxide is comparable to that of pristine single-layer WSe2 (Fig. 3a), implying that underlying WSe2 is hardly disordered by O3 exposure. Further work using a field-effect geometry may be needed to fully understand the effect of doping and oxide overlayers on the PL of single-layer WSe2.


11

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 24

Figure 4. XPS spectra of a W 4f core level of bulk WSe2 on Si3N4 before and after the O3 treatment at 100 ˚C for 2 hours and after argon ion cluster etching. The black solid lines are experimental data. The dashed lines colored in orange, green and pink are Lorentzian fits for the peaks of WSe2, WOx, and WSex, respectively. We observe no Raman peaks of crystalline oxides after oxidation of single-layer WSe2. To determine the chemical structure of the oxidation product, we performed x-ray photoelectron spectroscopy (XPS) of WSe2 after O3 exposure at 100 ˚C for 2 hours (see Methods). Figure 4 shows XPS spectra of a W 4f core level of pristine and oxidized bulk WSe2 supported on Si3N4 substrates. The W 4f core level spectrum of pristine WSe2 has a 4f7/2 and 4f5/2 doublet of WSe2 at 32.3 and 34.4 eV,45 while oxidized WSe2 shows weaker doublet peaks at 35.6 and 37.7 eV and stronger peaks at 31.7 and 33.9 eV. The larger binding energy doublet corresponds to 4f7/2 and 4f5/2 lines of either stoichiometric15, 18, 46 or non-stoichiometric46-48 tungsten oxide (WOx with x ≤ 3). The WOx doublet peaks disappears after in-situ argon ion cluster bombardment (see Methods), suggesting that the oxide film on WSe2 is completely etched away. The smaller binding energy doublet can be assigned as W-O bonds of tungsten dioxide (WO2)46, 49 or W-Se bonds of either stoichiometricWSe245 or non-stoichiometric WSex with x

Self-Limiting Layer-by-Layer Oxidation of Atomically Thin WSe2

Recommend Documents