Long-Range Lattice Engineering of MoTe2 by a 2D ... - ACS Publications

May 10, 2017 - Long-Range Lattice Engineering of MoTe2 by a 2D Electride. Sera Kim,. †. Seunghyun Song,. ‡. Jongho Park,. †,‡. Ho Sung Yu,. â€...
0 downloads 0 Views 912KB Size
Subscriber access provided by Eastern Michigan University | Bruce T. Halle Library

Communication

Long-Range Lattice Engineering of MoTe2 by 2D Electride Sera Kim, Seunghyun Song, Jongho Park, Ho Sung Yu, Suyeon Cho, Dohyun Kim, Jaeyoon Baik, Duk-Hyun Choe, Kee Joo Chang, Young Hee Lee, Sung Wng Kim, and Heejun Yang Nano Lett., Just Accepted Manuscript • Publication Date (Web): 10 May 2017 Downloaded from http://pubs.acs.org on May 10, 2017

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 17

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

Long-Range Lattice Engineering of MoTe2 by 2D Electride Sera Kim1†, Seunghyun Song2†, Jongho Park1,2, Ho Sung Yu1,2, Suyeon Cho2, Dohyun Kim1, Jaeyoon Baik3, Duk-Hyun Choe4, K. J. Chang4, Young Hee Lee1,2, Sung Wng Kim1* and Heejun Yang1* 1

Department of Energy Science, Sungkyunkwan University, Suwon 16419, Korea.

2

Center for Integrated Nanostructure Physics, Institute for Basic Science (IBS), Suwon 16419, Korea.

3

Pohang Accelerator Laboratory, Pohang University of Science and Technology, Pohang 790784, Korea.

4

Department of Physics, Korea Advanced Institute of Science and Technology, Daejeon 34141, Korea. †

These authors contributed equally to this work.

*Correspondence to: [email protected], [email protected] Doping two-dimensional (2D) semiconductors beyond their degenerate levels provides the opportunity to investigate extreme carrier density-driven superconductivity and phase transition in 2D systems. Chemical functionalization and the ionic gating have achieved the high doping density, but their effective ranges have been limited to ~1 nm, which restricts the use of highly doped 2D semiconductors. Here, we report on electron diffusion from the 2D electride [Ca2N]+⋅e− to MoTe2 over a distance of 100 nm from the contact interface, generating an electron doping density higher than 1.6×1014 cm-2 and a lattice symmetry change of MoTe2 as a consequence of the extreme doping. The longrange lattice symmetry change, suggesting a length scale surpassing the depletion width of conventional metal–semiconductor junctions, was a consequence of the low work function (2.6 eV) with highly mobile anionic electron layers of [Ca2N]+⋅e−. The combination of 2D electrides and layered materials yields a novel material design in terms of doping and lattice engineering. Keywords: MoTe2, electride, doping, phase transition, electron diffusion, work function

ACS Paragon Plus Environment

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

In two-dimensional (2D) layered materials, the carrier density can influence lattice symmetry1, superconductivity2,3, bandgap4 and electric conductivity properties. In particular, achieving high doping levels over conventional degenerate levels (~2.0×1013 cm-2) in a nondestructive manner has been reported as a critical prerequisite for optimal 2D device applications5-8. Although alkali metal doping9, organic molecular functionalization10, and Liintercalation11,12 have been introduced to induce high charge carrier concentrations, a carrier concentration of ~1014 cm-2 or ~1021 cm-3 is difficult to achieve through the abovementioned methods. While the ionic gating can induce a total charge density of ~1014 cm-2, in the case of high doping concentration ~1021 cm-3, the effective range of the doping is limited to ~1 nm from the top surface since the charge mirroring induces doping in the material. Clean and nondestructive means of high-level doping of 2D layered materials is challenging and has substantial potential in device applications and fundamental studies such as superconductivity and phase transition with the 2D materials. When two solid-state materials are in contact, the equilibrium state reflects strain effects from a lattice mismatch and balancing electron chemical potentials between contacting materials. The lattice strain effects have been investigated by transmission electron microscopy (TEM), showing a limited effective range of one or two atomic layers from the interface13. By contrast, the alignment of the electron chemical potentials, 'Fermi level aligning', demonstrates a charge transfer with a certain carrier density and screening length, which are influenced by the work function difference and carrier concentrations of contacting materials14-16. However, in the case of degenerate doping levels (1014 cm-2 or 1021 cm-3), the depletion width from the contact, scaled with the Debye length or screening length, remains at ~1 nm, which creates difficulties in the investigation and use of extremely doped 2D layered materials by contact. A group 6 transition metal dichalcogenide (TMD), MoTe2, has shown unique phase transition characteristics based on the small energy difference between its hexagonal (2H, semiconducting) and monoclinic (1T', metallic) phases1,17 and has been applied in lowcontact-resistance transistors6. It has been theoretically predicted that a structural phase transition from the hexagonal to the monoclinic phase can be triggered by doping the 2HMoTe2 with a carrier concentration higher than 1.6×1014 cm-2 or 2.2×1021 cm-3 (ref. 1). While ionic gating can theoretically induce enough charges to trigger the phase transition, electrostatic gating induced phase transition in MoTe2 has not been reported to the best of our

ACS Paragon Plus Environment

Page 2 of 17

Page 3 of 17

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

knowledge. We note that we could not realize the phase transition in MoTe2 using electrostatic gating methods, including ionic gating (where the chemical window might have limited the induced carrier density). Here we propose a different doping technology using the 2D nature of TMDs: placing MoTe2 in contact with recently reported single-crystalline 2D [Ca2N]+⋅e− electride that exhibits a low work function (~2.6 eV) with highly mobile anionic electron layers. This method, based on contact of two crystals, has benefits compared with previous studies such as alkali metalintercalation into TMDs9,18 that suffered an external impurity-induced strain effect19 and random and incomplete spatial distribution of the phase change5. The MoTe2/[Ca2N]+⋅e− contact interface could provide an opportunity to explore charge transfer, as opposed to the charge mirroring of gating, phenomena affected by the large work function difference and highly concentrated electron-driven phase transition of MoTe2. Here we report the long-range (up to ~100 nm) structural phase engineering of MoTe2 by the contact between MoTe2 and a 2D electride, [Ca2N]+⋅e−. The charge transfer between the two materials induced the lattice symmetry change from hexagonal to monoclinic phase in MoTe2, indicating the electron doping of MoTe2 with a density higher than 1.6×1014 cm-2 (ref. 1, 5) over 100 nm from the contact interface; this might be considered as violating the screening theory that offers the screening length of ~1 nm for an electron density of ~1014 cm-2 or 1021 cm-3. We used optical and X-ray photoemission spectroscopy and scanning probe microscopy techniques to explore the exceptional electron transfer or electric potential screening phenomenon near the junction. We synthesized 2H-MoTe2 and [Ca2N]+⋅e− single crystals using previously reported methods1,20. We first exfoliated the Ca2N using a “scotch tape” method onto 300 nm SiO2 substrate. Then we performed the second exfoliation of 2H-MoTe2 onto the same substrate where Ca2N had been previously exfoliated. We identified, from the subsequent exfoliations, the Ca2N/MoTe2 heterostructure with optical microscope. All sample preparation steps with as-grown single crystals of 2H-MoTe2 and [Ca2N]+⋅e− were carried out in an argon (Ar)-filled glove box to prevent the oxidation of [Ca2N]+⋅e− because its low work function makes the material susceptible to oxidation under ambient conditions20,21. Fig. 1a shows the distinctly different optical contrast and Raman signatures of an oxidized [Ca2N]+⋅e− from pristine [Ca2N]+⋅e−. A gold align-marker underneath the oxidized [Ca2N]+⋅e− becomes visible as the

ACS Paragon Plus Environment

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

[Ca2N]+⋅e− turns transparent after oxidation. After obtaining the [Ca2N]+⋅e− oxidation reference data, all characterizations (confocal Raman spectroscopy, scanning probe microscopy, and scanning photoemission microscopy) were conducted without exposing the samples to air. A schematic of our contact-driven lattice symmetry change is shown in Fig. 1b. The electron anionic layers in [Ca2N]+⋅e−, the origin of the low work function of the material20, is substantially modified by the physical contact. This generates unusually high electron transfer into MoTe2 and drives a structural phase transition from the hexagonal to the monoclinic phase of MoTe2. Thus, the electron transfer between [Ca2N]+⋅e− and MoTe2 single crystals allows a unique opportunity for rigorous study on contact between two materials with a large work function difference; the electron-density-dependent structural phase transition of MoTe2, with a threshold electron density higher than 1.6×1014 cm-2 (ref. 1), enables us to spatially probe the extent of charge transfer inside MoTe2. Following the 2H-MoTe2/[Ca2N]+⋅e− heterostructure preparation, Raman spectra were taken at different positions of the sample with a 532 nm excitation laser (Fig. 2). The Fig. 2a shows an optical image of the sample with white dashed lines outlining the underlying [Ca2N]+⋅e−. The red curve in Fig. 2b, taken at red spot in Fig. 2a, shows Raman spectrum distinctively different from that of 2H-MoTe2. Rather, the Raman spectrum taken at the MoTe2 sitting on top of [Ca2N]+⋅e− corresponded to that of monoclinic MoTe2 with a signature of modified [Ca2N]+⋅e− by the contact. The reference Raman spectra of monoclinic MoTe2 were previously studied with TEM and X-ray diffraction (XRD) to verify the crystal symmetry of MoTe2 (ref. 6, 22). We note that the Raman spectra of monoclinic MoTe2 (Fig. 2b.) exhibits different relative intensities and larger FWHM compared to the reference taken from a single crystal. We tentatively attribute this difference to the small domain sizes of monoclinic MoTe2; that the phase transitioned monoclinic is a polycrystalline. In the presence of defects or grain boundaries, phonon confinement effect can modify the Raman spectra23 (the position, FWHM, and relative intensity), where a similar effects are shown in MoS2 because of the relaxation of back scattering geometry24, 25 (Please refer to SI5 for a more detailed discussion regarding the phonon confinement effect of monoclinic MoTe2). Although the sample was not exposed to the ambient environment, the [Ca2N]+⋅e− in contact with MoTe2 (red curve in Fig. 2b) exhibited a different Raman spectrum from the pristine [Ca2N]+⋅e− but similar to the

ACS Paragon Plus Environment

Page 4 of 17

Page 5 of 17

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

oxidized [Ca2N]+⋅e− (brown vertical line in Fig. 2b, see also Supplementary Fig.1). This change in the [Ca2N]+⋅e− Raman signature implies that the electron loss in [Ca2N]+⋅e− is significant, akin to the redox of alkali metal such as Li in contact with another material. We attribute both the lattice symmetry change from hexagonal (blue curve in Fig. 2b) to monoclinic phase (red curve in Fig. 2b) of MoTe2 and the oxidation of [Ca2N]+⋅e− to the contact-driven electron transfer from [Ca2N]+⋅e− to MoTe2. Previous studies report charge transfer-driven phase transition in MoS2 through chemical Li-intercalation process11 and plasmonic hot electrons generation26. However, those previous works used irregular structures that cannot be combined with theory for a quantitative charge-transfer study. In addition to the emergence of monoclinic MoTe2, modulation of the Raman signatures of hexagonal MoTe2 is observed from the spectra. For instance, near the MoTe2/[Ca2N]+⋅e− interface (purple spot in Fig. 2a), the red shifts and broadening of 2H peaks were observed as shown in the purple spectrum in Fig. 2b, with little evidence of monoclinic phase. By contrast, the spectrum taken far away from the MoTe2/[Ca2N]+⋅e− interface (blue spot in Fig. 2a) is indistinguishable from the reference 2H-MoTe2 spectrum1,6 (blue curve in Fig. 2b). Since the thickness of the MoTe2 is more than 50 nm, we exclude the layer number dependence of the Raman spectra as the origin of the red shifts and broadening of 2H peaks; we attribute the change in Raman spectra to the different doping concentrations27,28 in hexagonal MoTe2. The red shift accompanied by the peak broadening was consistently observed from different locations varying in doping densities. To quantify the change in the 2H Raman signatures, 2H E2g peaks from more than 20 positions progressively away from MoTe2/[Ca2N]+⋅e− interface were fitted to a Lorentzian distribution to obtain the peak positions and full-width-at-halfmaximum (FWHM), as plotted in Fig. 2c. The larger redshift of the E2g mode is correlated with the greater peak broadening as a result of phonon renormalization under doping. This result is reminiscent of the phonon renormalization observed in MoS2 under electrostatic gating27,28. To quantitatively elucidate the redshift, we performed first-principles calculations on the Raman peak positions of E2g mode of hexagonal MoTe2 under different electron doping levels. The vertical dashed lines of the Fig. 2c indicate the peak location under varying levels of electron doping density ranging from 1013 cm-2 to 1.2×1014 cm-2 (ref. 27, 28). Along with the emergence of monoclinic MoTe2 on top of [Ca2N]+⋅e− by a doping density (> 1.6×1014 cm-2),

ACS Paragon Plus Environment

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

hexagonal MoTe2 areas laterally close to the contact interface also show a considerable electron doping with carrier densities up to ~1014 cm-2. The continuous shift in the Raman peak and electron density of MoTe2 around [Ca2N]+⋅e− supports the large electron transfer from the [Ca2N]+⋅e−. Despite the strong evidence of contact-driven phase transition by Raman spectroscopy in Fig. 2, the effective range of the influence of the contact could not be estimated by the optical technique. We prepared 2H-MoTe2/[Ca2N]+⋅e− samples with different 2H-MoTe2 thicknesses and probed the structural phase of MoTe2 at the top surfaces using two surface-sensitive techniques: friction force microscopy (FFM)29,30 and scanning photoemission microscopy (SPEM) (ref. 6). The thickness of the 2H-MoTe2 and the presence of phase transition at the top surface of the heterostructure could estimate the spatial range of contact-induced charge transfer or resulting phase transition in the vertical direction. We adopted FFM (E-Sweep, Seiko) to distinguish semiconducting (hexagonal) and metallic (monoclinic) phases on the top MoTe2 surface31 partially covering a [Ca2N]+⋅e−. FFM measures adhesion forces between metallic or semiconducting MoTe2 surfaces and the platinum tip by the torsional displacement of the cantilever; for 2D TMDs, the friction rapidly approaches the bulk value when the number of layer is over 10 (ref. 30), corresponding to the probing depth of the few nanometers. The metallic MoTe2 surface produces a stronger adhesion force or friction force with the metallic tip due to the attractive interaction (from image charges) between two metals32. Prior to the experiment, the baseline for the friction values (torsional displacements of the cantilever in units of millivolt) was established with exfoliated flakes from semiconducting hexagonal and metallic monoclinic MoTe2 single crystals; the effects of scanning speed, loading force, tip material, and sample geometry were reflected in the friction measurement for an unambiguous identification of MoTe2 phase (see Supplementary Fig. 6). Figure 3a shows a friction force image of a large 2H-MoTe2 flake covering the [Ca2N]+⋅e− flake outlined by the white dashed line. The topography corresponding to Fig. 3a is provided in the inset; the MoTe2 film was approximately 120 nm thick, whereas the underlying [Ca2N]+⋅e− was approximately 70 nm thick. We verified the underlying flake is indeed [Ca2N]+⋅e− by measuring the friction force of [Ca2N]+⋅e− at the small crack, black region in inset of Fig. 3a. A higher adhesion or friction force, described by the red colored region, was

ACS Paragon Plus Environment

Page 6 of 17

Page 7 of 17

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

identified above the underlying [Ca2N]+⋅e−, indicating that the surface electronic structure of the MoTe2 above the [Ca2N]+⋅e− flake is different from the other MoTe2 area without [Ca2N]+⋅e− underneath. The redlines outside of the white dotted line is due to the curvature affecting the friction data as they mostly corresponds to the step edges of exfoliated MoTe2. Moreover, the friction force of the region above the [Ca2N]+⋅e− was quantitatively consistent with the established baseline of metallic monoclinic MoTe2 (Fig. S6). This demonstrates that the top surface region underwent a phase transition from the semiconducting to the metallic phase. Given that the phase transition extended through the bulk and to the top MoTe2 surface, the phase engineering range in the vertical direction by the contact is estimated to have a film thickness of 120 nm. Beyond the friction force measurement based on the metallic and semiconducting nature of MoTe2, the chemical states of the surface atoms in two different structural phases were directly examined by scanning photoemission microscopy (SPEM). Through the vacuum transfer of MoTe2/[Ca2N]+⋅e− heterojunction samples to the SPEM chamber, the binding energies of Mo and Te 3d electrons with and without underlying [Ca2N]+⋅e− and their spatial distributions were obtained and are shown in Fig. 3b. The blue (without [Ca2N]+⋅e− underneath) and red curves (above [Ca2N]+⋅e−) in Fig. 3b correspond well with the reference binding energy data of hexagonal and monoclinic MoTe2 established with bulk MoTe2 single crystals6. Figures 3c-3e show the correlation between the AFM topography and SPEM intensity. Because the underlying [Ca2N]+⋅e− location was identified by AFM (white dashed lines in Fig. 3c-3e), the clear topological dependence of SPEM intensities on [Ca2N]+⋅e− locations (Fig. 3d, e) confirms the [Ca2N]+⋅e− contact-driven phase transition. Since the probing depth of x-ray photoemission is in the order of 1 nm, we can infer the phase transition propagation length from the thickness of the flake; the thicknesses of the flakes where phase transition occurs range from 60 to 120 nm, similar to the samples studied with FFM. We have demonstrated a contact-induced lattice symmetry change at the top MoTe2 surface, located ~100 nm away from the [Ca2N]+⋅e−; the friction force on the top MoTe2 surface matches the baseline of metallic monoclinic MoTe2, whose carrier density is 5.9×1014 cm-2. We referred to the Hall measurement result from as-synthesized single crystal of monoclinic MoTe2 to estimate the carrier density of the phase transitioned region, which was measured to

ACS Paragon Plus Environment

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

be 5.9×1014 cm-2 (Fig. 3a and Supplementary Fig. 7). Moreover, the lattice symmetry change on the top surface, revealed in surface chemical state analyses, confirms the presence of a large surface carrier density (>1.6×1014 cm-2); such carrier density is critical to the lattice symmetry change that has not been realized by conventional electric gating methods. In solid state physics, Thomas–Fermi screening limits the screening length (or, equivalently, the Debye length) to ~1 nm for a carrier concentration of 1014 cm-2 or 1021 cm-3 (see Supplementary Fig. 9, 10 and ref. 32). Thus, our finding of contact-driven long-range (100 nm) charge transfer with a density of 1014 cm-2 might be considered as a violation of the screening theory because [Ca2N]+⋅e− is positively charged by giving its electrons to MoTe2 in contact that should be screened in a short-range of ~1 nm by dense electrons in MoTe2 (ab initio calculation for an equivalent scenario is provided in ref. 33 concluding that the screening is confined within few layers at the interface). In order to model the charge exchange between [Ca2N]+⋅e− and MoTe2, we considered the interface structure, in which the (001) surface of [Ca2N]+⋅e− is in contact with the (001) surface of either hexagonal or monoclinic MoTe2. (mono-layered case is shown in Fig. 4a and multi-layered case is shown in Supplementary Fig. 9 below. Please refer to SI section 9 for more calculation details.) For the 2H-MoTe2/[Ca2N]+⋅e− heterostructure, first-principles density functional calculations, which provide the profile of electron densities in the ground state, indeed show that the electronic band structures of MoTe2 and [Ca2N]+⋅e− are mostly affected near the interface by the charge transfer, regardless of the layer thickness of MoTe2 (Fig. 4a and Supplementary Figs. 10, 11). Whereas the conduction band minimum (CBM) of hexagonal MoTe2 is empty before the contact (upper panel in Fig. 4a), strong electron doping occurs after the contact between the MoTe2 and the [Ca2N]+⋅e−, and the Fermi level is then located at 0.15 eV above the CBM (lower panel in Fig. 4a). The transferred charge density in the conduction band (up to the Fermi level) of the MoTe2 is estimated to be 1014 cm-2 (marked by the red symbol in Fig. 4b), which quantitatively matches with the carrier density deduced in our measurements. The observed long-range (~100 nm) lattice symmetry change driven by high electron density seems puzzling since the screening length should be in the order of 1 nm for such a high electron density; hence this cannot explain the long range symmetry change observed in this study. In additions, the first principle calculation also predicts that the charge transfer would be limited to at most a few layers at the interface. While the exact mechanism of how the

ACS Paragon Plus Environment

Page 8 of 17

Page 9 of 17

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

electrons diffuse from [Ca2N]+⋅e− across MoTe2 remains unclear, we consider a couple of scenarios that can help explain the observed phenomenon. When the electrons from [Ca2N]+⋅e− enter MoTe2 initially, they would have high kinetic energy owing to the large work function difference between the two materials (~1.2 eV). The highly mobile electrons can propagate through MoTe2 achieving a longer doping length than its equilibrium state. This is also compatible with the observation of high doping of 2H-MoTe2 with Raman spectroscopy. An alternative scenario is a sequential phase transition; first few layers at the interface undergoes a symmetry change due to heavy doping, cascading to the next layers eventually reaching the top layer. We note that a vertical gradient of 2H/1T’ ratio might be present in the sample, i.e. MoTe2 is mostly in 1T’phase at the interface and progressively become more 2H further from the interface. Our observation of long-range lattice symmetry change in MoTe2 in contact with [Ca2N]+⋅e−, suggests that the modulation of material property is feasible without altering chemical composition of the material. The electride/material hetero interface can be a platform for studying high carrier dynamics of materials as well as a tool of modulating the material properties.

ACS Paragon Plus Environment

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

Figure 1. [Ca2N]+⋅e− oxidation by air and MoTe2. a. Raman spectra and optical images of pristine (green) and oxidized (brown) exfoliated [Ca2N]+⋅e−. b. Illustration of [Ca2N]+⋅e− oxidation by doping electrons to MoTe2 in vertical contact. Electron anionic layers are represented by blue clouds between [Ca2N]+⋅e− layers. Long-range phase change from hexagonal to monoclinic MoTe2 occurs with the [Ca2N]+⋅e− contact-driven electron diffusion and the modification of the [Ca2N]+⋅e− lattice structure.

ACS Paragon Plus Environment

Page 10 of 17

Page 11 of 17

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 2. Evolution of the Raman spectra of MoTe2 in contact with [Ca2N]+⋅e−. a. Optical image of the MoTe2/[Ca2N]+⋅e− vertical heterostructure. b. Raman spectra of MoTe2 in contact with [Ca2N]+⋅e−. The Raman signature of MoTe2 shows a systematic change from the redshift through a phase transition to the monoclinic phase. The colors in the dots in '(A)' correspond to those in the Raman spectra. c. FWHM of the Eg mode vs. Raman shift of 2HMoTe2 collected from different samples exhibiting redshifts. The peak positions of the E2g mode under different electron doping densities obtained from first-principles calculations are indicated with magenta dotted vertical lines with the corresponding doping densities.

ACS Paragon Plus Environment

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

Figure 3. Phase transition probed at the top surface away from the MoTe2/[Ca2N]+⋅e− interface. a. friction force mapping showing different friction forces between MoTe2 (blue) and MoTe2/[Ca2N]+⋅e− regions (red). The white dashed line indicates the position of the underlying [Ca2N]+⋅e− identified from the topography (inset). b. Representative binding energy spectra of Mo and Te 3d electrons with (red) and without (blue) underlying [Ca2N]+⋅e− from SPEM. The red dot represents the location where the XPS spectra is taken. c. AFM topography of a [Ca2N]+⋅e−/MoTe2 heterostructure. The white dashed line indicates the [Ca2N]+⋅e− position. d-e. Intensity mapping at two different energy levels (231.4 eV for the hexagonal phase and 229.8 eV for the monoclinic phase) from SPEM.

ACS Paragon Plus Environment

Page 12 of 17

Page 13 of 17

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 4. First-principles calculations for the amount of charge transferred at the MoTe2/[Ca2N]+⋅e− interface. a. Contact-driven band structure changes. The local density of states along the vertical direction (right) is correlated with the atomic lattice models of hexagonal MoTe2 and [Ca2N]+⋅e− (left). The electron doping of MoTe2 with its Fermi level shifted by 0.15 eV above the conduction band minimum (CBM) is represented as a consequence of contact (bottom). b. Band alignment reflecting energy levels by DFT calculations. Fermi electrons in Ca2N have much higher chemical potential than that electrons in MoTe2. c. Calculated electron densities as a function of the Fermi level position with respect to the CBM of 2H-MoTe2. The amount of doping in 'a' is indicated by a red symbol.

ACS Paragon Plus Environment

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

■ ASSOCIATED CONTENT Supporting Information Supporting Information Available: [Ca2N]+⋅e− and its oxidation studied with optical microscope and Raman spectroscopy, phonon confinement effect on monoclinic MoTe2, friction force microscopy study on both phases of MoTe2, detailed description of DFT calculation involving monolayer/multilayer MoTe2/[Ca2N]+⋅e− interface, and non-linear TF screening calculations are included. This material is available free of charge via the Internet at http://pubs.acs.org. ■ AUTHOR INFORMATION Corresponding Authors * E-mail: [email protected], [email protected] Notes The authors declare no competing financial interests. ■ ACKNOWLEDGEMENTS This work was supported by Samsung Research Funding & Incubation Center of Samsung Electronics under Project Number SRFC-MA1701-01 (H.Y.), the Institute for Basic Science (IBS-R011-D1) (Y. H. L.), and Samsung Science and Technology Foundation under Grant No. SSTF-BA1401-08 (D.H.C. and K.J.C.).

ACS Paragon Plus Environment

Page 14 of 17

Page 15 of 17

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

■ REFERENCES 1

Keum D. H. et al. Nature Phys. 2015, 11, 482-486.

2

Lu, J. M. et al. Science 2015, 350, 1353-1357.

3

Saito, Y. & Iwasa, Y. ACS Nano 2015, 9, 3192-3198.

4

Ohta, T., Bostwick, A., Seyller, T., Horn, K. & Rotenberg, E. Science 2006, 313, 951-

954. 5

Kappera, R. et al. Nature Mater. 2014, 13, 1128-1134.

6

Cho, S. et al. Science 2015, 349, 625-628.

7

Zhang, Y., Ye, J., Matsuhashi, Y. & Iwasa, Y. Nano Lett. 2012, 12, 1136-1140.

8

Saito, Y. & Iwasa, Y. ACS Nano 2015, 9, 3192-3198.

9

Fang, H. et al. Nano Lett. 2013, 13, 1991-1995.

10

Kiriya, D., Tosun, M., Zhao, P., Kang, J. S. & Javey, A. J. Am. Chem. Soc. 2014, 136,

7853-7856. 11

Chhowalla, M. et al. Nature Chem. 2013, 5, 263.

12

Csanyi, G., Littlewood, P. B., Nevidomskyy, A. H., Pickard, C. J. & Simons, B. D.

Nature Phys. 2005, 1, 42-45. 13

Smith, A. M., Mohs, A. M. & Nie, S. Nature Nanotech. 2009, 4, 56-63.

14

Ohta, T. et al. Phys. Rev. Lett. 2007, 98, 206802.

15

Yu, Y. -J. et al. Nano Lett. 2009, 9, 3430-3434.

16

Cai, Y., Zhang, G. & Zhang, Y.-W. Scientific Reports 2014, 4, 6677.

17

Duerloo, K.-A. N., Li, Y. & Reed, E. J. Nature Commun. 2014, 5, 4214.

18

Eda, G. et al. Nano Lett. 2011, 11, 5111-5116.

19

Lin, Y.-C., Dumcenco, D. O., Huang, Y.-S. & Suenaga, K. Nature Nanotech. 2014, 9,

391-396. 20

Lee, K., Kim, S. W., Toda, Y., Matsuishi, S. & Hosono, H. Nature 2013, 494, 336-

ACS Paragon Plus Environment

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

340. 21

Zhao, S., Li, Z. & Yang, J. J. Am. Chem. Soc. 2014, 136, 13313-13318.

22

Song, S. et al. Nano Lett. 2016, 16, 188-193.

23

Campbell, I. H., & Fauchet, P. M. Solid State Commun. 1986, 58, 739-741.

24

Mignuzzi, S. et al. Phys. Rev. B 2015, 91, 195411.

25

Frey, G. L. et al. Phys. Rev. B 1999, 60, 2883.

26

Kang, Y. et al. Adv. Mater. 2014, 26, 6467-6471.

27

Pimenta, M. A., del Corro, E., Carvalho, B. R., Fantini, C. & Malard, L. M. Accounts

of Chemical Research 2015, 48, 41-17. 28

Chakraborty, B. et al. Phys. Rev. B 2012, 85, 161403.

29

Choi, J. S. et al. Science 2011, 333, 607-610.

30

Lee, C. et al. Science 2010, 328, 76-80.

31

Kim, S. et al. 2D Materials 2017, 4, 024004.

32

Buckley, D. H. J. Coll. Inter. 1977, 58, 36-53.

33

Brumme, T., Calandra, M. & Mauri, F. Phys. Rev. B. 2015, 91, 155436.

ACS Paragon Plus Environment

Page 16 of 17

Page 17 of 17

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

A graphic for the Table of Contents

ACS Paragon Plus Environment