Origins of Fermi Level Pinning between Molybdenum

Subscriber access provided by Nottingham Trent University

C: Surfaces, Interfaces, Porous Materials, and Catalysis 2

Origins of Fermi Level Pinning between Molybdenum Dichalcogenides (MoSe, MoTe) and Bulk Metal Contacts: Interface Chemistry and Band Alignment 2

Christopher M. Smyth, Rafik Addou, Christopher L. Hinkle, and Robert M. Wallace J. Phys. Chem. C, Just Accepted Manuscript • DOI: 10.1021/acs.jpcc.9b04355 • Publication Date (Web): 12 Aug 2019 Downloaded from pubs.acs.org on August 13, 2019

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 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 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.

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 49 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

The Journal of Physical Chemistry

Origins of Fermi Level Pinning between Molybdenum Dichalcogenides (MoSe2, MoTe2) and Bulk Metal Contacts: Interface Chemistry and Band Alignment Christopher M. Smyth†, Rafik Addou†,‡, Christopher L. Hinkle†,§, Robert M. Wallace*,†,1

†Department

of Materials Science and Engineering, University of Texas at Dallas, Richardson, Texas, 75080, USA ‡School

of Chemical, Biological and Environmental Engineering, Oregon State University, Corvallis, Oregon, 97331, USA §Department

1

of Electrical Engineering, Notre Dame, South Bend, Indiana, 46556, USA

Email: [email protected]; Phone: +1 (972) 883-6638

1 ACS Paragon Plus Environment

The Journal of Physical Chemistry 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 49

Abstract Nanometer scale films of contact metals (Au, Ir, Cr, and Sc), deposited on bulk, exfoliated MoSe2 and MoTe2 by electron beam evaporation under high vacuum (HV, < 2 × 10-6 mbar) and ultra-high vacuum (UHV, < 2 × 10-9 mbar), are compared to elucidate the metal-TMD interface chemistry and its relationship with the reactor ambient. Au does not react with MoSe2, but does react with MoTe2, regardless of reactor ambient. In contrast, the presence of an intermetallic is detected at the Ir–MoSe2 and Ir–MoTe2 interfaces when it is deposited in UHV and HV. The typically more reactive, low work function metals Cr and Sc completely reduce the TMD near the interface. Sc is completely oxidized during metallization in HV. These results highlight the reactive nature of interfaces formed between Mo-based TMDs and metals. Furthermore, band alignment between Au, Ir, and Sc and the Mo-based TMDs deviate significantly from the Schottky-Mott rule. These results elucidate the true chemistry of selected contact metal–TMD interfaces and the oxidizing effects that a higher deposition chamber base pressure has on the interface chemistry. Additionally, our work highlights the need to consider the true interface chemistry when engineering and modeling metal contacts to MoSe2 and MoTe2.


Page 3 of 49 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


Introduction Transition metal dichalcogenides (TMDs) are a large family of layered, two-dimensional materials with demonstrated utility in a wide variety of next-generation technologies, such as low power electronics,1-4 spintronics,5 photonics,6 optoelectronics,7,8 and catalysis.9 The most widely investigated TMDs are stable in a 2:1 ratio (i.e. MX2, M = transition metal, X = chalcogen), and each unit cell is composed of three alternating X–M–X layers.10 For example, MoSe2 and MoTe2 have been successfully implemented in high-performance field-effect transistors (FETs) with impressive subthreshold swing (140 mV/dec),11 ION/IOFF ratio (~1010), and room temperature mobilities (480 cm2/V s and ~200 cm2 V-1 s-1, respectively).2,12 MoTe2 is perhaps the most intriguing semiconductor of the three Mo-based TMDs due to its moderate band gap, ~1.0 eV (bulk) to 1.2 eV (single layer), which facilitates low power device operation and is promising for optoelectronic applications.13 The realization of commercially viable TMD-based electronic or spintronic devices requires the formation of low resistance contacts. Intolerably high contact resistance is a critical limiting factor to TMD-based device performance and must be mitigated.14-16 The low dimensionality of TMDs typically employed in devices prevents adaptation of traditional doping strategies (e.g., ion implantation). Molecular doping17,18 has provided up to 105 reduction in contact resistance (Rc), but most treatments are unstable over time. Employing a metal-insulator-TMD contact structure has significantly reduced the Rc and Schottky barrier height (SBH) and can also prevent spurious reactions between the metal and the TMD.19,20 However, the best Rc achieved in TMD-based devices employing an insulating interlayer at the contact interface (~10-5 Ω cm-2)19 is too high to for commercial viability. Perhaps the most promising strategy to minimize Rc is to contact the edge of the TMD sheet rather than the van der Waals (vdW) surface. The semi-metallic 1T’ phase of



Page 4 of 49

MoSe221 and MoTe222 have been exploited in edge contact schemes and enabled the lowest Rc (0.2 kΩ μm) in a TMD-based device to date.22 Three-dimensional metal contacts are most commonly employed in TMD-based devices to date due to the ease of fabrication. The highest performance bulk metal contacts in MoSe2- and MoTe2based devices have been reported with room temperature (RT) Rc values of ~400 and 2 k m20,23 (Ti and Ag, respectively), which are orders-of-magnitude higher than that required for state-ofthe-art logic circuits (< 0.1 k m).24 However, a number of studies have demonstrated strong EF pinning manifests in a number of metal-TMD systems due to a variety of gap state phenomena,2530 which complicates contact engineering when using bulk metals. Density functional theory (DFT)

calculations reveal strong hybridization, short interatomic bonds (1.7-2.4 Å), and negligible Schottky barriers between transition metals with partially filled d-orbitals (Sc, Ti, Cr, Ni, Pt) and MoSe2. In contrast, transition metals with either empty (Ag) or completely filled (Au, Al) dorbitals weakly hybridize with MoSe2 resulting in large SBHs and interatomic bonds at the interface.31,32 The aforementioned DFT studies suggests the contact polarity and performance can be indirectly inferred from the valence orbitals of the contact metal. However, theoretical calculations typically consider pristine TMD layers, which neglects the significant effects that interface defects have on the electrostatics at the metal-TMD interface. Defects commonly found in TMDs significantly alter the transport properties across the metalTMD interface due to the localized density of state variation making contact engineering difficult.28,30,33-37 Critically, most studies ignore the effects of chemical reactions that often occur at the metal-TMD interface during metallization or post-metallization anneals on the electrostatics of the interface. Low Rc in silicon electronics has been repeatedly engineered using silicide technology, which has required a detailed understanding of the intermetallic chemistry and


Page 5 of 49 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


structure as a function of processing conditions.38 Our previous work has shown an intermetallic forms spontaneously at the metal-TMD interface in most systems during metallization at RT.23,3941

Moreover, the interface chemistry can vary significantly with the deposition chamber

ambient.16,39-41 Therefore, the relationship between interface chemistry and metal-TMD electrostatics must be understood in greater detail to enable commercially viable metal contacts to TMDs. In this work, we explore the effects of the background pressure during metal deposition on the interface chemistry formed with bulk MoSe2 and MoTe2. Specifically, metals (Au, Ir, Cr, Sc) with work functions spanning a wide energy range (3.5-5.3 eV)42 are deposited to a thickness of ~1 nm to elucidate the relationship between the interface chemistry and the degree of Fermi level pinning. The metal-TMD interface chemistry formed under two different deposition chamber base pressures is characterized using X-ray photoelectron spectroscopy (XPS). The associated band alignments are extracted from valence band edge spectra obtained by XPS. Chemical and thermodynamic trends are discussed in relation to interface chemistry commonalities across metal– TMD systems and correlated generally with the extracted band alignments. All thermodynamic data is obtained from references. No unique thermodynamic calculations have been performed in this work. The relationship between reactivity and the metal growth mechanism is characterized using atomic force microscopy (AFM).

Methods Metal Deposition in UHV and HV and Characterizing the Interface Chemistry Bulk crystals of synthetic MoSe2 and MoTe2 employed in this work were purchased from HQ Graphene.43 All samples originated from the same MoSe2 and MoTe2 crystals (25-100 mm2 surface



Page 6 of 49

area), but individual pieces were sliced from these crystals and exfoliated for each respective metal deposition. We employed identical procedures to those described in our previous work39 to prepare and mount samples, outgas metal sources, determine metal deposition rates (0.3-0.6 nm min-1), and characterize films using XPS. The bulk MoSe2 and MoTe2 crystals were fixed on a 4” silicon wafer by carbon tape. The surfaces of as-exfoliated samples used for deposition under HV conditions are assumed to be chemically congruent with those characterized prior to deposition under UHV conditions considering all samples are mechanically exfoliated under laboratory ambient conditions before loading into the deposition chamber. Metal deposition is performed at nominally RT in ultra-high vacuum (UHV, base pressure < 2 × 10-9 mbar) via electron beam (ebeam) evaporation in a chamber attached to a UHV cluster tool described elsewhere.44 After metallization, samples were transferred to a separate chamber for XPS characterization without breaking vacuum via a transfer tube (base pressure 10-10 mbar). Metal deposition is also performed under high vacuum (HV, base pressure < 2 × 10-6 mbar) conditions in an elastomer-sealed Temescal BJD-1800 e-beam evaporator in the UT-Dallas cleanroom facility. Samples were then transferred as quickly as possible (air exposure ~5 min) to the UHV cluster tool for subsequent characterization by XPS. All metal films were deposited on the TMDs to target thicknesses of 1 nm. The method by which the Au 4f, Ir 4f, Cr 2p, Sc 2p, and Mo 3d core level spectra were obtained is described elsewhere.39 It is also noted here that the evaporation process avoids any lithography patterning polymer photoresist typical for device fabrication, which is well known to introduce surface contamination that is difficult to remove. X-ray Photoelectron Spectroscopy and Spectra Analysis XPS on all sample surfaces was carried out using a monochromated Al Kɑ source and an Omicron EA125 hemispherical analyzer with resolution of ± 0.05 eV. An analyzer acceptance angle of 8°,


Page 7 of 49 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


takeoff angle of 45°, and pass energy of 15 eV were utilized in this study. Therefore, the core level spectra obtained in this work reflect chemical state averages from a ~500 m diameter elliptical region of the sample surface. The analyzer was calibrated using sputter cleaned Au, Cu, and Ag foils, according to the standard procedure described in ASTM E2108.45 Spectra were deconvolved using the curve fitting software AAnalyzer.46 All stoichiometric values are calculated using the integrated intensity of the defined core levels after being corrected by the appropriate atomic sensitivity factors, which are unique to the hemispherical analyzer employed in this work (see Supporting Information for all relevant atomic sensitivity factors). Atomic Force Microscopy AFM images were obtained ex–situ from bare Mo-based TMDs immediately after exfoliation and again after metal deposition and subsequent XPS characterization using a Veeco Model 3100 Dimension V Atomic Probe Microscope in non-contact tapping mode in the cleanroom facility.47 WSxM software48 was employed to process images, obtain line profiles, and determine RMS roughness values reported here.

Results and Discussion Chalcogen-Dependent Interface Chemistry Formed between Molybdenum Dichalcogenides and Au Au/MoSe2 Interface Study Residual gas molecules exhibit an impingement rate in HV that is orders of magnitude greater than that in UHV according to the kinetic theory of gases. In a typical device fabrication process utilizing a standard cleanroom tool, metallization is most commonly performed in HV, which can result in dramatic interface chemistry differences from a metallization process performed in



Page 8 of 49

UHV.39-41 The binding energies of all reaction products detected in the molybdenum (Mo 3d5/2) and chalcogen (Se 3d5/2 or Te 3d5/2) core level spectra after depositing Au, Ir, Cr, or Sc can be found in Table 1. Table 1. Binding energies of chemical states detected in Mo 3d5/2, Se 3d5/2, and Te 3d5/2 core level spectra after exfoliation and subsequent metal deposition in UHV or HV. Metals

TMD

Core Level

Exfoliated (UHV)

UHV

HV

MoSe2

Mo 3d5/2 Se 3d5/2

229.16 (MoSe2) 54.71 (MoSe2)

228.95 (MoSe2) 54.52 (MoSe2)

Mo 3d5/2

228.53 (MoTe2)

228.30 (MoTe2)

Te 3d5/2

573.21 (MoTe2) 575.78 (TeO2)

572.70 (AuTex) 573.04 (MoTe2) 575.51 (TeO2)

Mo 3d5/2

229.21 (MoSe2)

228.35 (MoxIrySe2) 228.72 (MoSe2)

Se 3d5/2

54.77 (MoSe2)

54.26 (MoxIrySe2) 54.45 (MoSe2)

Mo 3d5/2

228.52 (MoTe2) 232.52 (MoOx)

228.15 (MoxIryTex) 228.28 (MoTe2)

Te 3d5/2

573.19 (MoTe2) 576.04 (TeO2)

572.97 (MoTe2) 573.25 (MoxIryTez)

Mo 3d5/2

229.18 (MoSe2)

Se 3d5/2

54.72 (MoSe2)

Mo 3d5/2

228.47 (MoTe2)

Te 3d5/2

573.16 (MoTe2) 575.80 (TeO2)

229.08 (MoSe2) 54.64 (MoSe2) 228.38 (MoTe2) 231.99 (MoOx) 572.60 (AuTex) 573.06 (MoTe2) 575.65 (TeO2) 228.29 (MoxIrySe2) 228.78 (MoSe2) 232.42 (MoOx) 54.36 (MoxIrySe2) 54.51 (MoSe2) 228.31 (MoxIryTez) 228.44 (MoTe2) 231.73 (MoOx) 573.18 (MoTe2) 573.92 (MoxIryTez) 575.66 (TeO2) 227.82 (Mo0) 228.98 (MoSe2) 54.32 (CrSex) 54.52 (MoSe2) 227.75 (Mo0) 228.27 (MoTe2) 572.61 (CrTex) 572.94 (MoTe2)

Mo 3d5/2

229.17 (MoSe2)

227.69 (Mo0) 229.00 (MoSe2)

229.10 (MoSe2)

Se 3d5/2

54.71 (MoSe2)

54.52 (MoSe2) 54.80 (ScSex)

54.20 (MoSe2)

228.53 (MoTe2)

227.72 (Mo0) 228.53 (MoTe2)

Au MoTe2

MoSe2 Ir MoTe2

MoSe2 Cr MoTe2

MoSe2 Sc MoTe2

Mo 3d5/2 Te 3d5/2

227.61 (Mo0) 228.75 (MoSe2) 54.20 (CrSec) 54.30 (MoSe2) 227.64 (Mo0) 228.30 (MoTe2) 572.78 (CrTex) 573.00 (MoTe2)

573.22 (MoTe2) 572.86 (ScTex) 576.19 (TeO2) 573.20 (MoTe2) van der Waals interface covalent interface

227.60 (Mo0) 228.21 (MoTe2) 232.40 (MoOx) 572.86 (MoTe2) 576.41 (TeO2)

The chemistry detected at the Au-MoSe2 interface is chemically congruent with the Au-MoS2 interface, which we have discussed in detail elsewhere.39 Therefore, the corresponding core level spectra are displayed in the Supporting Information. Significant concentrations of defects are expected on the surface of CVT MoSe2 crystals according to a recent STM study.49 However, the 8 ACS Paragon Plus Environment

Page 9 of 49 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


close fit of the Mo 3d and Se 3d core level spectra achieved with a single chemical state in each case indicates the defect densities which would give rise to alternative bonding variations in the MoSe2 crystals employed in this work are below the limit of XPS detection (Figure S1a). Reaction products that form during Au deposition on MoSe2 are also below the XPS detection limit indicating Au forms a vdW interface with MoSe2. Reactions between Au and MoSe2 are not thermodynamically favorable (ΔG°f,AuSe (-33.0 kJ/mol) >> ΔG°f,MoSe2 (-98.2 kJ/mol)) MoSe2.50,51 Reactions between Au and carbon- or oxygen-based species are below the limit of XPS detection. The C 1s and O 1s core level spectra obtained after exfoliation and subsequent metallization in UHV and HV are discussed in the Supporting Information. Au/MoTe2 Interface Study A previous XPS study ruled out reactions between Au and MoTe2. However, the hemispherical analyzer employed in the work exhibited a relatively low resolution of 0.9 eV.52 The spectrometer employed here exhibits a much higher 0.05 eV resolution, providing much higher sensitivity to minor chemical state changes that could indicate the formation of reaction products. After exfoliation, a Te:Mo ratio of 2.8 is detected in the MoTe2 crystal employed in this part of the experiment indicating a Te-rich chemistry, which is commonly observed in MoTe2 grown by chemical vapor transport (CVT).33 It is important to note the integrated intensities of the MoTe2 chemical states in the Te 3d5/2 and Mo 3d core levels corrected with the appropriate atomic sensitivity factors are employed to calculate the ratio. The much lower BE Te 4d core level is more bulk sensitive than the Mo 3d and Te 3d5/2 core levels and therefore, if employed in the ratio calculation, would underestimate the concentration of excess Te present in the MoTe2 at the surface. It is possible that excess Te accumulates near the surface of the bulk MoTe2 during the cooling stage of the CVT growth process, which could also explain the excess Te detected in this



Page 10 of 49

work. However, excess Te in bulk MoTe2 crystals (grown by CVT) after many exfoliations is reported in previous publications,53 which indicates interstitial or intercalated Te is energetically favorable in MoTe2.33 In addition to the MoTe2 chemical state, a chemical state is detected at 575.78 eV in the Te 3d5/2 core level indicating the presence of TeO2 on the exfoliated surface. Te oxidation is likely exacerbated by the excess Te within the XPS sampling depth of the MoTe2 surface, considering less defective, stoichiometric MoTe2 is quite resistant to oxidation.53 After depositing ~1 nm Au in either UHV or HV, additional chemical states are detected at low BE in the Te 3d5/2 and Mo 3d core levels (572.70 eV and 227.88 eV, respectively) indicating the formation of AuTex and a Mo-based metallic species, respectively (Figure 1a). The exothermic nature of reactions between Au and Te (ΔG°f,AuTe2 = -65.0 kJ/mol)54 rationalizes the formation of AuTex formed via the excess Te in the MoTe2 in this work. The AuTex chemical state exhibits a lower BE than the MoTe2 chemical state in the Te 3d5/2 core level because the Au-Te electronegativity difference (0.59)55,56 is less than that of Mo-Te (0.71).55 The observed persistence of TeO2 throughout the Au depositions is expected considering the endothermicity of the Au-O reaction and the sizeable Te-O bond dissociation energy (BDE, 391.0 kJ/mol).57 The low BE chemical state in the Mo 3d core level spectra obtained after Au deposition agrees with that of a metallic Mo reference indicating the presence of a metallic Mo-based species. AFM images show the Au also exhibits volmer weber-like growth on MoTe2 (Figure 1b). Similar to the Au–MoSe2 system, the Au 4f core level is detected ~0.25 eV higher than the Au reference after Au is deposited on MoTe2 in UHV or HV (Figure 1c). A second chemical state is detected at 84.83 eV and 84.98 eV in the Au 4f core levels obtained after Au is deposited on MoTe2 in UHV and HV, respectively, corresponding with the AuTex intermetallic.


Page 11 of 49 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. a) Te 3d5/2, Mo 3d, and c) Au 4f core level spectra obtained from bulk MoTe2 after exfoliation and subsequent Au deposition in UHV (in–situ) and HV (ex–situ). b) 500 nm2 AFM image obtained ex–situ after ~1 nm Au deposition on MoTe2. A line profile obtained from the AFM image is overlaid to clearly show the surface topography of the Au on MoTe2. Adventitious species are detected on the MoTe2 surface after exfoliation in the form of hydroxyls, hydrocarbons, and adventitious carbon (Figure S2a,b). The presence of TeO2 is corroborated by the chemical state detected below 531.0 eV after exfoliation and subsequent Au depositions, which is consistent with chalcogen-oxygen bonding. Reactions between Au and adventitious species on MoTe2 are below the limit of XPS detection. The O 1s and C 1s core level spectra obtained after 11 ACS Paragon Plus Environment


Page 12 of 49

MoTe2 exfoliation and subsequent Au deposition in UHV and HV are discussed in greater detail in the Supporting Information. One may expect fewer impurities at an Au-TMD interface formed in UHV compared with HV.16 Considering the kinetic theory of gases and that all depositions occurred at a rate less than 0.01 nm/s, it is quite obvious that there is a continuous supply of gas species at the surface throughout deposition under HV conditions. Moreover, the relatively high flux of residual gas species present in HV would also be expected to enable interfacial reactions at much higher metal deposition rates as well. Therefore, the presence of adventitious organics at the metal-TMD interface should be expected and necessitates further investigation of their effects on contact performance.

Covalent Interface Formed between Molybdenum Dichalcogenides and Ir Although Ir is not widely employed as a contact metal, it exhibits an unfilled d-orbital valence like other more commonly employed contact metals, including Co. Therefore, the interface chemistry formed between Ir and the molybdenum dichalcogenides investigated in this work provides insight into the chemistry that likely forms between the same TMDs and other transition metals with an unfilled d-orbital valence. Ir/MoSe2 Interface Study The formation energy of iridium selenide is not available to our knowledge. However, the fugacity of Se required for iridium selenide formation from the elemental constituents is higher than that of MoSe2 and therefore the Ir-Se reaction is inferred to be endothermic.54 Figure 2a shows the Mo 3d and Se 3d core level spectra obtained from bulk MoSe2 after exfoliation and subsequent Ir deposition in UHV and HV. After exfoliation, one chemical state is detected in each of the Mo 3d and Se 3d core level spectra each corresponding with bulk MoSe2. After depositing ~1 nm Ir on


Page 13 of 49 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


MoSe2 in UHV and HV, an additional chemical state is detected at low (high) BE in the Mo 3d (Se 3d) core level spectrum indicating the formation of a ternary MoxIr(1-x)Se2 species as a result of reactions between Ir and MoSe2. It is also possible that two separate IrSex and MoSex (i.e. Sedeficient molybdenum selenide) reaction products form, which would result in two unique chemical states at high BE from the bulk MoSe2 chemical state in the Se 3d core level spectrum. MoSex, which would form as Ir scavenges Se from the MoSe2, should exhibit a chemical state closer to elemental Se (i.e. at higher BE) than MoSe2 in the Se 3d core level. The most common stable allotropes of iridium selenide (IrSe2, Ir2Se3)54 would both exhibit chemical states in the Se 3d core level closer to elemental Se (i.e. higher BE) than the MoSe2. The smaller electronegativity difference between Se and Ir (0.95) compared with Se and Mo (1.18) suggests the Se-Ir bond is more covalent than the Se-Mo bond, which would manifest as a smaller BE offset of the Se2chemical state in IrSe2 than that in MoSe2.55 Therefore, the broad chemical state detected at high BE in the Se 3d core level could represent a convolution of IrSex and MoSex chemical states. However, it is not reasonable to arbitrarily distinguish between the two without more detailed characterization of the bonding environment at the Ir-MoSe2 interface (e.g., through studying the X-ray absorption fine structure). A small concentration of MoOx is detected after depositing Ir exsitu in HV, evidenced by the low intensity chemical state in the corresponding Mo 3d core level. MoOx formation in air is a highly exothermic process and therefore should be expected in this work considering the ~1 nm metal films employed could contain pinholes through which adventitious species can access the substrate. The reactions presumably involve interdiffusion that could also promote partial oxidation of the reactants products.



Page 14 of 49

Figure 2. a) Mo 3d, Se 3s, Se 3d, and b) Ir 4f core level spectra obtained from bulk MoSe2 after exfoliation and subsequent ~1 nm Ir deposition in UHV (in–situ) and HV (ex–situ). The Ir 4f core level spectra obtained after depositing Ir on MoSe2 in UHV and HV are dominated by an asymmetric doublet that exhibits a BE consistent with metallic Ir (Figure 2b). An additional chemical state is detected at 61.62 eV and 61.69 eV after deposition in UHV and HV, respectively, corresponding with the MoxIr(1-x)Se2 chemical state. After depositing Ir in HV, IrOx is also detected in the corresponding Ir 4f core level spectrum at 63.45 eV. This is due in part to both ex-situ air exposure-induced surface oxidation and also in-situ oxidation as Ir is deposited in HV. The formation of iridium oxide is spontaneous in the presence of excess oxidizing gases (ΔG°f,IrO2 = 137.0 kJ/mol).58 Ir/MoTe2 Interface Study The chemistry detected at the Ir-MoTe2 interface is congruent with that detected at the Ir-MoSe2 interface when Ir is deposited in UHV and HV. Therefore, the core level spectra obtained from MoTe2 after exfoliation and subsequent Ir deposition in UHV and HV are displayed and discussed in the Supporting Information.


Page 15 of 49 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


Covalent Interface and in-situ High Vacuum Oxidation of the Cr Contact on Molybdenum Dichalcogenides Cr/MoSe2 Interface Study The formation of chromium selenide as a result of reactions between Cr and MoSe2 is thermodynamically favorable considering the ΔG°f,MoSe2 (-98.2 kJ/mol)50 is more positive than the ΔG°f,Cr2Se3 (-175.1 kJ/mol).59 Cr completely reduces the MoSe2 at the interface to form metallic Mo and CrSex during deposition in either UHV or HV (Figure 3a), which is chemically congruent to the Cr-MoS2 interface detected in our previous work.39

Figure 3. a) Mo 3d, Se 3s, Se 3d, and b) Cr 2p3/2 core level spectra obtained from bulk MoSe2 after exfoliation and subsequent ~1 nm Cr deposition in UHV (in–situ) and HV (ex–situ).The Cr 2p3/2 spectra in b) are normalized to the metallic Cr chemical state, Cr0. The chemical states detected in the Cr 2p, O 1s, and C 1s core level spectra after Cr deposition on MoSe2 in UHV and HV are also chemically congruent with those detected in the Cr-MoS2 system discussed previously.39 Therefore, the corresponding spectra are displayed in Figures 3b and S5 but will not be discussed further in the main text. Cr/MoTe2 Interface Study 15 ACS Paragon Plus Environment


Page 16 of 49

Reactions between Cr and crystalline, stoichiometric 2H-MoTe2 are slightly endothermic considering the ΔG°f,CrTe (-70.3 kJ/mol)60 is slightly more positive than the ΔG°f,MoTe2 (-77.9 kJ/mol).61 The ΔG°f of chromium telluride compounds stable in Te:Cr ratios other than 1:1 are not known for comparison. It is therefore possible that the formation of a more stable chromium telluride compound (e.g., Cr2Te3) could be thermodynamically favorable compared with the persistence of MoTe2 when in contact with Cr. Nonetheless, the reaction between elemental Cr(s) and Te(s) is exothermic. The Te:Mo ratio detected after exfoliating the MoTe2 and before depositing Cr in UHV is 2.8. Therefore, the formation of CrTex at the Cr-MoTe2 interface is thermodynamically favorable due to the excess Te available to react exothermically with Cr. The chemical states detected at the CrMoTe2 interface are analogous with those detected at the Cr-MoSe2 and Cr-MoS2 interfaces. Therefore, the core level spectra obtained from MoTe2 after exfoliation and subsequent Cr deposition in UHV and HV are displayed and discussed in greater detail in the Supporting Information. It is important to note here that the low BE chemical state detected in the Mo 3d spectra obtained after depositing Cr in UHV and HV exhibits an asymmetric line shape with high BE tail (characteristic of a metallic species)62 and a BE close to that expected of metallic Mo (Figure S6a). Therefore, Cr not only reacts with the excess Te in MoTe2 but also completely reduces the MoTe2 near the interface to form metallic Mo and CrTex. The state detected at low BE in the corresponding Te 3d5/2 spectra exhibits a BE near that of elemental Te, which indicates the CrTex is a metallic species. Chromium telluride exhibits a metallic band structure and permanent magnetic anisotropy that could be useful in future memory technology.63 In addition, the metallic behavior of the CrTex reaction product at the Cr-MoTe2 interface may benefit carrier injection.


Page 17 of 49 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


Cr oxidation is more thermodynamically favorable than the formation of the most stable tellurium or molybdenum oxides (ΔG°f,TeO2 = -161.0 kJ/mol, ΔG°f,MoO3 = -167.0 kJ/mol),64,65 which rationalizes the absence of Mo-O or Te-O bonds after depositing Cr in HV. This suggests the native chromium oxide that forms in air is a good oxygen diffusion barrier that otherwise could result in oxidation of the underlying reaction products or MoTe2. In addition, a significant concentration of chromium oxide in the form of Cr-O and Cr-OH bonds should be expected in a typical “Cr contact” and the potential effects of such species on the contact performance should not be ignored. Figure 4 displays AFM images obtained from the MoTe2 samples after Cr is deposited in UHV and HV. Pinholes are clearly resolvable in the Cr films despite the covalent bonding between Cr and the MoTe2 according to the XPS results. DFT calculations predicting covalent bonding between certain early transition metals (e.g., Cr, Sc, Ti) and Mo-based TMDs29,31,32 consider a small number of atoms in their interface models due to computational constraints, which excludes the effects that long range lattice mismatch induced stress can have on the initial chemistry of such metal films from consideration. Cr is known to initially grow as clusters on Si(111)66 despite forming reaction products with the substrate. A few-nm thick Cr film, commonly used as an adhesion layer for other metals in TMD-based devices, could impact the contact resistance in cases where the Cr film is incomplete (as observed on MoSe2 and MoTe2 in this work). Contacts that employ an ‘adhesion layer’ must be carefully designed and fabricated to ensure the true interface structure and chemistry yields the desired contact electrostatics.



Page 18 of 49

Figure 4. 1 m2 AFM images and associated line profiles (below) obtained ex–situ from bulk MoSe2 and MoTe2 crystals after depositing ~1 nm Cr in HV.

Highly Exothermic Reactions between Sc and Molybdenum Dichalcogenides Sc/MoSe2 Interface Study Figure 5a shows the Mo 3d and Se 3d core level spectra obtained after depositing ~1 nm Sc on MoSe2 in UHV and HV. It is important to note the Se 3d core level is convoluted by the Sc 3p core level after depositing Sc so that the reader does not confuse this feature with a unique chemical state. The BE and intensity of the Sc 3p core level is carefully accounted for based on the intensity ratio and BE separation between the Sc 3p and Sc 2p core levels obtained from the reference Sc film.67 Sc completely reduces the MoSe2 in the vicinity of the interface when deposited in UHV resulting in the formation of metallic Mo and ScSex, as evidenced by the asymmetric chemical state detected at low BE in the corresponding Mo 3d spectrum and the additional chemical state detected at high


Page 19 of 49 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 in the corresponding Se 3d spectrum. Only the bulk MoSe2 chemical states are detected after depositing Sc in HV. The full width at half maximum of the MoSe2 state in the Se 3d core level increases from 0.48 eV after exfoliation to 0.57 eV after Sc is deposited in HV. This suggests Sc reacts with MoSe2 in–situ in HV, presumably via a similar reaction mechanism to that observed in–situ after depositing Sc on MoSe2 in UHV. In contrast, the Sc-Se bonds dissociate in the HV ambient due to the highly exothermic formation of scandium oxide. In addition, the BDESc-O is ~286 kJ/mol greater than the BDESc-Se, which corroborates the unfavorable persistence of ScSex in the presence of oxidizing gases.68 Se anions that are liberated in the scandium oxidation reaction are free to react with metallic Mo, which is liberated as Sc reduces MoSe2, and presumably form a disordered MoSex species. The disordered nature of the re-formed MoSex at the interface contributes to the symmetric broadening of the MoSe2 chemical states observed in the corresponding Mo 3d and Se 3d core levels. The re-formation of transition metal-chalcogen bonds at the contact metal interface could be advantageous in some applications, such as the formation of an ultra-thin high-κ dielectric layer.



Page 20 of 49

Figure 5. a) Mo 3d, Se 3s, Se 3d, Sc 3p, b) Sc 2p, and Mo 3p core level spectra obtained from MoSe2 after exfoliation and subsequent ~1 nm Sc deposition in UHV (in–situ) and HV (ex– situ). The spectra in b) are normalized to the Sc 2p3/2 core level. The Mo 3p core level and a Se Auger feature convolute the Sc 2p core level spectra obtained after depositing Sc in UHV and HV (Figure 5b). The chemical states detected in the Sc 2p core level spectra obtained from MoSe2 after exfoliation and subsequent Sc depositions in UHV and HV are chemically congruent with the states detected in the same core levels obtained from the Sc-MoS2 system and are therefore not discussed in detail here.39 Sc/MoTe2 Interface Study The thermodynamic favorability of reactions between Sc and MoTe2 cannot be directly evaluated at the time we prepared this manuscript due to a lack of thermodynamic data on scandium telluride compounds. However, we presume the Sc-MoTe2 reaction is exothermic considering the exothermic nature of the Sc-MoS2 and Sc-MoSe2 reactions. The chemical states detected at the Sc-MoTe2 interface after depositing Sc in UHV and HV are chemically congruent to the chemical states detected in the Sc-MoS2 system in our previous work.39 Therefore, the corresponding core level spectra are displayed and discussed in greater 20 ACS Paragon Plus Environment

Page 21 of 49 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


detail in the Supporting Information. It is important to note here that significant concentrations of TeO2 and MoOx are detected at high BE in the Te 3d5/2 and Mo 3d core levels obtained after Sc is deposited in HV (Figure S8a), which indicates the ScTex and metallic Mo that presumably form in–situ as Sc is deposited are completely oxidized. The results presented here indicate ScxOy, not metallic Sc, is deposited in an elastomer sealed HV deposition tool typically found in a cleanroom environment. Scandium oxide is a highly resistive compound that has been employed in previous work as the high- dielectric in high electron mobility transistors.69 The extremely low work function of Sc (3.5 eV) indicates it should form an Ohmic electron contact with MoSe2 and MoTe2 considering the corresponding electron affinities (3.80 eV and 3.75 eV, respectively). However, intermetallic and oxidized species that significantly affect the band alignment are detected at the Sc-MoSe2 and Sc-MoTe2 interfaces in this work. The highest performance n-type MoTe2 transistors have recently been demonstrated by utilizing Sc contacts, but a hexagonal boron nitride interlayer was required to prevent detrimental reactions between Sc and MoTe2.23 The degraded contact performance observed when a ScTex intermetallic is present between a Sc contact and MoTe2 is likely due in part to the intermetallic exhibiting a resistivity 2.5× greater than metallic Sc.70 The reaction products formed between Sc and the Mobased TMDs therefore have a significant impact on the associated Rc. It is reasonable to expect the impact of interfacial reactions on contact resistance to extend to other contact metal-TMD systems as well, which will be discussed in subsequent publications.

Trends in Chemical State Assignment and Reactivity Low Work Function Metals: Cr and Sc



Page 22 of 49

Electronegativity trends can be invoked to explain the BEs of certain chemical states detected in this work. The formation of CrSex is associated with the appearance of a chemical state in the Se 3d core level with a lower BE than the Se state in MoSe2. The difference in electronegativity between Mo and Se is significantly smaller than that between Cr and Se and therefore one would expect the BE of the Se state in CrSex to be less than that in MoSe2. This is consistent with the Se chemical state detected at the Cr-MoSe2 interface formed in UHV and HV. Oxygen is significantly more electronegative than Se. It is therefore reasonable to expect the metal chemical state BE in a metal oxide species to be higher than that in a metal selenide species. This is consistent with the BEs of metal chemical states in various reaction products observed in this study. The CrTex chemical states in the Cr 2p and Te 3d5/2 core levels exhibits BEs near the elemental Cr and Te constituents, which is consistent with previous reports of metallic chromium telluride compounds.60 Electrons can move freely between atomic constituents in metallic compounds. Therefore, BE shifts that would indicate cationic or anionic oxidation states are not observed. The ScSex and ScTex chemical states detected in the respective chalcogen core levels after depositing Sc in UHV exhibit BEs near elemental Se and Te, respectively. The Sc–Te bond should exhibit ionic character considering the significant electronegativity difference (0.81) and the electron accumulation on the Te anion predicted by DFT.71 However, it is possible that the ScSex and ScTex chemical states are convolved of two components; an elemental chalcogen state and a scandium chalcogenide state. This would imply a lesser degree of charge transfer occurs from Sc to the chalcogen in ScSex and ScTex than from Mo to the chalcogen in MoSe2 or MoTe2 and therefore scandium chalcogenide species with chalcogen:Sc ratios < 2. This is a reasonable interpretation of our results considering stable phases of Sc-rich chalcogenide species have been synthesized and are stable at RT.72


Page 23 of 49 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


High Work Function Metals: Au and Ir The Au-Te electronegativity difference (0.44) is small compared with Mo-Te (0.06), which indicates the Au-Te bond is more ionic than the Mo-Te bond. It is therefore reasonable that the chalcogen anion state associated with a AuTex species exhibits a slightly smaller BE than elemental Te. Iridium forms ternary compounds with MoSe2 and MoTe2. The associated chalcogen and molybdenum chemical states are detected with BEs closer to elemental Se and Mo than the MoSe2 chemical states. In addition, the Ir-Se electronegativity difference is less than that of Mo-Se. Therefore, the reaction between Ir and MoSe2 oxidizes the Se and reduces the Mo, which manifests as a BE shift of the intermetallic chemical states towards elemental Se and Mo in the corresponding core levels. In contrast, the Mo and Te states associated with the MoxIr(1-x)Te2 compound exhibit BE shifts away from elemental Mo and Te. The Ir-Te electronegativity difference (0.10) is slightly larger than that of Mo-Te (0.06). Therefore, the formation of the MoxIr(1-x)Te2 compound increases the oxidation (reduction) of the Mo cation (Te anion) and manifests as BE shifts of the MoxIr(1x)Te2

chemical states away from elemental Te and Mo.

Reactivity Let us now consider the relationship between the reactivity of each metal-TMD system investigated in this work with metal work functions. Figure 6a shows a scatter plot relating the concentration of intermetallic species formed at each metal-TMD interface with the work functions of the contact metals. Only the chalcogen core level intensities detected after depositing metals in UHV are plotted because the air exposure after metal deposition in HV perturbs the interface chemistry. In decreasing reactivity, the contact metals follow the trend Ir > Sc > Cr > Au according



Page 24 of 49

to the intermetallic concentration formed by each metal averaged across the three Mo-based TMDs.

Figure 6. a) Reactivity of the metal-TMD systems discussed in this work. Metal-MoS2 data is included from our previous work for reference. The reactivity is gauged according to the intensity ratio of the contact metal–chalcogenide (e.g., ScSex) intermetallic chemical state in the corresponding chalcogen core level (MoSe2: Se 3d, MoTe2: Te 3d5/2) to the sum of all chemical states in the same chalcogen core level. These are plotted versus the Ir, Au, Cr, and Sc work functions for metal depositions in UHV (in-situ). The ‘intensity’ refers to the integrated photoelectron intensity. b) A scatter plot showing the known Gibbs free energy for each metal chalcogenide alloy and relevant metal oxides as a function of metal vacuum work function. The specific compounds are listed on the plot. The dotted lines represent the ΔG°f,MoSe2, ΔG°f,MoTe2, ΔG°f,MoO3, ΔG°f,TeO2, and ΔG°f,SeO. The metal-TMD reaction thermodynamics can indicate the energetic favorability for certain species to form. The ΔG°f of stable intermetallic, metal oxide, and chalcogen oxide compounds in each metal-chalcogen system are plotted as a function of the metal work function in Figure 6b. The formation of Ir chalcogenide and Au chalcogenide are thermodynamically unfavorable according to the more positive ΔG°f compared with that of MoSe2 or MoTe2. The ΔG°f,Ir2Se3 and ΔG°f,Ir2Te3 were not available when this manuscript was prepared. However, the fugacity of a gaseous reactant over different metals can alternatively be used to gauge the energetic favorability of reactions which include a common reactant. The fugacity represents the effective chemical potential between a metal and a pure gaseous reactant. Fugacity and reaction favorability roughly 24 ACS Paragon Plus Environment

Page 25 of 49 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


share a direct relationship. The fugacities of Ir2Se3 and Ir2Te3 are greater than that of AuSe or AuTe, but smaller than that of MoSe2 or MoTe2 under partial pressures of Se(g) or Te(g), respectively. Therefore, the formation of iridium selenide and iridium telluride are not thermodynamically favorable compared to the persistence of MoSe2 and MoTe2, respectively.54 The presumably unfavorable reactions between Ir and the Mo-based TMDs contradict the significant reactions observed in this work. Ir is generally considered a noble transition metal, but its unfilled 5d-orbital facilitates energetically favorable reactions in some cases. Ir is the only contact metal investigated in this work with a d-orbital valence, which likely enhances orbital overlap with the Mo and therefore reactivity with the Mo-based TMDs. In general, the reactivity trends observed in the other metal-MoSe2 and metal-MoTe2 systems agree with the trends predicted by the ΔG°f,Mo-based TMD and ΔG°f,intermetallic. In other words, the reactivity of a metal-TMD system increases with increasingly negative ΔG°f,intermetallic (Au < Cr < Sc). In addition, the reactivity increases as the chalcogen-Mo BDE decreases, which is consistent with the increasing TMD instability as the Z of the chalcogen increases.73 When metals are deposited in HV, transition metal and chalcogen oxidation is highly exothermic (except for Se) and is observed in the Ir-, Cr-, and Sc-TMD systems investigated in this work. The ΔG°f’s of MoSe2, MoTe2, and associated intermetallics are significantly more positive than that of molybdenum oxide, iridium oxide, chromium oxide, and scandium oxide (Figure 6b) in agreement with the formation of IrOx, CrxOy, and ScxOy when the metals are deposited in HV. Therefore, reactive contact metals should be deposited in the highest available vacuum and capped with a sufficiently thick inert metal film to minimize the oxygen concentration throughout the metal contact.



Page 26 of 49

Implications for Contact Resistance Native defects in Mo-based TMDs can cause strong EF pinning with many metals resulting in a discrepancy between the predicted and experimentally derived SBH.15,23,28,30,41 Therefore, inconsistencies between predictions and experimental observations are typical if the SchottkyMott rule25,27 is employed to predict SBHs between metals and TMDs. The presence of reaction products at the contact metal-TMD interface will also affect the SBH and therefore the Rc.74 Even when a metal-TMD junction exhibits a vdW interface (e.g., Au-MoSe2), the Rc is often much higher if the metal is deposited in HV rather than in UHV.16 Given the orders of magnitude higher rate of gas impingement in HV than in UHV, the deposition chamber ambient will directly affect the interface chemistry even when higher metal deposition rates are employed.74 To consider the effect of intermetallic and oxide species at the metal-TMD interface on the band alignment, band diagrams of all metal-TMD systems investigated in this work are derived from the XPS measurements and plotted in Figure 7 for comparison (see Supporting Information for details regarding band diagram construction). The electronic structure of a metal film approaches that of the bulk metal at sub-nm thickness.29,31,32,75 Therefore, the band alignment formed between the ~1 nm thick metal films deposited on MoSe2 and MoTe2 in this work are expected to reflect the band alignment induced by bulk metal films. The Schottky-Mott rule states that the Schottky barrier height is directly proportional to the energy difference between the work function of the metal and the electron affinity or ionization energy of the contacted semiconductor. The degree of Fermi level pinning between a semiconductor and various metal contacts is often quantified by the pinning factor (S), where S = 1 when the system obeys the Schottky-Mott rule perfectly. Metal-semiconductor systems that do not obey the Schottky-Mott rule (e.g., metal-Ge; i.e., strong EF pinning manifests) exhibit S < 1.


Page 27 of 49 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


All of the exfoliated TMDs exhibit a EF in the upper half of the band gap, which is due in part to the abundant Mo d-orbitals predominantly contributing to the density of states in the conduction band edge of Mo-based TMDs. Strong EF pinning (i.e., significant disagreement with the Schottky-Mott rule) is measured in the metal-MoSe2 and metal-MoTe2 systems investigated in this work (S = 0.14 and S = 0.18, respectively), which is likely due in part to the significant concentrations of intermetallic species detected at most interfaces studied here. The degree of orbital hybridization between the contact metal and TMD also significantly affects the band alignment. Strong EF pinning in a number of metal-Mo TMD systems has been reported.30 The largest EF shift induced by metal deposition of any metal-TMD system investigated here is observed in the Ir-TMD systems, which is likely induced by the strong d-orbital hybridization that occurs between Ir and Mo. In contrast, Au, Cr, and Sc exhibit s-orbital valences, which inhibits orbital hybridization with Mo. However, chalcogen valence shells are comprised of a mixture of s- and p-orbitals, which enhance hybridization with Cr and Sc. In general, the electron SBH decreases with decreasing metal work function. The electron SBHs of the Sc-TMD and Ir-TMD systems are far from the expected SBHs derived using the Schottky-Mott rule. Therefore, the intermetallic at the Sc-TMD and Ir-TMD interfaces contributes significantly to EF pinning. It is possible to prevent reactions between reactive contact metals and semiconducting TMDs by inserting an extremely thin barrier layer between the metal and the TMD. In a previous report, physical characterization was employed to demonstrate a single hBN layer prevents reactions between Sc and MoTe2 and dramatically improves the electron injection efficiency of the Sc contact,23 which verifies our hypothesis that reaction products at the metal-TMD interface significantly affect the interfacial electrostatics. It is possible that other two dimensional materials,



Page 28 of 49

such as graphene, could prevent reactions between bulk contact metals and TMDs, but there are no reports to date that use physical characterization to confirm an atomically abrupt interface is achieved by inserting graphene at the metal-TMD interface. Reactive contact metals, such as Cr and Sc, will likely scavenge oxygen from other interlayer materials (e.g., TiO2, Al2O3), further complicating the electrostatics of the junction and perhaps degrading the injection efficiency depending on the oxide that forms. A typical device fabrication procedure will include steps which expose the channel material to air. Adventitious carbon is detected on all exfoliated TMD crystals in this work. As a result, it is reasonable to expect exfoliated TMDs used in electronic devices to have an appreciable concentration of surface adsorbates, unless special processing to desorb/remove such species is undertaken. The effects of carbon present at the contact metal-TMD interface, which are not found to be carbidic, should be considered in addition to the true chemistry of the interface when evaluating electrical properties of both the contact and the device.


Page 29 of 49 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 7. The band alignment of metal-MoSe2 and metal-MoTe2 systems obtained from the initial valence band offset of the exfoliated TMD and the bulk TMD core level shift after depositing ~1 nm metal in UHV, which is ascribed to a EF shift (see Supporting Information for band diagram construction procedure and Figure S9 for corresponding valence band spectra used to derive the band diagrams in Figure 7). Metals that react with Mo-based TMDs (Ir, Cr, Sc, Ti) distort the TMD structure, especially in the monolayer limit. A sacrificial TMD layer (or layers) may be necessary in certain devices that rely on the integrity of a monolayer channel for successful operation. Using a similar philosophy, Rc has been significantly reduced in graphene transistors when carbon from graphene diffuses into a Ni or Co contact.76 However, it has yet to be determined if an analogous mechanism in TMD devices is desirable to improve Rc. Further investigation is necessary to determine the most desirable interface chemistry (vdW gap, interfacial covalent bonding, complete underlying TMD consumption, etc.) and contact architecture (edge, top, mixed, vdW, covalent) to achieve an acceptable Rc.

Conclusions



Page 30 of 49

In this work, we have investigated the interface chemistry formed between various contact metals (Au, Ir, Cr, and Sc) and exfoliated, bulk TMDs (MoSe2 and MoTe2) as a function of the deposition chamber ambient. The metal-TMD reaction thermodynamics are correlated with the degree of reactivity detected in each metal-TMD system. The XPS-derived band alignment formed between each metal and TMD show the interfacial intermetallic and oxide species cause strong EF pinning according to pinning factors of S = 0.14 and S = 0.18 calculated from the metal-MoSe2 and metalMoTe2 systems, respectively. The deposition chamber ambient is shown to have a significant effect on the compounds that are present at the metal-TMD interface and within the metal contact away from the interface. Reaction products are detected at all Ir-, Cr-, and Sc-TMD interfaces investigated here (except the Sc-MoSe2 system) regardless of the deposition chamber ambient. Au forms a vdW interface with MoSe2. In contrast, Au forms an intermetallic when deposited on MoTe2, which is likely facilitated by the excess Te detected in all MoTe2 samples investigated in this work. Cr is partially oxidized and Sc is completely oxidized when deposited in HV. In addition, the pinholes observed in the Cr film deposited on MoSe2 and MoTe2 in this work highlight the need to carefully design and fabricate contact structures including a Cr adhesion layer to achieve the desired performance. Our findings elucidate the critical relationships between metalTMD interface chemistry and band alignment. Critically, the interface chemistry between many contact metals and Mo-based TMDs varies significantly depending on the deposition chamber ambient, contact metal, and quality of the TMD.

Conflict of Interest. The authors declare no competing financial interest. Acknowledgement. The authors thank Prof. S. McDonnell at the University of Virginia for experimental advice and useful discussions. This work was supported in part by NSF Award No.


Page 31 of 49 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


1407765 under the US/Ireland UNITE collaboration and by the Semiconductor Research Corporation (SRC) as the NEWLIMITS Center and NIST through award number 70NANB17H041. Supporting Information. Calculating stoichiometry using atomic sensitivity factors; Detailed discussion of redundant core level spectra and AFM images obtained from the Au-MoSe2, IrMoTe2, Cr-MoTe2, and Sc-MoTe2 systems as well as the evolution of oxygen and carbon chemical states in each metal-TMD system; Constructing the metal-TMD band alignment from core level and valence band spectra

References (1) Cho, S.; Kim, S.; Kim, J. H.; Zhao, J.; Seok, J.; Keum, D. H.; Baik, J.; Choe, D. H.; Chang, K. J.; Suenaga, K.; et al. Phase Patterning for Ohmic Homojunction Contact in MoTe2. Science. 2015, 349, 625-628. (2) Chamlagain, B.; Li, Q.; Ghimire, N. J.; Chuang, H.-J.; Perera, M. M.; Tu, H.; Xu, Y.; Pan, M.; Xaio, D.; Yan, J.; et al. Mobility Improvement and Temperature Dependence of MoSe2 FieldEffect Transistors on Parylene-C Substrate. ACS Nano. 2014, 8, 5079-5088. (3) Liu, T.; Xiang, D.; Zheng, Y.; Wang, Y.; Wang, X.; Wang, L.; He, J.; Liu, L.; Chen, W. Nonvolatile and Programmable Photodoping in MoTe2 for Photoresist-Free Complementary Electronic Devices. Adv. Mater. 2018, 30, 1804470. (4) Zhang, F.; Zhang, H.; Krylyuk, S.; Milligan, C. A.; Zhu, Y.; Zemlyanov, S. Y.; Bendersky, L. A.; Burton, B. P.; Davydov, A. V.; Appenzeller, J. Electric-Field Induced Structural Transition in Vertical MoTe2- and Mo1-xWxTe2-Based Resistive Memories. Nat. Mater. 2019, 18, 55-61. (5) Wang, T.-H.; Jeng, H.-T. Wide-Range Ideal 2D Rashba Electron Gas with Large Spin Splitting in Bi2Se3/MoTe2 Heterostructure. npj Comput. Mater. 2017, 3, 5. (6) Bie, Y.-Q.; Grosso, G.; Heuck, M.; Furchi, M. M.; Cao, Y.; Zheng, J.; Bunandar, D.; NavarroMoratalla, E.; Zhou, L.; Efetov, D. K.; et al. A MoTe2-Based Light-Emitting Diode and Photodetector for Silicon Photonic Integrated Circuits. Nat. Nanotechnol. 2017, 12, 1124. (7) Jiang, C.; Liu, F.; Cuadra, J.; Huang, Z.; Li, K.; Rasmita, A.; Srivastava, A.; Liu, Z.; Gao, W.-B. Zeeman Splitting via Spin-Valley-Layer Coupling in Bilayer MoTe2. Nat. Commun. 2017, 8, 802.



Page 32 of 49

(8) Kumar, N.; He, J.; He, D.; Wang, Y.; Zhao, H. Valley and Spin Dynamics in MoSe2 TwoDimensional Crystals. Nanoscale. 2014, 6, 12690-12695. (9) Liu, X.; Yang, H.; He, J.; Liu, H.; Song, L.; Li, L.; Luo, J. Highly Active, Durable Ultrathin MoTe2 Layers for the Electroreduction of CO2 to CH4. Small. 2018, 14, 1704049. (10) Wang, Q.-H.; Kalantar-Zadeh, K.; Kis, A.; Coleman, J. N.; Strano, M. S. Electronics and Optoelectronics of Two-Dimensional Transition Metal Dichalcogenides. Nat. Nanotechnol. 2012, 7, 699-712. (11) Pradhan, N. R.; Rhodes, D.; Feng, S.; Xin, Y.; Memaran, S.; Moon, B.-H.; Terrones, H.; Terrones, M.; Balicas, L. Field Effect Transistors Based on Few-Layered α-MoTe2. ACS Nano. 2014, 8, 5911-5920. (12) Conan, A.; Bonnet, A.; Zoaeter, M.; Ramoul, D. Dependence of the Total Mobility in a OneBand Model Application to n-Type MoTe2. Phys. Status Solidi B. 1984, 124, 403-410. (13) Huang, H.; Wang, J.; Hu, W.; Liao, L.; Wang, P.; Wang, X.; Gong, F.; Chen, Y.; Wu, G.; Luo, W. Highly Sensitive Visible to Infrared MoTe2 Photodetectors Enhanced by the Photogating Effect. Nanotechnology. 2016, 27, 445201. (14) Allain, A.; Kang, J.; Banerjee, K.; Kis, A. Electrical Contacts to Two-Dimensional Materials. Nat. Mater. 2015, 14, 1195-1205. (15) Das, S.; Chen, H. Y.; Penumatcha, A. V.; Appenzeller, J. High Performance Multilayer MoS2 Transistors with Scandium Contacts. Nano Lett. 2013, 13, 100-105. (16) English, C. D.; Shine, G.; Dorgan, V. E.; Saraswat, K. C.; Pop, E. Improved Contacts to MoS2 Transistors by Ultra-High Vacuum Metal Deposition. Nano Lett. 2016, 16, 3824-3830. (17) Kiriya, D.; Tosun, M.; Zhao, P.; Kang, J. S.; Javey, A. Air-Stable Surface Charge Transfer Doping of MoS2 by Benzyl Viologen. J. Am. Chem. Soc. 2014, 136, 7853-7856. (18) Yu, L.; Zubair, A.; Santos, E. J. G.; Zhang, X.; Lin, Y.; Zhang, Y.; Palacios, T. HighPerformance WSe2 Complementary Metal Oxide Semiconductor Technology and Integrated Circuits. Nano Lett. 2015, 15, 4928-4934. (19) Kaushik, N.; Karmakar, D.; Nipane, A.; Karande, S.; Lodha, S. Interfacial n-Doping Using an Ultrathin TiO2 Layer for Contact Resistance Reduction in MoS2. ACS Appl. Mater. Interfaces. 2016, 8, 256-263. (20) Ghiasi, T. S.; Quereda, J.; van Wees, B. J. Bilayer h-BN Barriers for Tunneling Contacts in Fully-Encapsulated Monolayer MoSe2 Field-Effect Transistors. 2D Mater. 2019, 6, 015002. (21) Cheng, F.; Hu, Z.; Xu, H.; Shao, Y.; Su, J.; Chen, Z.; Ji, W.; Loh, K. P. Interface Engineering of Au(111) for the Growth of 1T’-MoSe2. ACS Nano. 2019, In Press, DOI: 10.1021/acsnano.8b09054.


Page 33 of 49 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


(22) Sung, J. H.; Heo, H.; Si, S.; Kim, Y. H.; Noh, H. R.; Song, K.; Kim, J.; Lee, C.-S.; Seo, S.Y.; Kim, D.-H.; et al. Coplanar Semiconductor–Metal Circuitry Defined on Few-Layer MoTe2 via Polymorphic Heteroepitaxy. Nat. Nanotechnol. 2017, 12, 1064-1070. (23) Mleczko, M. J.; Yu, A. C.; Smyth, C. M.; Chen, V.; Shin, Y. C.; Chatterjee, S.; Tsai, Y.-C.; Nishi, Y.; Wallace, R. M.; Pop, E. Contact Engineering High Performance n-Type MoTe2 Transistors. Nano Lett. 2019 Accepted; DOI:10.1021/acs.nanolett.9b02497. (24) Jena, D.; Banerjee, K.; Xing, G. H. 2D Crystal Semiconductors: Intimate Contacts. Nature. 2014, 13, 1076-1078. (25) Kim, G.-S.; Kim, S.-H.; Park, J.; Han, K. H.; Kim, J.; Yu, H.-Y. Schottky Barrier Height Engineering for Electrical Contacts of Multilayered MoS2 Transistors with Reduction of MetalInduced Gap States. ACS Nano. 2018, 12, 6292-6300. (26) Townsend, N. J.; Amit, I.; Craciun, M. F.; Russo, S. Sub 20 meV Schottky Barriers in Metal/MoTe2 Junctions. 2D Mater. 2018, 5, 025023. (27) Liu, Y.; Guo, J.; Zhu, E.; Liao, L.; Lee, S.-J.; Ding, M.; Shakir, I.; Gambin, V.; Huang, Y.; Duan, X. Approaching the Schottky-Mott Limit in van der Waals Metal-Semiconductor Junctions. Nature. 2018, 557, 696-700. (28) McDonnell, S.; Addou, R.; Buie, C.; Wallace, R. M.; Hinkle, C. L. Defect Dominated Doping and Contact Resistance in MoS2. ACS Nano. 2014, 8, 2880-2888. (29) Gong, C.; Colombo, L.; Wallace, R. M.; Cho, K. The Unusual Mechanism of Partial Fermi Level Pinning at Metal-MoS2 Interfaces. ACS Nano. 2014, 14, 1714-1720. (30) Kim, C.; Moon, I.; Lee, D.; Choi, M. S.; Ahmed, F.; Nam, S.; Cho, Y.; Shin, H.-J.; Park, S.; Yoo, W. J. Fermi Level Pinning at Electrical Metal Contacts of Monolayer Molybdenum Dichalcogenides. ACS Nano. 2017, 11, 1588-1596. (31) Ҫakir, D.; Peeters, F. M. Dependence of the Electronic and Transport Properties of MetalMoSe2 Interfaces on Contact Structures. Phys. Rev. B. 2014, 89, 245403. (32) Pan, Y.; Li, S.; Ye, M.; Quhe, R.; Song, Z.; Wang, Y.; Zheng, J.; Pan, F.; Guo, W.; Yang, J.; et al. Interfacial Properties of Monolayer MoSe2-Metal Contacts. J. Phys. Chem. C. 2016, 120, 13063-13070. (33) Zhu, H.; Wang, Q.; Cheng, L.; Addou, R.; Kim, J.; Kim, M. J.; Wallace, R. M. Defects and Surface Structural Stability of MoTe2 Under Vacuum Annealing. ACS Nano. 2017, 11, 1100511014. (34) Guo, Y.; Liu, D.; Robertson, J. 3D Behavior of Schottky Barriers of 2D Transition-Metal Dichalcogenides. ACS Appl. Mater. Interfaces. 2015, 7, 25709-25715.



Page 34 of 49

(35) Addou, R.; McDonnell, S.; Barrera, D.; Guo, Z.; Azcatl, A.; Wang, J.; Zhu, H.; Hinkle, C. L.; Quevedo-Lopez, M.; Alshareef, H. N.; et al. Impurities and Electronic Property Variations of Natural MoS2 Crystal Surfaces. ACS Nano. 2015, 9, 9124-9133. (36) Chen, J.; Feng, Z.; Fan, S.; Shi, S.; Yue, Y.; Shen, W.; Xie, Y.; Wu, E.; Sun, C.; Liu, J.; et al. Contact Engineering of Molybdenum Ditelluride Field Effect Transistors Through Rapid Thermal Annealing. ACS Appl. Mater. Interfaces. 2017, 9, 30107-30114. (37) Bampoulis, P.; van Bremen, R.; Yao, Q.; Poelsema, B.; Zandvliet, H. J. W.; Sotthewes, K. Defect Dominated Charge Transport and Fermi Level Pinning in MoS2/Metal Contacts. ACS Appl. Mater. Interfaces. 2017, 9, 19278-19286. (38) Zhang, S.-L.; Smith, U. Self-Aligned Silicides for Ohmic Contacts in Complementary MetalOxide-Semiconductor Technology: TiSi2, CoSi2, and NiSi. J. Vac. Sci. Technol. A. 2004, 22, 1361. (39) Smyth, C. M.; Addou, R.; McDonnell, S.; Hinkle, C. M.; Wallace, R. M. Contact Metal-MoS2 Interfacial Reactions and Potential Implications on MoS2-Based Device Performance. J. Phys. Chem. C. 2016, 120, 14719-14729. (40) Smyth, C. M.; Addou, R.; McDonnell, S.; Hinkle, C. L.; Wallace, R. M. WSe2-Contact Metal Interface Chemistry and Band Alignment under High Vacuum and Ultra High Vacuum Deposition Conditions. 2D Mater. 2017, 4, 025084. (41) McDonnell, S.; Smyth, C. M.; Hinkle, C. L.; Wallace, R. M. MoS2-Titanium Contact Interface Reactions. ACS Appl. Mater. Interfaces. 2016, 8, 8289-8294. (42) Michaelson, H. B. The Work Function of the Elements and Its Periodicity. J. Appl. Phys. 1977, 48, 4729-4733. (43) Bulk MoSe2 and MoTe2 crystals, HQ Graphene, 2016; www.hqgraphene.com (44) Wallace, R. M. In-Situ Studies of Interfacial Bonding in High-κ Dielectrics for CMOS Beyond 22 nm. In Physics and Technology of High-K Gate Dielectrics 6, Kar, S., Landheer, D., Houssa, M., Misra, D., VanElshocht, S., Iwai, H., Eds.; The Electrochemical Society: Pennington, NJ, 2008; 16, pp 255-271. (45) ASTM E2108 - 10 Standard Practice for Calibration of the Electron Binding-Energy Scale of an X-Ray Photoelectron Spectrometer, 2000. (46) Herrera-Gomez, A.; Hegedus, A.; Meissner, P. L. Chemical Depth Profile of Ultrathin Nitrided SiO2 Films. Appl. Phys. Lett. 2002, 81, 1014-1016. (47) University of Texas at Dallas Cleanroom http://www.utdallas.edu/research/cleanroom/ (Accessed March 25, 2018). (48) Horcas, I.; Fernández, R.; Gómez-Rodríguez, J. M.; Colchero, J.; Gómez-Herrero, J.; Baro, A. M. WSXM: A Software for Scanning Probe Microscopy and a Tool for Nanotechnology. Rev. Sci. Instrum. 2007, 78. 34 ACS Paragon Plus Environment

Page 35 of 49 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


(49) Amani, M.; Taheri, P.; Addou, R.; Ahn, G. H.; Kiriya, D.; Lien, D.-H.; Ager III, J. W.; Wallace, R. M.; Javey, A. Recombination Kinetics and Effects of Superacid Treatment in Sulfurand Selenium-based Transition Metal Dichalcogenides. Nano Lett. 2016, 16, 2786-2791. (50) Ake, O.; Nolang, B.; Osadchii, E. G.; Ohman, L-O,; Rosen, E. In Chemical Thermodynamics of Selenium; Mompean, F. J.; Perrone, J.; Illemassene, M., Eds.; NEA Data Bank: Issy-lesMoulineaux, France, 2004; Vol. 7. (51) All ΔG°f values given in this work are per anionic constituent. (52) Bortz, M. L.; Ohuchi, F. S.; Parkinson, B. A. An Investigation of the Growth of Au and Cu on the van der Waals Surfaces of MoTe2 and WTe2. Surf. Sci. 1989, 223, 285-298. (53) Zhu, H. Surface and Interface Characterization of 2D Materials: Transition Metal Dichalcogenide and Black Phosphorous. Ph.D. Thesis, The University of Texas at Dallas, Dallas, TX, 2017. (54) Nekrasov, I. A. Geochemistry, Mineralogy, and Genesis of Gold Deposits. Balkema Publishers: Brookfield, VT, 1996, 254-255. (55) Chemical bonding overview, www.chemhume.co.uk; 2015. (56) All electronegativity values reported here are referenced from the Pauling electronegativity scale. (57) Luo, Y-R. Comprehensive Handbook of Chemical Bond Energies; CRC Press, Taylor & Francis Group: Boca Raton, FL, 2007. (58) Westrum, E. F.; Carlson, H. G. Low Temperature Heat Capacities and Thermodynamic Functions of Some Palladium and Platinum Group Chalcogenides. II. Dichalcogenides; PtS2, PtTe2, and PdTe2. J. Chem. Phys. 1961, 35, 1670-1676. (59) Mohanty, B. C.; Malar, P.; Osipowicz, T.; Murty, B. S.; Varma, S.; Kasiviswanathan, S. Characterization of Silver Selenide Thin Films Grown on Cr-Covered Si Substrates. Surf. Interface Anal. 2009, 41, 170-178. (60) Viswanathan, R.; Sai Baba, M.; Raj, D. D. A.; Balasubramanian, R.; Saha, B.; Mathews, C. K. Thermochemistry of Metal-Rich Chromium Telluride and its Role in Fuel-Clad Chemical Interactions. J. Nucl. Mater. 1989, 67, 94-104. (61) Mallika, C.; Sreedharan, O. M. Standard Molar Gibbs Energy of Formation of MoTe2 from E.M.F. Measurements. J. Chem. Thermodyn. 1988, 20, 769-775. (62) Doniach, S.; Sunjic, M. Many-Electron Singularity in X-ray Photoemission and X-ray Line Spectra from Metals. J. Phys. C: Solid State Phys. 1970, 3, 285-291.



Page 36 of 49

(63) Roy, A.; Guchhait, S.; Pramanik, T.; Hsieh, C.-C.; Rai, A.; Banerjee, S. K. Perpendiular Magnetic Anisotropy and Spin Glass-like Behavior in Molecular Beam Epitaxy Grown Chromium Telluride Thin Films. ACS Nano. 2015, 9, 3772-3779. (64) Reed, T. B. Free Energy of Formation of Binary Compounds: An Atlas of Charts for High Temperature Chemical Calculations. The MIT Press: Cambridge, MA, 1971. (65) Selected Values of Chemical Thermodynamic Properties. In CRC Handbook of Chemistry and Physics 70th Edition; Weast, R. C.; Lide, D. R.; Astle, M. J.; Beyer, W. H.; CRC Press Inc.: Boca Raton, FL, 1989-1990. (66) Galkin, N. G.; Goroshko, D. L.; Sergey, A.; Turchin, T. V.; Self-Organization of CrSi2 Nanoislands on Si(111) and Growth of Monocrystalline Silicon with Buried Multilayers of CrSi2 Nanocrystallites. J. Nanosci. Nanotechnol. 2008, 8, 557-563. (67) Smyth, C. M.; Walsh, L. A.; Bolshakov, P.; Catalano, M.; Schmidt, M.; Sheehan, B.; Addou, R.; Wang, L.; Kim, J.; Kim, M. J.; et al. Engineering the Interface Chemistry for Scandium Electron Contacts in WSe2 Transistors and Diodes. 2D Mater. 2019, In Press, DOI: 10.1088/20531583/ab2c44. (68) Waldner, P. Thermodynamic Modeling of the Cr-Fe-S System. Metall. Mater. Trans. A. 2014, 45, 798-814. (69) Polyakov, A. Y.; Smirnov, N. B.; Govorkov, A. V.; Danilin, V. N.; Zhukova, T. A.; Luo, B.; Ren, F.; Gila, B. P.; Onstine, A. H.; Abernathy, C. R.; et al. Deep Traps in Unpassivated and Sc2O3–Passivated AlGaN/GaN High Electron Mobility Transistors. Appl. Phys. Lett. 2003, 83, 2608. (70) Maggard, P. A.; Corbett, J. D. The Synthesis, Structure, and Bonding of Sc8Te3 and Y8Te3. Cooperative Matrix and Bonding Effects in the Solid State. Inorg. Chem. 1998, 37, 814-820. (71) Li, Y.-X.; Putungan, D. B.; Lin, S.-H. Two-Dimensional MTe2 (M = Co, Fe, Mn, Sc, Ti) Transition Metal Tellurides as Sodium Ion Battery Anode Materials: Density Functional Theory Calculations. Phys. Lett. A. 2018, 382, 2781-2786. (72) Maggard, P. A.; Corbett, J. D. Sc9Te2: A Two-Dimensional Distortion Wave in the ScandiumRichest Telluride. J. Am. Chem. Soc. 2000, 122, 838-843. (73) Brainard, W. A. The Thermal Stability and Friction of the Disulfides, Diselenides, and Ditellurides of Molybdenum and Tungsten in Vaccum (10-9 to 10-6 Torr). 1969, NASA TN D5141. (74) Freedy, K. M.; Giri, A.; Foley, B. M.; Barone, M. R.; Hopkins, P. E.; McDonnell, S. Titanium Contacts to Graphene: Process-Induced Variability in Electronic and Thermal Transport. Nanotechnology. 2018, 29, 145201. (75) Jaegermann, W.; Pettenkofer, C.; Parkinson, B. A. Cu and Ag Deposition on Layered p-Type WSe2: Approaching the Schottky Limit. Phys. Rev. B. 1990, 42, 7487-7496. 36 ACS Paragon Plus Environment

Page 37 of 49 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


(76) Leong, W. S.; Gong, H.; Thong, J. T. L. Low-Contact-Resistance Graphene Devices with Nickel-Etched-Graphene Contacts. ACS Nano. 2014, 8, 994-1001.

TOC Graphic



Figure 1. a) Mo 3d, Se 3s, Se 3d, Au 5p3/2 and b) Au 4f core level spectra obtained from bulk MoSe2 after

exfoliation and subsequent Au deposition in UHV (in–situ) and HV (ex–situ). c) 500 nm2 AFM image obtained ex–situ after ~1 nm Au deposition on MoSe2. A line profile obtained from the AFM image is overlaid to show the surface topography. 119x114mm (300 x 300 DPI)

ACS Paragon Plus Environment

Page 38 of 49

Page 39 of 49 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 2. a) Te 3d5/2, Mo 3d, and c) Au 4f core level spectra obtained from bulk MoTe2 after exfoliation and subsequent Au deposition in UHV (in–situ) and HV (ex–situ). b) 500 nm2 AFM image obtained ex–situ after ~1 nm Au deposition on MoTe2. A line profile obtained from the AFM image is overlaid to clearly show the surface topography of the Au on MoTe2. 129x150mm (300 x 300 DPI)



Figure 3. a) Mo 3d, Se 3s, Se 3d, and b) Ir 4f core level spectra obtained from bulk MoSe2 after exfoliation and subsequent ~1 nm Ir deposition in UHV (in–situ) and HV (ex–situ). 169x85mm (300 x 300 DPI)


Page 40 of 49

Page 41 of 49 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 4. a) Te 3d5/2, Mo 3d, and b) Ir 4f core level spectra obtained from bulk MoTe2 after exfoliation and subsequent ~1 nm Ir deposition in UHV (in–situ) and HV (ex–situ). 169x85mm (300 x 300 DPI)



Figure 5. a) Mo 3d, Se 3s, Se 3d, and b) Cr 2p3/2 core level spectra obtained from bulk MoSe2 after exfoliation and subsequent ~1 nm Cr deposition in UHV (in–situ) and HV (ex–situ).The Cr 2p3/2 spectra in b) are normalized to the metallic Cr chemical state, Cr0. 169x85mm (300 x 300 DPI)


Page 42 of 49

Page 43 of 49 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 6. a) Cr 2p3/2, Te 3d5/2, and Mo 3d core level spectra obtained from bulk MoTe2 after exfoliation and subsequent ~1 nm Cr deposition in UHV (in–situ) and HV (ex–situ). The spectra in the left panel of a) are normalized to the Te 3d5/2 core level. 124x82mm (300 x 300 DPI)



Figure 7. 1 μm2 AFM images and associated line profiles (below) obtained ex–situ from bulk MoSe2 and MoTe2 crystals after depositing ~1 nm Cr in HV. 244x200mm (96 x 96 DPI)


Page 44 of 49

Page 45 of 49 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 8. a) Mo 3d, Se 3s, Se 3d, Sc 3p, b) Sc 2p, and Mo 3p core level spectra obtained from MoSe2 after exfoliation and subsequent ~1 nm Sc deposition in UHV (in–situ) and HV (ex–situ). The spectra in b) are normalized to the Sc 2p3/2 core level. 169x85mm (300 x 300 DPI)



Figure 9. a) Te 3d5/2, Mo 3d, b) Sc 2p, and Mo 3p core level spectra obtained from MoTe2 after exfoliation and subsequent ~1 nm Sc deposition in UHV (in–situ) and HV (ex–situ). 169x85mm (300 x 300 DPI)


Page 46 of 49

Page 47 of 49 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 10. a) Reactivity of the metal-TMD systems discussed in this work. Metal-MoS2 data is included from our previous work for reference. The reactivity is gauged according to the intensity ratio of the contact metal–chalcogenide (e.g., ScSex) intermetallic chemical state in the corresponding chalcogen core level (MoSe2: Se 3d, MoTe2: Te 3d5/2) to the sum of all chemical states in the same chalcogen core level. These are plotted versus the Ir, Au, Cr, and Sc work functions for metal depositions in UHV (in-situ). The ‘intensity’ refers to the integrated photoelectron intensity. b) A scatter plot showing the known Gibbs free energy for each metal chalcogenide alloy and relevant metal oxides as a function of metal vacuum work function. The specific compounds are listed on the plot. The dotted lines represent the ΔG°f,MoSe2, ΔG°f,MoTe2, ΔG°f,MoO3, ΔG°f,TeO2, and ΔG°f,SeO. 150x70mm (300 x 300 DPI)



Figure 11. The band alignment of metal-MoSe2 and metal-MoTe2 systems obtained from the initial valence band offset of the exfoliated TMD and the bulk TMD core level shift after depositing ~1 nm metal in UHV, which is ascribed to a EF shift (see Supporting Information for band diagram construction procedure and Figure S9 for corresponding valence band spectra used to derive the band diagrams in Figure 11). 61x50mm (300 x 300 DPI)


Page 48 of 49

Page 49 of 49 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


ToC Figure 82x44mm (300 x 300 DPI)


Origins of Fermi Level Pinning between Molybdenum

Recommend Documents