Research Article Cite This: ACS Appl. Mater. Interfaces 2019, 11, 32144−32150
www.acsami.org
Impact of Etch Processes on the Chemistry and Surface States of the Topological Insulator Bi2Se3 Adam T. Barton,†,∥ Lee A. Walsh,†,‡,∥ Christopher M. Smyth,† Xiaoye Qin,† Rafik Addou,† Christopher Cormier,† Paul K. Hurley,‡ Robert M. Wallace,† and Christopher L. Hinkle*,†,§ †
Department of Materials Science and Engineering, University of Texas at Dallas, Richardson, Texas 75080, United States Tyndall National Institute, University College Cork, Lee Maltings Complex, Cork T12R5CP, Ireland § Department of Electrical Engineering, University of Notre Dame, Notre Dame, Indiana 46556, United States Downloaded via NOTTINGHAM TRENT UNIV on September 6, 2019 at 12:37:05 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
‡
ABSTRACT: The unique properties of topological insulators such as Bi2Se3 are intriguing for their potential implementation in novel device architectures for low power and defect-tolerant logic and memory devices. Recent improvements in the synthesis of Bi2Se3 have positioned researchers to fabricate new devices to probe the limits of these materials. The fabrication of such devices, of course, requires etching of the topological insulator, in addition to other materials including gate oxides and contacts which may impact the topologically protected surface states. In this paper, we study the impact of He+ sputtering and inductively coupled plasma Cl2 and SF6 reactive etch chemistries on the physical, chemical, and electronic properties of Bi2Se3. Chemical analysis by X-ray photoelectron spectroscopy tracks changes in the surface chemistry and Fermi level, showing preferential removal of Se that results in vacancy-induced n-type doping. Chlorine-based chemistry successfully etches Bi2Se3 but with residual Se−Se bonding and interstitial Cl species remaining after the etch. The Se vacancies and residuals can be removed with postetch anneals in a Se environment, repairing Bi2Se3 nearly to the as-grown condition. Critically, in each of these cases, angle-resolved photoemission spectroscopy (ARPES) reveals that the topologically protected surface states remain even after inducing significant surface disorder and chemical changes, demonstrating that topological insulators are quite promising for defect-tolerant electronics. Changes to the ARPES intensity and momentum broadening of the surface states are discussed. Fluorine-based etching aggressively reacts with the film resulting in a relatively thick insulating film of thermodynamically favored BiF3 on the surface, prohibiting the use of SF6-based etching in Bi2Se3 processing. KEYWORDS: topological insulator, bismuth selenide, surface states, process integration, Dirac cone, etch chemistry
■
damage.7,8 However, there is a limit to this defect tolerance. Previous studies have shown that at some critical defect concentration, the disorder in the near-surface region can significantly alter the band structure, effectively forming a different material at the extreme limit, and causing the surface states to recede to the next topmost quintuple layer (in the case of 3D TIs).9−11 Before Bi2Se3 can be integrated into process flows that include etching steps, the effects of processinduced damage must be understood. While there have been a few studies focused on the impact of (nonmagnetic) defects on the surface states of TIs, those studies did not investigate specifically how the damage impacts the structure and chemistry of the surface.12,13 Furthermore, there have been no publications on the impact of dry-etching Bi2Se3 with reactive precursors that are common in advanced device processing. Because of the ease of sputtering through
INTRODUCTION
Topological insulators (TIs) are a unique class of material that not only possess narrow, semiconducting “bulk” band gaps but also have gapless edge or surface states (in 2D or 3D TIs, respectively)1−3 that are protected by time-reversal symmetry.4,5 These topologically protected states have their spin locked to the direction of the momentum, potentially enabling spin-polarized current and defect-tolerant transport. Bi2Se3 is the most widely studied TI, and recent progress in the growth of Bi2Se3, including the reduction of parasitic bulk conduction caused by point defects,6 has finally positioned researchers to synthesize unique devices based on the topologically protected states. During device processing, there are often a number of etching and sputtering steps that are used to form and isolate the devices. Although necessary for complex device fabrication, these destructive processes can also have the adverse effect of inducing damage in the underlying materials. Research has shown that the surface states of 3D TIs are robust to nonmagnetic defects and should persist even after significant © 2019 American Chemical Society
Received: June 17, 2019 Accepted: August 16, 2019 Published: August 16, 2019 32144
DOI: 10.1021/acsami.9b10625 ACS Appl. Mater. Interfaces 2019, 11, 32144−32150
Research Article
ACS Applied Materials & Interfaces van der Waals chalcogenides with ionized noble gasses, we first investigate the impact of Bi2Se3 damaged by He+ ion sputtering. If Bi 2 Se 3 is utilized in a van der Waals heterostructure, ionized sputtering with noble gasses may be sufficient for shaping the device features. However, if more robust materials are utilized in the device structure (i.e., Fe, Co, Ni, etc.), it may be necessary to use more reactive etching processes.14−16 Therefore, the use of inductively coupled plasma (ICP) etching using Cl2 and SF6 reactive species was also investigated. Using angle-resolved photoemission spectroscopy (ARPES), X-ray photoelectron spectroscopy (XPS), scanning tunneling spectroscopy (STS), and atomic force microscopy (AFM), we investigate the surface chemistry, physical damage, and changes in the electronic structure induced by these destructive processes and provide a guide for minimizing and, in some cases, even reversing some of the adverse effects.
■
same decapping procedure as previously mentioned was used in the ARPES preparation chamber to recover the bare Bi2Se3 surface. For the XPS characterization of the ICP-etched films (the XPS tool is in the same building as the ICP etchers at UT Dallas), the samples were transferred directly from the ICP etcher to the XPS system, while allowing the same amount of atmospheric exposure as the ARPES samples. This resulted in minor atmospheric oxidation.6 XPS Characterization and Peak Fitting. XPS for all samples (except the SF6-etched sample as described later) was carried out using a monochromated Al Kα source (E = 1486.7 eV) and an Omicron EA125 hemispherical analyzer with a resolution of ±0.05 eV. An analyzer acceptance angle of 8°, takeoff angle of 45°, and pass energy of 15 eV were utilized in this study. This tool is described in detail elsewhere.20 The analyzer was calibrated using sputter-cleaned Au, Cu, and Ag foils, as is outlined in ASTM E2108. The stoichiometries extracted from XPS are calculated using the appropriate relative sensitivity factors for the Bi 5d, Se 3d, O 1s, Cl 2p, F 1s, and S 2s core levels (1.259, 0.722, 0.711, 0.77, 1.000, and 0.399, respectively).21 The stoichiometry ratios calculated are accompanied by a ±0.2 error. Owing to the need for a charge neutralizer, the XPS for the SF6etched sample was scanned in an Omicron ultrahigh vacuum (UHV) system equipped with scanning tunneling microscopy (STM) and XPS running at a base pressure of ∼1 × 10−10 mbar. A monochromatic Al Kα source (E = 1486.7 eV) and an Omicron Argus detector operating with a pass energy of 15 eV have been used with a spot size of 630 μm2. The analyzer spectrometer energy scale was calibrated using sputter-cleaned Au, Ag, and Cu specimens as indicated in the ASTM E2108 procedure. The core-level spectra were deconvoluted using the curve-fitting software AAnalyzer.22 Metallic chemical states were fit with the asymmetric double Lorentzian line shape, while nonmetallic chemical states were fit with Voigt line shapes. An active Shirley background subtraction was employed in fitting all spectra.22 In order to accurately detect the presence of additional features in any of the core levels, all fits were performed with comparison to reference samples, that is, using the bare Bi2Se3 after removal of the Se cap and oxidized Bi2Se3 after one week of atmospheric exposure. The peak separations and full-width at halfmaximum of the reference peaks are kept constant to maintain consistency. STS Characterization. Room-temperature STS spectra were obtained using a VT-STM with an etched tungsten tip in the Omicron UHV system. STS data were collected at a single position and averaged from at least 30 sweeps to obtain a single curve.23 Then, a direct differentiation of the I−V curves was plotted in a linear or log scale. ARPES Measurements. ARPES measurements were performed at beamline 5-4 of the Stanford Synchrotron Light Source (SSRL) with a Scienta R4000 electron analyzer.18 All of the data shown was acquired using 23 eV photon energy with total energy and angular momentum resolutions of 10 meV and 0.25°, respectively. The decapping process was accomplished by heating the samples to 170 °C in UHV for 1 h. The temperatures during the measurement were acquired close to the sample position with an accuracy of ±2 °C. The samples were then cooled down to room temperature before transfer to the ARPES analysis chamber. The ARPES measurements were performed at 20 K in a UHV chamber with a base pressure better than 4 × 10−11 mbar. Measurements on each sample were completed within less than 10 h of Se cap removal. Post Sputter/Etch Se Anneals. The samples etched with the Cl2/Ar precursors were reintroduced into the MBE chamber used for growth and annealed under a Se flux pressure of 4.5 × 10−7 mbar at a temperature of 250 °C for 1 h. The base pressure was maintained below 1 × 10−9 mbar during the anneal, and the sample was transferred to the XPS system immediately after growth without Se capping. The transferring of samples was timed such that each sample used for comparison saw the same amount of atmospheric exposure (
