Structure and Stability of the Twofold Surface of Icosahedral Al-Pd-Mn by Low-Energy Electron Diffraction and X-Ray Photoemission Spectroscopy

We have used low-energy electron diffraction and x-ray photoemission spectroscopy to investigate the structure of the twofold surface of icosahedral Al-Pd-Mn. The regrowth of the surface by annealing after sputtering took place in two distinct stages. The first stage was the appearance of a fine-grained surface phase with icosahedral, or near-icosahedral, symmetry. For higher annealing temperatures (above 800 K) a bulk terminated face-centered icosahedral surface was observed. Disciplines Biological and Chemical Physics | Materials Science and Engineering | Physical Chemistry Comments This article is from Physical Review Letters 78, no. 6 (1997): 1050–1053, doi:10.1103/PhysRevLett.78.1050. Authors Z. Shen, Cynthia J. Jenks, James W. Anderegg, D. W. Delaney, Thomas A. Lograsso, Patricia A. Thiel, and A. I. Goldman This article is available at Iowa State University Digital Repository: http://lib.dr.iastate.edu/ameslab_pubs/20 VOLUME 78, NUMBER 6 P H Y S I C A L R E V I E W L E T T E R S 10 FEBRUARY 1997 Structure and Stability of the Twofold Surface of Icosahedral Al-Pd-Mn by Low-Energy Electron Diffraction and X-Ray Photoemission Spectroscopy Z. Shen,* C. J. Jenks,* J. Anderegg,* D. W. Delaney, T. A. Lograsso, P. A. Thiel,* and A. I. Goldman Ames Laboratory, USDOE and Iowa State University, Ames, Iowa 50011 (Received 7 November 1996) We have used low-energy electron diffraction and x-ray photoemission spectroscopy to investigate the structure of the twofold surface of icosahedral Al-Pd-Mn. The regrowth of the surface by annealing after sputtering took place in two distinct stages. The first stage was the appearance of a fine-grained surface phase with icosahedral, or near-icosahedral, symmetry. For higher annealing temperatures (above 800 K) a bulk terminated face-centered icosahedral surface was observed. [S0031-9007(97)02382-X] PACS numbers: 61.44.Br, 61.14.Hg, 68.35.Bs With the availability of large single grains of quasicrystalline alloys, such as the icosahedral phase of Al-Pd-Mn, several experimental probes have been newly applied to the study of quasicrystalline structures. Over the past two years, the study of surface structures and chemistry by techniques such as scanning tunneling microscopy (STM) [1–5], low-energy electron diffraction (LEED) [4–7], and photoelectron spectroscopy [8,9] has emerged as one of the most active areas in quasicrystal research. The heightened interest in quasicrystalline surfaces has been motivated, in part, by reports of intriguing properties such as oxidation resistance [10–12], low surface friction [13,14], superior wear resistance, and other attractive tribological characteristics [13]. All of these properties are ultimately related to the physics and chemistry of the surface on an atomic scale. Therefore, a basic understanding of the intrinsic surface structure of quasicrystalline alloys is prerequisite to understanding how these surfaces interact with their environment. There are some very basic issues about quasicrystalline surfaces that have yet to be resolved or even addressed. The very nature of the surfaces themselves, as well as the effects of various surface preparation techniques, are the subject of debate. For example, STM measurements on a sample of Al-Pd-Mn prepared by sputtering and annealing in ultrahigh vacuum have revealed well-defined relatively flat terraces with quasicrystalline order within the plane of the surface [2]. In contrast, STM measurements of surfaces of Al-Pd-Mn prepared by in situ cleavage revealed significant atomic-scale roughness [3]. These latter measurements provide some support for a cluster-based approach to quasicrystalline structure advocated by several groups in recent years [15], and have raised concerns regarding the effects of ion bombardment and high temperature annealing upon the surface since selective evaporation and sputtering can significantly change the surface stoichiometry. Indeed, as pointed out by Ebert et al., quasicrystalline phases are complex chemically ordered phases whose surface structures need not be the same as in the bulk [3]. In this Letter, we directly address these issues through x-ray photoemission spectroscopy (XPS) and LEED measurements conducted on a sample of icosahedral Al-PdMn oriented with a twofold axis perpendicular to the surface. After sputtering the surface with argon, no LEED pattern was observed and we found substantial depletion of the aluminum at the surface. After annealing the sample at temperatures above 800 K, a well-ordered, bulk terminated quasicrystalline surface with the same composition as the bulk was recovered. At intermediate annealing temperatures, however, the surface is best characterized as a fine-grained, nanocrystalline (or nanoquasicrystalline) precursor of the icosahedral phase. We also found that XPS measurements of the width of the Mn 2p3y2 peak can provide a good indication of the presence of the face-centered icosahedral (FCI) phase at the surface, as proposed in previous work [9]. The LEED experiments were performed in a stainlesssteel ultrahigh vacuum (UHV) chamber sbase pressure , 3 3 10211 Torrd also equipped with provisions for Auger electron spectroscopy (AES), ion sputtering, and annealing. Supporting measurements by x-ray photoemission spectroscopy (XPS) were performed in a second chamber under the same conditions as the measurements described below, albeit at a higher base pressure (upper limit 4 3 10210 Torr). Our sample, a flat wafer approximately 12 mm 3 15 mm 3 2 mm in size, was cut from a single grain of a boule, grown via the Bridgman method [16] using a starting composition of Al72Pd19.5Mn8.5, and oriented by the x-ray Laue technique so that a twofold axis was normal s60.2±d to the surface. Inductively coupled plasma atomic-emission spectroscopy analysis of a small piece adjacent to our sample indicated a bulk composition of Al71Pd19.8Mn9.2. The phase purity was verified by scanning electron and Auger microscopies to within 0.5% by volume. For further details regarding our methods of quasicrystalline sample preparation, outside of UHV, we refer the reader to Ref. [17]. After polishing and characterization, the sample was fixed onto a thin Ta plate s20 mm 3 25 mmd using two 1050 0031-9007y97y78(6)y1050(4)$10.00 © 1997 The American Physical Society VOLUME 78, NUMBER 6 P H Y S I C A L R E V I E W L E T T E R S 10 FEBRUARY 1997 Ta strips. The sample could be resistively heated and liquid-nitrogen cooled via the Ta plate. A thermocouple (W-5% ReyW-26% Re) was spot welded on the tantalum plate for real-time control of the sample temperature. To confirm the sample temperature measured by the thermocouple, an infrared thermometer (IR gun) was also used. The difference between the IR-gun reading of the sample temperature and thermocouple reading was less than 20 K. After initial cleaning cycles by ion bombardment and annealing, up to a maximum temperature of 900 K, LEED data at several temperature steps were taken after sputtering for 40 min at room temperature and annealing at the desired temperature for 2–4 h. Unless indicated otherwise, all of the LEED data were collected at temperatures at or below 120 K. After sputtering, but prior to annealing at temperatures above 600 K, no LEED pattern from the twofold surface was observed. Furthermore, AES measurements indicated that the surface composition was Al6163Pd3363Mn661, well away from both the known stoichiometry for the FCI phase of Al-Pd-Mn and the composition of the bulk sample. The shift in composition upon sputtering has been reported previously [5]. We also point out that the line shape of the Mn 2p3y2 peak measured by XPS under these conditions was quite broad (see Fig. 1). This last point is notable since recent XPS measurements on clean, well-ordered, fivefold surfaces of Al-Pd-Mn suggest that the sharpness of the Mn peak may be used as an indicator of the presence of quasicrystalline order at the surface of the sample [9]. As discussed in Ref. [9], the shape and width of the Mn 2p3y2 peak can be related to both the position and population/distribution of Mn states near the Fermi energy. Therefore, the Mn 2p3y2 peak can be a sensitive probe of the structural and chemical environment of the Mn sites in the structure. After resputtering the sample and annealing at approximately 600 K, a LEED pattern exhibiting twofold FIG. 1. XPS profiles of the Mn 2p3y2 peak for samples that were (a) sputtered and then annealed at (b) 600 K and (c) 900 K. symmetry (Fig. 2) was observed. The diffraction spots, however, are quite broad and the pattern itself clearly does not correspond to the LEED pattern expected for a bulk terminated twofold FCI surface (see discussion below). We shall refer to this pattern as “rhombic” because rhombi are apparent in the LEED pattern at certain energies as shown in Fig. 2. Twofold, threefold, and fivefold axes, corresponding to the orientation of the underlying bulk alloy, have been superimposed on the pattern in Fig. 2 in order to show that this intermediate phase is orientationally coherent with the bulk quasicrystal. The surface composition after the 600 K anneal was measured, by AES, to be Al6862Pd2762Mn562, closer to, but still well away from, the bulk composition. Furthermore, the line shape of the Mn 2p3y2 peak, measured by XPS, under these conditions remained quite broad (Fig. 1). The breadth of the LEED spots suggests that the in-plane domain size for this phase is on the order of approximately 40 Å, so that the surface layer is best described as nanocrystalline, or, perhaps, nanoquasicrystalline [18]. After sputtering and annealing at temperatures above 800 K, the LEED pattern changed dramatically in several ways. First of all, as shown in Fig. 3(a) (taken after a 900 K anneal), the diffuse spots of Fig. 2 were replaced by a new sharp LEED pattern, also with twofold symmetry, but in a rectangular rather than rhombic pattern. In addition, faceting of the surface was observed, especially close to the edges of the sample, at temperatures above 700 K. The faceting is evidenced by Fig. 4 which shows two new s0, 0d spots in addition to the twofold s0, 0d spot (normally positioned at the center of the screen but moved here by sample rotation). The LEED patterns associated with the new s0, 0d beams have threefold and fivefold symmetry. We point out that the faceting observed here is consistent with a higher surface energy for the twofold surface relative to surfaces perpendicular to the threefold and fivefold directions. After annealing the sample at 900 K the rectangular LEED pattern dominated the surface scattering except at FIG. 2. LEED pattern of the rhombic phase taken at an electron energy of 110 eV. The axes denoted correspond to the twofold, threefold, and fivefold directions in the twofold plane for an icosahedral quasicrystal.