Homer D. Hagstrum

Orbital Energy Spectra of Electrons in Chemisorption Bonds: O, S, Se on Ni(100)

This work presents the first experimental determinations of orbital energy spectra of electrons in chemisorption bonds. Our method, ion neutralization spectroscopy, determines a transition density function which is essentially the local density of states in the surface region of the solid. For the particular surface formed by chemisorption of the chalcogens in ordered, surface‐crystalline arrays on Ni(100), the local density of states includes a strong component from the orbital energy spectrum of the electrons in the bonds of the surface molecules formed from the adsorbate atom and Ni atoms presented by the substrate lattice. The orbital levels of electrons in the surface bonds are broadened resonances or virtual states in whose energy range local wavefunction magnitude and hence transition probability for ion neutralization are enhanced. These orbital energy spectra, coupled with measurement of work‐function change on adsorption, suggest, in many instances, plausible identifications of local bonding str...

Orbital energy spectra of electrons in chemisorption bonds: O, S, Se ON Ni(110) and Ni(111)

Abstract In an extension of previous work on Ni(100), the method of ion-neutralization spectroscopy has been used to determine orbital energy spectra of electrons in the chemisorption bonds holding O, S, or Se in ordered arrays on the Ni(110) and Ni(111) faces. Observations of surface crystalline structure by LEED and of thermal stability properties of the structures formed were made. Work-function changes on adsorption were also measured. These results are used in an attempt to assess the nature of local bonding structures. It is concluded for Ni(110), as for Ni(100), that the evidence points strongly toward reconstruction for the O structures, but not for the S and Se structures. The evidence is less convincing for Ni(111), and no conclusion about reconstruction is drawn. Comparisons between Ni(100) and Ni(110) indicate that when the structure is non-reconstructed the chalcogen bonds to as many Ni atoms as possible, that is, produces pyramidal rather than diatomic or “sidebridge” surface molecules.

Cleaning of Silicon Surfaces by Heating in High Vacuum

Results from experiments with four independent surface‐sensitive phenomena agree in demonstrating that an atomically clean surface can be produced on silicon single crystals by heating to 1550°K or above for several minutes in high vacuum. The four phenomena studied are: 1. oxidation and gas adsorption, 2. Auger neutralization of He+ ions, 3. external photoelectric emission, and 4. field electron emission in a Muller microscope. Evidence is given as to the nature of the silicon surface before cleaning, properties of the clean surface, and behavior upon contamination with gases and during subsequent recleaning. The surfaces were found to form a permanent p‐type layer several microns deep as a result of the vacuum heating.

Instrumentation and Experimental Procedure for Studies of Electron Ejection by Ions and Ionization by Electron Impact

Four different instruments are described in this paper. Three have been or are being used in studies of electron ejection from metal surfaces by positive ions. A fourth has been used in the study of the ion optics employed and in studies of ionization and dissociation of diatomic molecules by electron impact. Basic design features and performance characteristics are discussed with particular emphasis placed upon the steps taken to assure proper ion‐beam characteristics at the final focus. Experimental procedures are also described among which those having to do with the vacuum and adsorption of gas on the target are considered to be especially important. Auxiliary apparatus and the determinations of contact potential and work function are discussed briefly. The interpretation of data obtained with these instruments is also examined.

The determination of energy-level shifts which accompany chemisorption

Abstract Considerable confusion exists concerning a reasonable procedure for determining energy shifts of atomic levels on adsorption of a free gas atom to a solid surface and concerning the basic concepts involved. This paper discusses the ionization limit with respect to which energy levels of the adsorbed complex should be referenced, how this limit may be defined and what the best method is for determining its energy level. It is also demonstrated that one cannot sidestep the determination of the ionization limit by measuring in the same apparatus the kinetic energies of electrons ejected from the free atom and from the adsorbed atom.

Hydrogen chemisorption on the 100 (2 × 1) surfaces of Si and Ge

Abstract This paper combines a theoretical study of the Si(100) surface having a monolayer of atomic hydrogen chemisorbed to it with an experimental study of the analogous Ge(100) and Ge(110) surfaces. In the theoretical work the underlying (100) silicon surface is taken to be reconstructed according to the Schlier-Farnsworth-Levine pairing model with the hydrogen located on the unfilled tetrahedral bonds of this structure. Self-consistent calculations of the electronic potential, charge density, spectrum, and occupied surface density of states are carried out. The force on the hydrogen atoms is then calculated using the Hellman-Feynman theorem. This force is found to be close to zero, confirming that the hydrogen atoms are indeed at the equilibrium position for the chosen silicon geometry. Features in the calculated photoemission spectrum for the Si(100) 2 × 1 : H surface are discussed in terms of related features in the photoemission spectrum of Si(111) : H, but are found not to agree with the previously measured photoemission spectrum of Si(100) 2 × 1 : H. Measured photoemission and ion-neutralization spectra for Ge(100) 2 × 1 : H agree in their major features with what is calculated for Si(100) 2 × 1 : H, however, suggesting that the Ge(100) 2 × 1 : H surface is reconstricted according to the pairing model. Similarly, measured spectra for clean Ge(100) 2 × 1 agree with calculations for the row dimerized Si(100) surface.

Electronic Characterization of Solid Surfaces Determination of the energy levels of electrons at surfaces is now possible over a wide energy range

At this point it would be presumptuous to suggest either that we understand very much about the electronic structure of solid surfaces or that we can specify in detail exactly what each of our tools for studying such structure is telling us. I think it is fair to say, however, that FES, UPS, and INS do make it possible for us to determine energy level spectra which can with some confidence be ascribed to the resonances of electrons in surface orbitals. It is true that INS is the more surface-selective of the two electron spectroscopies capable of producing data over at least a 10-ev energy range. We have seen intriguing differences between INS and UPS which, when we come to understand them, will most certainly reveal important characteristics of surface electronic structure and greatly expand our ability to distinguish electronic states in the surface from those in the selvedge. Possibly it is not too much to hope that the combined use of INS and UPS with incidence angle as an independent variable will give us information on the geometrical extent of surface orbitals, as well as the net electrical charge and the electric potential gradient in the region of the surface in which the orbitals lie.

Production and Demonstration of Atomically Clean Metal Surfaces

Ion bombardment as a means of cleaning solid surfaces has been tested by applying it to the metal tungsten. The test, which showed the resulting surface to be clean, is a sensitive one because atomically clean tungsten is very reactive to the common gases. The test is definitive since it is performed on a material for which there is overwhelming evidence that another means, namely heating, does produce an atomically clean surface. The phenomenon used in this work to observe surface conditions during the cleaning procedures is the Auger‐type ejection of electrons by slowly moving positive ions. This is again shown to be a sensitive means of detecting surface contamination. Data on the electron release by ions from heavily contaminated metal surfaces are reported.

Oxygen Adsorption on Silicon and Germanium

Data are presented which yield relative magnitudes of the initial, room‐temperature sticking probabilities S0 of nitrogen on clean tungsten and oxygen on clean silicon and germanium. Taking S0 to be 0.35 for N2 on W, we obtain 1×10−2 and 8×10−4 for S0 of O2 on Si and Ge, respectively. Data are also given concerning the temperatures at which the atomically clean surface is thermally regenerated from the oxygenated surface for both Si and Ge.

Ion neutralization studies of the (111), (1̄1̄1̄) and (110) surfaces of GaAs

Abstract The (111), (111) and (110) faces of GaAs have been studied by the Auger neutralization process in three conditions: (1) atomically clean and stoichiometric, (2) after exposure of the clean surfaces to CO, H 2 , N 2 and O 2 , and (3) after high temperature evaporation. The (111) and (110) faces have the same form of electron energy distribution shifted relative to each other indicating that among the characteristics detectable by the ion neutralization process the surfaces differ only in photothreshold with gF(111) ≅ φ(110) + + 0.12 eV. The distribution for the (111) face differs appreciably from those for the (111) and (110) faces indicating a different transition probability as a function of energy near the top of the valence band. Energy broadening differs among the surfaces in an analogous manner. O 2 , but not h 2 , CO or N 2 , is adsorbed by these surfaces and produces the characteristic change observed for other materials, namely, a loss of fast electrons from the distribution. O 2 sticking probability is estimated for the three faces as: (111), 10 −5 ; (110), 7 × 10 −5 ; (111), 10 −4 . Heating of the (111) face causes the electron energy distribution to shift in a manner consistent with, first, faceting to (110) and then return to a (111) surface character. Changes in form of the electron energy distributions accompany heating of the (110) and (111) faces.