The M5 photoelectron line of Ag metal measured in coincidence with the M5–N45N45 Auger-electron spectral line

The M₅ (3d_5/2) photoelectron line of Ag metal was measured in coincidence with the M₅–N₄₅N₄₅ (¹G or ³F) Auger-electron line by Auger-photoelectron coincidence spectroscopy (APECS). The M₅ photoelectron line of Ag metal measured in coincidence with either the ¹G or the ³F Auger-electron line is unshifted, within the accuracy of the experiment. As the theory [M. Ohno, J. Electron Spectrosc. Relat. Phenom. 124 (2002) 53] predicts, the energy shift and asymmetrical narrowing of the coincidence photoelectron line compared to the singles one is much smaller than that in Cu metal.     

Analysis of the Auger-photoelectron coincidence spectroscopy (APECS) spectra of metallic Pd,Ag, and Sn

The M45-level photoelectron spectrum of Ag metal measured in coincidence with the M45–VV(¹G) Auger-electron line is analyzed by taking into account the possibility of the M4–M5–VV Coster–Kronig (CK)-transition preceded Auger transition. We denote the atomic shells Mx(Mxy) and Nxy (x, y = 4,5) by MX(MXY) and V, respectively. The M4–M5 CK-transition rate is very small. The M45–VV Auger-electron spectra of metallic Pd and Sn measured in coincidence with the M4 (or M5)-level photoelectron line are analyzed. The M4–M5 CK-transition rates are also very small in metallic Pd and Sn. The coincidence Auger-electron line previously interpreted as the M4–M5–VV (or M4–M5V–VVV) Auger-electron line is largely due to the inelastically scattered M5-level photoelectron background beneath the M4-level photoelectron line. The APECS spectrum of Pd metal shows the first evidence of the M5V–VVV transition of the localized M5V shakeup two-hole state. The intensity ratio of the inelastically scattered Auger-electron background to the M5–VV Auger-electron main line of Ag metal measured in coincidence with the inelastically scattered M5-level photoelectron background beneath the M4-level photoemission line increases, as compared to that measured in coincidence with the M5-level photoelectron main line. This is because when the probability of the photoelectron being inelastically scattered increases, that of the Auger electron emitted by the same ionized atom, being inelastically scattered increases. In other words the photoelectrons and the Auger electrons are originated from the deeper atomic sites (longer pathlength).     

Calculation of the yield of Xe<Superscript><Emphasis Type="Italic">i</Emphasis>+</Superscript> photoions upon the decay of <Emphasis Type="Italic">M</Emphasis><Subscript>45</Subscript> vacancies through intermediate states in coincidence with energy-selected Auger electrons

A theoretical model is proposed for calculating the yield of photoions detected in coincidence with energy-selected Auger electrons. This model provides a correct explanation for experimental data on the yield of photoions upon ionization of the M₄₅ shell of xenon atoms. In the framework of the proposed model, the intermediate-coupling approximation is used to calculate both the structure of the terms of the N^?2 two-hole states generated through the M₄₅NN Auger decays and the probabilities of branching due to decays of these states into three-hole states upon the Auger, Coster—Kronig, and super-Coster—Kronig transitions. The probabilities of branching for subsequent branches of the cascade decay are calculated in the configuration-average approximation. The results of these calculations are in good agreement with the experimental data.     

Many-electron effects in the Auger-photoelectron coincidence spectroscopy spectra of the late 3d-transition metals

The valence hole created by the L2–L3 M45 Coster–Kronig (CK) transition may hop away from the ionized atomic site before the L3-hole decays. Then when the third (Auger) electron emitted by the L3-hole decay is measured in coincidence with the photoelectron emitted by the initial L2-level electron ionization, the coincidence spectrum becomes similar or identical to the singles spectrum of the secondary (Auger) electron emitted by the L3-hole decay as if it decayed as an initial single core hole. Thus the coincidence spectrum is essentially governed by the valence-hole dynamics of both the intermediate states and the final states of the L2–L3 (M45) CK-transition preceded Auger transition. In the present paper the Auger-photoelectron coincidence spectroscopy (APECS) spectra of Fe, Co, and Ni metals reported by C.P. Lund et al. (Phys. Rev. B55 (1997) 5455) are analyzed in light of the delocalization and localization of the valence hole(s) created by the CK transition or the CK-transition preceded Auger transition.     

The Ag M5N45N45 Auger photoelectron coincidence spectra of disordered Ag0.5Pd0.5 alloy

An effect of disorder broadening (DB) on the Ag M₅N₄₅N₄₅ Auger spectra in the random substituted Ag_0.5Pd_0.5 has been investigated by Auger photoelectron coincidence spectroscopy (APECS). Data were collected for the Ag M₅N₄₅N₄₅ Auger line coincident with the Ag 3d_5/2 photoelectron line (and its higher and lower binding energy sides). It is shown that the broadening of the Ag M₅N₄₅N₄₅ line is directly associated with the presence of disorder broadening of the Ag 3d_5/2 photoelectron line. The APECS experiment is used to demonstrate the broadening in a novel way.     

On the Auger-photoelectron coincidence spectroscopy spectra of Cu(100), Cu metal and CuOx/Cu surface

The M₃–VV Auger-photoelectron coincidence spectroscopy (APECS) spectrum of Cu(100) and the L₃–VV APECS spectra of Cu metal and CuOx/Cu surface are analyzed in detail. The narrowing and energy shift of the photoelectron line in the M₃–VV APECS spectrum is well predicted by the present theory. The spectrum shows the presence of the M₂–M₃(V)–VV(V) decay in which a hole in the 4s band hops away prior to the decay of M₃ hole. The analysis of the L₃ photoelectron spectra of Cu metal measured in coincidence with the ³F or ¹G Auger line raises a question concerning the presence of two different core–hole states upon the L₃ level ionization recently proposed by Thurgate and Jiang [Surf. Sci. 466 (2000) L807]. The analysis of the L₃–VV APECS spectrum of CuOx/Cu shows that the final-state charge–transfer interaction plays an important role in CuO.     

Complete breakdown of the one-electron picture of auger spectra of solids

The very broad M₄₅?N₂₃N₄₅ (3d^-1 → 4p^-14d^-1) Auger spectra for metals in the range Pd to Te arise from complete breakdown of the one-electron picture of the final state double vacancy due to 4p^-14d^-1 ? 4d^-3?f giant Coster-Kronig fluctuation and decay processes. This is an example of atomic-like effects giving rise to level shifts and broadening which are large in comparison with typical widths of d-bands in solids. The 4p^-14d^-1 two-hole spectral function can be approximated by a 4p^-1 single-hole spectral function perturbed by a static, sharp 4d^-1 hole. The resulting Auger line-shapes are different from single-hole 4p^-1 line-shapes from XPS and XES. Analogous effects also occur for double vacancies involving one or two 4s-holes.     

Many-body effect in the N23-VV Auger-photoelectron coincidence spectroscopy (APECS) spectra of Ag metal

The coincidence N23-VV Auger-electron spectroscopy (AES) spectra and N23 photoelectron spectroscopy (PES) spectra of Ag metal are analyzed. Here NX is the notation for atomic shell Nx (X = 2, 3). The band-like feature in the coincidence N23-VV AES spectra is much more intense than that in the coincidence M45-VV ones because the potential in the delocalized two-hole state is less attractive than that in the localized one. The partial N23-VV super Coster–Kronig (sCK) transition rate depends critically on both the final-state potential and the sCK-electron kinetic energy (KE) because the KE is low, whereas the partial M45-VV Auger-transition rate is fairly independent of them because the KE is very high. As a result, the partial sCK-transition rate to the band-like state is enhanced compared to that to the atomic-like localized state. The low KE tail in the coincidence N23-VV AES spectra which is likely due to the sCK transition involving more than two electrons, is more enhanced than that in the coincidence M45-VV ones. This is due to the enhancement of the partial sCK-transition rate by the presence of extra holes in the final state. The sharp peaks of small intensity on the lower KE side of the main line in the coincidence N2 PES spectrum are tentatively attributed to the shakeup satellites.     

Many-body effect in the coincidence L3 and M3 photoelectron spectroscopy spectra of Cu metal

The coincidence L₃ and M₃ photoelectron spectroscopy (PES) main lines of Cu metal are calculated by a many-body theory. There is no peak-energy shift between the singles PES main line and the coincidence one. The asymmetric narrowing of the coincidence PES main line on the low kinetic energy (KE) side is very small. This is in accord with recent experimental findings. In Cu metal, the shakeup satellite intensity is small and the main-line satellite separation energy is much larger than the core–hole lifetime width. The interference via the final-state interaction is negligible. In the PES main line, the imaginary part of the self-energy by shakeup excitations, which is very small compared to the core–hole lifetime width, decreases very slowly in linear with photoelectron KE. The branching ratio of Auger decay of a single hole state then increases very slowly in linear with photoelectron KE so that the deviation of the coincidence PES main line from the singles one is very small. The 939 eV structure seen only in the coincidence L₃ PES spectrum of Cu metal is attributed to the enhancement of the inelastic peak of a smaller energy loss due to electrons of a smaller average emission depth measured in coincidence with the elastic Auger peak. The structure will not be enhanced in the singles PES spectrum. The background subtraction in the coincidence spectrum cannot be the same as that in the singles one. Such consideration is necessary before we can conclude about the asymmetric narrowing on the low KE side. A unique capability of APECS by which one can determine the photoelectron KE dependent part of the imaginary part of the self-energy is pointed out.     

ESCA studies of some copper and silver selenides

Photoelectron and Auger spectra have been obtained for the copper and silver selenides CuSe, Cu ₂Se, Ag ₂Se, and AgCuSe as well as from CuS, Ag ₂O, Ag ₂S, Cu, and Ag. Binding-energy values, chemical shifts, and peak-shapes are reported for the Cu 3 d, Ag 4 d and Se 3 p electrons. Absence of multiplet splitting and shake-up structure is discussed in relation to the magnetic properties. It is shown that chemical shifts are much better revealed in the Auger spectra (Cu L₃M_4,5M_{4, 5} and Ag M₅N_{4, 5}N_{4, 5}) than in the direct photoelectron ones. In addition the use of the Auger parameter to characterize the series under study is emphasized. Finally the valence-band spectra have been examined and the electronic structures are interpreted.     

EELS characterization of TiN grown by the DC sputtering technique

Titanium nitride thin films were deposited on monocrystalline silicon (mc-Si) substrates by direct current reactive magnetron sputtering. Auger electron spectra (AES) of deposited films at different nitrogen partial pressures, show the typical N KL₂₃L₂₃ and Ti L₃M₂₃M₂₃ Auger transition overlapping. Also, changes in the Ti L₃M₂₃M₄₅ Auger transition peak are observed. X-ray diffraction and high resolution electron microscopy (HRTEM) of a golden color TiN/mc-Si sample, reveal a preferential polycrystalline columnar growth in the 〈111〉 orientation. This sample was also analyzed by electron energy-loss spectroscopy (EELS). The N/Ti elemental ratio is slightly different to the value determined by AES. Atomic distribution around the N atoms is in agreement with that expected from the N atom in the fcc unit cell of TiN. This distribution was obtained via an extended energy-loss fine structure (EXELFS) analysis from EELS spectra.     

Observation of resonance recombination lines in electron excited auger spectra of Gd

Combined measurements of electron excited N_4,5 Auger spectra and photoelectron emission on clean and oxidized Gd lead to a distinction between Auger lines originating from 4d → continuum and 4d → 4? resonance excitations. Several Auger structures are identified as due to the direct recombination of 4d⁹4?⁸ states with the 4f and valence electrons. The shape of the most prominent Auger line for oxidized Gd agrees perfectly with the Fano profile of the 4? photoemission intensity.     

The tungsten Auger transitions : 150–200 eV

The low-energy Auger spectrum of tungsten, induced by electron bombardment, has been studied. The second-derivative detection mode was employed to improve the resolution of fine structure. A comparison of N(E) spectra and yield curves with theoretical Auger rates and line energies showed that N_4,5N_6,7X processes dominate the spectrum, where X denotes N_6,7 or O_2,3. The core—core—valence (X = O_4,5) processes occur with low probability. The N₄N₅V transition appears to be energetically allowed in tungsten, and to occur at a sufficiently high rate to alter the initial N₅ : N₄ hole-states ratio for the above Coster—Kronig and super-Coster—Kronig processes. For primary excitation energies above 425 eV, there is also a small contribution (about 12% at maximum) from N₃N₅V processes, which occurs around 166 eV.     

Study of L2,3M4,5M4,5 Auger spectra of Zn and Cu in molecular ZnCl2 and (CuCl)3 vapours

L_2,3M_4,5M_4,5 Auger electron spectra of Zn and Cu have been measured in molecular ZnCl₂ and (CuCl) ₃ vapours. The spectra have been analyzed and compared with the corresponding free-atom spectra. It is found that the main features of the spectra are atomic-like. The energies are shifted by 0.55 eV in ZnCl₂ and by 3.2 eV in (CuCl)₃ towards higher kinetic energy compared with the corresponding free-atom spectra. For the intensity ratios between the L₃ and L₂ groups, the values 2.8 and 3.7 are obtained for Zn and Cu, respectively. These intensity ratio, together with energy considerations based on free-atom Dirac—Fock calculations and observed Auger shifts, indicate that the L₂L₃M_4,5 Coster—Kronig process is energetically possible in (CuCl)₃ molecular clusters but not in ZnCl₂. The satellite structure in the spectra studied also supports this conclusion.     

Semi-empirical elemental sensitivities for quantitative auger electron spectroscopic analysis

A semi-empirical equation is presented which may be used to calculate the relative sensitivities of the elements for Auger electron spectroscopic analysis. The equation involves the product of a large number of theoretically or empirically derived terms, including ionization cross section. Coster—Kronig transitions, backscattered electron contribution, individual Auger transition probabilities, atomic density, electron mean free path, analyser transmission, modulation effects, detector efficiency, and peak asymmetry. The calculated sensitivities are compared with experimental relative elemental sensitivities from the Handbook of Auger Electron Spectroscopy. The close agreement obtained with experiment suggests that the semi-empirical theory is now capable of providing elemental sensitivities for analysis, in a predictive way, where reliable empirical sensitivities are missing.     

M4,5N4,5N4,5 auger electron spectrum of Ba metal

The Al Kα excited M_4,5N_4,5N_4,5 Auger spectrum of Ba has been measured from the metallic sample evaporated on a Ag substrate. The spectrum has been decomposed into individual line components after the background subtraction. The decomposed spectrum has been compared with the theoretical spectrum calculated for the 4d^?2 final state configuration in the mixed coupling scheme applying jj-coupling for the initial state and intermediate coupling for the final state. The most prominent structure of the spectrum shows the two 4d-hole coupling, but the structure which is caused by the Auger transitions M_,45N_2,3V has also been observed. The screening of the core holes in Ba metal seems to be produced by (5d6s) electrons. The simple excited atom model HF-calculations give an Auger kinetic energy shift (metal-free atom) of 16.7 eV, which is comparable to the experimental value 14–18 eV.     

Influence of surface environment on Auger peaks : The Palladium M45N45N45 transition

The fine structure of the M₄₅N₄₅N₄₅ Palladium Auger transition was investigated. The energy of the peaks and the changes of their shapes due to surface modifications, either by alloying or by oxidation are studied. It is found that the fine structure is due to the overlapping of two “band-like” transitions of the type M₄VV and M₅VV.     

The 4d auger,coster-kronig,and recombination spectra of the lanthanides

A complete set of electron-excited 4d-basedAuger spectra of the lanthanide metals from lanthanum to lutetium (except the unstable promethium) is presented, in both differential and integral forms. It is believed that the set is more representative of clean surfaces of the lanthanides than any published hitherto. With the help of binding energy and electron loss measurements made in this laboratory and elsewhere, values of the various possible Auger, Coster—Kronig, and direct recombination transition energies are calculated, and for each transition a “centre of gravity” is derived based on relative intensities of final state multiplets, electron occupation of core levels, etc. By using arguments based on trends in the spectra across the series, on theoretically and experimentally derived values of U_eff for the difference 4fⁿ⁺¹→4f^n?1, and on plausibility, values of U_eff for 4fⁿ→ 4f^n?2 as well as for the other final state hole pair configurations are allocated. The relaxed transition energies so calculated are then compared with the experimental energies, from which it is possible in most cases to make assignment of the spectral features to the various transitions. As a result it is found that there are some significant disagreements with the theoretical rates of McGuire for lanthanide free atoms. The reasons for these disagreements are discussed, and an empirical model based on effective 4f and 5d populations and on the restrictions imposed by spin alignment is used to resolve the differences qualitatively.     

A study of LaH3 by auger electron appearance potential spectroscopy

The Auger electron appearance potential spectra (AEAPS) of M_4,5 levels of lanthanum in the elemental state and in LaH₃ are presented. AEAPS measures the total differential secondary electron yield from the surface constituents as a function of incident electron energy. In the presence of a 3d hole, the 4f level is pulled down below the Fermi energy, E_F. In the metallic La spectrum this is seen as a shoulder on the low energy side of the main peak, but in the LaH₃ spectrum this shoulder disappears and is replaced by two peaks, again on the low energy side. The high binding energy satellite reported in the M₅ level XPS spectra in La insulating compounds has been attributed to the lowering of the 4f level above the ligand band. Since no high binding energy satellite is observed in the M_4,5 level AEAPS of LaH₃, the lowest energy peak is assigned as the transition of the 3d electron to the 4f level pulled down below the hydrogen induced band. The other peak is related to the partially filled hydrogen induced band below E_F. Tentative energy level diagrams for these transitions are proposed.     

Inelastic effects and structure in the auger electron emission spectra of V2O5(010) and V(100) surfaces: Study of chemical shifts

Characteristic ionization losses and plasma losses occurring in the AES spectra of V₂O₅ (010) and V(100) surfaces are discussed. The former are used to evaluate chemical shifts of the VL₂ and VL₃ levels in V₂O₅. Values of approximateley 2.5 eV are found. These are smaller than the values obtained with ESCA (± 5 eV). It is described how the electron beam used in AES is thought to be responsible for this effect. ESCA data from partly reduced V₂O₅ samples tend to confirm the proposed model, based on oxygen loss and decomposition of V₂O₅ single crystals under the influence of the electron beam.The plasma losses of the larger Auger peaks are discussed. It is shown how some fine structure in the spectra can be partly explained by their presence. The plasma losses were simulated with numerical techniques based on the use of a signal averager.Signal averaging, curve fitting and related numerical techniques improve the resolution of AES spectra. Spectra of V₂O₅(010) and V(100) obtained in this manner are discussed. With respect to the transitions involving the valence band it is shown that the complex valence band structure is one of the causes of the observed discrepancies between theoretical and experimental AES data. Furthermore there is an uncertainty concerning the way in which ionization correction should be applied in this case. This correction is thought to increase with the degree of localization of the valence electrons.Auger peak intensities in function of the primary energy were found to show a maximum at about 3.5 times the critical potential, as was expected from theory if a moderate amount of backscattering is taken into account. Finally the intensity variations of the Auger peaks under continuous electron bombardment show the rate of oxygen loss at a V₂O₅ surface due to the primary beam.