Resonant Auger spectroscopy of poly(4-hydroxystyrene)

https://doi.org/10.1016/S0368-2048(01)00303-6Get rights and content

Abstract

The electronic structure of poly(4-hydroxystyrene) (PHS), which presents strong similarities with that of phenol, has been studied via resonant Auger spectroscopy. Spectator shifts and relative intensities of participator channels are extracted from our data. Symmetry considerations on the occupied and unoccupied molecular orbitals (MO) are helpful in the assignment of the resonating valence band structures of PHS. However they are clearly not sufficient to explain the MO intensities when close-lying intermediate electronic states are involved, as interference effects may play an important role.

Introduction

Due to the availability of photon energy tunable radiation, delivered by synchrotron sources, resonant Auger electron spectroscopy (RAES) has become an important technique to study the electronic properties of condensed matter. Not only have classical topics of solid-state physics been addressed by these techniques, e.g. strong electron correlation in systems like 3d transition metals [1] and their oxides [2], [3], but molecular systems have also attracted the attention of many research groups, e.g. C60 (in relation with the issue of intermolecular electron hopping [4], [5]), small molecule condensates [6], [7], [8], and more especially adsorbates on surfaces (with much emphasis laid on the coupling issue between molecule and substrate [9], [10]). On the other hand, polymers have not been the matter of such systematic studies. To our knowledge only the RAES study on polystyrene by Kikuma and Tonner [11] has been published, despite the fact that these authors clearly demonstrate the usefulness of this core-decay spectroscopy in the fingerprinting of ‘key’ chemical units.

The resonant Auger regime is observed when the photon energy is set at, or scanned across, a resonance in the X-ray absorption spectrum, due to a well-localized core-state. (In the following these resonances in the near edge X-ray absorption fine structure (NEXAFS) spectrum, are called NEXAFS states, or core-excited states.) It constitutes a non-radiative X-ray scattering process in which the incoming X-ray photon with energy excites a collection of atoms (e.g. a molecule) through dipole interaction D to an intermediate electronic state ∣m〉, which then decays with emission of the Auger electron due to Coulomb interaction Q. The amplitude of the one-step process, ending in a final electronic state ∣f〉, is given by the generalized Kramers–Heisenberg formula [12], after nuclear and electronic degrees of freedom are separated by the Born–Oppenheimer approximation:F0fR=m 〈f∣Q∣m〉〈m∣D∣0〉hν−Em−E0−iΓm/2where E0 and Em are the energies of ground ∣0〉 and intermediate ∣m〉 electronic states and Γm the lifetime width of the core-excited state.

The cross-section of the resonant Auger process is written:σRAES(ε, hν)=f ∣FR0f(hν)∣2δ((Ef−E0)−hν+ε)where Ef is the energy of the final state, and ε the energy of the Auger electron.

The decay of the intermediate state leaves the molecule with two vacancies in the outer shell. The electron lifted to the bound unoccupied state may remain as a spectator in the Auger process: then the final state is a two-holes one-electron (2h1e) final state. The 2h1e spectrum is found shifted to higher kinetic energy (KE) with respect to off-resonance ‘normal’ Auger — the energy difference is termed ‘spectator shift’ — because of the screening of the two-holes state by the extra electron.

The excited electron may also participate to the Auger decay, leaving the system with one hole in the occupied valence band (hence it is termed participator). This 1h final state is equally reached via direct photoemission [13], although the selection rules are different. The writing of the RAES amplitude shows also clearly that the intensity of a particular valence band orbital in a participant process depends critically on intramolecular matrix elements, that is (i) on the degree of localization of the excited wavefunction to the core-hole and (ii) on the overlap of particular occupied orbitals with that of the excited core electron.

Due to the fact that this resonant X-ray scattering (RXS) process is described by a one-step formalism, energy conservation implies, in the case of an isolated resonance, that the KE of spectator and participant lines disperse linearly with [14]. While Raman (i.e. linear) dispersion is firmly established in atomic systems [15], in the molecular case the energy dispersion of the emitted electrons may appear more complex, due to the closeness of the vibronic levels, as shown in the theoretical treatment of RXS from an excited molecular state carried out by Gel’mukhanov and Ågren [16]. It was predicted that narrow band excitation can lead to a strong non-linear dependence of the center of gravity of the RXS spectrum, the parameters defining the potential surfaces of the states involved in the scattering event being key ingredients.

Site selectivity of RAES (i.e. when NEXAFS resonances can be used to select specific core-hole orbitals) has already been discussed in the literature, e.g. in the case of furan (C4H4O) [8]. Indeed, this small heterocyclic aromatic molecule (of C2v symmetry) possesses two non-equivalent C atoms (bonded or non-bonded to oxygen). The monosubstituted benzene molecules, aniline, phenol (see Fig. 1a) and fluorobenzene, present also non-equivalent C sites: one C atom connected to the substituent and five ‘unconnected’ sites in C2v (Cs) symmetry. But unlike furan, for which the π*(3b1) NEXAFS state splits into two components separated by ∼1 eV, the non-equivalent sites of monosubstituted benzene give rise to a manifold of π* core-excited levels, consisting of a group of quasi-degenerate levels and of a well-isolated structure [17]. As a matter of fact, the effect of substitution of a H atom of benzene, by a more electronegative atom or radical, has the following consequences. Firstly, the broken D6h symmetry lifts the two-fold degeneracy of the lowest unoccupied molecular orbital (LUMO) π*(e2u) of benzene; secondly, a net negative charging of the substituent leads to an increased ionization energy for the carbon site to which it is attached. Then the NEXAFS spectrum of monosubstituted benzene is dominated by two well-separated π* resonances at the C 1s threshold. The one at lower energy contains the quasi-degenerate π* transitions originating from the unconnected carbons. The second NEXAFS peak is due to a single isolated transition originating from the connected carbon atom (a schematic energy diagram of phenol is given in Fig. 1b). With an increasing electronegativity of the substituent, the energy separation between unconnected and connected carbon transitions increases (it is 1.5 eV, 1.9 eV and 2.2 eV, for aniline, phenol and fluorobenzene, respectively); by contrast, the energy variations with respect to the core-sites become smaller for the unconnected carbons [17].

The monosubstituted benzene compounds offer the opportunity of studying the influence of the chemical shifts on the RXS process. In a resonant inelastic X-ray scattering (RIXS) study of aniline ice, it was demonstrated that core-excited levels with small (sub eV) chemical shifts give resonant spectra that interfere strongly [18]. This observation is a strong motivation to examine site selectivity and possible interference effects — in relation to the energy separation of the various NEXAFS states — in the case of the non-radiative version of RXS, that is RAES. In order to achieve this goal, we have chosen a polymer, poly(4-hydroxystyrene) (PHS) [19], which consists of a saturated polyethylene (PE) backbone where a phenolic group substitutes one of the hydrogen atoms at every other carbon (see Fig. 1a). Given that the basic unit of PHS (C8H8O) contains six carbons from the phenolic group for only two carbons from the PE backbone, the electronic structure — especially the NEXAFS spectrum — is expected to be dominated by the pendant group. We have measured RAES spectra both at C and O K-edges. The data we have acquired are sufficiently detailed to be confronted with a theoretical analysis that would fully exploit the formalism of , . However, in the absence of such a theoretical backup, we outline here what can easily be ascertained from the experimental data. Despite the fact we are doomed to work out the RAES data in terms of a superposition of participant and spectator contributions (i.e. a ‘normal Auger’ shifted to higher KE), following the approach of Menzel et al. [6] and Kikuma and Tonner [11], useful quantities can be extracted, such as spectator shifts at NEXAFS resonances and (relative) intensities of participator channels. Criteria on the ‘spatial localization’ of occupied and unoccupied ground state orbitals are proposed, which are helpful in the assignment of the resonating valence structures of PHS. However symmetry considerations are clearly not sufficient to explain MO cross-sections at resonance when close-lying intermediate states are involved.

Section snippets

Experimental method

The experiments were performed at the LURE-SuperAco storage ring in Orsay (France), at beam line SB7, using a Dragon-type (spherical grating, 800 l/mm) monochromator placed behind a bending magnet [20]. Linearly polarized radiation was used. The newly built ultra-high vacuum experimental station ABS6/CSEAL is equipped with a SCIENTA200 hemispherical electron analyzer for X-ray photoemission (XPS) and Auger spectroscopy and an electron multiplier for X-ray absorption measurements (NEXAFS was

NEXAFS spectroscopy

The NEXAFS spectra of PHS at the C and O 1s edges are shown in Fig. 2, Fig. 3, respectively. For comparison, the C and O 1s spectra of gas-phase phenol, measured by inner-shell electron energy loss spectrometry (ISEELS) [23] (with a resolution of 0.7 eV) are also given on a common excitation energy scale.

The features labeled from 1 to 3 in the carbon K-edge spectra of PHS (Fig. 2) have a one-to-one correspondence in the ISEELS spectrum of phenol. Moreover the energies of all spectral features

Summary

Resonant Auger spectroscopy experiments have been carried out at the C and O K-edges of poly(4-hydroxystyrene) (PHS). At both edges, the NEXAFS spectra of this polymer strongly mimic those of gas-phase phenol, emphasizing the dominance of the phenolic group in the electronic structure. Given the similarities with phenol (and other monosubstituted benzene molecules of the kind), here we can study the relationship between site selectivity and core-excited state energy separation in the

Acknowledgements

The authors thank Dr. M. Schott and Prof. F. Abel for enlightening discussions on polymer science and for their precious advice concerning the deposition of thin polymer films. They also thank R. Delaunay for his friendly help during the building of CSEAL.

References (34)

  • P.A. Brühwiler et al.

    Chem. Phys. Lett.

    (1992)
  • W. Wurth et al.

    J. Electron. Spectrosc. Relat. Phenom.

    (1993)
  • T. Åberg et al.
  • Y. Luo et al.

    Phys. Rev. A

    (1995)
  • The spectator shift we find for PS confirms the value given by Kikuma and Tonner (2.9 eV, Ref. [8]) who also used the...
  • M. Carbone et al.

    Surf. Sci.

    (1999)
  • In gas-phase phenol IP is 538.9 eV (see Ref. [20] and references...
  • M. Weinelt et al.

    Phys. Rev. Lett.

    (1997)
  • M. Finazzi et al.

    Phys. Rev. B

    (1999)
  • M. Finazzi et al.

    Phys. Rev. B

    (1999)
  • A.J. Maxwell et al.

    Phys. Rev. Lett.

    (1997)
  • D. Menzel et al.

    J. Chem. Phys.

    (1992)
  • M. Mauerer et al.

    J. Chem. Phys.

    (1993)
  • C. Keller et al.

    Phys. Rev. Lett.

    (1998)
    C. Keller et al.

    Phys. Rev. B

    (1999)
  • F. Bournel et al.

    Phys. Rev. B

    (2000)
  • J. Kikuma et al.

    J. Electron. Spectrosc. Relat. Phenom.

    (1996)
  • The amplitude of the normal (non-resonant) photoemission process, which is F0fNR=〈fD0〉, adds to the resonant Auger...
  • Cited by (0)

    View full text