HAL Id: jpa-00220021
https://hal.archives-ouvertes.fr/jpa-00220021
Submitted on 1 Jan 1980
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
Reflectance, fluorescence yield, and photoelectron spectroscopy of CaF2 in the 10 to 100 eV range
R. Williams, D. Nagel, M. Kabler
To cite this version:
R. Williams, D. Nagel, M. Kabler. Reflectance, fluorescence yield, and photoelectron spectroscopy of CaF2 in the 10 to 100 eV range. Journal de Physique Colloques, 1980, 41 (C6), pp.C6-439-C6-442.
�10.1051/jphyscol:19806114�. �jpa-00220021�
JOURNAL DE PHYSIQUE Colloque C6, supplément au n° 1, Tome 41, Juillet 1980, page C6-439
Résumé. — On a mesuré dans le domaine de 10 à 100 eV le pouvoir réflecteur et les spectres de fluorescence et photoélectronique de monocristaux de fluorine en utilisant le rayonnement d'un anneau de stockage. Le spectre présente des caractéristiques attribuées à des excitons et à la structure de bande. Dans les trois types de spectre les maximums liés aux états de la bande de conduction associés aux niveaux atomiques 3d sont très importants.
Réflectance, fluorescence yield, and
photoelectron spectroscopy of CaF
2in the 10 to 100 eV range
R. T. Williams, D. J. Nagel and M. N. Kabler Naval Research Laboratory, Washington, DC 20375, U.S.A.
Abstract. — Reflectivity, fluorescence and photoelectron spectra from CaF2 single crystals hâve been measured using synchrotron radiation in the 10 to 100 eV range. The spectra contain features ascribed to excitons and to the band structure. Peaks due to conduction-band states derived from the 3d atomic levels are prominent in ail three spectra.
1. Introduction. — The work being reported hère is part of an ongoing study encompassing the alkaline earth fluorides BeF2, MgF2, CaF2, SrF2 and BaF2, in which complementary measurements of réflectance, fluorescence yield, and photoelectron émission are being made. Both the fundamental electronic structure of the perfect crystal and properties of defects and relaxed exciton states near the surface are of interest.
In this paper, we will discuss preliminary data for CaF2, which has been emphasized in our early measurements because it is the best characterized alkaline earth fluoride both theoretically [1, 2] and experimentally [3-10]. Part of the présent results are simply presented as corroboration of prior published data. However, because of the wide spectral range available to us, the combination of photoelectron energy analysis and tunable photon energy, and the ability to résolve fluorescence spectra as well as uv excitation curves, this work also provides new expéri- mental data.
2. Expérimental apparatus. — The investigation was conducted using vacuum ultraviolet radiation from the SURF-II électron storage ring facility of the National Bureau of Standards. The monochromator being used is a horizontally dispersing toroidal grating instrument [11]. The resolution for most of this work was 3 Â, set by the exit aperture. The first-order flux available on the sample in the 20 to 60 eV range was typically 5 x 109photon/s per 3Â bandpass, but was considerably lower near the extrema of the 10-100 eV range. The spectral distribution of first through fifth order radiation was measured during the course of our photoelectron studies.
Détails of the expérimental grating characterization will be published separately. For the data presented hère, second order has been removed during the data réduction, but no order-sorting filters were used.
Given thèse factors, the data below 20 eV are subject to considérable uncertainty.
Experiments were conducted in an ultrahigh vacuum chamber which has a base pressure of 4 x 10" 9 torr before baking, and hydrocarbon partial pressure below 1 0- 1 2 torr. CaF2 samples were cleaved in air about 30 min. before pumpdown, and mns were usually made without subséquent bakeout of the entire System. However, a heatable/coolable sample stage allowed baking of the sample when required, and the System is equipped with an argon sputter- etching gun. The measurement apparatus will be discussed briefly as results are presented.
3. Réflectance. — The reflectivity of CaF2 has been measured at several angles of incidence using photodiodes employing either aluminum or anodized aluminum photocathodes. A spectrum for near- normal (3°) incidence is shown in the middle of figure 1. The réflectance scale is logarithmic.
Réflectance of CaF2 for énergies up to 40 eV at near-normal incidence has been reported pre- viously [4, 5]. In the présent data for near-normal incidence, the peak near 35 eV is almost 8 times stronger than the neighboring peak near 25 eV.
However, in référence [4] the ratio is about 1.6, and in référence [3] it is about 3. The reason for this dis- crepancy is not known at présent, but dominance of the 35 eV peak is also found in our fluorescence measurements, to be discussed.
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:19806114
C6-440 R. T. WILLIAMS, D. J. NAGEL AND M. N. KABLER
PHOTON ENERGY (eV)
Fig. 1. - The upper curve is the excitation spectrum of self- trapped exciton fluorescence in CaF, (spectrum shown in the inset figure), a t T w 140 K. The middle curve is normal-incidence reflectance, shown on a logarithmic scale. The lower curve is proportional to the total photoelectron yield from a CaF2 single crystal sample subject to charging.
4. Fluorescence yield. - Using a cooled S-20 photo- multiplier and photon counting, visible and near-uv fluorescence was measured under excitation with VUV radiation at normal incidence. By means of a monochromator, luminescence spectra such as that shown in the inset of figure 1 were measured. The luminescence was identified with self-trapped excitons in CaF2 [13]. The incident flux was monitored using an aluminium photocathode which intercepted the outer annular portion of the beam and transmitted the central portion to the sample. This monitor photocathode was in turn calibrated against an NBS transfer standard photodiode [12] placed at the sample position. Dividing the observed fluorescence signal by the VUV flux so determined yielded the fluorescence excitation curve at the top of figure 1.
This is similar to those by Bourdillon and Beau- mont [9] for CaF, and by Beaumont, Bourdillon, and Kabler [14] for alkali halides. As in their work, a striking anticorrelation of the fluorescence yield and the reflectivity is found. There is qualitative correspondence between most features in the two spectra of figure 1 but, as noted in reference [9],
the magnitude of reflectance is not enough to account entirely for the dips in fluorescence yield. This is especially true on the high-energy side of the 35 eV structure.
Another effect which might, in principle, contri- bute to the anticorrelation of fluorescence and reflectance is photoelectron emission. However, comparison with total photoelectron yield (lower curve in figure 1, to be discussed) illustrates that because of the charging of bulk samples, photo- electron .emission is not correlated with the main fluorescence features. Note that all such measurements of fluorescence excitation in CaF, have been obtained for single-crystal samples. A second possibility is non-radiative decay of excitons near the surface, causing a dead layer with respect to fluorescence.
Several physical mechanisms for this quenching could be contributing, and further experiments are needed to sort out the possibilities. Fluorescence spectra similar to the figure 1 inset were measured for various excitation energies in an effort to detect spectral changes as a function of the depth of radiation penetration, but no significant dependence has been found in preliminary analysis of the data. Fluorescence yield measured for higher energy excitation (and correspondingly greater penetration depth) exhibits anticorrelation between near-uv fluorescence and the calcium K-shell absorption structure near 4 keV [lo].
In that case, it was proposed that re-emission of the energy via core level fluorescence competes effectively with near-uv fluorescence from self-trapped excitons.
5. Photoelectron emission. - Ultraviolet photo- electron spectra were measured using a double-pass cylindrical mirror analyzer (PHI model 15-255).
Bulk insulator samples were used for the worE -being reported, so that it was necessary to counteract the charging of samples due to photoemission. For the data presented here, a thoriated iridium filament operated between 0 and - 3 V was used as a neutraliz- ing source.
A second setup employing a spiraltron detector and simple retarding grid was used to measure total photoelectron yield under the charged conditions that occurred during fluorescence measurements.
The resulting curve at the bottom of figure 1 exhibits a valence band peak near 60 eV, indicative of a 40 eV shift due to charging. It was possible to eliminate this shift by use of the neutralizing filament.
The photoelectron energy distribution curve (EDC) for a crystal illuminated by 90 eV photons is shown in figure 2. This may be compared with data of Poole et al. [7] for evaporated films. Because of sample charging we cannot measure binding energies directly, although if one peak in the EDC can be fixed from other data, the relative energies of the others are known within 1 eV.
Figure 3 presents partial yield spectra for different energies of the analyzed electrons as photon energy
REFLECTANCE, FLUORESCENCE YIELD, AND PHOTOELECTRON SPECTROSCOPY O F CaF, C6-441
PHOTOELECTRON ENERGY (eV) Fig. 2. - The photoelectron d~stribut~on curve for single-crystal CaF, is' shown for 90 eV incident photon energy. The peaks are labeled by the atomic level notation for the initial states.
PHOTON ENERGY (eV)
Fig. 3. - Partial yield spectra of single-crystal CaF2 are shown for four values of electron kinetic energy. The electron energy passband is 0.5 eV.
is scanned. The width of the photoelectron passband .was 0.5 eV. Starting from the valence band peak at left, features of the 5 eV spectrum can be identified with each of the initial-state atomic levels marked in figure 2.
Since the incident photon flux and photoemission probability vary with photon energy, the energies of peaks in these partial yield spectra can be shifted
relative to each other due to charging, even with the use of a neutralizing filament. However, analysis of peaks excited by higher order radiation and compa- rison to the more exact relative peak energies from EDC's ,shows that the uncertainty is at most 3 eV.
"5- e ~ " PHOTOEMISSION REFLECTANCE
.'--...: CO F2
... z,
\..-*:\,.
I 8 I 0 J
2 0 4 0 60 8 0 - 100
PHOTON ENERGY (eV)
Fig. 4. - The 5 eV photoelectron partial yield curve and the 110 reflectance curve are overlaid for comparison. Because of uncer- tainty in the absolute photoelectron energy due to charging, the photoelectron curve has been translated along the energy axis so that the features near 35 eV line up.
In figure 4 the 5 eV photoelectron spectrum and the reflectance spectrum are compared by shifting the photoelectron spectrum by a constant energy to align the prominent peaks corresponding to excitation of the Ca++3p core levels. This places the valence- band photoelectron peak about 6.5 eV above the valence band edge. The 3p-derived photoelectron peak is then also about 7 eV above the threshold for core transitions to the lowest conduction band edge, i.e. the weak structure in reflectivity just above the sharp 27.7 eV core exciton. However, because the photoelectron peak strongly resembles the 35 eV reflectance peak and the 29 eV threshold of core transitions is comparatively weak in reflectivity, we presume that the photoelectron peak near 35 eV arises predominantly from Ca++3p-3d excitations.
The 3d conduction bands evidently provide a high density of final states about 5 eV above the vacuum energy, which agrees reasonably well with theory [I, 51 and with the separation of the 29.1 and 34.5 eV reflectance features. The quite different behavior of the F-2p and C a + + 3 p peaks as the final-state energy is changed in figure 3 suggests that nearly-free- electron-like conduction bands are the most important final states reached by the F-2p valence band elec- trons, while more localized d-like conduction bands are the dominant final states for Ca++3p core elec- trons. This would appear to be at least qualitatively consistent with the transition matrix elements involved.
Other phenomena evident in the complete set of data for alkaline earth fluorides will be described in a subsequent publication.
C6-442 R. T. WILLIAMS, D. J. NAGEL AND M. N. KABLER
Acknowledgments,-We wish to thank S. C. Ebner, in the course of this investigation. We also wish to L. R. Hughey, R. P. Madden, and other members of thank M. C. Peckerar and J. Schreurs for use of the the SURF-I1 staff and far-uv physics groups at cylindrical mirror analyzer.
NBS for providing research facilities and assistance
DISCUSSION
Question. - Reply. - R. T . WILLIAMS.
Have you observed the excitation spectrum for The reflectance is measured using aluminium oxide the phosphorescence ? photocathodes which are quite stable in vacuum and are calibrated against NBS transfer standard diodes over the 10-100 eV range. No reflectance standards Reply. - R. T . WILLIAMS.
No. This will be a difficult experiment because the intensity of phosphorescence continuing beyond the self-trapped exciton luminescence is low.
Question. - D. Y. SMITH.
How accurate are the absolute values of refl ectance ? Have you employed a reflectance standard or calibrat- ed your spectrometer ?
have been used. Our s-mples are cleaved crystals, with the result that there is significant diffuse scattering by cleavage steps on the surface. We have not yet made corrections for this, so that the absolute reflec- tivity shown here is somewhat too low, and has in fact been found to depend on the angle subtended by the photocathode. In addition, our data below 20 eV are uncertain because we have not yet incorporated filters to remove higher orders from this region of low first-order intensity, although numerical removal of second order has been done.
References
[I] ALBERT, J. P., JOUANIN, C. and GOUT, C., Phys. Rev. B 16 (1977) 925 ; Phys. Rev. B 16 (1977) 4619.
[2] STAROSTEN, N. V. and GANIN, V. A., Fiz. Tverd. Tela 15 (1973) 3404; Sov. Phys., Solid State 15 (1974) 2265.
[3] STEPHAN, G., LE CALVEZ, Y., LEMONIER, J. C. and ROBIN, Sov.
Phys. Chern. Solids 30 (1960) 601.
[4] RUBLOW, C. W., Phys. Rev. B 5 (1971) 662.
[5] HAYES, W., KUNZ, A. B. and KOCH, E. E., J. Phys. C 4 (1971) L-200.
[6] LE COMTE, A. and ROBIN, S., Opt. Acta 19 (1972) 203.
[7] POOLE, R. T., SZAJMAN, J., LECKEY, R. C. G., JENKINS, J. C.
and LIESEGANG, J., Phys. Rev. B 12 (1975) 5872.
[8] FRANDON, J., LAHAYL, B. and PRADAL, F., Phys. Stalus Solidi B 53 ( I 972) 565.
[9] BOURDICLON, A. J. and BEAUMONT, J. H., J. Phys. C 9 (1976) L-479.
[lo] BIANCONI, A., JACKSON, D. and MONAHAN, K., Phys. Rev.
B 17 (1978) 2021.
[I 11 MADDEN, R. P. and EDERER, D. L., J. Opt. SOC. Am. 62 (1972) 722.
1121 CANFIELD, L. R., JOHNSTON, R. G. and MADDEN, R. P., Appl. Opt. 12 (1973) 1611.
[13] WILLIAMS, R. T., KABLER, M. N., HAYES, W. and STOT~, J. P., Phys. Rev. B 14 (1976) 725.
1141 BEAUMONT, J. H., BOURDILLON, A. J. and KABLER, M. N., J. Phys. C 9 (1976) 2961.
