Measurement of electron energy losses in strontium

Measurement of electron energy losses in strontium

Volume 2 7A. number 8 PHYSIC8 9 LETTERS September1968 quite similar to those observed in alhali halides by Bussolati et al. A comparative examin...

173KB Sizes 1 Downloads 33 Views

Volume 2 7A.

number 8

PHYSIC8

9

LETTERS

September1968

quite similar to those observed in alhali halides by Bussolati et al. A comparative examination allows us to draw the following two conclusions: a) the annihilation rates are prevailingly determined by the negative ion; b) to each negative ion one can associate, in a first approximation, the same values of the decay rates. This is true whatever the structure of the crystal may be.

des. The indexes 0,1,2,3 individuate the components in order of increasing lifetimes; those having the same index shall presumably have the same origin. For instance, if one attributes the various components to the annihilation of positrons from the first levels of a bound system, the same index shall individuate the same level. The rough model proposed by Bussolati et al. suggests that the components with index 1 and 2 may be ascribed to the annihilation of positrons from the ground and the first excited level, respectively, of the system e+-anion; as regards the 75 and 73 components, the proposed model is unable to explain their origin. We do not report in the table the “tail” component (whose intensity is lower than 1%) as its origin shall be attributed to positron annihilation in regions other than the homogeneous interior of the crystal [5,6]. We do not intend now to discuss a particular model for framing our present results but only to examine if they give further suggestions on positron annihilation in ionic media. The inspection of table 1 shows that the general trend of the time spectrum and the values of the lifetimes are

References A. Dupasquier and L. Zappa, Nuovo 1. C. Bussolati, Cimento 52B (1967) 529. Sov. Phys. 2. V.I. Gol’dsnskii and E. P..Prokop’ev, Solid State 8 (1966) 409. This article gives references to earlier works. S. Cova and L. Zappa, Nuovo Cimento 3: C. Bussolati, 50B (1967) 256. and L. Zappa, Nuovo Cimento 52B 4. M. Bertolaccini (1967) 487. 5. H.Weisberg and S.Berko, Phys.Rev.154 (1967) 249. to be published. 6. R. Paulin and G.Ambrosino,

*****

MEASUREMENT

OF

ELECTRON

ENERGY

LOSSES

IN STRONTIUM*

B. M. HARTLEY Department

of Physics,

University

Received

of Western Australia

11 July 1968

Surface and volume plasmon energies of 3.6 f 0.05 eV and 7.8 f 0.1 eV are measured for strontium. Energy losses due to NI and NHIH ionization are identified and a peak due to a polarization wave in the NHIII band is identified in the characteristic electron energy loss spectrum.

Electron energy loss spectra have been measured for metallic strontium over a range of bombarding energies using the reflection technique described previously [1,2]. Pure strontium was mounted as the target material and scraped in a vacuum of about 5 X 10-T Torr to produce a clean surface. No measurements of surface condition could be made but measured energy losses for each spectrum indicate that there is probably no significant contamination. During the running * Work supported by the U.S. Army Research the University of Western Australia.

Office and

of the spectra the strontium surface was renewed by continuous scraping. Fig. 1 shows the energy loss spectrum of strontium for primary energy 1000 V. The two lower energy losses in this spectrum are evaluated at 3.6 f 0.05 eV and 7.8 f 0.1 eV. The 3.6 eV energy loss is interpreted as being due to excitation of surface plasmons and the 7.8 eV loss as due to excitation of volume plasmons. This interpretation is consistent with the identification of similar peaks in calcium and barium [3]. Because of the natural width of the energy losses multiple volume and surface losses are difficult 499

Volume

27A, number 8

PHYSICS

LETTERS

io’. . .-

.

10

0 III1

20

III1 E

NERGV

Fig. 1. Characteristic

30

III1

LOSS

50

40

IIII

IIII

III

(eV)

electron energy loss for strontium.

spectrum

to observe; a small peak however can be detected with energy loss about 11.4 eV and this is interpreted as a surface plus volume loss corresponding to the sum of the two previous energy losses. Calculation of the plasmon energy for strontrium yields a free electron value of 7.02 eV. The effect of ion core shielding is to lower the plasmon energy. A calculation based on the polarizability of rubidium yields a value of 5.97 for the plasmon energy for strontium. It is suggested that an intraband transition between 4 and 5 eV could cause the discrepancy between the calculated and measured plasmon energies and also explain the large value of 2.1 for the ratio between the volume and surface plasmon energies. An intraband transition at this energy would also tend to broaden the volume plasmon loss. The measured width of the volume plasmon peak is 1.6 f 0.2 eV, compared with the surface plasmon width of 0.1 + 0.2 eV. *****

500

1968

Structure in the region above 19 eV can be explained in terms of excitations of the NIIIII and NI levels. It can be shown that the positions of maxima in the energy loss function, due to ionization processes, do not occur at the ionization energy but are displaced to higher energies. The minima which precede the ionization peaks give a measurement of the ionization energy corresponding to the absorption edge [4]. The NIIIII ionization edge is, therefore, evaluated at 19.98 f 0.1 eV compared with the tabulated atomic level of 19.9 f 0.3 eV [8]. The structure of the energy loss function near the NI ionization energy makes the measurement of a minimum impossible. The peak measured at 39 f 1 eV is as expected slightly higher than the atomic ionization energy which is tabulated at 37.7 f 0.3 ev. The broad peak at 29.5 f 0.5 eV is interpreted as being due to the excitation of a polarization wave in the NIIIII band. Further details of this interpretation will be given in a future publication [4]. Briefly the NIIIII oscillator strength is sufficient to depress the value of ~1 to negative values; as ~1 become positive with increasing energy, the possibility of energy loss in the form of a plasmon oscillation or polarization wave occurs. This behaviour is similar to the double plasmon found in the optical properties of graphite

1

I

9 September

FL The encouragement and help given by Dr. J. B. Swan is gratefully acknowledged.

References 1. J. B. Swan, Phys.Rev. 135 (1964) A1467. 2. B. M.HartIey and J. B. Swan, Phys. Rev. 144 (1966) 295. 3. J. R. Robins and P. E. Best, Proc. Phys. Sot. 79 (1962) 110. 4. B. M. Hartley, to be published. 5. G. A. Taft and H.R.PhiIipp, Phys.Rev. 138 (1965) A197. 6. C.Kunz, Z.Phsyik 196 (1966) 311; B. M. Hartley, Phys. Letters 24A (1967) 396. 7. D.Pines, Rev.Mod.Phys. 28 (1956) 184. 8. J. A. Bearden and A. F. Burr, Rev. Mod. Phys. 39 (1967) 125.