HAL Id: jpa-00218664
https://hal.archives-ouvertes.fr/jpa-00218664
Submitted on 1 Jan 1979
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.
CONVERSION ELECTRON MEASUREMENTS OF CORE POLARIZATIONS IN IRON METAL
K. Burin, A. Manalio, J. Parellada, M. Polcari, G. Rothberg
To cite this version:
K. Burin, A. Manalio, J. Parellada, M. Polcari, G. Rothberg. CONVERSION ELECTRON MEA-
SUREMENTS OF CORE POLARIZATIONS IN IRON METAL. Journal de Physique Colloques,
1979, 40 (C2), pp.C2-188-C2-189. �10.1051/jphyscol:1979266�. �jpa-00218664�
CONVERSION ELECTRON MEASUREMENTS OF CORE POLARIZATIONS IN IRON METAL+
K. Burin, A.A. Manalio, J. Parellada, M.R. Polcari and G.M. Rothberg
Department of Materials and Metallurgical Engineering, Stevens Institute of Technology, Hoboken, NJ 07030, U.S.A.
Abstract.-The polarizations of 2s and 3s/4s electron spin densities at the 57Fe nucleus in iron me- tal are being remeasured. The present result for the 2s electrons is tZV'2s+^0^/^Is^.^°^—'Zl = "0.0077
±0.0040. These results are shown to agree with observed multiplet structure in X-ray photoemission.
The two spectroscopies give different results for 3s electrons, however.
In 1972 first results were published /1,2/ of a direct measurement of the spin dependence of the electron density at the 57Fe nucleus in iron metal and Fe203. In our laboratory we are carrying out a new improved version of this experiment /3-6/. Here only a brief description will be given.
In essence the Mossbauer effect is used to excite either the (-1/2 to -3/2) or the (1/2 to 3/2) transition in an absorber nucleus. The former occurs at maximum negative resonance velocity and gives rise to line 1 of the normal six line spectrum, and the latter occurs at maximum positive velocity and gives rise to line 6. These transitions are detec- ted by observing the internal conversion electrons.
The electrons corresponding to line 1 have initial spin direction up in the atom, i.e., parallel to the majority d electron spin, and those corresponding to line 6 have initial spin down. Since the inter- nal conversion coefficient is proportional to the electron density at the nucleus, iji2(0), it is pos- sible to determine the spin dependence of ij; (0) from the intensities of lines 1 and 6. The electrons are further separated in an energy analyzer before detection so that the spin densities may be deter- mined for the different atomic shells individually.
The results for the n atomic shell are ex- pressed in terms of
where N is the electron count rate and a the in-
n n ternal conversion coefficient of the s electrons if
the p electrons are neglected. At present the value
obtained in our ongoing remeasurement of iron metal is, for the L shell, 6 = -0.0071±0.0037. The Pi/2,3/2 electrons make a small but non-negligible contribution to this result. If for lack of a better estimate they are assumed to have zero polarization, one finds 62 g = -O.0O77±O.0O4O. A 2p polarization as large as 1 percent would change 6» by only 0.001.
This result agrees with the earlier value /2/ of 62 g = -0.0063±0.0015.
These values are about a factor of three lar- ger in magnitude than various theoretical calcula- tions /7,8/ and correspond to a contribution to the internal magnetic field at the nucleus of about -1600 kG. At our present state of understanding of hyperfine interactions in solids there does not seem to be any positive contribution large enough to re- sult in the well known value for the total internal field in iron metal of -330 kG at 298 K. It is im- portant, therefore, to reexamine in detail the assumptions upon which the interpretation of the ex- periment is based. We have considered several influ- ences but only the one reported here has so far pro- ved significant.
It is observed that in X-ray photoelectron spectroscopy (XPS) of 2s and 3s electrons in 3d tran- sition metal compounds /9,10/ the spectra show a satellite structure attributed to the exchange in- teraction between the remaining unpaired s electron and the unfilled 3d shell. For example, the ground states of the ions F e3 + and M n2 + are 6S due to the (core) 3s23p63ds configuration. Creation of a Is, 2s or 3s hole while keeping the remaining orbitals fro- zen results in 7S and 5S ion states, the former when a spin down electron is removed and the latter when fSupported by the National Science Foundation under
grant DMR77-09911.
JOURNAL DE PHYSIQUE Colloque C2, supplément au n° 3, Tome 40, mars 1979, page C2-188
Résume.- Les polarisations de spin des électrons 2s et 3s/4s au noyau 57Fe du fer métallique ont été remesurées. Pour les électrons 2s on a trouvé ]j;2s-t'(0)/,'4s4-(0)~1 = ~0,0077±0,0040. Ces résultats sont conformes à l'observation de la structure du multiplet en photoémission des rayons X. Cependant, les deux spectroscopies donnent des résultats différents pour les électrons 3s.
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1979266
a spin up electron is removed. The two states differ in energy because of the exchange interaction. The relative intensities are just the ratios of densi- ties of final states, i.e., 7/5. This simple model works well for the 2s multiplet of Mn2+ compounds where the predicted intensity ratio of 1.4 is found.
The data on Fe3+ give an energy splitting in agre- ement with the model but are not adequate for accu- rate intensity ratios. We will assume this model is correct for Fe3+ also. The Is multiplet is also expected to have the same intensity ratio but has not been observed, perhaps due to too small a split- t ing
.
In the Gssbauer measurements lines 1 and 6 correspond, in the case of ~ e in FezO3, to the ~ + remaining ion in, respectively, the 5~ or 7~ state.
One might expect, then, that Nnl/NnB = 0.71 for n = I or 2, since in equation (I) the electronic part of the internal conversion matrix element is essentially the same as in photoemission except that the :electromagnetic field of the nucleus is substi- tuted for the external photon field and the coupling is MI instead of El. The results for the Is, 2s and 3s electrons from Fez03 and iron metal show, however, ratios equal to 1.0 within about 1%. For the 1s and 2s electrons these different results from XPS and ME may be reconciled as follows. In the Gssbauer expe- riment on Fe203, assuming ~ e ~ + , all the initial and final values of nuclear, ion, and outgoing electron angular momenta and Z-components of angular momenta are known
.
Thus the I and 6 transitions are each from a single initial state to a single final state and the ratio of densities of final states is 1:l and not 5:7. Iron metal has a more complicated elec- tronic structure, but in this case also a single final state seems to be involved.The 3s intensities cannot be explained by a simple density of final states argument because it has been found that in the 3s XPS spectra of Mn2+
and ~ e compounds the ratios of the ~ + 7 ~ : 5 ~ multiplet lines are greater than 2 instead of equal to 1.4.
This is attributed to electron-electron correlations between 3s, p and d electrons /11,12/. Expressing
these correlations by a configuration interaction calculatioa amounts to mixing into the primary 3 ~ 3 ~ ~ 3 d configuration other configurations capable of giving 7~ and 5~ spectroscopic terms, such as for example 3s23p43d6(5~) or 3~3~~3d'('S). Since inter- nal conversion and photoemission are one-electron processes, these other configurations are not exci- ted. However, since they are present in the final
state wavefunction the amplitude af the primary con- figuration, C is reduced below the value 1.0. The
P'
transition rate is proportional to
c2
which forP'
Mn2+ is calculated to be 1.0 ('5) and 0.63
('s).
The XPS ratio of intensities is then (7/5) (1.010.63) = 2.2 in agreement with experiment. Using the same argument as for the 2s case in the Gss-auer experi- ment one should find N31/N36 = 0.63 instead of the observed 1.0 if the same C are used.P
Thus the 3s spectra are not understood. One pos- sibility is that the observed structure in XPS is not due to multiplet splitting, which seems unlikely in view of other evidence not discussed here. Another possibility is that the two experiments may be obser- ving different final states; for example there may be some charecteristic times that allow for different relaxation processes in the electronic structure.
References
/I/ Song, C.J., Trooster, J., Benczer-Koller, N.
and Rothberg, G.M., Phys. Rev. Lett.
2
(1972) 1165./ 2 / Song, C.J., Trooster, J., Benczer-Koller, N.
Phys. Rev. B z (1974) 3854.
131 Polcari, M.R. et al., "Proc. Conf. Phys. of Transition Metals", Toronto, Canada (1977).
/4/ Parellada, J. et al., "Proc. Int. Conf.
MGSS-
bauer Spectr.", Bucharest, Romania (1977) 13.
/5/ Polcari, M.R., Dissertation, Stevens Institute of Technology (1978).
/6/ Parellada, J., Dissertation, Univ. of Barcelona, Spain (1978), Extended English abstract availa- ble.
/7/ Duff, K.J. and Das, T.P., Phys. Rev. B E (1975) 3870.
181 Callaway, J. and Wang, C.S., Phys. Rev. B G (1977) 2095.
/9/ Wertheim, G.K., Hiifner, S. and Guggenheim, H. J., Phys. Rev. BL (1973) 556.
/lo/ Kowalczyk, S.P. et al., Phys. Rev.
BL
(1973)4009.
/I11 Bagus, P.S., Freeman, A.J. and Sasaki, F., Phys.
Rev. Lett.
30
(1973) 850./12/ Viinikka, E.K. and U h m , Y., Phys. Rev. B s (1975) 4168.
