Photoemission Fermi Surface Topology Studies Of Magnetic Alloys
Photoemission Fermi Surface Topology Studies of Magnetic Alloys
M. Hochstrasser,1 , R.F. Willis1 , F.O. Schumann2 , J.G. Tobin2 , and Eli Rotenberg3
1The Pennsylvania State University, University Park, State College, Pennsylvania 16802, USA
2Lawrence Livermore National Laboratory, University of California, Livermore, California 94550, USA
3Advanced Light Source, Ernest Orlando Lawrence Berkeley National Laboratory,
University of California, Berkeley, California 94720, USA
ABSTRACT
Cuts through the Fermi Surfaces of CoNi and FeNi alloy films epitaxially grown on fcc Cu(100)
reveal topological features in angle-resolved photoemission. Changing the stoichiometry of these
pseudomorphic films permits us to observe changes in the Fermi surfaces. Increasing Co or Fe
content, increases the ferromagnetism from that of pure Ni. The average moment increases linearly
in the case of CoNi alloys but is arrested and sharply declines in FeNi. This work attempts to
identify the electronic features underpinning this difference.
INTRODUCTION
Measurements with the surface magneto-optic Kerr effect (SMOKE) and magnetic dichroism of
the Fe, Co, and Ni 3p core levels in photoemission with linearly polarized light (XMLD) have
shown that the elemental moments remain constant and finite, while the magnetization increases
with increasing Co and Fe content [1]. In the FeNi alloys, this linear increase of the average
magnetic moment, shows a sharp decrease for richer Fe concentrations [2]. Mössbauer
spectroscopy and SQUID magnetometry have shown that in the Fe-rich FeNi alloys an
antiferromagnetic phase emerges, which coexists with the ferromagnetic phase [3]. It is this
‘mixed’ phase, rather than any collapse in the magnitude of the magnetic moments that is, believed
to be responsible for the observed decrease in the ferromagnetic order.
RESULTS AND DISCUSSION
States at the Fermi energy are recorded in the high symmetrical k-plane for films 8 ML thick to
avoid any contribution from the Cu substrate. Patterns 1(a), (b), (c) and (d) show data for a (110)
cut through Brillouin zone in extended k-space. The alloy concentrations are from left to right: pure
Co (a), Co Ni (b), Co Ni (c), pure Ni (d). In this (110) plane, one feature remains
60
40
40
60
characteristic of the whole concentration range and sharply defined. It connects the L points in the
fcc Brillouin zone more or less continuously throughout momentum space. This feature is due to
the sp band and forms a ‘dogbone’ hole pocket, similar to that in copper.
As the number of holes in the alloy increases going from pure Ni towards pure Co, emission from
new states occurs at the center of the zone as well as across the neck of this ‘dogbone’ structure. In
simple terms, with increasing Co concentration the Fermi level shifts to lower energies and cuts
through the d-band in the center of the zone near the Γ point, as suggested by the pseudomorphic
bandstructures of Cu, Ni, Co, and Fe [4]. Also with increasing Co concentration, the exchange
splitting of the minority and majority d-bands increases in a way that would indicate that this
intensity surrounding the zone center is due mainly to minority spin polarized d-states. The
‘dogbone’ structure, a characteristic feature due to the sp-band collects some d-character but does
not change dramatically. This ‘dogbone’ is a characteristic feature of these fcc 3d materials. Across
the neck of this dogbone an increase of intensity is observed with increasing number of holes
suggesting increasing hybridization with the sp-band close to the Fermi level in this particular
region of k-space. This enhances electron scattering between d and sp-states.
a
b
c
d
Co
Co60Ni40
Co40Ni60
Ni
Figure 1. Fermi contour maps in the [110] plane obtained in the photon energy range between 80 eV and 195
eV of a pure fcc Co film (a) fcc alloy films Co Ni (b), fcc Co Ni (c), and a pure fcc Ni film (d), all
60
40
40
60
grown epitaxially on Cu(100).
The FeNi alloys show a similar evolution of states at the Fermi energy with increasing Fe content.
Again, we observe a strong contribution due to the sp-band and a clearly defined ‘dogbone’
feature. Enhanced intensity in the center of the Brillouin zone and across the neck of this dogbone
again appears with increased emptying of the d-bands. We observe well defined states in this
momentum space, which is a strong indication that the muffin-tin potential on the different atomic
sites, Fe, Ni, respectively Co is very similar. Any increased diffuseness in the patterns appears
mainly in the d-states, which is expected since the lifetimes of these states is less due to increased
scattering out of these states.
Γ
U
Γ
e
Γ
Γ
W
d
X
X
X
c
K
W
U
W
U
a
b
X
W
(a) Copper
(b) Nickel
(c) Cobalt
(d) Ni band structure
Figure 2. Fermi contour maps taken with a photon energy of hν=160 eV of a (a) Cu(100), (b) pure fcc Ni film
on Cu(100), and (c) pure fcc Co film on Cu(100), (d) calculated bandstructure at E in the (100) plane.
F
Figure 2 compares Fermi surface photoemission plots from single crystalline Cu (a), and pure fcc
films of Ni (b), and Co (c) on Cu(100). All the films have a thickness of 8 ML and the photon
energy was fixed at hν = 190 eV while the angle was varied between 0 and 30° off normal and
±120° in the surface plane. At this specific photon energy, a sector of the energy sphere just
encloses the (100) plane at the zone boundary, for emission angles smaller than ±20°. Also shown
in Figure 2d is the calculated bandstructure in the (100) plane for Ni [5]. For Cu, contributions
from the sp bands on the spherical parts of the Fermi surface are observed, whereas in Ni and Co
the d-hole pockets at the X points make an appearance. The sp band feature in Cu begins to shrink
towards the Γ-point, while the d-hole pockets grow around the X-point, in good agreement with
the calculated bandstructure.
CONCLUSION
We have measured contour maps of states in momentum space at the Fermi energies of a series of
binary alloys of the magnetic transition metals. These alloys possess the same fcc crystallographic
structure as that of the Cu(100) substrate when grown as ultrathin epitaxial layers. This implies that
the position of the Fermi level is a function of the hole concentration in the d-band, which we vary
by alloying the various elements. This is why we expect the Fermi surface contours in k-space to
evolve gradually with alloy stoichiometry, greatly aiding identification. We observe sharply
defined sp-states throughout k-space characterized by the same ‘dogbone’ structure as that
observed in copper. Emptying the d-band leads to states appearing localized in the dogbone region
and as a diffuse region at the zone center. Mixing of d- and sp-states, mainly of minority spin
polarization, occurs in specific regions of k-space from which derive the “spanning wavevectors”
responsible for Fermi surface oscillations and coupling between magnetic layers in spin-valve
heterostructures [6]. Future studies will concentrate on these Fermi surface spin-polarized
“hotspots”.
ACKNOWLEDGMENT
This work was funded by a grant from the Department of Energy, Office of Basic Energy
Science, DOE instrumentation grant # DE-PG02-96er 45595.
REFERENCES
1. F.O. Schumann, S.Z. Wu, G.J. Mankey, and R.F. Willis, Phys. Rev. B, 56, 2668 (1997).
2. F.O. Schumann, R.F. Willis, K.G. Goodman, and J.G. Tobin, Phys. Lett., 79, 5166 (1997).
3. J.W. Freeland, I.L. Grigorov, and J.C. Walker, Phys. Rev. B, 57, 80 (1998).
4. F.J. Himpsel, J.E. Ortega, G.J. Mankey, and R.F. Willis, Adv. in Physics 47, 511 (1998) and
References therein.
5. C.S. Wang and J. Callaway, Phys. Rev. B. 15, 298 (1977).
6. S.S.P. Parkin, N. Moore, and K.P. Roche, Physical Reviev Letter, 64, 2304 (1990).
Principal investigator: Dr. Roy F. Willis, Department of Physics, The Pennsylvania State University. Email:
willis@physics.psu.edu. Telephone: 814-865-6101.
