Journal of Electron Spectroscopy and Related Phenomena 183 (2011) 80–93
Contents lists available at ScienceDirect
Journal of Electron Spectroscopy and
Related Phenomena
journal homepage: www.elsevier.com/locate/elspec
Review
Understanding interface properties from high kinetic energy photoelectron
spectroscopy and first principles theory
Sari Granroth a,d , Weine Olovsson b , Erik Holmström c , Ronny Knut d , Mihaela Gorgoi e ,
Svante Svensson d , Olof Karis d,∗
a
Department of Physics and Astronomy, University of Turku, FIN-20014 Turku, Finland
Department of Materials Science and Engineering, Kyoto University, Sakyo, Kyoto 606-8501, Japan
Instituto de Física, Universidad Austral de Chile, Valdivia, Chile
d
Department of Physics and Materials Science, Uppsala University, SE-751 21 Uppsala, Sweden
e
Helmholtz Zentrum Berlin, BESSY II, Albert-Einstein-Str. 15, 12489 Berlin, Germany
b
c
a r t i c l e
i n f o
Article history:
Available online 24 August 2010
PACS:
79.60.Jv
71.20.Be
73.20.−r
Keywords:
High kinetic energy photoemission
Multilayers
Interface properties
Alloying
Electronic structure
Ab initio theory
a b s t r a c t
Advances in instrumentation regarding 3rd generation synchrotron light sources and electron spectrometers has enabled the field of high kinetic energy photoelectron spectroscopy (HIKE) (also often denoted
hard X-ray photoelectron spectroscopy (HX-PES or HAXPES)). Over the last years, the amount of investigations that relies on the HIKE method has increased dramatically and can arguably be said to have given
a rebirth of the interest in photoelectron spectroscopy in many areas. It is in particular the much increased
mean free path at higher kinetic energies in combination with the elemental selectivity of the core level
spectroscopies in general that has lead to this fact, as it makes it possible to investigate the electronic
structure of materials with a substantially reduced surface sensitivity. In this review we demonstrate
how HIKE can be used to investigate the interface properties in multilayer systems. Relative intensities
of the core level photoelectron peaks and their chemical shifts derived from binding energy changes are
found to give precise information on physico-chemical properties and quality of the buried layers. Interface roughening, including kinetic properties such as the rate of alloying, and temperature effects on the
processes can be analyzed quantitatively. We will also provide an outline of the theoretical framework
that is used to support the interpretation of data. We provide examples from our own investigations
of multilayer systems which comprises both systems of more model character and a multilayer system
very close to real applications in devices that are considered to be viable alternative to the present read
head technology. The experimental data presented in this review is exclusively recorded at the BESSY-II
synchrotron at the Helmholtz-Zentrum Berlin für Materialien und Energie. This HIKE facility is placed at
the bending magnet beamline KMC-1, which makes it different from several other facilities which relies
on undulators as the source. We will therefore also briefly describe some of the salient design features
of this facility.
© 2010 Elsevier B.V. All rights reserved.
Contents
1.
2.
3.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The HIKE facility at the KMC-1 beamline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Interface properties in multilayer systems from HIKE and ab initio theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.
First principles calculations of core-level shifts in an inherently disordered system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.1.
Computational methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.2.
Interface mixing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.3.
Simulation of spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.4.
First principles calculations of Ni/Cu multilayers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
∗ Corresponding author at: Department of Physics and Astronomy, Uppsala University, Box 516, 75120 Uppsala, Sweden.
E-mail address: olof.karis@fysik.uu.se (O. Karis).
0368-2048/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.elspec.2010.08.004
81
81
82
82
83
83
83
84
S. Granroth et al. / Journal of Electron Spectroscopy and Related Phenomena 183 (2011) 80–93
4.
3.2.
Interface properties in a model multilayer system: Ni/Cu multilayers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.
Alloying of Pt capped Fe/V multilayers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.
Towards applications: Interface properties in all-Heusler multilayers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1. Introduction
The inherent surface sensitivity of photoelectrons with kinetic
energies in the range of 50–1000 eV is the explanation why the
photoelectron spectroscopy has been key to many important contributions in the field of surface science (for example, see Ref.
[1]). However, as the kinetic energy of the emitted photoelectron is increased the inelastic mean free path is also increased
and can approach tens of nm for kinetic energies in the range of
4 keV and higher. Incidentally, the use of high energies to record
photoelectron spectra is not at all new. In the very early days
Prof. Kai Siegbahn and his co-workers were using Mo anodes as
excitation sources where the K˛1 and K˛2 X-ray lines are found
around 17.4 keV. Prof. Siegbahn’s interest in photoelectron binding energies originated from his interest in nuclear spectroscopy
and internal conversion, where part of the energy released in the
decay of the nucleus is taken by a core electron that is emitted.
To obtain the correct values for the nuclear transitions they thus
required accurate values for the electron binding energies of the
elements of interest [2,3]. The only detectors available to these
pioneering electron spectroscopists were Gieger–Müller (G–M)
counters that in many respects were a nuisance. Most importantly the window transmitted electrons only down to ∼4 keV.
It was hence impossible to realize anything else than electron
spectroscopy with high kinetic energy electrons. Despite the obstacles, electron binding energies and even chemical shifts were
determined with a surprising accuracy already in 1957–1958
[4–7].
The availability of new types of spectrometers with electrostatic lenses, new detectors and most importantly new narrow
band excitations sources led to enormous advances in the field of
electron spectroscopy during the 1960s. At this point the technique
was already identified to have a huge impact on the investigation of surface electronic structure. In the early stages of exploiting
synchrotron radiation as an excitation source for electron spectroscopies, Ingolf Lindau and Piero Piannetta pointed out the potential
of using monochromatized hard X-rays from a synchrotron for
photoelectron spectroscopy and also did some pioneering investigations at SLAC [8]. Despite this fact, there are not many reports of
high energy photoemission until the beginning of the 21st century.
Again this is simply a consequence of instrumentation advances.
During almost three decades, photoelectron spectroscopy had
matured into a very high resolution technique with a total resolution of some tens of meV or even better. With the availability
of new bright synchrotron sources and monochromators with a
resolving power in the range of 105 combined with electron energy
analyzers capable of analyzing electron energies in the range of
10 keV with meV resolution, technology had finally reached the
necessary level of perfection to fully enable the potential pointed
out by Lindau et al. We refer the reader to one of the excellent
reviews and facility reports that has been published over the last
few years for a more complete picture of the wealth of activities ongoing within the field of high X-ray energy photoemission
[9–18].
High Kinetic Energy photoelectron spectroscopy (HIKE) has
lately attracted large interest and rapidly developed into a promising tool to address electronic properties of buried interfaces and
81
84
88
89
91
92
92
bulk layers [19–23,14,24–28], as it is one of the few methods that
enable non-destructive bulk sensitive studies. At this point it could
be pointed out that hard X-ray photoelectron spectroscopy (HX-PES
or HAXPES) does not necessarily imply high kinetic energy of the
emitted photoelectron (i.e., the HIKE range), as tender and hard Xrays provide access to much deeper core levels. The great advantage
of HIKE is the accurate measurement of shifts in core-level binding
energies of bulk atoms, which reflect changes in chemical environment and give us information about intermixing of interface atoms
and alloying of the multilayers.
Multilayers have attracted interest in many fields because of
their numerous practical applications and interesting properties.
Nowadays, the development of technology is strongly correlating
with the advance of nanodevices, built up by multilayers and thin
films or superlattices. The thickness, composition and interface
structure of the layers are used to tailor magnetic, mechanical and
optical properties of the devices.
In this paper we review some of our work on analyzing
interface properties in multilayer systems using high kinetic
energy photoemission. We also describe the theoretical framework that is used to calculate and model experimental data.
The paper is organized as follows. We begin by briefly describing the essential features of the HIKE facility at the KMC-1
beamline at BESSY-II in Berlin. The remainder of the paper is
devoted to our multilayer studies. We first present our investigations on Cu/Ni multilayers, i.e., the system that we first
investigated using our methodology. We have later revisited this
system and we summarize our recent findings here. The following
section is devoted to the description of the theory of calculating chemical shifts in alloys and disordered systems. We then
finally present results from two systems, Fe/V multilayers and
Co2 MnGe/Rh2 CuSn multilayers, that have not been presented elsewhere to date.
2. The HIKE facility at the KMC-1 beamline
The HIKE experiments were carried out on the HIKE experimental station at the KMC-1 bending magnet beamline in BESSY, Berlin.
The beamline is equipped with a high resolution double crystal
monochromator which consists of three sets of crystals, Si(1 1 1),
Si(3 1 1) and Si(4 2 2), that can be changed within some minutes.
The resolution as a function of energy is given in Fig. 1. Despite the
fact that the source is a dipole on the BESSY-II storage ring operating at 1.7 GeV, it is possible to obtain good working conditions in
terms of resolution and flux between 1.7 and 12 keV, where there is
of course always a trade-off between these quantities. As illustrated
in Fig. 2, we note in particular that we have been able to record Ni
core level spectra from Ni metal at 12 keV excitation energy. In Ref.
[21] we used this possibility to show how differences in the screening of the core lead to differences in the satellite structure at the 1s
and 2p levels respectively. In that study, the total experimental resolution was limited by the Si(4 2 2) crystals which is approximately
1.2 eV at 12.6 keV photon energy.
In order to achieve necessary working conditions for HIKE it is
beneficial to design beamlines where concepts of soft X-ray beamlines regarding optical elements and beam paths to the experiment
(windowless design up to end-station) are married to state-of-the-
82
S. Granroth et al. / Journal of Electron Spectroscopy and Related Phenomena 183 (2011) 80–93
Fig. 1. The three sets of crystals available at the KMC-1 beamline covers a large
energy range with good resolution.
art high resolution X-ray monochromators. The KMC-1 beamline
comprises a minimum number of optical elements with focusing
optics that ensures that as much as possible of the radiation from
the dipole is used.
The HIKE end-station at KMC-1 is equipped with a Gammadata Scienta R4000 hemispherical electron analyzer. The analyzer
is modified for high transmission and high resolution at electron
kinetic energies up to 10 keV. As also pointed out by other authors
(see, e.g., the review by Takata [18]), it is important to match the
photon penetration depth to the electron escape depth as much as
possible implying that grazing X-ray incidence should be employed.
Typically grazing angles in the order some degrees are used. At the
same time electrons should be detected at normal emission to maximize the bulk contribution to the signal [18]. As shown in Fig. 3 this
geometry is well accommodated at HIKE experimental station. The
R4000 analyzer is positioned at 90◦ with respect to the beam. The
experimental chamber comprises a four-axis Omniax manipulator
with a He cryostat. The sample stage is also equipped with a boron
nitride heater which allows for sample heating well over 800 ◦ C, a
feature that is essential to the investigations presented here. More
detailed information on the experimental set up can be found in
Refs. [19,23].
Fig. 2. Ni 1s spectrum of Ni metal recorded at 12 keV excitation energy at the HIKE
facility at KMC-1. The data revealed differences in the screening for this 1s core hole
relative to the much studied 2p and 3p levels [21].
Fig. 3. The HIKE end-station at the KMC-1 beamline at BESSY-II comprises an modified Scienta R4000 electron energy analyzer, and a fluorescence detector for X-ray
absorption. The samples are introduced by through a load-lock chamber. A small
preparation chamber allows for some limited sample preparation (ion gun for sputtering and a diamond file scraper).
3. Interface properties in multilayer systems from HIKE
and ab initio theory
In this section we describe our work on characterization of interface properties in multilayer systems using high kinetic energy
photoemission. We begin by outlining a theoretical framework
capable of calculating chemical shifts and model spectra also for
a disordered system.
3.1. First principles calculations of core-level shifts in an
inherently disordered system
In this part we will demonstrate the possibility to perform first
principles calculations of core-level binding energy shifts (CLSs) in
metallic materials, especially considering interface structures in
disordered systems. It is the sensitivity of the CLS to the chemical environment of an atom that makes it useful as a tool for the
structural characterization of materials. In order to facilitate highly
accurate computations of the binding energy shifts, there are many
possible contributions which need to be considered in a theoretical
model. For example, one needs to take into account inter- and intraatomic charge transfer, the screening of the final-state core–hole,
and the re-distribution of charge due to bonding and hybridization
[29]. We solve this problem by using the so-called complete screening picture which takes the different effects into account by directly
including both initial (ground state) and final (core-ionized) state
effects in the same computation scheme. The complete screening
picture was from the start used in a thermodynamical approach,
utilizing Born–Haber cycles to determine the CLS between the free
atom and the atom inside the metal matrix [30]. The most important assumption here is that the symmetric part of the core-line in
the experimental spectrum corresponds to the full valence electron
relaxation in the presence of the core–hole [30]. A number of recent
first principles studies based on the complete screening picture
have been made; CLS in disordered alloys [31–33], the broadening
of the core-line due to disorder [34,35], surface CLS [36–38] and the
S. Granroth et al. / Journal of Electron Spectroscopy and Related Phenomena 183 (2011) 80–93
83
structural characterization of materials [39–42]. The model has also
been extended to core–core–core Auger kinetic energy and parameter shifts [43,44], with recent ab initio calculations performed for
AgPd alloys [45,46]. Brief summaries of work by some of the authors
can be found in Refs. [46,47]. Further discussion and more details
in general about binding energy shifts calculations can be found in
e.g. Refs. [29,48–50].
3.1.1. Computational methodology
All the present calculations are performed within density functional theory (DFT) [51,52]. A surfaces and interfaces Green’s
function technique is utilized, which correctly treats the boundary conditions occurring at an interface between two semi-infinite
slabs. Further, the atomic sphere approximation (ASA) is utilized
and the problem of disorder is solved using the coherent potential
approximation (CPA). The exchange-correlation function was set
to the local density approximation (LDA) parameterized according to Perdew et al. [53]. By applying the surface and interface
Green’s function technique it is now straightforward to describe a
system with an interface as a layer resolved binary alloy A(1−c) Bc (n),
with the concentration c depending on the layer n. For a detailed
description of the methodology, see Refs. [54–56] and references
therein.
cs ,
Within the complete screening picture the core level shift, ECLS
can be determined by considering the total energies of a system in
its ground (initial) state and core-ionized (final) state. For instance,
it is possible to consider the so-called generalized thermodynamic
chemical potential (GTCP) [31],
∂Etot
=
∂˛
,
(1)
˛→0
where Etot is the total energy of the system with ˛ concentration of
core-ionized atoms. The theoretical shift of a core-level i can now
be readily taken as the difference:
cs
= i = i − Ref
,
ECLS
i
(2)
Ref
i
denotes the reference system – usually chosen as the
where
bulk metal. This scheme is similar to experiment, with the GTCPs
corresponding to the binding energies.
In our calculations the final state is modeled by the removal of
the core-electron and its insertion into the valence band, ensuring the overall charge neutrality. A common approximation is to
replace the core-ionized atom of atomic number Z with the next
element (Z + 1) in the Periodic Table. However, in this way CLSs of
separate core-levels cannot be distinguished.
It is easy to generalize Eq. (2) for different kinds of shifts. In order
to obtain the layer resolved interface CLS (ICLS) which is our main
task here, it is possible to write,
EICLS (n) = IF (n) − bulk ,
(3)
where the chemical potential IF (n) corresponds to the specific
interface layer n and bulk to the pure bulk metal.
In order to compute CLSs it is also possible to apply the
Slater–Janak transition state model which in practice considers
eigenenergy shifts for fully screened partially occupied core-levels.
It was shown that the results are quite similar with the total energy
approach in the complete screening picture [33]. For further references and discussion of the transition state method see e.g. [57].
A computationally fast method is to directly consider the eigenenergy difference for a particular level i in the ground state. This ‘initial
state’ shift can be used as a first order approximation to the CLS,
is
ECLS
= −εi = −εi + εRef
,
i
(4)
here the negative sign is due to the convention of positive binding
energies.
Fig. 4. Concentration profile for a B/A5 /B system with interface alloying parameters
= 1.50, 0.75, and 0.00. The positions of the ideal interfaces are marked with vertical
dashed lines.
3.1.2. Interface mixing
Interface roughness for the thin film nanosystem is described
by a partial intermixing of the two constituent interface materials.
Consider a single interface between two metals A/B. The concentration profile around the interface due to intermixing can be modeled
by a layer resolved binary alloy composition profile A(1−c) Bc (n). The
probability of intermixing at each interface is given by a cumulative
distribution function
1
[X, ] = √
2
X
e−(x
2 /2 2 )
dx,
(5)
−∞
centered at each interface. In Eq. (5), X is the distance from the interface (centered in the middle between the atomic layers on each side
of the interface) and is the standard deviation which determines
the width of the interface mixing probability. This model is a simplification, since the inherent surface diffusion of the elements is
strongly dependent on the material combination.
The concentration profile c(n, ) of the multilayer interface is
then obtained as a sum of interface mixing probabilities in the
sample [58]
c(n, ) =
N
i=0
(−1)i (n − ni , ),
(6)
where ni is the position of interface i. Once the concentration profile
is obtained, the interface core level shifts can be calculated directly
using the method that was outlined above.
Fig. 4 illustrates the layer resolved concentration profile of the
B/A5 /B interface, described by the binary alloy A(1−c) Bc (n), for three
different values of the mixing parameter .
3.1.3. Simulation of spectra
With a fixed concentration profile determined by the parameter
, the layer dependent ICLS can be calculated by means of Eq. (3). In
order to obtain an estimation of photoelectron intensity as function
of energy the EICLS (n) must be averaged over all layers. We have
used the following model to obtain this average
I(E) =
1
g(EICLS (n), )c(n)
N
(7)
n
where g(E, ) is a normalized Gaussian function with standard
deviation = 0.1 eV at energy E. This choice of broadening of the
layer dependent ICLS energies is based on the estimated overall
resolution of the experiment (excitation and spectrometer) and
disorder broadening due to local environment effects.
84
S. Granroth et al. / Journal of Electron Spectroscopy and Related Phenomena 183 (2011) 80–93
Fig. 6. Experimental CLS as function of annealing temperature compared to calculated values when the degree of intermixing was assumed to be = 0.75. Theoretical
data for a 50–50 bulk alloy and Cu thicknesses 2, 3, 4 and 5 ML are shown.
Fig. 5. Cu 2p3/2 ICLSs in the Ni/CuN /Ni fcc (1 0 0) system for different values of the
interface alloying parameter . The symbols represent core-level shifts of individual
atomic layers and the continuous curves are the corresponding broadened spectra
according to Eq. 7. Filled symbols correspond to the central layer(s), while open
symbols correspond to layers out towards the interface. For comparison, the long
dashed and dotted lines show the bulk and diluted alloy shifts respectively.
3.1.4. First principles calculations of Ni/Cu multilayers
In this section we present first principles results of the layer
resolved Cu 2p3/2 interface core level shifts as a function of interface quality and embedded film thickness, N = 1–10 MLs in the Ni/Cu
multilayers. The concentration profiles which model the different interface qualities are controlled by a single parameter. For
the investigation we chose = 0, 0.75 and 1.5, representing the
ideal interface and more intermixed configurations. These values
are in line with previously determined sensible values of the intermixing parameter for other metallic multilayers [59–61]. ICLSs are
included for layers with at least ∼2% Cu atoms. Since the method
that we used to obtain the layer resolved shifts cannot account for
volume relaxations, a fixed theoretical lattice constant equal to the
surrounding metal is applied in all multilayer systems. If the influence from interface specific effects is small, it is expected that the
dispersion of the ICLSs is largely determined by the local coordination [41], following the trend of the corresponding disordered
alloys in bulk for the fixed metal volume.
In Fig. 5 the Cu 2p3/2 interface CLSs in a Ni/CuN /Ni fcc (1 0 0)
system are shown for the different interface qualities and thicknesses. The Cu/Ni multilayers in the calculations consist of a Cu
layer embedded between two semi-infinite Ni crystals, which form
a Ni/Cu/Ni trilayer sandwich. This setup allows us to calculate one
repetition of Ni/Cu/Ni. The geometries at the interfaces were not
relaxed and the lattice structure of Ni was used throughout the
trilayer.
The positive Cu shifts originates from the inner layers in the
spacer (filled symbols), while negative corresponds to layers fur-
ther out closer to the interface (empty symbols). The dotted lines
correspond to the CLS of dilute Cu in bulk Ni and the long-dashed
line denotes the CLS of pure bulk Cu in the lattice structure of Ni.
Starting with the perfect interfaces ( = 0) at the bottom in Fig. 5,
a fast but smooth increase of the shifts is observed as a function
of Cu layers, the atoms in the center of the slab reach a bulk metal
value when the spacer thickness is around 4 layers. A characteristic
shift at the sharp interface for N > 1 is already found at the two layer
system. The maximum spread is about 0.22 eV between the inner
and outer layers. Switching on the interface alloying, = 0.75, the
increase in ICLSs as a function of Cu thickness is not as fast, and
the bulk value is reached at about N = 5. The outer layers are now
represented by mixed alloys and give more negative shifts, closer
to the dilute limit. The total range of the shifts has also increased to
0.4 eV. When the interface alloying is increased, = 1.50 in the top
of Fig. 5, the shifts increases slowly but smoothly as a function of
spacer thickness. The spread in this case is now doubled compared
to the case of ideal interfaces. It is interesting to notice that the CLS
of the Cu layers in the middle of the slab approach the pure bulk
Cu value whereas the CLS of the most diffused Cu layers are close
to the dilute bulk limit of disordered CuNi alloys.
In Fig. 6 we show a comparison of the experimentally determined CLS [42] and the calculations. We can see that the calculation
is describing the destruction of the interfaces when the samples
were annealed above 250 ◦ . The exception is the sample that was
determined to have a Cu thickness of 2 ML that is better described
by the calculated results for 3 ML Cu. We believe that since the
inherent error of the Cu thickness determination in the sputtering
process is on the order of ±0.5 ML it is possible that the average Cu
thickness in the sample was closer to 3 ML.
An interesting option for future studies using interface CLSs is to
obtain the underlaying concentration profiles by performing Monte
Carlo simulations [62]. In order to study local environment effects
in more detail, it would also be of interest to perform supercell calculations as in the case for the disorder broadening of the spectral
core-line [34,35].
3.2. Interface properties in a model multilayer system: Ni/Cu
multilayers
Cu/Ni multilayers are technologically interesting because of the
relatively easy growth [63] and of numerous applications. They are
e.g. used, in applications related to magneto-optical recording or
sensors [64] or underbump metallization (UMB) to maintain solder wettability in flip chip interconnections [65]. Knowledge of the
S. Granroth et al. / Journal of Electron Spectroscopy and Related Phenomena 183 (2011) 80–93
Fig. 7. The Cu 2p3/2 photoemission spectra of Pt capped Ni5 Cu5 measured using
2010 eV photon energy. A reference spectrum of bulk Cu (100 ML Cu) is included.
The inset presents the Cu 2p3/2 core-level shift (CLS) as a function of temperature.
Adapted from Ref. [70].
bulk electronic structure of alloys is especially desirable to explain
the magnetic properties of both NiCu alloy [66,67] and the ordered
compounds [41,67–69]. The idea is that the intermixing at the interfaces changes the local magnetic moments due to hybridization
and the effect may be detectable in the total magnetization of the
sample [67]. Despite the numerous studies of nanostructured materials, some phenomena, such as diffusion and reactions at interfaces
are still quite poorly understood and difficult to predict within the
ultra-thin limit.
The first HIKE results on Ni/Cu multilayers were combined with
theoretical studies and the results convincingly demonstrated the
potential of HIKE to study the interface roughening and thus also
the alloying process [42]. In these studies the heating was carried
out in coarse steps. All measurements were done at room temperature after heating the samples to preset temperatures at a constant
rate of 10 K/min, and keeping them this temperature for 5 min. The
Cu 2p3/2 spectra of Pt capped Ni5 Cu5 are shown in the Fig. 7 together
with a reference spectrum of bulk Cu 2p3/2 .
There are no remarkable changes in the Cu 2p spectra measured
between 20 and 200 ◦ C. The most obvious changes in the binding
energy and in the asymmetry of the line seemed to take place above
200 ◦ C. At 250 ◦ C the spectrum has shifted about 0.2 eV towards
lower binding energy when compared to the bulk Cu 2p. After heating the sample up to 300 ◦ C the shift has become slightly smaller.
The small positive shift at lower temperatures as a comparison to
the binding energy of bulk Cu 2p was consistent with the theoretical
model described by a Gaussian distribution function with standard
85
deviation as discussed above in Section 3.1.2. It was concluded
that these changes are due to the alloying and the distribution of
the atoms becoming more homogenous. The inset in Fig. 7 presents
the variation of binding energy shift of the Cu 2p spectra as a function of temperature. As evidenced by the data, a dramatic change
of the core level binding energy occurs in the narrow temperature
range between 200 and 250 ◦ C. We have later on made extended
studies to more carefully investigate the intermixing process that
explores the role of segregation on the opposite shift of the Cu 2p
level at 300 ◦ C in a more detailed manner [70]. We summarize the
results of this extended study below.
Three sets of samples were prepared on MgO(0 0 1) substrates
using UHV-based dc magnetron sputtering. The multilayer structures included a Fe/Pt/Cu seeding buffer layer on top of which
the bilayer Ni/Cu fcc (1 0 0) unit was repeated [71]. The uppermost Pt or Ni capping layer was grown to protect the sample.
The overall geometry of the multilayer structures is the following: MgO/Fe5.6 /Pt39.2 /Cu45 /[Ni8.8 /CuN ]n /Ni8.8 /X (subscripts give the
thickness in Å). This study focuses on the samples where Cu film
thickness N in the bilayers was 9 or 3.6 Å (5 or 2 ML). More details are
given in Ref. [70]. The reference spectrum of bulk Cu was obtained
from a 200 nm thick Cu film grown on the same substrate and buffer
layer as described above.
Most spectra were measured with photon energy 2010 eV where
the overall resolution is 0.26 eV. Spectra were also taken using the
third order Bragg angle around 6 keV. This implied a higher photon
energy resolution and increased the inelastic mean free path (IMFP)
of photoelectrons. With 2010 eV photon energy the sampling depth
was estimated to cover approximately two Ni/Cu interfaces. This
photon energy was used to gain the best possible energy resolution and statistics within the time limits of the experiments. Using
6030 eV photon energy the sampling depth increased up to about
7 nm (6–7 interfaces in the case of Ni5 Cu5 /Pt). As explained in Section 2, the spectra were measured in normal emission to maximize
signal and bulk sensitivity.
The studied samples may conceptually be divided into groups
depending on cap material: In two multilayer samples, a Pt cap was
used to prevent oxidation or other contamination of the sample. In
one of the multilayers a thicker Ni layer had the same function. This
Ni cap was very gently sputtered away before the measurements.
From now on we will refer to the samples as Ni5 Cu5 /Pt, Ni5 Cu2 /Pt
and Ni5 Cu5 according to the thickness of the repeated Ni and Cu
layers (in ML) and the presence or absence of a Pt cap. The interfacial
quality was manipulated by heating the samples to temperature
between 80 and 530 ◦ C. The difference between heating steps was
usually about 20 ◦ C but at higher temperatures we used steps of
50 ◦ C. The samples were heated at a rate of 10 or 15 ◦ C/min to the
final temperature. After heating, the samples were cooled down
close to room temperature at which point spectra were measured.
The energy scale of all the spectra was calibrated by setting
the binding energy of Fermi edge to 0 eV. A background was fitted and subtracted using a Shirley background [72] as a model. The
full-width-at-half-maximum (FWHM) was estimated by fitting the
spectra with Voigt line shape with fixed Lorentzian width (constant
life time) at every temperature.
One of the central questions in the HIKE experiments of Ni/Cu
multilayers was to study the behavior in the temperature range
between 200 and 300 ◦ C where, according to the first experiments,
the most rapid changes in the Cu 2p spectra occurred. Using the
Ni5 Cu5 /Pt and Ni5 Cu2 /Pt samples, we followed a similar heating
procedure as before. However, the heating was carried out by using
smaller (10–30 ◦ C) steps between 80 and 350 ◦ C. The resulting Cu
2p3/2 spectra of Ni5 Cu5 /Pt are presented in Fig. 8a. Similar behavior
was observed for the Cu 2p binding energy in Ni5 Cu2 /Pt.
Compared to our previous investigation (conf. Fig. 7) a pronounced shoulder is now clearly present already in the spectrum
86
S. Granroth et al. / Journal of Electron Spectroscopy and Related Phenomena 183 (2011) 80–93
Fig. 8. The Cu 2p3/2 photoelectron spectra of Ni5 Cu5 Pt (a) and Ni5 Cu5 (b) measured using 2010 eV photon energy. The black line corresponds the binding energy of bulk Cu
2p3/2 . Adapted from Ref. [70].
recorded at 250 ◦ C. The intensity of the shoulder keeps increasing
as the temperature increases and at 300 ◦ C the peak has shifted
about −0.45 eV when compared to the Cu 2p3/2 spectrum recorded
at room temperature. This is more than double the shift observed in
the first experiments. Also the shape of the Cu 2p spectra in Fig. 8a
is clearly more asymmetric between 200 and 300 ◦ C than it was
in the previous studies. The binding energy of the room temperature Cu 2p3/2 spectrum is very close to the 2p3/2 binding energy of
bulk Cu (932.65 eV). The presented data reveal that the interfaces
are destroyed as a result of the heating and the sample is better
described as an alloy at higher temperatures. However, the large
shift of Cu 2p core-level and the dramatic changes in the asymmetry
of the spectra strongly suggest that we are studying a ternary alloy
instead of a binary NiCu alloy. The reported experimental binding energy shift of Cu 2p in Pt1−x Cux alloys varies between −0.3
and −0.6 eV [73,74] while it is maximum −0.25 eV in the case of
Ni1−x Cux alloys [75,76]. Also the Ni 2p core-level spectra were measured at each temperature. A most reasonable expectation would
be to observe a binding energy shift caused by alloying also for this
case. After careful calibration Ni 2p showed at most a −0.1 eV shift
between room temperature and higher temperatures. The binding
energy of Pt 4f level did not change significantly in course of the
alloying process. However, some narrowing of the Pt 4f photoemission line was observed as a function of temperature and at lower
temperatures an asymmetry was seen on the high binding energy
side.
By analyzing the Cu 2p3/2 spectra of Ni5 Cu2 /Pt sample where
the bilayers ideally only contain interface Cu atoms, we will obtain
information on the chemical shift of Cu atoms at the interface.
The Cu 2p binding energy at room temperature was approximately
0.1 eV lower than the binding energy of such spectra of Ni5 Cu5 /Pt
sample. At 250 ◦ C a shoulder appears on the low binding energy
side of Cu 2p3/2 and at 300 ◦ C we observe about −0.4 eV shift. The
changes in the asymmetry of the Cu 2p3/2 spectra of Ni5 Cu2 /Pt are
consistent with the results of Ni5 Cu5 /Pt even if the changes in spectral shape and binding energy are not as pronounced as in Ni5 Cu5 /Pt
due to lack of bulk Cu atoms in Ni5 Cu2 /Pt. Also the Ni 2p and Pt 4f
spectra of Ni5 Cu2 /Pt were consistent to those of Ni5 Cu5 /Pt. As stated
above, this gives an estimate of the interface core-level shift (ICLS)
of Cu 2p in the situation where four of the twelve nearest neighbours of Cu are Ni atoms and the rest of them are Cu atoms. Since
the interface is never strictly perfect on the atomic scale, this 0.1 eV
shift can be considered as an average ICLS in a situation where most
of the interface is sharp but still some islands may exist.
The obvious way to confirm the formation of a ternary PtNiCu
alloy was to repeat the experiments with a sample without Pt capping. We chose to study a Ni5 Cu5 sample where Pt capping layer
was replaced by 2 nm thick Ni(1 0 0) layer. The protective Ni layer
was gently sputtered using low current and short sputtering times.
Overview spectra were measured after every cycle and the sputtering was continued until no signature of oxide was found in the
Ni 2p spectrum. At this point the intensity of Cu 2p relative to the
intensity of Ni 2p increased slightly, consistent with the expectation of removing a surface oxide layer with a thickness of about
3 unit cells. Resulting Cu 2p3/2 spectra, which were measured with
2010 eV photon energy, are presented in Fig. 8b. This time changes
were observed also in Ni 2p3/2 spectra (Fig. 9) at the higher heating
temperatures.
Ni 2p starts to shift towards lower binding energies at about
200 ◦ C. The shift continues as the temperature increases indicative of a progressive alloying. At 530 ◦ C the Ni 2p3/2 has shifted
about −0.2 eV. The shift at 300 ◦ C is comparable to the results of
the Ni5 Cu5 /Pt and in line with reported Ni 2p shifts in NiCu alloys
[75,77]. The calculated Ni 2p core-level shift in a NiCu alloy is
approximately the same as the experimentally observed shift. The
binding energy shift of Cu 2p is still negative but has decreased
considerably being maximum −0.1 eV at temperatures between
320 and 400 ◦ C. Now, the shift is 50% smaller than observed in the
earlier experiments [42]. It is also interesting that Cu 2p starts to
shift back towards higher binding energies when the temperature
reaches 400 ◦ C, and is less than −0.05 eV at 530 ◦ C. We had indentified a similar behavior in the first set of Cu 2p spectra (conf. Fig. 7)
[42] but this shift was observed already at 300 ◦ C.
Fig. 9. The Ni 2p3/2 photoelectron spectra of Ni5 Cu5 measured using 2010 eV photon
energy. Adapted from Ref. [70].
S. Granroth et al. / Journal of Electron Spectroscopy and Related Phenomena 183 (2011) 80–93
Fig. 10. The Cu 2p3/2 photoelectron spectra of Ni5 Cu5 measured using 6030 eV photon energy. Adapted from Ref. [70].
To summarize our observations to this point, we identified that
depending on the heating procedure the Pt cap will alloy. This effect
was less severe in first study [42] due to the overall amount of heat
deposited was smaller (fewer temperatures). When the Pt cap is
replaced by Ni, we initially observe something that is qualitatively
similar to our first results but the shifts are smaller and the change
of sign in the shift of the Cu 2p occurs at a lower temperature.
To shed more light on this problem, we analyzed the relative
intensities of the Cu 2p, Ni 2p, and Pt 4f core levels as a function
of temperature for the Ni5 Cu5 /Pt and Ni5 Cu2 /Pt samples [70]. The
most obvious observation found was that the intensity of Cu relative to the intensity of Pt increases in both samples, Ni5 Cu5 /Pt
and Ni5 Cu2 /Pt, as a function of temperature. The intensity of Cu
relative to the intensity of Pt in Ni5 Cu5 /Pt increases very fast and
almost twice us much as in Ni5 Cu2 /Pt between 250 and 270 ◦ C. In
Ni5 Cu2 /Pt we see only small changes in the intensity of Cu relative the intensity of Ni. Since the intensities of Cu and Ni relative
to the intensity of Pt were increasing approximately at the same
rate in a relation to each other, and changes in ICu /INi curve were
very small in the Ni5 Cu2 /Pt, we could conclude that the Pt cap was
diffusing into the layers and intermixing with both Cu and Ni. The
analysis further revealed that the intermixing seems to start already
around 150 ◦ C. The intensity of Cu in the case of the Ni5 Cu5 /Pt sample increases faster than the intensity of Ni relative to the intensity
of Pt, suggesting that the relatively larger amount of Cu accelerates
the copper segregation to the surface.
By performing a similar intensity analysis for the Ni5 Cu5 sample
(Cu and Ni 2p intensity ratios) and comparing this to the Ni5 Cu5 /Pt
results we found that the Cu segregation is slightly faster in the
presence of Pt [70]. Using the additional information obtained from
the intensity analysis we are thus able to conclude that spectra
contain a signal due to Cu that is segregated to the surface of the
samples and that this segregation is accelerated when the Pt cap
starts to diffuse into the multilayer structure.
To obtain spectra that are more representative of the interior of
the multilayer, we also measured a set of Cu 2p3/2 spectra of Ni5 Cu5
at 6030 eV photon energy. These data are presented in Fig. 10. The
available photon flux at this higher excitation energy is considerably lower than at 2010 eV, often making a complete data series at
this energy intractable due to time constraints. As can be seen from
Fig. 10, the Cu 2p3/2 shifts towards lower binding energy having
about −0.2 eV shift at 450 ◦ C. This shift is reached at higher temperature than what was observed earlier [42], which we attribute
to differences in alloying process depending on the relative proximity of the layers probed to the surface. Spectra recorded at 6 keV
have much larger contribution from deeper interfaces where alloy-
87
ing may not be so fast and pronounced as it is closer to the surface.
The spectrum measured at 530 ◦ C has slightly shifted backwards
because of incipient Cu segregation to more layered structure after
alloying. These more bulk sensitive measurements give us an estimate of the Cu 2p binding energy shift in a NiCu alloy where the
concentrations of Ni and Cu are approximately the same and the
reorganization starts to have effect on the shifts only at higher
temperature.
In order to provide a more quantitative picture of the interface
modifications, we attempted a simple fitting model for Cu 2p3/2 to
estimate of the rate of alloying and to obtain a simple picture of the
chemical environment of atoms.
The idea of the model was to describe the alloying process by
describing the spectra as comprised of two components. A high
binding energy component represents the bulk Cu atoms or Cu
atoms that have more other Cu atoms than Ni or Pt atoms as a nearest neighbor (denoted Cu–Cu). Depending on the capping material
the other component represents those Cu atoms that have more Ni
(Cu–Ni) or Pt (Cu–Pt) atoms than Cu atoms as a nearest neighbors.
Despite the simplicity of the model, it captures the important
features of the alloying process. In particular we find the intensity
ratio for the fitted components shows a temperature dependence
consistent with what we expect from other intensity analysis. The
shifts between the two features in every sample also follow the
binding energy shifts found in the literature reasonably well. Considering the overall energy resolution of the experiments and the
uncertainty of fits, one realizes that this model will essentially capture the change of environment for an atom which moves from a
bulk like site (mostly Cu coordination) into a site where it can be
described as a dilute atom in an alloy.
Our HIKE study of Ni/Cu multilayers gives information on the
alloying process and the intermixing of capping material, segregation effects and the influence of heating and thickness of the
individual layers in the multilayer on these phenomena. The first
observation is that the Pt capped Ni/Cu multilayers form ternary
and binary alloys in the course of heating process. Obvious changes
in asymmetry and binding energy of Cu 2p were observed and the
intermixing of Pt with Ni and Cu was convincingly demonstrated.
The calculated core-level shifts of CuNiPt alloys presented here
support our conclusions related to Pt intermixing with Cu and Ni
and provide new information on the shifts in ternary alloys. Our
observations suggest that Pt diffusion is more pronounced in the
environment with more Cu atoms than Ni atoms (conf. Fig. 11).
This is supported by the fast changes in intensity ratios of Cu relative to Pt in the Ni5 Cu5 /Pt [70], which shows that the number of
Cu atoms in the multilayers plays an important role in intermixing. The histograms in Fig. 11 are based on the intensities of the
two-peak-fit components of Ni5 Cu5 /Pt and Ni5 Cu5 measured with
2010 eV photon energy and they present how the chemical environment of Cu atoms changes as a function of temperature. The
percentages of CuNi and CuPt alloys can be estimated by comparing the intensity changes of two fitted components in the spectra of
these samples. Considering the results presented in the histograms
we can also derive more quantitative information on the chemical environment of Ni and Pt atoms. For example, at 240 ◦ C more
Ni atoms are in an alloy than Pt atoms. At 300 ◦ C the situation has
changed and the Ni layers seem to be more organized.
The core-level binding energies in Ni1−x Cux , Pt1−x Cux and
Pt1−x Nix alloys reported in the literature are also consistent with
our experiments. Several experimental binding energy shifts (EB )
have been compiled and are compared to our results. For Ni1−x
Cux alloys the following binding energy shifts for Cu 2p and Ni 2p
have been measured: EB Ni : 0–0.6 eV [75,77], EB Cu : 0–0.25 eV
[75,76] as a function of concentration. EB Cu is close to zero when
Cu concentration is higher than 50%. EB Ni decreases as a function
of decreasing Ni concentration. EB Cu in Pt1−x Cux alloys varies
88
S. Granroth et al. / Journal of Electron Spectroscopy and Related Phenomena 183 (2011) 80–93
Fig. 11. Schematic illustration of intermixing and alloying of Ni/Cu multilayers and Pt cap as a function of temperature. The changes in the chemical environment are
described by histograms that are based on the two-peak-fit models of Ni5 Cu5 /Pt and Ni5 Cu5 . Adapted from Ref. [70].
between −0.3 and −0.6 eV [73,74] as a function of decreasing x
while EB Pt stays close to 0 eV [73,74,78]. In Pt1−x Nix alloy the
EB Ni changes between 0 and −0.2 eV and EB Pt between 0 and
0.3 eV as a function of decreasing Ni and Pt concentrations [79].
Studies of Pt on Cu(1 0 0) or Cu(1 1 1) observe intermixing of Pt and
Cu around 300 ◦ C [80,81]. For 0.8–3 ML thick Ni film on Pt (1 1 1)
alloying temperature between 450 and 700 K has been reported
[82]. The changes in the core-level binding energies and the relative intensities give valuable information on surface roughening
and alloying as a function of temperature. When it comes to multilayers it is especially important to pay attention to the changes
caused by heating since in the magnetron sputtered Cu/Ni samples
the substrate temperature plays an important role in the formation of a superlattice structure. One of the most studied properties
of binary alloys has been surface segregation which plays an important role also in the present study. In NiCu alloys the Cu segregation
is known to be a pronounced effect since the surface free energy
of Cu is significantly lower than that of Ni. The heat of formation
for the alloy is slightly endothermic and Cu segregation is strongly
exothermic. According to the X-ray photoelectron forward scattering studies of Ni–Cu–Ni (1 0 0) structures the Cu segregation in
1 ML Cu/1 ML Ni sample starts already below room temperature
and is a rapid process at about 450 K [83]. It is also known that
under equilibrium conditions the surfaces of NiCu and CuPt alloys
are enriched in Cu [84,85]. Erdélyi et al. have studied the interplay of Ni layer-by-layer dissolution and Cu segregation of finite
Ni (1 1 1) layers on semi-finite Cu (1 1 1) substrate by Auger electron spectroscopy and calculations of concentration profiles [86].
Their reported results of the time evolution of segregation and concentration profiles support our findings on Cu segregation based
on the intensity variations and decrease in Cu 2p chemical shift at
higher temperatures. This study also confirms that the temperature
and the concentration of Cu in multilayers have an important role
in the kinetics. In our experiments the segregation of Cu towards
the surface was evident in the case of Ni capped specimen. We
note that our results also allows to address the influence of the
total deposited power on the alloying which will ultimately allow
direct connections to thermodynamical data. The diffusion of Pt
made it more difficult to estimate the segregation in the Pt capped
samples, but the results indicate that Cu segregation occurs also
for these and that it may have a role on the Pt intermixing as
well.
3.3. Alloying of Pt capped Fe/V multilayers
Fe/Cr multilayers have been extensively studied because they
display one of the largest giant magneto-resistance (GMR) effects
[87]. V is next to Cr in the periodic table and also has a bcc structure like Cr. High quality Fe/V superlattices have been prepared
using a variety of techniques and Fe/V multilayers are good candidates for coupling research. Especially the effect of the thickness
of the V layer on magnetic properties has been the key question in
several studies [88–93]. Inspired by the results of changes in binding energy, cap diffusion and segregation of Cu in Ni/Cu multilayers
we continued the HIKE research by investigating Fe/V multilayers
capped with Pt and in this study we report the preliminary results
of changes in V 2p, Fe 2p and Pt 4f core-level spectra of two different multilayer samples as a function of alloying temperature.
The Pt capped Fe6 V6 (8.6 Å Fe/9.1 Å V) and Fe6 V2 (8.6 Å Fe/3.0 Å V)
(20 repetitions in each case) were grown on MgO substrate using a
S. Granroth et al. / Journal of Electron Spectroscopy and Related Phenomena 183 (2011) 80–93
89
Fig. 12. The V 2p3/2 photoelectron spectra of Fe6 V2 (a, left) and Fe6 V6 (b, right) measured using 4000 eV photon energy. The insets show the binding energy range of V 2p3/2
and Pt 4p3/2 core-levels at different temperatures.
UHV-based magnetron sputtering system. The thickness of the Pt
cap was 13 Å for both samples.
The small, about 5%, lattice mismatch between Fe and V(0 0 1),
allows epitaxial growth and formation of sharp interfaces [94]. To
investigate the interface roughening and alloying between Fe and
V the experiments were performed in a similar way to the in the
case of Ni/Cu studies. Fe/V multilayer samples were heated up to
selected temperatures between room temperature and 500 ◦ C. The
sample was kept at the heating temperature for 5 min after which
it was let to cool down before measuring the spectra. X-ray diffraction (XRD) and atomic force microscopy (AFM) measurements were
performed to monitor multilayer quality and the quality of the protective cap, before the HIKE experiments. Fig. 12a and b show the
V 2p core-levels in Pt capped Fe6 V2 and Fe6 V6 samples as a function of temperature. The inset in both figures shows the overview
spectra in the energy range of Pt 4p3/2 and V 2p3/2 .
V 2p spectra of both samples broaden and shift towards higher
binding energies as the temperature increases and the alloying of
Fe/V interfaces proceeds. The binding energy change of the V 2p3/2
is approximately +0.2 eV in Fe6 V2 and +0.4 eV in Fe6 V6 , respectively.
The largest changes in binding energy seem to take place between
300 and 400 ◦ C. In both samples, the broadening of V 2p core-level
spectrum is obvious and cannot be explained only by the effect of
background increase due to the intensity changes of Pt 4p core-level
(insets in Fig. 12a and b). When we compare the spectra of Pt 4p3/2
and V 2p3/2 the changes in intensity strongly suggest that diffusion
of Pt cap and segregation of V can be observed in these samples
similarly than we did in the case of Pt capped Ni/Cu multilayers
[70]. First the intensity of Pt 4p increases relative to the intensity of
V 2p between room temperature and 300 ◦ C and between 300 and
400 ◦ C increase of V intensity starts which we assume to be caused
by the segregation of V. The intensity changes for Fe6 V6 are not
as pronounced as in the case for Fe6 V2 , but on the other hand the
binding energy shift in V 2p is larger in Fe6 V6 . The binding energy
shift caused by alloying in Fe 2p is smaller than those observed for
V 2p spectra but a small positive shift can be distinguished in both
samples as Fig. 13 shows.
Also in the case of Fe 2p the shift in Fe6 V6 is somewhat larger
than in Fe6 V2 . The intensity changes of Fe 2p relative to V 2p and Pt
4f are small. Also the Pt 4f core-level shifts towards higher binding
energy as a function of temperature. All the spectra we present and
discuss here are taken using 4000 eV photon energy. The intensity
changes and binding energy shifts show that the interface roughening has already started around 200 ◦ C and instead of sharp interface
structure we are studying an alloy at temperatures higher than that.
Also strong signs of Pt diffusion and V segregation are visible. We
believe that this is critical information regarding the sample fabrication temperatures [89] and that the bulk sensitive core-level
spectra of Fe and V measured by HIKE bring new information on
the interface roughness and alloying of Fe/V multilayers when compared with previous studies about the topic [94,95]. An extended
account of this investigation will be published elsewhere.
3.4. Towards applications: Interface properties in all-Heusler
multilayers
The ever-increasing demand for high-density storage of data has
been made possible by several enabling technologies like the discovery of the GMR effect that gave us much increased sensitivity in
the read heads used in magnetic storage devices. Today’s standard
Fig. 13. The Fe 2p3/2 photoelectron spectra of Fe6 V2 and Fe6 V6 measured using
4000 eV photon energy.
90
S. Granroth et al. / Journal of Electron Spectroscopy and Related Phenomena 183 (2011) 80–93
is however based on the tunneling magnetoresistance effect (TMR)
due to the much superior change in magnetoresistance found for
TMR. The TMR effect is typically an order of magnitude larger than
the GMR effect. There are, however, some alarming facts on the
horizon when the quest for further increases in storage density will
lead to further reductions on the dimensions of the read head sensor. As the bit density is increased beyond 300 Gbit/in.2 it has been
predicted that current-perpendicular-to-plane (CPP) GMR heads
will have a better performance than read heads based on TMR [96].
Over the past few years there has been a number of reports on the
use of full-Heusler alloys as magnetic layer(s) in metallic CPP structures [97,98], including some read heads applications [99–101].
While somewhat higher values of MR have been reported and certain improvements in the electrical performance of read heads have
been achieved, output of CPP-GMR read heads remains to be far
inferior to tunneling magneto-resistance read heads. One reason
is that the high spin polarization is difficult to maintain in a GMR
structure due to interface related phenomena such as intermixing
and roughening.
Recently Nikolaev et al., have shown that CPP-GMR heads based
on all-Heusler structures are interesting alternatives to achieve a
CPP-GMR based read head exhibiting a MR of 7% and a RA of
4 m m2 [102,100]. These performance parameters are achieved
within the temperature constrained of the actual reader manufacturing. One of the critical factors for the performance of an
all-Heusler GMR structure is sufficiently high spin polarization
which is in reality limited due to defects and specific details
of the interface structure regarding roughness and intermixing.
It is therefore central to characterize degree of disorder at the
interfaces and investigate its effect on spin polarization and ultimately current spin asymmetry. Ambrose and Mryasov [103,104]
proposed selection criteria for choosing a combination of ferromagnetic and non-magnetic Heusler alloys to maximize interface
spin asymmetry by the right band matching at the interfaces.
This approach is an attempt to achieve higher MR in Heusler
alloy CPP-GMR structures. It is essential to understand the nature
of the interfaces between the ferromagnetic and non-magnetic
layers to describe the interface contribution to the spin dependent scattering that ultimately deteriorates the performance of the
device. We have used high kinetic energy photoemission to investigate the chemical roughness and the degree of intermixing [105]
depending on preparation conditions and post-treatment in a set
of all-Heusler GMR-type structures with varying thickness of the
magnetic and non-magnetic layers. High kinetic energy photoemission allows investigations of modification of interfacial properties
by measurements of the electronic structure as manifested in the
photoelectron spectra and the modification thereof upon changes
of layers and interfaces.
In Fig. 14 we show the first data obtained on all-Heusler multilayers using the HIKE facility at KMC-1, BESSY-II. [105]. The
ferromagnetic Heusler is in this case Co2 MnGe (CMG), while the
nonmagnetic Heusler is Rh2 CuSn (RCS). The samples were prepared
by cosputtering from elemental targets using a commercial magnetotron sputtering system. The CMG/RCS multilayers were deposited
on commercial Si wafers with a sputtered seed comprised of 12 Å
Ta/20 Å Ru/10 Å Co0.7 Fe0.3 alloy. It has been demonstrated that it
is possible to fabricate (1 1 0)-textured films with some amount of
highly ordered L21 phase, using this method [102,103]. We present
data from three different multilayer structures varying in thickness of the constituent layers. Two samples were manufactured
to provide a signature of the interface contribution. These samples
contain layers where the Co2 MnGe and the Rh2 CuSn layers are only
Fig. 14. Core-level spectra obtained from Rh2 CuSn/Co2 MnGe (CMG/RCS) multilayers annealed at different temperatures. The changes on the different core levels upon heating
are complex and a straight-forward interpretation is difficult. The trend however suggest that the thin layers are very succeptible to inter-diffusion while an increased order
may be obtained for the thicker layers at higher annealing temperatures.
S. Granroth et al. / Journal of Electron Spectroscopy and Related Phenomena 183 (2011) 80–93
two unit cells thick, respectively. The samples are denoted (CMG)
6 Å/(RCS) 18Å and (CMG) 18 Å/(RCS) 6 Å, respectively. It should be
noted that the growth for such very thin layers is known to be nonoptimal in terms of structural quality. A third set comprises thicker
layers of the geometry (CMG) 18Å/(RCS) 18 Å implying that each
layer is approximately 6 units cells thick.
In this first compilation of results we present data at only two
temperatures, 20 and 500 ◦ C, respectively. All presented data is
recorded using a excitation energy of 4 keV using the Si(3 1 1) crystals. A more extended study will be published elsewhere [105]. We
start by noting that the modifications in the samples upon heating,
as manifested by the changes in core-level data, are substantial.
Substantial chemical shifts for a given core level for the different multilayer samples suggest that we have contributions to the
chemical shift from electronic structure effects due to the multilayer structure. We will first consider the changes occurring in the
non-magnetic Rh2 CuSn layers and begin with the Cu 2p spectra.
There is no shift of the binding energy for the (CMG) 18 Å/(RCS)
18 Å and (CMG) 6 Å/(RCS) 18 Å samples, only a change in line profile (a narrowing and broadening, respectively), while the (CMG)
18 Å/(RCS) 6 Å exhibits a shift to lower binding energy. For the Rh
3d spectra we find much larger changes upon heating. Firstly, the Rh
core level shifts to lower binding energy for all situations. There is
also a large change in the line profiles for all cases. Here the (CMG)
18 Å/(RCS) 18 Å and (CMG) 18 Å/(RCS) 6 Å narrows substantially,
albeit in different manners. The Rh 3d (CMG) 18 Å/(RCS) 18 Å core
level loses a high binding energy component, coinciding in energy
with the 3d core level for the (CMG) 18 Å/(RCS) 6 Å sample. It also
narrows on the low binding energy side (∼306.5eV), coinciding in
energy with where we find an asymmetry developing for the (CMG)
6 Å/(RCS) 18 Å sample.
The sign of the observed chemical shifts tells us that there are
modifications to the electronic structure that are not shifts of the
Fermi energy. Instead we find that the chemical environment is
modified in a manner where initial and final state effects contribute
leading to differences both in sign and in magnitude for Cu and Rh
in the non-magnetic layer. It is instructive to consider the effect of
introducing a core ionized impurity into the layer to get an intuitive
impression of the final state effect. This concept is well-developed
as a way to understand shifts in core level spectra and is commonly
referred to as the (Z + 1) approximation when the core ionized atom
is approximated by the element with one unit higher nuclear charge
[30,43,49]. A core ionized Cu impurity would in this model thus be
considered to be a Zn impurity in the system. Depending on the local
chemical environment we find a high Cu 2p binding energy for the
situation where we expect the Rh2 CuSn layer to be most ordered,
i.e., for the (CMG) 18 Å/(RCS) 18 Å and a high binding energy component also for the (CMG) 6 Å/(RCS) 18 Å though this core level is
significantly broadened. We can thus picture the modifications of
the Cu 2p core level as a measure of how ordered the local environment is. We find that a similar conclusion holds for the Rh data
though the shift of the Rh line is negative towards lower binding energy. Here we find the lowest binding energy for the (CMG)
18 Å/(RCS) 18 Å sample. This line narrows substantially and the centroid moves to lower binding energy suggesting that we achieve
more order in the Rh2 CuSn layer upon annealing at least at this
thickness.
Applying the same type of analysis to the magnetic Co2 MnGe
layer reveals that we observe the lowest binding energy for both
Co 2p as well as Ge 2p for what initially corresponds to the most
ordered situation, the (CMG) 18 Å/(RCS) 18 Å. Upon annealing the
Ge core level data changes little with almost no discernible shift
for the (CMG) 18 Å/(RCS) 18 Å sample, a small shift of the centroid
and a narrowing for the (CMG) 18 Å/(RCS) 6 Å and a broadening
and negative shift of the Ge line for (CMG) 6 Å/(RCS) 18 Å. The data
obtained for the Co 2p core level show a shift to lower binding
91
energy upon annealing to 500 ◦ C, but little change in the line profiles
for the (CMG) 18 Å/(RCS) 18 Å and (CMG) 18 Å/(RCS) 6 Å samples.
For the (CMG) 6 Å/(RCS) 18 Å sample on the other hand, we find
that the Co core level predominantly modified by a broadening on
the low binding energy side.
The picture that evolves from this data set is already quite
involved and complicated. It is not straightforward to provide a
unique interpretation of the data without model support. However,
already from this preliminary study we can make some tentative
conclusions regarding the stability of the constituent layers. The
modification of the core level spectra upon annealing can be understood as the consequence of strong intermixing occurring for thin
Co2 MnGe layers. The data for the (CMG) 6 Å/(RCS) 18 Å sample suggests that the interface quality is poor already for the as-grown
samples and deteriorates further upon annealing. For the other two
samples, the data are consistent with increased order in the layers
upon annealing and especially the (CMG) 18 Å/(RCS) 18 Å sample
appears to be quite robust for post-annealing even up to 500 ◦ C
which is very important as it would allow post-annealing at much
higher temperatures than what is presently used to achieve a higher
degree of structural order which will improve performance of the
final devices.
This project is ongoing and more HIKE experiments are underway which will be corroborated with characterization of both
transport and magnetic properties as a function of temperature.
4. Conclusions
We have demonstrated how high kinetic energy photoelectron spectroscopy can be used to investigate interface properties
of multilayer samples with different compositions and degree of
complexity. The bulk sensitivity and non-destructive character of
the HIKE method were exploited to observe interface roughening
and intermixing. Destroying of interfaces and first signs of alloying were seen already at low temperatures for the capped Fe/V and
Cu/Ni samples, which brings valuable information related to sample preparation. Another important result connected especially to
multilayer fabrication process is the diffusion of protective capping
layers into the sample. Analysis of the relative core-level binding
energies and peak intensities gives information on the physical
and chemical phenomena occurring in the multilayers. However,
exact information about the interface quality is hard to obtain without theoretical support to model the measured data. For the Cu/Ni
system we have however deployed a simple two-peak-fit model
for Cu 2p spectra that can be used to qualitatively and even at
some level also quantitatively estimate the interface roughening.
The relative intensity variations of core-level photoelectron peaks
of Cu, Ni and Pt and the components of the two-peak-fit model
of Cu 2p as a function of temperature give information about the
kinetics.
We have also demonstrated how theoretical layer resolved core
level binding energy shifts can be used as a tool for determining
interface qualities by only using a single parameter to describe the
intermixing profiles in thin film nanosystems, allowing for conceptually simple and efficient computations.
We have presented detailed layer resolved calculated Cu 2p3/2
core-level binding energy shifts for Cu layers embedded in Ni that
may serve as a basis for modeling more complicated interface structures on the sub-nanometer scale.
We have also shown how chemical shifts can be analyzed even
for relatively complex systems. We have used HIKE to investigate
interface properties in an all-Heusler (Co2 MnGe/Rh2 CuSn) based
multilayer. The material combination is considered to be a viable
alternative to TMR structures in future read heads for magnetic
memory applications. Even though the results are complex we
find that data suggest that the individual layers exhibit different
92
S. Granroth et al. / Journal of Electron Spectroscopy and Related Phenomena 183 (2011) 80–93
propensity for alloying depending on the thickness and for thicker
layers we can even achieve an increased order in the layers upon
annealing. This finding has great impact on the manufacturing
process as it is known that the performance of the device will
ultimately depend on high interface quality and high structural
perfection of the individual layers. These qualities are often mutually exclusive since interface intermixing normally occurs at lower
temperatures than what is required to achieve structurally perfect
layers.
Acknowledgments
The authors are indebted for Gabriella Andersson, Panos Panogitis, and Marcello Moreno for providing us with the Cu/Ni and
Fe/V multilayers used in our studies. We are very thankful to Oleg
Mryasov, who supplied us with the high quality all-Heusler multilayers. The authors are grateful to the Swedish Research Council
(VR), the Knut and Alice Wallenberg Foundation (KAW), and the
Göran Gustafsson Foundation in Science and Medicine for financial
support. EH likes to acknowledge financial support from FONDECYT
grant 11070115, UACH DID grant SR-2008-0, and anillo ACT/24 of
the Bicentennial Program of Science and Technology, Chile.
References
[1] N. Mårtensson, A. Nilsson, Applications of Synchrotron Radiation, Springer
Series in Surface Science, Springer-Verlag, Berlin/Heidelberg, 1995, pp.
65–126, and references therein.
[2] K. Siegbahn, in: G. Ekspong (Ed.), Nobel Lectures in Physics 1981–1990, World
Scientific Publishing Company, 1993, 1981 Nobel Lecture, Also Available from
http://www.nobel.se.
[3] K. Siegbahn, Proceedings of the Workshop on Hard X-ray Photoelectron Spectroscopy – HAXPES, vol. 547, 2005, pp. 1–7, doi:10.1016/j.nima.2005.05.007,
http://www.sciencedirect.com/science/article/B6TJM-4GBD5BP7/2/532b3736f091b1bf078697788e6c6f41.
[4] C. Nordling, E. Sokolowski, K. Siegbahn, Phys. Rev. 105 (5) (1957) 1676–1677.
[5] E. Sokolowski, C. Nordling, K. Siegbahn, Arkiv Fysik 12 (1957) 301–318.
[6] C. Nordling, E. Sokolowski, K. Siegbahn, Arkiv Fysik 13 (1958) 483–500.
[7] E. Sokolowski, C. Nordling, Arkiv Fysik 14 (1959) 557–564.
[8] I. Lindau, P. Pianetta, S. Doniach, W.E. Spicer, Nature 250 (5463) (1974)
214–215, doi:10.1038/250214a0.
[9] C.S.
Fadley,
J.
Electron.
Spectrosc.
178–179
(2010)
2–32,
doi:10.1016/j.elspec.2010.01.006, http://www.sciencedirect.com/science/
article/B6TGC-4Y9SVPN-2/2/b037486e67d9de22116f0393ece06f3e.
[10] J. Zegenhagen, B. Detlefs, T.-L. Lee, S. Thiess, H. Isern, L. Petit, L.
André, J. Roy, Y. Mi, I. Joumard, J. Electron. Spectrosc. 178–179 (2010)
258–267, doi:10.1016/j.elspec.2009.09.008, http://www.sciencedirect.com/
science/article/B6TGC-4XBR4JN-2/2/57f83bd2f42aa69fe6255f125a9751ed.
[11] L.
Kövér,
J.
Electron.
Spectrosc.
178–179
(2009)
241–257,
doi:10.1016/j.elspec.2009.12.004, http://www.sciencedirect.com/science/
article/B6TGC-4Y0D64X-1/2/62eabf344705acbdf8d93670b02ebdac.
[12] R. Claessen, M. Sing, M. Paul, G. Berner, A. Wetscherek, A. Muller, W. Drube,
New J. Phys. 11 (2009), doi:10.1088/1367-2630/11/12/125007.
[13] S. Ouardi, B. Balke, A. Gloskovskii, G.H. Fecher, C. Felser, G. Schönhense, T.
Ishikawa, T. Uemura, M. Yamamoto, H. Sukegawa, W. Wang, K. Inomata, Y.
Yamashita, H. Yoshikawa, S. Ueda, K. Kobayashi, J. Phys. D 42 (8) (2009),
doi:10.1088/0022-3727/42/8/084010.
[14] G. Panaccione, G. Cautero, M. Cautero, A. Fondacaro, M. Grioni, C. Henriquet,
G. Monaco, M. Mulazzi, F. Offi, L. Paolasini, G. Paolicelli, P. Pittana, M. Sacchi, G. Stefani, P. Torelli, Nucl. Instrum. Methods B 246 (1) (2006) 106–111,
doi:10.1016/j.nimb.2005.12.021.
[15] G.R. Castro, J.R. Rubio-Zuazo, Nucl. Instrum. Methods A 547 (1) (2005) 64–72,
doi:10.1016/j.nima.2005.05.013.
[16] Y. Takata, M. Yabashi, K. Tamasaku, Y. Nishino, D. Miwa, T. Ishikawa, E. Ikenaga,
K. Horiba, S. Shin, M. Arita, K. Shimada, H. Namatame, M. Taniguchi, H. Nohira,
T. Hattori, S. Södergren, B. Wannberg, K. Kobayashi, Nucl. Instrum. Methods
A 547 (1) (2005) 50–55, doi:10.1016/j. nima.2005.05.011.
[17] S. Thiess, C. Kunz, B.C.C. Cowie, T.L. Lee, M. Renier, J. Zegenhagen, Solid State
Commun. 132 (9) (2004) 589–594, doi:10.1016/j.ssc.2004.09.021.
[18] Y. Takata, Very High Resolution Photoelectron Spectroscopy, Springer-Verlag,
Berlin, Germany, 2006, pp. 373–397, doi:10.1007/3-540-68133-7 2.
[19] M. Gorgoi, S. Svensson, F. Schäfers, G. Öhrwall, M. Mertin, P. Bressler, O.
Karis, H. Siegbahn, A. Sandell, H. Rensmo, W. Doherty, C. Jung, W. Braun,
W. Eberhardt, Nucl. Instrum. Methods A 601 (1–2) (2009) 48–53, Special
issue in honour of Prof. Kai Siegbahn, doi:10.1016/j.nima.2008.12.244,
http://www.sciencedirect.com/science/article/B6TJM-4VGPHJW1/2/eefae1c20c5d68dfb3c6eac4bebfe199.
[20] E. Johansson, M. Odelius, M. Gorgoi, O. Karis, R. Ovsyannikov, F. Schafers,
S. Svensson, H. Siegbahn, H. Rensmo, Chem. Phys. Lett. 464 (4–6) (2008)
192–197, doi:10.1016/j.cplett.2008.09.016.
[21] O. Karis, S. Svensson, J. Rusz, P.M. Oppeneer, M. Gorgoi, F. Schafers,
W. Braun, W. Eberhardt, N. Martensson, Phys. Rev. B 78 (23) (2008)
233105, doi:10.1103/PhysRevB.78.233105, http://link.aps.org/abstract/
PRB/v78/e233105.
[22] J. Rubio-Zuazo, G. Castro, Surf. Interface Anal. 40 (11) (2008) 1438–1443,
doi:10.1002/sia.2920.
[23] F. Schäfers, M. Mertin, M. Gorgoi, Rev. Sci. Instrum. 78 (12) (2007) 123102,
doi:10.1063/1.2808334.
[24] N. Kamakura, M. Taguchi, K. Yamamoto, K. Horiba, A. Chainani, Y. Takata, E.
Ikenaga, H. Namatame, M. Taniguchi, M. Awaji, A. Takeuchi, K. Tamasaku,
Y. Nishino, D. Miwa, T. Ishikawa, Y. Ueda, K. Kobayashi, S. Shin, J. Electron.
Spectrosc. 144–147 (2005) 841–843, doi:10.1016/j.elspec.2005.01.203.
[25] G. Panaccione, G. Cautero, M. Cautero, A. Fondacaro, M. Grioni, P. Lacovig,
G. Monaco, F. Offi, G. Paolicelli, M. Sacchi, N. Stojic, G. Stefani, R. Tommasini, P. Torelli, J. Phys.: Condens. Matter 17 (17) (2005) 2671–2679,
doi:10.1088/0953-8984/17/17/015.
[26] M. Sacchi, F. Offi, P. Torelli, A. Fondacaro, C. Spezzani, M. Cautero, G. Cautero,
S. Huotari, M. Grioni, R. Delaunay, M. Fabrizioli, G. Vanko, G. Monaco, G. Paolicelli, G. Stefani, G. Panaccione, Phys. Rev. B 71 (15) (2005) 155117–155121,
doi:10.1103/PhysRevB.71.155117.
[27] Y. Teraoka, A. Yoshigoe, Surf. Sci. 507–510 (June) (2002) 797–802.
[28] D. Iusan, R. Knut, B. Sanyal, O. Karis, O. Eriksson, V.A. Coleman,
G. Westin, J.M. Wikberg, P. Svedlindh, Phys. Rev. B 78 (8) (2008)
085319, doi:10.1103/PhysRevB.78.085319, http://link.aps.org/abstract/
PRB/v78/e085319.
[29] M. Weinert, R.E. Watson, Phys. Rev. B 51 (23) (1995) 17168–17180.
[30] B. Johansson, N. Mårtensson, Phys. Rev. B 21 (1980) 4427,
http://link.aps.org/abstract//v21/p4427.
[31] I.A. Abrikosov, W. Olovsson, B. Johansson, Phys. Rev. Lett. 87 (17) (2001)
176403, doi:10.1103/PhysRevLett.87.176403.
[32] W. Olovsson, I.A. Abrikosov, B. Johansson, J. Electron. Spectrosc. 127 (2002)
65.
[33] W. Olovsson, C. Göransson, L.V. Pourovskii, B. Johansson, I.A. Abrikosov, Phys.
Rev. B 72 (6) (2005) 064203, doi:10.1103/PhysRevB.72.064203.
[34] T. Marten, W. Olovsson, S.I. Simak, I.A. Abrikosov, Phys. Rev. B 72 (2005)
054210.
[35] T. Marten, I.A. Abrikosov, W. Olovsson, B. Johansson, R.J. Cole, G. Beamson,
S.R. Haines, P. Weightman, Phys. Rev. B 79 (1) (2009) 012201.
[36] M. Aldén, H.L. Skriver, B. Johansson, Phys. Rev. Lett. 71 (15) (1993) 2449–2452.
[37] M. Aldén, I.A. Abrikosov, B. Johansson, Phys. Rev. B 50 (8) (1994) 5131–5146.
[38] K. Kokko, M. Ropo, M.P.J. Punkkinen, P. Laukkanen, M. Alatalo, L. Vitos, J. Kollár,
B. Johansson, Surf. Sci. 601 (23) (2007) 5419–5423.
[39] W. Olovsson, E. Holmström, A. Sandell, I.A. Abrikosov, Phys. Rev. B 68 (2003)
045411.
[40] W. Olovsson, L. Bech, T.H. Andersen, Z. Li, S.V. Hoffmann, B. Johansson, I.A.
Abrikosov, J. Onsgaard, Phys. Rev. B 72 (2005) 075444.
[41] W. Olovsson, E. Holmström, J. Wills, P. James, I.A. Abrikosov, A.M.N. Niklasson,
Phys. Rev. B 72 (15) (2005) 155419, doi:10.1103/PhysRevB.72.155419.
[42] E. Holmström, W. Olovsson, I. Abrikosov, A. Niklasson, B. Johansson, M.
Gorgoi, O. Karis, S. Svensson, F. Schäfers, W. Braun, G. Öhrwall, G. Andersson, M. Marcellini, W. Eberhardt, Phys. Rev. Lett. 97 (2006) 266106,
doi:10.1103/PhysRevLett.97.266106.
[43] N. Mårtensson, R. Nyholm, H. Calén, J. Hedman, Phys. Rev. B 24 (4) (1981)
1725–1738.
[44] N. Mårtensson, P. Hedegård, B. Johansson, Phys. Scr. 29 (1984) 154.
[45] W. Olovsson, I.A. Abrikosov, B. Johansson, A. Newton, R.J. Cole, P. Weightman,
Phys. Rev. Lett. 92 (2004) 226406.
[46] W. Olovsson, T. Marten, E. Holmström, B. Johansson, I.A. Abrikosov, J. Electron. Spectrosc. 178–179 (2009) 88–99, doi:10.1016/j.elspec.2009.10.007,
http://www.sciencedirect.com/science/article/B6TGC-4XMKB4R1/2/6fe305978b8ea99a66424f51113ab167.
[47] W. Olovsson, C. Göransson, T. Marten, I.A. Abrikosov, Phys. Status Solidi (b)
243 (2006) 2447.
[48] W.F. Egelhoff Jr., Surf. Sci. Rep. 6 (1987) 253.
[49] N. Mårtensson, A. Nilsson, J. Electron. Spectrosc. 75 (1995) 209–223.
[50] P. Bagus, F. Illas, G. Pacchioni, F. Parmigiani, J. Electron. Spectrosc. 100 (1–3)
(1999) 215–236.
[51] P. Hohenberg, W. Kohn, Phys. Rev. 136 (3B) (1964) B864–B871,
doi:10.1103/PhysRev.136.B864.
[52] W. Kohn, L. Sham, Phys. Rev. 140 (4A) (1965) A1133–A1138,
doi:10.1103/PhysRev.140.A1133.
[53] J.P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 77 (18) (1996) 3865–3868,
doi:10.1103/PhysRevLett.77.3865.
[54] H.L. Skriver, N.M. Rosengaard, Phys. Rev. B 43 (12) (1991) 9538–9549,
doi:10.1103/PhysRevB.43.9538.
[55] I.A. Abrikosov, H.L. Skriver, Phys. Rev. B 47 (24) (1993) 16532–16541,
doi:10.1103/PhysRevB.47.16532.
[56] A.V. Ruban, H.L. Skriver, Comput. Mater. Sci. 15 (2) (1999) 119–143, doi:
10.1016/S0927-0256(99)00003-8, http://www.sciencedirect.com/science/
article/B6TWM-3WN6VBX-2/2/a7e3ecced3dc1050b2b8c4aaf3ff803b.
[57] C. Göransson, W. Olovsson, I.A. Abrikosov, Phys. Rev. B 72 (2005) 134203.
[58] E. Holmström, L. Nordström, L. Bergqvist, B. Skubic, B. Hjörvarsson, I.A.
Abrikosov, P. Svedlindh, O. Eriksson, Proc. Natl. Acad. Sci. 101 (14) (2004)
S. Granroth et al. / Journal of Electron Spectroscopy and Related Phenomena 183 (2011) 80–93
[59]
[60]
[61]
[62]
[63]
[64]
[65]
[66]
[67]
[68]
[69]
[70]
[71]
[72]
[73]
[74]
[75]
[76]
[77]
[78]
[79]
[80]
[81]
[82]
[83]
4742–4745, doi:10.1073/pnas.0305313101, arXiv:http://www.pnas.org/
content/101/14/4742.full.pdf+html, http://www.pnas.org/content/101/14/
4742.abstract.
G. Andersson, A.M. Blixt, V. Stanciu, B. Skubic, E. Holmström, P. Nordblad, J.
Magn. Magn. Mater. 267 (2003) 234.
B. Skubic, E. Holmström, O. Eriksson, A. Blixt, G. Andersson, B. Hjörvarsson, V.
Stanciu, Phys. Rev. B 70 (9) (2004) 094421.
B. Skubic, E. Holmström, A. Bergman, O. Eriksson, Phys. Rev. B 77 (2008)
144408.
T. Marten, O. Hellman, A.V. Ruban, W. Olovsson, C. Kramer, J.P.
Godowski, L. Bech, Z. Li, J. Onsgaard, I.A. Abrikosov, Phys. Rev. B 77
(12) (2008) 125406, doi:10.1103/PhysRevB.77.125406, http://link.aps.org/
abstract/PRB/v77/e125406.
H.C. Barshilia, K. Rajam, Surf. Coat. Technol. 155 (2–3) (2002) 195–202,
doi:10.1016/S0257-8972(02)00008-7.
T. Mizuguchi, S. Terada, T. Miyauchi, A. Matsuzono, IEEE Trans. Magn. 34 (4)
(1998) 1504–1506, doi:10.1109/20.706597.
T. Leng, G. Kirkpatrick, A. Tay, L. Li, Proceedings – Electronic Components and
Technology Conference 51st, 2001, pp. 939–944.
P. Heimann, H. Neddermeyer, M. Pessa, Phys. Rev. B 17 (1978) 427–435,
doi:10.1103/PhysRevB.17.427.
A. Niklasson, B. Johansson, H. Skriver, Phys. Rev. B 59 (1999) 6373–6382,
doi:10.1103/PhysRevB.59.6373.
A. Ney, A. Scherz, P. Poulopoulos, K. Lenz, H. Wende, K. Baberschke, F. Wilhelm, N.B. Brookes, Phys. Rev. B 65 (2) (2001) 024411,
doi:10.1103/PhysRevB.65.024411.
Z. Yang, R. Wu, Surf. Sci. 496 (1–2) (2002) L23–L28, doi:10.1016/S00396028(01)01489-3,
http://www.sciencedirect.com/science/article/B6TVX442HRJ6-3/2/3d8ada7723ba5499c4aa94e18a988d72.
S. Granroth, R. Knut, M. Marcellini, G. Andersson, S. Svensson, O.
Karis, M. Gorgoi, F. Schäfers, W. Braun, W. Eberhardt, W. Olovsson,
E. Holmström, N. Mårtensson, Phys. Rev. B 80 (9) (2009) 094104,
http://link.aps.org/abstract/PRB/v80/
doi:10.1103/PhysRevB.80.094104,
e094104.
T. Mewes, M. Rickart, A. Mougin, S. Demokritov, J. Fassbender, B.
Hillebrands, M. Scheib, Surf. Sci. 481 (1–3) (2001) 87–96, doi:10.1016/S00396028(01)01000-7.
D.A. Shirley, Phys. Rev. B 5 (1972) 4709.
G. Kleiman, V. Sundaram, J. Rogers, M. de Moraes, Phys. Rev. B 23 (7) (1981)
3177–3185, doi:10.1103/PhysRevB.23.3177.
Y.-S. Lee, K.-Y. Lim, Y.-D. Chung, C.-N. Whang, Y. Jeon, Surf. Interface Anal.
30 (1) (2000) 475–478, doi:10.1002/1096-9918(200008)30:1<475::AIDSIA817>3.0.CO;2-W.
P. Barbieri, A. De Siervo, M. Carazzolle, R. Landers, G. Kleiman, J. Electron.
Spectrosc. 135 (2004) 113–118, doi:10.1016/j.elspec.2004.02.121.
P. Steiner, S. Hufner, Acta Metall. 29 (11) (1981) 1885–1898,
doi:10.1016/0001-6160(81)90114-0.
P. Steiner, S. Hüfner, N. Mårtensson, B. Johansson, Solid State Commun. 37 (1)
(1981) 73–77, doi:10.1016/0038-1098(81)90892-9.
E.-J. Cho, S.-J. Oh, J. Korean Phys. Soc. 31 (2) (1997) 323–328.
E. Choi, S.-J. Oh, M. Choi, Phys. Rev. B 43 (8) (1991/03/15) 6360–6368,
doi:10.1103/PhysRevB.43.6360.
M. Walker, C. Parkinson, M. Draxler, C. McConville, Surf. Sci. 584 (2–3) (2005)
153–160, doi:10.1016/j.susc.2005.03.050.
Y. Shen, D. O’Connor, K. Wandelt, R. MacDonald, Surf. Sci. 357–358 (1–3)
(1996) 921–925, doi:10.1016/0039-6028(96)00292-0.
C. Su, H. Ho, C. Shern, R. Chen, Surf. Sci. 499 (2002) 103–104,
doi:10.1016/S0039-6028(01)02022-2.
J. Egelhoff, W.F., J. Vac. Sci. Technol. A 7 (1989) 2060–2064,
doi:10.1116/1.575970.
93
[84] N. Saeki, R. Shimizu, Surf. Sci. 71 (2) (1978) 479–490, doi:10.1016/00396028(78)90346-1.
[85] A.M. Peter, A. Dowben (Eds.), Surface Segregation Phenomena, CRC Press,
1990.
[86] Z. Erdélyi, C. Girardeaux, Z. Tokei, D. Beke, C. Cserhati, A. Rolland, Surf. Sci. 496
(2002) 129–140, doi:10.1016/S0039-6028(01)01571-0.
[87] E.E. Fullerton, M.J. Conover, J.E. Mattson, C.H. Sowers, S.D. Bader,
Appl. Phys. Lett. 63 (12) (1993) 1699–1701, doi:10.1063/1.110689,
http://link.aip.org/link/?APL/63/1699/1.
[88] G. Harp, M. Schwickert, M. Tomaz, T. Lin, D. Lederman, E. Mayo, W. O’Brien,
IEEE Trans. Magn. 34 (4) (1998) 864–866, doi:10.1109/20.706293.
[89] E. Nordström, B. Kalska, L. Häggström, P. Blomqvist, R. Wäppling, J. Alloys
Compd. 348 (1–2) (2003) 208–213, doi:10.1016/S0925-8388(02)00835-6,
http://www.sciencedirect.com/science/article/B6TWY-4760CSD4/2/979688a7a4037d126dd6341a08a5644c.
[90] B. Kalska, P. Blomquist, L. Haggström, R. Wäppling, Europhys. Lett. 53 (3)
(2001) 395–400, http://stacks.iop.org/0295-5075/53/395.
[91] B. Kalska, L. Häggström, P. Blomquist, R. Wäppling, J. Magn. Magn. Mater.
226–230 (Part 2) (2001) 1782–1784, doi:10.1016/S0304-8853(00)01014-3,
http://www.sciencedirect.com/science/article/B6TJJ-44F777994/2/ce316c8c9b4cccbdd5e99c38eeb2a9a1.
[92] B. Kalska, L. Haggström, P. Blomquist, R. Wäppling, J. Phys.: Condens. Matter
12 (5) (2000) 539–548, http://stacks.iop.org/0953-8984/12/539.
[93] O. Eriksson, L. Bergqvist, E. Holmström, A. Bergman, O.L. Bacq, S. FrotaPessoa, B. Hjörvarsson, L. Nordström, J. Phys.: Condens. Matter 15 (5) (2003)
S599–S615, http://stacks.iop.org/0953-8984/15/S599.
[94] P. Isberg, B. Hjörvarsson, R. Wäppling, E.B. Svedberg, L. Hultman, Vacuum 48 (5) (1997) 483–489, doi:10.1016/S0042-207X(97)00003-1,
http://www.sciencedirect.com/science/article/B6TW4-3SP9110H/2/ab63294d2a0f19234ec06a05b47ff80d.
[95] T. Igel, R. Pfandzelter, H. Winter, Solid State Commun. 116 (9) (2000)
477–481, doi:10.1016/S0038-1098(00)00371-9, http://www.sciencedirect.
com/science/article/B6TVW-41FKW1D-2/2/588c3666403bae1cc49961409ba61b5b.
[96] K. Nakamoto, H. Hoshiya, H. Katada, K. Hoshino, N. Yoshida, M. Shiimoto, H.
Takei, Y. Sato, M. Hatatani, K. Watanabe, M. Carey, S. Maat, J. Childress, IEEE
Trans. Magn. 44 (1) (2008) 95–99, doi:10.1109/TMAG.2007.911022.
[97] T. Furubayashi, K. Kodama, H. Sukegawa, Y.K. Takahashi, K. Inomata, K.
Hono, Appl. Phys. Lett. 93 (12) (2008) 122507, doi:10.1063/1.2990647,
http://link.aip.org/link/?APL/93/122507/1.
[98] K. Yakushiji, K. Saito, S. Mitani, K. Takanashi, Y.K. Takahashi, K.
Hono, Appl. Phys. Lett. 88 (22) (2006) 222504, doi:10.1063/1.2207987,
http://link.aip.org/link/?APL/88/222504/1.
[99] J. Childress, M. Carey, S. Maat, N. Smith, R. Fontana, D. Druist, K. Carey, J. Katine,
N. Robertson, T. Boone, M. Alex, J. Moore, C. Tsang, IEEE Trans. Magn. 44 (1)
(2008) 90–94, doi:10.1109/TMAG.2007.911019.
[100] K. Nikolaev, P. Anderson, P. Kolbo, D. Dimitrov, S. Xue, X. Peng, T. Pokhil, H.
Cho, Y. Chen, J. Appl. Phys. 103 (7) (2008) 507–533, doi:10.1063/1.2839344.
[101] M. Saito, N. Hasegawa, Y. Ide, T. Yamashita, Y. Nishiyama, M. Ishizone,
S. Yanagi, K. Honda, N. Ishibashi, D. Aoki, H. Kawanami, K. Nishimura, J.
Takahashi, A. Takahashi, Digest of the Intermag 2005, Paper No. FB-02,
2005.
[102] K. Nikolaev, P. Kolbo, T. Pokhil, X. Peng, Y. Chen, T. Ambrose, O. Mryasov, Appl.
Phys. Lett. 94 (22) (2009) 222501, doi:10.1063/1.3126962.
[103] T. Ambrose, O. Mryasov, Half-Metallic Alloys. Fundamentals and Applications,
Springer-Verlag, Berlin, Germany, 2005, pp. 187–220.
[104] T.F. Ambrose, O.N. Mryasov, GMR spin valve structure using Heusler alloy,
United States Patent 6,876,522 (2005).
[105] S. Granroth, et al., Characterization of interface properties in all-Heusler GMR
structures, unpublished data.