Vij D.R. Handbook of Applied Solid State Spectroscopy
Подождите немного. Документ загружается.
11. X-Ray Photoelectron Spectroscopy
example, a comparison between the Mg KD and Al KD sources for Cu spectra
are shown in Figure 11.4. Note that the Cu 2p and Cu 3p peak positions
remain unchanged in both spectra, whereas the Auger peaks are shifted on the
binding energy scale due to the energy difference between the Mg KD and Al
KD lines. Cathode filament alignment is crucial in dual-anode sources
because a misaligned cathode filament may result in “cross-talk,” a spectral
defect characterized by additional ghost peaks (i.e., the appearance of Zr LĮ
excited peaks when using the Mg KD anode) [8, 11, 12].
Figure 11.4 Cu XP spectra illustrating the shift in Auger peak positions with the change from
Mg to Al anodes in the X-ray source. Adapted from [11].
These direct X-ray sources produce a high intensity photon flux reaching
However, several disadvantages are associated with these sources.
Bremsstrahlung radiation can induce sample damage. Emission from multiple
X-ray fluorescence lines leads to satellite peaks in XP spectra. Finally, the
XPS system’s resolution is limited by the natural linewidth of the source,
typically 1–2 eV. To alleviate the problems associated with direct application
from X-ray sources, monochromators are often employed in XPS systems.
Figure 11.5 shows one configuration for a monochromator used in XPS.
Monochromators are made with crystals such as quartz, silicon, and
germanium, which are used to select desired emission lines and/or remove the
Bremsstrahlung continuum [8, 11]. The lattice spacing within these crystals is
such that the characteristic X-rays emitted from the source will satisfy the
Bragg relationship given below, equation (11.2), where n is the diffraction
order, O the wavelength of light, d the lattice plane spacing, and T the angle of
incidence:
2sinndO T (11.2)
Furthermore, the dual-anode X-ray source can also be used in furthering
spectral interpretation, especially in identifying Auger electron peaks. For
490
the sample surface, resulting in high photoelectron throughput to the detector.
11.3 Instrumentation
intensity output at the desired wavelength. Advances in manufacturing X-ray
optics, such as silicon and germanium crystal monochromators, facilitates the
use of XPS in providing information regarding lateral and spatial variation in
surface composition, chemical state information such as oxidation states, and
chemical bonding configurations of the atoms based on chemical shifts [8,
11]. Use of a focused X-ray gun and a well-engineered monochromator
crystal can also reduce the size of the spot size on the sample from the mm to
the Pm range. Perhaps the most common configuration for X-ray generation is
an Al X-ray source and a quartz crystal monochromator.
Figure 11.5 Schematic of components in a monochromatic XPS source with a rotating anode.
The development of XPS has been aided by the advent of third generation
synchrotron radiation [9]. Synchrotron radiation is generated through
magnetic field pulses that induce electrons to move in circular orbits at speeds
near that of light. Continuous energy radiation is emitted as these electrons
move in circular orbits to provide photons ranging from 10 eV to 10 keV.
X-ray beams of specific energies may be selected using monochromators.
However, the high monetary and time costs involved in using synchrotrons
greatly limit accessibility for routine analysis. Additionally, the high intensity
of synchrotron radiation may increase X-ray beam-induced sample damage.
The use of synchrotron radiation for XPS analysis has recently been reviewed
elsewhere [13, 14].
In some applications, particularly those involving investigation of
coordination environment, an ultraviolet light source is used in the place of
the X-ray source. This complementary technique to XPS is called ultraviolet
photoelectron spectroscopy (UPS) [15].
Specific crystal materials are chosen for mechanical integrity to withstand
the X-ray intensity. These materials must possess small mosaicity and high
491
11. X-Ray Photoelectron Spectroscopy
Photoelectrons are produced as the incident X-rays strike the sample surface.
The photoelectrons are then directed into the electron energy analyzer usually
through a collecting lens (electrostatic or electromagnetic) to be sorted
according to their respective energies. Although the linewidths of the
measured photoelectron peaks depend mostly on the X-ray source’s natural
linewidth for non-monochromatic systems, the quality of the analyzer dictates
the sensitivity and resolution of the X-ray photoelectron spectrometer. The
most common types of electron energy analyzers for XPS are the cylindrical
mirror analyzer (CMA) (see Figure 10.7 in Chapter 10, Auger Electron
Spectroscopy) and the concentric hemispherical analyzer (CHA), which is
also called the hemispherical sector analyzer or spherical capacitor analyzer
(see Figure 10.8, Chapter 10). The CHA is the most frequently employed
analyzer in XPS due to its high resolution, which is important in obtaining
chemical state information [8–11].
Both types of analyzers are also commonly used in Auger electron
spectroscopy. As described in Chapter 10, the cylindrical mirror analyzer has
high transmission efficiency, and it is also compact in size and easy to use.
However, the CMA is very sensitive to sample-to-analyzer distance in energy
scale linearity and complicates the precise determination of Auger or
photoelectron energy. In the hemispherical sector analyzer, the photoelectrons
ejected by the sample are focused using electron optics into the entrance slits,
and the photoelectrons are separated as they travel through the electric field
between the inner and outer spheres. As the photoelectrons exit the analyzer
they enter electron detectors such as electron multiplier tubes, multichannel
plates, or anode electrodes. The electron detectors act as signal transducers to
yield a current, which is then processed to produce XP spectra.
11.4 SAMPLE SELECTION AND PREPARATION
To enable analysis of the true sample surface, cleaning and etching are
near surface chemistry and topology. However, in some cases the sample for
XPS analysis needs to be cleaned, outgassed, and reduced in size prior to
introduction into the vacuum chamber. Additional care is generally given to
sampling. The goal of sampling is to select a portion of the sample that is
representative of its entirety and to avoid undesired sample modification and
2
cleaning [8].
Sample mounting is commonly accomplished by affixing the specimen to
the sample holder using one of a variety of commercially available electrically
conductive adhesive tapes. The contamination introduced by the adhesive tape
11.3.3 Electron Energy Analyzer
492
typically avoided prior to XPS analysis because these processes will alter the
of filtered solvents, chemical etching, CO “snow,” plasma cleaning, and ion
sputtering
contamination. When necessary, samples may be cleaned using combinations
11.4 Sample Selection and Preparation
powderlike samples may sometimes be analyzed without affixing them to the
sample holder, at other times it may be necessary to secure the small
specimens on the sample stage. This type of sample mounting may be
achieved by evacuating between the sample and sample stage to create
airlocks, and by using silver paint, silver epoxy, indium foil, or conductive
tape [8].
11.4.1 Sample Charging
As photoelectrons are emitted by the sample after X-ray bombardment, the
sample becomes positively charged. The charged surface possesses a potential
that may alter the velocity of the ejected photoelectrons as they are directed
into the electron energy analyzer and distort the energy measurement of the
electrons. With conductive samples, this charging effect may be compensated
by electrically connecting the sample to the spectrometer. The electrical
Therefore, the kinetic energy of the photoelectrons can be accurately
measured if the spectrometer is set to measure the kinetic energy of the
entering electrons with respect to the spectrometer’s own Fermi level [8–10,
17–19]. Figure 11.6a depicts the energy levels for conducting samples.
E
elec
Sample Spectrometer
E
Vac
E
F
I
spectrometer
hQ
E
elec
E
KE,2
E
KE,1
E
Vac
E
F
E
core
I
sample
E
BE
Sample Spectrometer
E
elec
E
Vac
E
F
I
spectrometer
hQ
E
elec
E
KE,2
E
KE,1
E
Vac
E
F
E
core
I
sample
E
BE
E
charge
(a) Conducting (b) Insulating
E
elec
Sample Spectrometer
E
Vac
E
F
I
spectrometer
hQ
E
elec
E
KE,2
E
KE,1
E
Vac
E
F
E
core
I
sample
E
BE
Sample Spectrometer
E
elec
E
Vac
E
F
I
spectrometer
hQ
E
elec
E
KE,2
E
KE,1
E
Vac
E
F
E
core
I
sample
E
BE
E
charge
(a) Conducting (b) Insulating
Figure 11.6 Schematic for energy levels between the instrument and (a) conducting or (b)
insulating samples. Here E
elec
, E
Vac
, E
F
, E
core
are the energy levels of the free electron, the
vacuum, the Fermi level, and the core level, respectively, from which the photoelectron was
generated, hv is the energy of the incident X-ray photon, E
BE
the binding energy
of the electron, E
KE,
1
the kinetic energy of the electron, E
KE,2
the measured kinetic energy of
the electron, I
sample
the sample’s work function, I
spectrometer
the spectrometer’s work function,
and E
charge
the magnitude of the sample charging. Adapted from [20].
may be characterized prior to sample analysis by the collection of a
contamination depth profile [16]. Additionally, although small particles and
493
contact equilizes the Fermi level of the sample and spectrometer.
11. X-Ray Photoelectron Spectroscopy
Additional procedures are needed to treat insulating samples to account for
the charging effect. Electrically connecting the sample and spectrometer is
obviously not possible. As a result the Fermi level of the sample and
binding energies calculated from the photoelectrons’ kinetic energies. The
relative energy levels involved are illustrated in Figure 11.6b. The two main
problems in analyzing insulating materials through XPS stem largely from
variable surface potential of insulating materials caused by differential
potential of the insulating material as charge accumulates, which leads to
difficulties in matching measured photoelectron binding energies to tabulated
respect to location and time, in the surface potential of the insulating material,
which may greatly distort the spectral lineshapes. Because the aforementioned
problems contribute to lower spectral resolution despite the use of
monochromatic X-ray beams, the surface potential of insulating materials
must be controlled.
To control the surface potential of insulating samples, the accumulated
sample charges need to be compensated. The techniques involved in
compensating the sample charges are known as “charge neutralization” or
“charge compensation.” The goal of charge neutralization is to provide
uniform and stable surface potential. Traditionally, flood guns are used to
provide a source of low energy electrons to take the place of the ejected
photoelectrons in XPS, as shown in Figure 11.7.
Figure 11.7 The use of a flood gun in XPS. Adapted from [8].
The use of a non-monochromatic X-ray source may help in minimizing
problems associated with insulating samples. The non-monochromatic X-ray
beams are divergent and are therefore broader than monochromatic X-ray
beams [8, 10, 11]. As a result, the non-monochromatic X-ray beams may
bombard other surfaces in the sample chamber to create copious numbers of
low energy secondary electrons that may be used to compensate the emitted
494
spectrometer may be different with insulting samples, leading to error in the
values. The second problem is attributed to the heterogeneity, both with
charging. The first problem results from the lack of knowledge about the surface
11.4 Sample Selection and Preparation
photoelectrons. Additionally, the differential charging may also be reduced
through placing additional sources of low energy electrons on the sample
holder in proximity to the insulating sample. For example, the placement of a
metallic wall surrounding the sample was shown to reduce differential
charging [19]. Other techniques to reduce differential charging may include
placing the sample on a conductive platform and mixing the sample with
another substance, such as carbon black powder in a 50/50 ratio [8, 21].
There are several considerations one makes in achieving ideal charge
x To achieve uniform surface potential, the low energy electrons
emitted by flood guns should exhibit narrow energy distribution.
x The flood gun should be positioned such that the low energy electron
beam emitted by the flood gun is normal to the plane of the sample
surface to avoid deflection.
x
sample.
Charge neutralization by the application of an external bias has recently
been investigated. The external bias technique, involving applying a 0–10-V
bias to the sample during data collection, may be used to increase the
separation of overlapping peaks to improve spectral resolution [18]. The
external bias alleviated differential sample charging through controlling the
flow of low energy electrons to the sample. The researchers successfully
utilized the external bias to resolve Au 4f peaks from gold nanoparticles in a
SiO
2
matrix and Au 4f peaks from gold metal strips, resulting in a facile
method for analyzing heterostructures [18].
11.4.2 X-Ray Beam Effects
conducting XPS experiments. Materials such as metals and ceramics are
usually stable under X-ray irradiation. However, the radiation can induce
alterations in certain polymers, such as poly vinyl chloride (PVC) and soft
biological specimens, especially when the samples are in thin films [8].
Alterations of the sample induced by X-ray may be physical and/or chemical.
Some physical structural changes may include creation of point defects,
changes in near surface crystal structure, and changes in the surface
topography. Point defects are of concern when analyzing metallic oxides and
multielement samples [8]. The possible X-ray-induced point defects include
lattice and surface vacancies and dislocations that may complicate spectral
interpretation [8].
X-ray-induced chemical changes are of particular concern when using XPS
to investigate the effects of chemical treatment on a sample. Decoupling of X-
ray-induced chemical changes from those resulting from the chemical
495
neutralization for insulating samples [8]:
The flood gun electron energy should be set such that it is high enough
to eliminate space charge effect, but low enough to avoid overcharging
Because of the penetrative nature of X-ray, sample damage may be of concern in
11. X-Ray Photoelectron Spectroscopy
treatment is challenging. In addition to chemical changes brought about by the
X-ray irradiation itself, the secondary electrons that may be generated as a by-
product of the irradiation could also cause additional chemical changes.
Specifically, both secondary electron and X-ray irradiation may cause
chemical changes, including changes in oxidation states and the cleavage of
chemical bonds. Additionally, the surface chemical composition may be
altered after X-ray irradiation. The radiation may enhance the surface
reactivity and increase adsorption of background gas molecules, break
chemisorption bonds to increase desorption of surface species, or induce
migration of surface species on the surface. Generally, the use of
monochromatic X-ray sources may reduce the amount of sample damage
because much of the energy in the original X-ray beam does not contribute to
the useful spectral information. The monochromator attenuates the beam,
making it a more effective X-ray source from a damage standpoint [8, 11].
11.5 SPECTRAL ANALYSIS
As Figure 11.4 demonstrates, XP spectra are generally plotted with respect to
binding energy, increasing from right to left. The binding energy scale may be
thought of as the reverse of kinetic energy, which would increase from left to
right (see equation (11.1)). Typically, the peaks in the range from 0 eV to ~15
eV in binding energy are associated with valence electrons. These valence
electron peaks are followed by characteristic peaks of the core-level electron
ejection at higher binding energies. The valence electron orbitals are often
delocalized or hybridized such that the ejection of the valence electrons may
result in peaks so closely spaced as to present a valence band structure.
The valence band spectra may provide additional information to the core
spectra to yield valuable information in differentiating similar compounds.
For example, valence spectra were used to differentiate industrial lubricants
on an aluminum surface that exhibit extremely similar core spectra [22].
These valence band peaks are often useful for investigating not simply the
chemical environment of an atom but also changes in coordination around the
atom. An example of this type of analysis is the differentiation of
molybdenum in the 6+ oxidation state in a tetrahedral versus octahedral
coordination environment [23]. Researchers have successfully identified
sodium molybdate, which contains tetrahedral Mo
6+
, and ammonium
heptamolybdate, which contains octahedral Mo
6+
, using valence band XPS
[23]. These valence band peaks are often investigated using an ultraviolet
light source in UPS [15].
In the low binding energy region, the spectral background may be
attributed to the Bremsstrahlung radiation and secondary electron cascade
background; at higher binding energy the background may be attributed
mainly to inelastic electron scattering, and the steplike rise character may be
attributed to photoelectrons losing energy as they reach the electron detector
496
11.5 Spectral Analysis
[8, 10]. The increasing background from low to high binding energy that is
seen in Figure 11.4 is significantly reduced with the use of monochromatic
sources. Prior to analysis such as peak-fitting and peak area quantification, XP
background removal schemes found in commercially available XPS systems
[16, 21, 24–26]. Other background removal schemes may involve principle
component analysis, polynomial approximation, and the use of reflected
electron energy loss spectroscopy (REELS) data [26].
spectra. The major Auger series that are commonly observed in XPS include
the KLL, the LMM, and two types of the MNN series [8, 10, 11]. Because the
Auger electron’s peak position on the binding energy scale is dependent upon
the energy of the excitation source, while the photoelectron’s peak position on
the binding energy scale is independent of the source energy, the
identification of Auger electron peaks may be accomplished by switching the
X-ray source anode as is illustrated in Figure 11.4.
Although the binding energies of the photoelectrons are usually adequate
for elemental identification, XPS offers information that enables identification
of atoms of the same element in different chemical states. Surrounding
species can cause changes in the binding energies of the core electrons; these
changes are called “chemical shifts.” Since the chemical shifts caused by
surrounding species are generally less than 10 eV and the linewidth of the
peaks are usually ~1 eV, complications such as spectral overlaps of the atoms
in different chemical states, difficulties in eliminating spectral background,
and further difficulties in analyzing insulating samples may arise [8]. Possible
use of an internal reference.
Figure 11.8 contains an XP spectrum of the C 1s region for poly-
methylmethacrylate (PMMA), which illustrates the differentiation of the
carbon atoms in various chemical states. Curve-fitting identified four peaks,
labeled 1–4 in the spectrum, in the C 1s binding energy region near 285 eV.
The four peaks are assigned to specific carbon atoms within PMMA. For
carbon, as with most atoms, the atom in the most electron-withdrawing
environment will have the highest binding energy. For PMMA, this atom
corresponds to the carbon in the ester group (number 4 carbon in the inset to
Figure 11.8). The next highest binding energy, ~287 eV, corresponds to the
carbon bound to the oxygen of the ester, number 3 in the inset to Figure 11.8.
The carbon adjacent to the carbonyl group is in the next most electron-
withdrawing environment and corresponds to the peak labeled 2 in Figure
11.8. Although the two carbons assigned as number 1 carbon are in slightly
different chemical environments, they are similar enough that the
corresponding peaks are unresolved and fit as a single peak. In addition, the
aforementioned rationale is corroborated by the relative intensities of the
peaks—the areas of peaks 2, 3, and 4 are relatively similar, implying similar
497
spectral backgrounds are commonly removed with either linear or Shirley
Auger electron peaks may also be observed, often in series in the XP
solutions include curve-fitting, careful choice in background correction, and
11. X-Ray Photoelectron Spectroscopy
percentage of carbon atoms in the sample volume analyzed, and peak 1 has
approximately twice the total area of the other peaks.
Figure 11.8 C 1s region XP spectrum for polymethylmethacrylate (PMMA). The inset shows
the repeat unit for PMMA along with identification of the peaks in the spectrum.
Other factors affecting the binding energies of the ejected photoelectrons
may include sample size, morphology, and thermal history, as well as the
surrounding matrix, means of charge compensation, and application of
external bias [8, 9, 18].
11.5.1 Core Level Splitting
When the photoelectrons are emitted from levels with the orbital angular
momentum quantum number (l) greater than 0, a doublet in the spectrum may
momentum quantum number (j = l + s, j =| l – s|) values. These two states are
angular momentum vector of the unpaired electron resulting from the ejection
of the photoelectron are aligned or not. Examples of the spin-orbit coupling
phenomenon are shown in Figure 11.9, which is discussed in more detail
below.
The energy difference between the doublet, ǻE
j
, is dependent on the spin-
orbit constant, ȟ
nl
, which is related to the expectation value of the inverse of
498
be observed because of spin-orbit coupling. Spin-orbit
energetically different depending on whether the spin vector and orbital
coupling arises when
an
orbital possesses more than one possible state with different total angular
11.5 Spectral Analysis
radius cubed, ¢1/r
3
², of the particular orbital involved [8]. The relative peak
areas of the two doublet peaks are ultimately dependent on the atomic
photoemission cross-section, ı. The atomic photoemission cross-section is
related to the atomic number Z, principal quantum number n, orbital angular
momentum quantum number l, the total angular momentum quantum
number j, and how close the X-ray energy is to the photoemission threshold
[8, 11].
Table 11.2 lists common area ratios of doublet splitting. The core level
splitting resulting in doublet formation is also illustrated in Figure 11.9. The C
1s shell photoelectron, with only one possible j value, exhibits a single
spectral peak, whereas the Cu 2p, Ag 3d, and Au 4f photoelectrons exhibit
binding energy difference between the doublets is specific to the metal and
orbital involved, 'E
j
can be a helpful tool in element identification.
Figure 11.9 Core level splitting patterns for (a) C 1s, (b) Ag 3d, (c) Au 4f, and (d) Cu 2p.
Table 11.2 Summary for core level splitting relative intensities.
Subshell j Values
Area Ratio;
(2j
1
+ 1):(2j
2
+ 1)
s 1/2
p 1/2, 3/2 1:2
d 3/2, 5/2 2:3
f 5/2, 7/2 3:4
499
doublet peaks at approximately the ratio indicated in Table 11.2. Since the