The effect of Pt on Ni3Al surface oxidation at low-pressures

Article abstract of DOI:10.1016/j.susc.2006.09.014

The fully-oxidized surface that forms on (1 1 1) oriented Ni₃Al single crystals, with and without Pt addition, at 300-900 K under oxygen pressures of ca. 10^-7 Torr was studied using XPS, AES, and LEIS. Two main types of surfaces form, depending upon oxidation temperature. At low-temperature, the predominant oxide is NiO, capped by a thin layer of aluminum oxide, which we refer to generically as Al_xO_y. At high-temperature (i.e., 700-800 K), NiO is replaced by a thick layer of Al_xO_y. By comparing samples that contain 0, 10 and 20 at.% Pt in the bulk, we find that the effect of Pt is to: (1) reduce the maximum amount of both NiO and Al_xO_y; and (2) shift the establishment of the thick Al_xO_y layer to lower temperatures. Platinum also decreases the adsorption probability of oxygen on the clean surface.

Full text of DOI:10.1016/j.susc.2006.09.014

Surface Science 601 (2007) 146–154
www.elsevier.com/locate/susc
The effect of Pt on Ni3Al surface oxidation at low-pressures
a,b,
a
a
a
a,b,c
F. Qin
*, J.W. Anderegg , C.J. Jenks , B. Gleeson a,c, D.J. Sordelet , P.A. Thiel
a
Ames Laboratory, Iowa State University, Ames, IA 50011, United States
Department of Chemistry, Iowa State University, Ames, IA 50011, United States
Department of Materials Science and Engineering, Iowa State University, Ames, IA 50011, United States
b
c
Received 24 May 2006; accepted for publication 11 September 2006
Available online 29 September 2006
Abstract
The fully-oxidized surface that forms on (111) oriented Ni3Al single crystals, with and without Pt addition, at 300–900 K under oxy-
gen pressures of ca. 10ꢀ7 Torr was studied using XPS, AES, and LEIS. Two main types of surfaces form, depending upon oxidation
temperature. At low-temperature, the predominant oxide is NiO, capped by a thin layer of aluminum oxide, which we refer to generically
as AlxOy. At high-temperature (i.e., 700–800 K), NiO is replaced by a thick layer of AlxOy. By comparing samples that contain 0, 10 and
20 at.% Pt in the bulk, we find that the effect of Pt is to: (1) reduce the maximum amount of both NiO and AlxOy; and (2) shift the estab-
lishment of the thick AlxOy layer to lower temperatures. Platinum also decreases the adsorption probability of oxygen on the clean
surface.
ꢀ 2006 Elsevier B.V. All rights reserved.
Keywords: Oxidation; Aluminum oxide; Nickel oxides; Nickel aluminide; Platinum; X-ray photoelectron spectroscopy; Low-energy ion scattering (LEIS);
Auger electron spectroscopy
1. Introduction
preferential substitution of Pt at Ni sites in the ordered L12
crystal structure of Ni3Al [9]. In other words, the addition
of Pt increases the Al/Ni ratio at the Ni3Al surface.
Certain properties of the intermetallic alloy Ni3Al, such
as high-strength at elevated temperatures and low-density,
make it an attractive constituent phase in materials used
for high-temperature components in turbine engines [1–
3]. High-temperature oxidation of Ni3Al at atmospheric
pressure typically leads to the formation of a complex
group of oxides, including Al2O3, NiO, and NiAl2O4 [4–
6]. Recent studies have shown that adding Pt to Ni3Al
can significantly improve its oxidation resistance at ele-
vated temperatures by promoting the preferential forma-
tion of an adherent, Al2O3-rich scale [7,8]. It has been
suggested that the effect of Pt is, at least in part, to kineti-
cally favor the formation of aluminum oxide by decreasing
the nickel available for oxidation. This is attributable to the
The structure and composition of the oxide formed dur-
ing the initial stages of oxidation are thought to be an
important factor in determining the resistance of Ni–Al al-
loys to corrosion and spallation [10,11]. Ultrahigh vacuum
(UHV) conditions can probe this regime very well because
of the low-oxygen pressure (low-flux) at which oxidation is
carried out—typically in the range of 10ꢀ8 to 10ꢀ5 Torr.
There have been numerous oxidation studies on both
Ni3Al and NiAl at low-pressures. Some useful generaliza-
tions can be drawn from those studies. First, it is well-
established that an amorphous layer of alumina (with
short-range order) develops when the sample is oxidized
at low-temperatures. A crystalline Al2O3 layer develops
when the sample is oxidized at high-temperatures. For
˚
Ni3Al(111), this 5–7 A crystalline layer has been identified
*
Corresponding author. Address: Ames Laboratory, Iowa State Uni-
as c-like Al2O3 [12–14]. For both alloys, the transition
occurs around 700–800 K [10,11]. It has been proposed
that this transition marks the onset of long-range diffusion
versity, 220 Spedding Hall, Ames, IA 50011, United States. Tel.: +1 515
294 0905; fax: +1 515 294 4709.
E-mail address: flqin@iastate.edu (F. Qin).
0039-6028/$ - see front matter ꢀ 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.susc.2006.09.014

F. Qin et al. / Surface Science 601 (2007) 146–154
147
of Al between the bulk and surface, on the time scale of
oxidation [15,16].
The Auger beam current is 1.8 lA from sample to ground
without bias.
Two other chemical species have also been identified. At
room temperature, the oxide is preceded by chemisorbed
oxygen at low-coverage, distinguishable from the oxide
by its adsorption kinetics. In addition, an aluminum oxide
with Al in an oxidation state lower than that in Al2O3 exists
on Ni3Al surfaces. Several possible explanations have been
given for this state [17–20].
The single-crystal samples were prepared by the Materi-
als Preparation Center at the Ames Laboratory [24]. Based
upon electron probe microanalysis (EPMA) of different
areas on each sample, the compositions of the samples used
for XPS and LEIS were Pt0Ni77.4 0.3Al22.6
(hereafter
0.3
called the 0%-Pt sample), Pt9.7 0.1Ni65.3 0.3Al25.0
0.3
(called 10%-Pt), and Pt20.4 0.1Ni54.7 0.2Al24.9
(called
0.2
Second, aluminum oxide is favored thermodynamically
over the Ni-containing oxides, NiO and NiAl2O4, but the
latter are faster-growing and hence favored kinetically
[12,20]. One of the nickel-containing oxides, NiAl2O4, is
not known to form on either Ni3Al or NiAl at low-pres-
sures [10,12,16,17,21]. On NiAl surfaces, NiO has not been
observed at any temperature; however, on Ni3Al surfaces,
NiO has been observed by some authors, in addition to
the alumina layer described above [14,17,22]. Presumably,
NiO forms on Ni3Al but not on NiAl because the former
has a higher Ni concentration.
Despite these previous works, the chemical nature of
oxide formed upon saturation and how it changes with oxi-
dation temperature have not been systematically examined.
In this work, our goal was to systematically probe the
amount, chemical nature, and spatial distribution of the
oxide that forms upon saturation at an oxygen pressure
of about 10ꢀ7 Torr, as a function of oxidation temperature.
In this way, we hoped to gain insight into the initial stages
of the formation of the protective phase. Information is
derived from X-ray photoelectron spectroscopy (XPS),
low-energy ion scattering (LEIS), and Auger electron spec-
troscopy (AES). We are particularly interested in the effect
of Pt, and therefore we compare samples of (111) oriented
Ni3Al single crystals with and without Pt addition.
20%-Pt). The compositions, and standard deviations in
composition, were similar for the samples used in AES–
LEED. The standard deviations quoted above reflect small
and nonsystematic spatial variations in composition, which
in turn indicate good spatial homogeneity.
Pressures in the chamber were measured by a vendor-
calibrated nude ion. Exposures are reported in terms of
Langmuir (1 L = 10ꢀ6 Torr s) and have not been corrected
for ion gauge sensitivities (relative to N2, the calibration
gas) or flux to the sample. The oxidation conditions were
as follows. In the AES–LEED chamber, oxygen was intro-
duced by backfilling the chamber to a constant pressure of
2 · 10ꢀ7 Torr. An exposure of about 1000 L corresponded
to saturation at this pressure. In the XPS–LEIS chamber,
oxygen was introduced through a tube that was 1 mm in
internal diameter and about 2 mm from the sample. Com-
parison of the oxygen uptake curves from XPS and AES
suggested that the local oxygen pressure at the sample in
the XPS–LEIS chamber was also in the low 10ꢀ7 Torr
range. Spectra were recorded after evacuating the chamber.
On the time scale of our experiments, the spectra remained
invariant after evacuation.
In the XPS data, overlap between lines of Pt, Ni, and Al
prevented use of certain peaks that are considered stan-
dards. In particular, the Al2p could not be used because
it overlapped completely with the Pt4f5/2, and both were
close to the Ni3p. These issues led us to choose the Al2s,
2. Experimental description
The methods of sample preparation and characteriza-
tion were the same as described elsewhere [23]. XPS and
LEIS were performed in one chamber, while AES was car-
ried out in a separate chamber, on parallel sets of samples.
XPS and LEIS measurements were performed by means of
a hemispherical analyzer with multi-channel detector. For
XPS, the incident X-rays were monochromatized Al Ka
(hm = 1486.6 eV) radiation operated at 15 kV and 250 W.
The XPS instrument was a Perkin Elmer PHI 5500,
multi-technique system, fitted with an Omni Focus III
lens system and run with PHI-Access software. The XPS
energy scale was calibrated using Au4f7/2 and Cu2p3/2 at
84.0 eV and 932.6 eV, respectively. The energy resolution
was 0.65 eV at pass energy of 58.7 eV, judging from
Ag3d5/2 photoelectrons. For LEIS we used a beam of
He+ at 1 keV and at a current of approximately
3 · 10ꢀ8 A/cm2. Data were acquired in two modes, fixed
and variable angle. In the fixed mode, the emission angle
was 45ꢁ. In the variable angle mode, the emission angle
varied from 15 to 70ꢁ. The Auger beam energy is 5 keV.
Al2s
O1s
Ni2p
Al0
Ni3s
Al2O3
Ni2+
Al3+
NiO
900 K
800 K
700 K
500 K
300 K
Clean
Metal
124
116
108 534 532 530 528 880870 860 850
Binding Energy (eV)
Fig. 1. XPS spectra of 0%-Pt Ni3Al(111), clean and after oxidation to
saturation at various temperatures.

148
F. Qin et al. / Surface Science 601 (2007) 146–154
Ni2p3/2, and Pt4d5/2 from among the metal peaks. The first
two are shown in Figs. 1–3. Experimental spectra were ana-
lyzed by fitting to component peaks using PHI’s MultiPak
software. The binding energies and full width at half max-
ima (FWHM) were kept constant for a given elemental or
oxidic peak throughout the fitting of all experimental data.
In XPS, the composition of the oxide film was calculated
from the intensities (peak areas) of O1s, Ni2p3/2, Al2s and
Pt4d5/2 spectra using the following relationship [25]:
tivity factor [26] appropriate to the spectrometer for ele-
ment A. The sum is over all of the constituents of the
solid. In AES, intensities were determined from peak-to-
peak height and published atomic sensitivity factors [27]
were employed in quantification.
For the LEIS results reported here, the incident beam
impinged on the surface at an angle of 40ꢁ with respect
to the surface normal. The scattering angle was 135ꢁ.
In order to calculate relative atomic concentrations, the
differential scattering cross-section for each element, given
our scattering conditions, was calculated using the ZBL
screened Coulomb potential [28]. The broad background
arising from inelastic processes and subsurface scattering
was removed prior to determining the peak areas.
IA=I1A
P
XA
¼
;
ð1Þ
Ii=I1i
i¼A;B
where XA, IA and I1A , respectively, represents the atomic
concentration, XPS intensity and published atomic sensi-
3. Experimental results and interpretation
Figs. 1–3 show a series of XPS spectra obtained on the
three different samples. Fig. 1 represents the 0%-Pt sample,
Fig. 2 the 10%-Pt sample, and Fig. 3 the 20%-Pt sample.
The bottom spectrum represents the clean alloy, while
higher spectra show the result of oxidizing to saturation.
The temperature of oxidation increases, going upward,
from room temperature to 900 K. For each sample and
each temperature, three photoemission peaks are shown:
Al2s, O1s, and Ni2p. For the Pt-containing samples, the
Pt4d5/2 peak was also examined, but is not shown. It exhib-
its no evidence of oxidation.
Al2s
O1s
Ni2p
Al2O3
Al0
Ni3s
Ni2+
Al3+
NiO
900 K
800 K
700 K
500 K
300 K
Both the Ni2p and Al2s peaks indicate oxidation of the
Ni and Al, because the metal peaks broaden toward higher
binding energy, sometimes leading to a discernible peak, in
going from the bottom spectrum upward. For Al2s, the
oxide peak is usually clear, as can be seen in Figs. 1–3. For
Ni2p, the oxide (Ni2+) intensity is always relatively small,
nestled between the metallic Ni2p3/2 peak (853.2 eV) and
its satellite peak (858.6 eV). The Ni2+ line is expected to be
a doublet for NiO at 854.1 and 855.9 eV [29] and a single
peak for NiAl2O4 at 857 eV [6,30]. The best fit using a single
peak led to a peak energy of 855.5 eV, in disagreement with
the position expected for NiAl2O4. The fit was significantly
better with a doublet at the positions expected for NiO.
(Fits also allowed different relative intensities of satellite
peaks for the oxide and the metal.) Analysis of XPS peak
intensities, below, further rules out the presence of the spinel
phase.
The O1s spectra in Figs. 1–3 are fitted well with three
components having FWHM of 1.9 eV at 530.3 eV (NiO
[17,31–33]), 531.4 eV, and 532.2 eV (Al2O3 [17,34,35]).
The oxygen component at 531.4 eV could have multiple
assignments. It could be oxygen associated purely with
the nickel. For instance, other authors have assigned a
peak at about this binding energy to Ni2O3 or to ‘‘defect’’
nickel oxide on surfaces of pure Ni [36,37]. Alternatively, it
could be oxygen associated both with nickel and alumi-
num, perhaps in the spinel phase, NiAl2O4. Third, it could
be associated only with the aluminum in a form that is
commonly designated AlOx (i.e., a type of aluminum oxide
Clean
Metal
124
116
108 534 532 530 528
880 870 860 850
Binding Energy (eV)
Fig. 2. XPS spectra of 10%-Pt Ni3Al(111), clean and after oxidation to
saturation at various temperatures.
Al2s
O1s
Ni2p
Al2O3
Al0
Ni3s
Al3+
Ni2+
NiO
900 K
800 K
700 K
500 K
300 K
Clean
Metal
124
116
108 534 532 530 528 880870 860 850
Binding Energy (eV)
Fig. 3. XPS spectra of 20%-Pt Ni3Al(111), clean and after oxidation to
saturation at various temperatures.

F. Qin et al. / Surface Science 601 (2007) 146–154
149
with an aluminum oxidation state intermediate between
Al2O3 and Al metal) [11,16,17,20,38]. To distinguish be-
tween these possibilities, we checked whether the intensities
of the (weak) Ni2+ peaks in the Ni2p region were correlated
with the intensity of the O1s peak at 531.4 eV, as a function
of oxidation temperature. The result is shown in Fig. 4 for
0%-Pt, the alloy where the Ni2+ peaks are strongest. The
trend is similar in the other two alloys. The anti-correlation
between the intensities of Ni2+, and O1s at 531.4 eV, clearly
rules out a form of the oxide associated with nickel, such as
Ni2O3 or spinel, in agreement with conclusions by other
authors for this alloy [17]. Therefore, the O1s line at
531.4 eV is assigned to AlOx.
qualitatively if the Al2s and Ni2p oxidic lines are used in-
stead of the O1s components.
Fig. 5(a) shows the total oxygen concentration at satura-
tion as a function of oxidation temperature, measured with
XPS. The curves are depressed for higher Pt content, mean-
ing that higher Pt content leads to less oxygen incorpora-
tion. This is true at all temperatures except 700 K.
Furthermore, Pt changes the shapes of the curves. For
0%-Pt, the curve shows two maxima, at around 500 and
800 K. For 10%-Pt and 20%-Pt, there is only one maxi-
mum, at around 700 K. The AES data for the two Pt-con-
taining alloys, shown in Fig. 5(b), are complementary and
confirm certain trends in the data. One is the presence of a
single maximum in total oxygen concentration at 700 K for
both of the Pt-containing samples. Another is the fact that
oxygen concentration is lower for 20%-Pt than for 10%-Pt.
The variation in total oxygen concentration is broken
down into contributions from NiO and AlxOy, based on
the O1s spectra, in Fig. 6(a)–(c). In Fig. 6(a), this deconvo-
lution shows that the maxima in the total oxygen concen-
tration for the 0%-Pt sample are due, first, to a peak at
500 K in the NiO concentration, and second, to a peak at
800 K in the AlxOy concentration. In each case, the varia-
tion in concentration of the other oxide is small around the
maximum, so the peak in total content is due to the varia-
tion of a single constituent. Fig. 6(b) and (c) show that one
effect of Pt is to shift the rise in AlxOy concentration to low-
er temperatures, so that the sum of the two curves gives a
single maximum at 700 K. The rise in AlxOy concentration
begins at about 700 K in the absence of Pt, and at about
500 K in the presence of Pt. Hence, the presence of Pt shifts
the increase in Al oxidation downward by about 200 K.
This shift is not progressive with Pt concentration, but
rather is constant for both the 10%-Pt and 20%-Pt samples.
Pt also reduces the maximum amount of NiO, from 34 at.%
to 25 at.% to 18 at.%, in going from Fig. 6(a) to (b) to (c).
This reduction is not balanced by an equal gain in AlxOy,
so the net effect is to decrease the total oxygen content at
500 K (Fig. 5). Pt also reduces the maximum concentration
In the remainder of this paper, we focus on the total
nickel oxide and aluminum oxide concentrations from the
O1s data. Since both Al2O3 and AlOx are aluminum oxi-
des, the sum of their contribution (which we denote as
‘‘AlxOy’’) is employed. All of the trends are reproduced
50
Ni2+ doublet at 854.1
eV and 855.9 eV
40
30
20
531.4 eV
10
0
100
300
500
700
900
1100
Temperature (K)
Fig. 4. Compositions of the Ni2p3/2 doublet in XPS at 854.1 eV and
855.9 eV, and of the O1s peak at 531.4 eV, at saturation, as a function of
temperature for the 0%-Pt sample. The y-axis shows the concentrations of
the species associated with these features, based on their intensities.
XPS
AES
50
40
30
20
10
0%-Pt
50
10%-Pt
10%-Pt
40
30
20
20%-Pt
20%-Pt
700
Temperature (K)
10
100
300
500
900
1100
100
300
500
700
900
1100
Temperature (K)
Fig. 5. Oxygen composition at saturation, as a function of temperature. (a) XPS data (using O1s spectra); (b) AES data (using O KLL line). For XPS, all
points represent experiments that were repeated 2 or 3 times except 300 K for 0%-Pt sample and 500 and 800 K for 10%-Pt and 20%-Pt samples. In some
cases, error bars are so small that they are obscured by the symbol. For AES, there was only 1 run for all points.

150
F. Qin et al. / Surface Science 601 (2007) 146–154
55
45
35
25
15
55
a
b
0%-Pt
O
10%-Pt
45
35
O
25
NiO
NiO
15
AlxOy
5
5
AlxOy
-5
100
-5
300
500
700
900
1100
100
300
500
700
900
1100
Temperature (K)
Temperature (K)
55
45
35
25
15
5
c
20%-Pt
O
NiO
AlxOy
-5
100
300
500
700
900
1100
Temperature (K)
Fig. 6. Concentrations of oxygen, NiO and AlxOy at saturation: (a) 0%-Pt; (b) 10%-Pt; (c) 20%-Pt. All data are based on the O1s line in XPS. NiO and
AlxOy concentrations were calculated by deconvoluting the O1s spectra as described in the text.
of AlxOy produced at higher temperatures, for instance
from 45 at.% at 800 K in Fig. 6(a), to 23 at.% at 800 K in
Fig. 6(b). In summary, Pt reduces the extent of oxidation
of both Ni and Al.
These observations serve to explain the trends in Fig. 5.
The total oxygen uptake is suppressed by Pt because the ex-
tents Ni and Al oxidation are reduced by Pt. At 700 K, the
uptake curves for the two Pt-containing samples cross-over
the curve for the 0%-Pt alloy simply because the maximum
in the amount of AlxOy formation is shifted to lower tem-
perature by Pt.
There is evidence that aluminum oxide is always the
principal constituent in the topmost surface layer, even
when nickel forms the predominant oxide. This is demon-
strated by the angle-resolved XPS data of Fig. 7, and by
the LEIS data of Fig. 8. In Fig. 7, a higher value of the
1.6
1.8
Al
1.4
AlxOy
1.4
O
1.2
1
1
Pt
0.6
0.8
0.6
Ni
0.2
5
15
25
35
45
55
65
75
5
15
25
35
45
55
65
75
Take-off Angle (º)
Take-off Angle (º)
Fig. 7. XPS angle resolved data after saturation oxidation of 10%-Pt Ni3Al(111) at 300 K. The concentrations of each species are normalized to the value
at 70ꢁ, which is the signal with the highest bulk contribution. (a) Total elemental concentration ratios. (b) Oxide concentration ratios from the O1s data.

F. Qin et al. / Surface Science 601 (2007) 146–154
151
take-off angle corresponds to a lower contribution of the
surface signal to the total signal. Fig. 7(a) shows the total
signal originating from each of four elements—Ni, Pt, Al,
and O—as a function of angle. For each element, the an-
gle-resolved data are normalized to the value at 70ꢁ, which
is the value closest to the bulk limit. For the 10%-Pt sample
oxidized at 300 K, the Al and O signals are above one,
while the Ni and Pt signals are below one, at low-angles.
This means that the surface is enriched in Al and O, and
depleted in Pt and Ni, relative to bulk concentrations.
Yet, Fig. 6(b) also shows that the oxide is mainly NiO at
300 K, suggesting that even at room temperature, the
kinetics of Al oxidation are sufficiently fast to form a thin
surface layer of AlxOy. Qualitatively, the trends shown in
Fig. 7(a) are robust. For all samples and all temperatures,
the normalized total Al2s and O1s intensities exceed one at
low-angles, while the Ni2p3/2 and Pt4d5/2 (if present) are
below one at low-angles. This means that AlxOy is always
the principal constituent in the outermost layer, even under
conditions where it is not the predominant oxide, and
regardless of whether Pt is present.
These conclusions are strengthened by the XPS data of
Fig. 7(b), which show the angular variation of the O1s sig-
nals originating separately from NiO and AlxOy, deconvo-
luted as described for Figs. 1–3, and normalized to their
values at 70ꢁ. Again, the deviations from one at low-angle
indicate that AlxOy is on top, while the NiO lies deeper.
However, it is not clear from these data whether the NiO
beneath the AlxOy is ‘‘pure’’ or is intermixed with metal.
The qualitative trends shown in Fig. 7(b) are observed
for all samples and all temperatures.
Al. Because LEIS mainly probes the top one or two atomic
layers, this confirms that AlxOy is enriched at the surface.
Returning to Fig. 7(b), the AlxOy portion of the
O1s signal in angle-resolved XPS is always enhanced at
15ꢁ, relative to its value at 70ꢁ. However, the degree of
enhancement—the ratio of the signals at 15ꢁ and 70ꢁ—is
not constant. This degree of enhancement is presented in
Fig. 9 as a function of oxidation temperature, for all three
samples. To interpret these data, we assume that AlxOy is
confined to a region at the surface where it is separate from
the other components (NiO and the metals). Then this ratio
should not depend upon the total area of AlxOy at the sur-
face, but rather upon the average thickness of the AlxOy re-
gions, with a lower ratio corresponding to thicker regions.
In Fig. 9, for 0%-Pt, the ratio is high through 700 K, but
drops sharply between 700 and 800 K. This correlates with
the sharp rise in AlxOy concentration in Fig. 6(a). For the
Pt-containing samples, the ratio drops at lower tempera-
tures, with most of the change occurring between 500
and 700 K. Again, this correlates with the temperature
range where AlxOy concentration rises in Fig. 6(b) and
(c). Thus, Pt reduces the temperature at which a thicker
AlxOy layer becomes established. The curves in Fig. 9 do
not turn upward again between 800 and 900 K, where the
AlxOy layer disappears. This signals that as the oxide dis-
appears it does not become thinner; islands shrink laterally,
not vertically.
Platinum also slows the kinetics of oxygen uptake. This
can be seen from Fig. 10, which displays the oxygen con-
centration derived from AES using the O KLL (510 eV)
line intensity versus oxygen exposure at four different tem-
peratures. The initial slope of the curve, which is propor-
tional to the initial sticking coefficient, S0, is slightly but
consistently lower for 20%-Pt than for 10%-Pt. The sticking
coefficient, S, is the probability that an incident molecule
adsorbs on the surface. The initial sticking coefficient, S0,
is the value in the limit of zero coverage [39]. Note that
Fig. 8 shows the surface atomic concentrations derived
from LEIS as a function of oxidation temperature, for
10%-Pt. (The data are similar for 20%-Pt.) Among the
three metals, Al is by far the most abundant at all temper-
atures. For instance, at 300 K, the ratio of Al:Ni is about
3.5. By contrast, in the bulk sample this ratio is 0.4, and
for the clean surface it is 0.7 from LEIS [23], showing that
oxidation strongly enhances the surface concentration of
1.7
0%-Pt
80
20%-Pt
1.5
1.3
1.1
0.9
O
60
40
Al
10%-Pt
20
Ni
0
Pt
100
300
500
700
900
1100
-20
100
Temperature (K)
300
500
700
900
1100
Fig. 9. AlxOy at an XPS take-off angle of 15ꢁ after saturation oxidation of
0%-Pt, 10%-Pt, and 20%-Pt samples. All the data points are normalized to
the value at a take-off angle of 70ꢁ, which contains the highest bulk
component. Contribution of AlxOy is calculated by deconvoluting the O1s
spectra as described in the text.
Temperature (K)
Fig. 8. LEIS data after saturation oxidation of 10%-Pt Ni3Al(111).
Experimental data were reproduced as follows: 1 run at 300, 500, and
800 K, 3 runs at 700 K, and 2 runs at 900 K.

152
F. Qin et al. / Surface Science 601 (2007) 146–154
30
30
25
20
15
10
5
10%-Pt
a
c
b
300 K
500 K
25
20
15
10
5
10%-Pt
20%-Pt
20%-Pt
0
0
0
20
40
60
80
100
0
20
40
60
80
100
30
25
20
15
10
5
30
25
20
15
10
5
10%-Pt
d
700 K
1000 K
10%-Pt
20%-Pt
40
20%-Pt
60
0
0
0
20
60
80
100
0
20
40
80
100
Exposure (L)
Fig. 10. Oxygen composition (from AES data) as a function of exposure at (a) 300 K; (b) 500 K; (c) 700 K; (d) 900 K.
at 300 K, adsorption seems to occur in two regimes, with a
break at about 30 L. This probably reflects the transition
from chemisorbed oxygen to oxide, reported by other
authors for surfaces of NiAl [15,40] and Ni3Al [41]. In
those reports, the transition point ranged from 5 to 15 L
at room temperature. This transition is also known on sur-
faces of pure Al (e.g., [42–44]) and pure Ni (e.g., [45,46]).
Also, the maximum amount of AlxOy is significantly
lower for both the Pt-doped samples. The maximum
amount of oxygen as AlxOy, for instance, decreases upon
going from 0%-Pt to 20%-Pt (Fig. 6). The origin of this
effect is unclear.
Second, Pt shifts the enhancement of AlxOy to lower
temperatures. This is evident both in the temperature at
which the surface concentration of AlxOy begins to rise
steeply (Fig. 6), and in the temperature at which AlxOy
thickens (Fig. 9). In the absence of Pt, oxidation of Al
accelerates strongly between 700 and 800 K and the oxide
thickens in this same interval. Other authors have sug-
gested that 700–800 K is the temperature at which long-
range diffusion of Al from the bulk to the surface begins,
on the time-scale of low-pressure oxidation, both for NiAl
and Ni3Al [15,16]. In this context, the natural interpreta-
tion is that Pt accelerates diffusion of Al to the surface. A
higher diffusion coefficient for Al in the presence of Pt
has also been inferred by Gleeson et al. based on diffusion
studies of NiAl at 1400 K [7,9]. Moreover, Hayashi et al.
[50] conducted a systematic interdiffusion study on Pt-mod-
ified c-Ni and c0-Ni3Al alloys at 1423 K and found that the
4. Discussion
4.1. Effects of Pt
The primary goal of our study was to elucidate the effect
of Pt on oxidation of Ni3Al. We find that Pt exerts two
main effects on the oxygen-saturated surface: (1) it de-
creases the maximum amount of both NiO and AlxOy;
and (2) it shifts the formation of AlxOy to lower tempera-
tures. It also slows oxygen adsorption on the clean surface.
Interpretation is given below.
First, Pt decreases the amount of NiO (Fig. 6). This may
simply reflect the fact that Pt replaces Ni in the bulk lattice,
and so reduces the amount of Ni available for oxidation [9].
There is even more replacement of Ni by Pt at the surface
than in the bulk, since Pt segregates at the clean surface
[23,47]. The effect of Ni concentration is clear from a
comparison of the literature for NiAl and Ni3Al surface
oxidation. At the low-oxygen pressures used in UHV
experiments, NiO has never been observed on the more
Ni-deficient alloy, NiAl [20,40,48,49]. However, NiO has
been reported by some authors on Ni3Al at P < 10ꢀ5 Torr
and at various temperatures [14,17,22] and its presence is
confirmed in this study.
Ni
~
main-term ternary interdiffusion coefficient for Al, DAlAl, is
up to a factor of three greater than the corresponding
binary Ni–Al interdiffusion coefficient.
Finally, Pt slows the rate of initial oxygen uptake. This
could also reflect the substitution of Pt for Ni at the clean
surface. Other authors have proposed that in binary Ni–Al
alloys, Ni is the more active site for oxygen dissociation,
serving to increase S0 above that for pure aluminum, which
is quite low. For instance, S0 is one on Ni(111) [45] but
only 0.005 on Al(111) at 300 K [42]. The initial sticking

F. Qin et al. / Surface Science 601 (2007) 146–154
153
coefficient of oxygen on Pt is structure-dependent, but is
also low on the (111) surface, with values of 0.04–0.08 at
300 K [51].
which is given as DGꢂNiO ¼ ꢀ235797 þ 86:195 T (J/mol)
[54]. Formation of NiO is thermodynamically possible only
if PO > PꢁO . Eq. (2) can alternatively be used to calculate
2
2
the maximum temperature at which NiO is stable under a
constant PO of 3 · 10ꢀ8 Torr which reflects the current
2
4.2. Model of oxidation
experimental conditions. In view of the lack of experimen-
tal data for aNi in the Ni–Al and Ni–Al–Pt systems at
low-temperatures, we assume ideal solution behavior,
i.e., aNi = xNi = 0.75, and calculate Tmax = 1253 K. Thus,
NiO should be stable, relative to gaseous oxygen and
metallic nickel, at the pressures used in our experiments.
The remaining possibility is that NiO dissociates and re-
leases dissolved oxygen at temperatures above 500 K. This
scenario is consistent with a careful study of the thermal
decomposition of chemisorbed oxygen at the surface of
Ni(110) by Holloway and Outlaw, who found that the oxy-
gen disappeared into the bulk at about 600 K when the
sample was heated in UHV [55].
The concentration of AlxOy remains low and constant to
a certain temperature (700 K for 0%-Pt, 500 K for 10%-Pt
and 20%-Pt), then rises; simultaneously, it thickens. This is
the second main regime. As discussed above, we believe
that the rise in AlxOy reflects the onset of long-range diffu-
sion of Al from the bulk, on the time scale of oxidation at
these oxygen pressures.
At still higher temperature (800 K for 0%-Pt and 10%-
Pt, 700 K for 20%-Pt), the concentration of aluminum
oxide begins to fall. Other authors have reported the com-
plete disappearance of metastable aluminum oxide from
NiAl and Ni3Al surfaces in UHV at high-temperatures,
1100 K to 1500 K [15,17,48,49,56]. Thermal instability of
AlxOy has been observed on other metal substrates as well,
including pure Al, at temperatures as low as 1000 K [57–
59]. Two interpretations have been given: Desorption [59]
or dissolution [57] of oxygen. This occurs well below the
melting point of bulk a-Al2O3, 2345 K [15].
The LEIS data of Fig. 8 show that as the oxide disap-
pears at high-temperature, the islands of AlxOy shrink lat-
erally, not vertically. This is reminiscent of results from
Bardi et al. [10], who concluded that during the reverse pro-
cess—growth—island thickness remains constant, at least
on the binary Ni3Al surface.
Based upon the experimental data, we propose that oxi-
dation occurs in two broad regimes as a function of oxida-
tion temperature. In the first regime, NiO is the most
abundant oxide. It is replaced by AlxOy in the second
regime.
First, in the temperature range 300–500 K, both Ni and
Al oxidize. NiO predominates, but it lies beneath a layer
that is mainly AlxOy. It is not clear whether the NiO is pure
or intermixed with metallic species. The amount of NiO
consistently peaks at 500 K. This description does not de-
pend upon whether Pt is present. Our observation of NiO
after low-pressure exposure to oxygen is generally compat-
ible with previous reports, although the wide variety of
temperatures, pressures, and annealing conditions used in
previous work makes an exact comparison difficult
[14,17,22].
Some Al probably diffuses from underlying layers to the
surface, or Ni diffuses away from it, even at 300 K. This is
because the Al:Ni ratio, measured with LEIS, increases by
a such a large factor–5·– after oxidation. Exchange be-
tween the surface and underlying layers may be limited to
the top few layers of the alloy, perhaps in a Cabrera–Mott
mechanism. This is a logarithmic growth rate mechanism,
in which the driving force for the oxide formation is an
electric field set up in the oxide film resulting from a con-
tact potential difference between metal and adsorbed oxy-
gen. This field enables the metal ions to move through
the thin oxide film independent of temperature [52,53]. This
is a minor modification to previous hypotheses for the bin-
ary alloy, which said that at room temperature there is
mainly or exclusively lateral diffusion of Al [14,41].
Next, in going from 500 to 700 K, less and less NiO
forms. The temperature at which this decline begins
(500 K) does not depend on the presence of Pt, even though
the oxidation of Al does, suggesting that loss of NiO is not
due to competition with AlxOy formation. This is also sug-
gested by the data for the 0%-Pt alloy in Fig. 6(a), which
shows significant reduction in the amount of NiO between
500 and 700 K, while the amount of AlxOy remains approx-
imately constant. One must then consider other mecha-
nisms by which the amount of NiO at saturation is
reduced, one being its destabilization with increasing tem-
perature. The oxygen partial pressure (in atm) at which
there is equilibrium between Ni in the alloy and NiO can
be calculated as:
5. Conclusions
One main effect on oxidation of adding Pt to Ni3Al re-
vealed by this work is that Pt reduces the NiO content. This
is consistent with the suppression of nickel oxides (both
NiO and NiAl2O4) observed in previous work conducted
at much higher temperatures and pressures. The explana-
tion for both may simply be a reduction in the availability
of Ni. (This explanation may also account for the decrease
in sticking coefficient in the presence of Pt, in our data.) An
effect for which we have no explanation at present, how-
ever, is the reduction in the amount of AlxOy formed in
the presence of Pt. Another main effect is the shift in AlxOy
growth to lower temperatures, which is consistent with a
ꢀ
ꢁ
1
a2Ni
2DGꢂNiO
RT
POꢁ
¼
exp
;
ð2Þ
2
where aNi is the chemical activity of Ni in the alloy and
DGꢂNiO is the standard Gibbs energy of formation of NiO,

154
F. Qin et al. / Surface Science 601 (2007) 146–154
[26] C.D. Wagner, W.M. Riggs, L.E. Davis, J.F. Moulder, Handbook of
X-Ray Photoelectron Spectroscopy, Physical Electronics, Inc., Eden
Prairie, MN, 1979.
[27] L.E. Davis, N.C. MacDonald, P.W. Palmberg, G.E. Raich, R.E.
Weber, Handbook of Auger Electron Spectroscopy, Physical
Electronics Industries, Inc., Eden Prairie, MN, 1979.
[28] J.F. Ziegler, J.P. Biersack, U. Littmark, The Stopping and Range of
Ions in Solids, Pergamon, New York, 1985.
Pt-induced increase in the rate of diffusion of Al from the
bulk, to form surface AlxOy. Importantly, the effects of
Pt revealed in this study are consistent with proposed
explanations of its effects under more technologically-
relevant oxidizing conditions [7,9].
Acknowledgement
[29] N. Kitakatsu, V. Maurice, C. Hinnen, P. Marcus, Surf. Sci. 407
(1998) 36.
This work was supported by the Director, Ames Labo-
ratory, U.S. Department of Energy, Contract No. W-405-
Eng-82.
[30] Q. Zhong, F.S. Ohuchi, J. Vac. Technol. A 8 (1990) 2017.
[31] G.T. Tyuliev, K.L. Kostov, Phys. Rev. B 60 (1999) 2900.
[32] S. Evans, J. Pielaszek, J.M. Thomas, Surf. Sci. 55 (1976) 644.
[33] D. Cappus, C. Xu, D. Ehrlich, B. Dillmann, J.C.A. Ventrice, K. Al-
Shamery, H. Kuhlenbeck, H.-J. Freund, Chem. Phys. 177 (1993) 533.
[34] J.C. Fuggle, L.M. Watson, D.J. Fabian, Surf. Sci. 49 (1975) 61.
[35] C. Chen, S.J. Splinter, T. Do, N.S. McIntyre, Surf. Sci. 382 (1997)
L652.
[36] T. Fleisch, N. Winograd, W.N. Delgass, Surf. Sci. 78 (1978) 141.
[37] P.R. Norton, R.L. Tapping, J.W. Goodale, Surf. Sci. 65 (1977) 13.
[38] Y.G. Shen, D.J. O’Connor, R.J. McDonald, Surf. Interface Anal. 18
(1992) 729.
References
[1] Y. Lu, W. Chen, R. Eadie, Intermetallics 12 (2004) 1299.
[2] V.K. Sikka, S.C. Devvi, S. Viswanathan, R.W. Swindeman, M.L.
Santella, Intermetallics 8 (2000) 1329.
[3] N.S. Stoloff, C.T. Liu, S.C. Devvi, Intermetallics 8 (2000) 1313.
[4] G.C. Wood, B. Chattopadhyay, Corros. Sci. 10 (1970) 471.
[5] J.D. Kuenzly, D.L. Douglass, Oxid. Met. 8 (1974) 139.
[39] K.W. Kolasinski, Surface Science: Foundations of Catalysis and
Nanoscience, John Wiley and Sons Ltd., Chichester, West Sussex,
England, 2002.
`
[6] E.W.A. Young, J.C. Riviere, L.S. Welch, Appl. Surf. Sci. 31 (1988)
370.
[7] B. Gleeson, W. Wang, S. Hayashi, D. Sordelet, Mater. Sci. Forum
461–464 (2004) 213.
[8] C. Leyens, B.A. Pint, I.G. Wright, Surf. Coat. Technol. 133–134
(2000) 15.
[9] S. Hayashi, T. Narita, B. Gleeson, Mater. Sci. Forum 522–523 (2006)
229.
[10] U. Bardi, A. Atrei, G. Rovida, Surf. Sci. Lett. 239 (1990) L511.
[11] U. Bardi, A. Atrei, G. Rovida, Surf. Sci. 268 (1992) 87.
[12] S.G. Addepalli, B. Ekstrom, N.P. Magtoto, J.-S. Lin, J.A. Kelber,
Surf. Sci. 442 (1999) 385.
[40] H. Isern, G.R. Castro, Surf. Sci. 211–212 (1989) 865.
[41] A. Rosenhahn, J. Schneider, C. Becker, K. Wandelt, Surf. Sci. 433–
435 (1999) 705.
[42] J. Trost, H. Brune, J. Wintterlin, R.J. Behm, G. Ertl, J. Chem. Phys.
104 (1998) 1740.
[43] H. Brune, J. Wintterlin, J. Trost, G. Ertl, J. Chem. Phys. 99 (1993)
2128.
[44] R. Michel, J. Gastaldi, C. Allasia, C. Jourdan, Surf. Sci. 95 (1980)
309.
[45] P.H. Holloway, J.B. Hudson, Surf. Sci. 43 (1974) 141.
[46] P.H. Holloway, J.B. Hudson, Surf. Sci. 43 (1974) 123.
[47] P.Y. Hou, K.F. McCarty, Scripta Mater. 54 (2006) 937.
[48] T.T. Lay, M. Yoshitake, W. Song, Appl. Surf. Sci. 239 (2005) 541.
[49] W. Song, M. Yoshitake, Thin Solid Films 464–465 (2004) 52.
[50] S. Hayashi, D.J. Sordelet, L.R. Walker, B. Gleeson, Acta Mater.,
submitted for publication.
[51] A.C. Luntz, M.D. Williams, D.S. Bethune, J. Chem. Phys. 89 (1988)
4381.
[52] N. Cabrera, N.F. Mott, Rep. Prog. Phys. 12 (1948–1949) 163.
[53] R. Franchy, Surf. Sci. Rep. 38 (2000) 195.
[13] A. Rosenhahn, J. Schneider, C. Becker, K. Wandelt, Appl. Surf. Sci.
142 (1999) 169.
[14] C. Becker, J. Kandler, H. Raaf, R. Linker, T. Pelster, M. Draeger, M.
Tanemura, K. Wandelt, J. Vac. Sci. Technol. A 16 (1998) 1000.
[15] P. Gassmann, R. Franchy, H. Ibach, Surf. Sci. 319 (1994) 95.
[16] A.M. Venezia, C.M. Loxton, Surf. Sci. 194 (1988) 136.
[17] M. Garza, N.P. Magtoto, J.A. Kelber, Surf. Sci. 519 (2002) 259.
[18] D.R. Jennison, C. Verdozzi, P.A. Schultz, M.P. Sears, Phys. Rev. B
59 (1999) R15605.
[19] F. Qin, N.P. Magtoto, J.A. Kelber, Surf. Sci. 565 (2004) L277.
[20] R.M. Jaeger, H. Kuhlenbeck, H.-J. Freund, M. Wuttig, W. Hoff-
mann, R. Franchy, H. Ibach, Surf. Sci. 259 (1991) 235.
[21] F. Qin, N.P. Magato, J.A. Kelber, D.R. Jennison, J. Mol. Catal. A
(2005) 83.
[54] H.H. Kellogg, J. Chem. Eng. Data 14 (1969) 41.
[55] P.H. Holloway, R.A. Outlaw, Surf. Sci. 111 (1981) 300.
[56] Y.G. Shen, D.J. O’Connor, R.J. MacDonald, Surf. Interface Anal.
17 (1991) 903.
[22] M. Haerig, S. Hofmann, Appl. Surf. Sci. 125 (1998) 99.
[23] F. Qin, C. Jiang, J.W. Anderegg, C.J. Jenks, B. Gleeson, D.J.
Sordelet, P.A. Thiel, submitted for publication.
[24] Materials Preparation Center, Ames Laboratory US-DOE, Ames,
IA, USA, available from: <www.mpc.ameslab.gov>.
[25] D. Briggs, M.P. Seah, Practical Surface Analysis, Wiley, New York,
1990.
[57] S.M. Bedair, F. Hofmann, J.H.P. Smith, J. Appl. Phys. 39 (1968)
4026.
[58] F. Schmithu¨sen, G. Cappello, M.D. Boissieu, M. Boudard, F. Comin,
J. Chevrier, Surf. Sci. 444 (2000) 113.
[59] C. Jourdan, R. Michel, J. Gastaldi, J. Derrien, Philos. Mag. A 41
(1980) 443.