Surface Science 506 (2002) 228–234
www.elsevier.com/locate/susc
Characterization of MgO(1 0 0) thin film growth on Mo(1 0 0)
Y.D. Kim, J. Stultz, D.W. Goodman
*
Department of Chemistry, Texas A & M University, College Station, TX 77842-3012, USA
Received 23 October 2001; accepted for publication 21 January 2002
Abstract
Various MgO thin films grown on Mo(1 0 0) have been characterized by using metastable impact electron spectroscopy (MIES), ultraviolet photoelectron spectroscopy, low energy electron diffraction (LEED), and CO as a probe
molecule to titrate defects. Using MIES and LEED, the as-grown MgO(1 0 0) films (thickness ¼ 15 monolayers
(MLs)) are found to be highly defective. Annealing at 1150 K significantly reduces the density of defects on the
MgO(1 0 0) surfaces. CO adsorbs on the as-grown MgO(1 0 0) film at 90 K, whereas no CO adsorption was detected on
the annealed MgO(1 0 0) film, in agreement with the MIES and LEED results. For MgO thin films with a thickness of
15 ML, no surface charging was observed during the UPS/MIES measurements. Ó 2002 Elsevier Science B.V. All
rights reserved.
Keywords: Photoelectron spectroscopy; Magnesium oxides; Carbon monoxide; Surface defects
1. Introduction
Geometric and electronic structures of oxide surfaces as well as their interactions with gas molecules have been widely studied both experimentally
and theoretically because of their importance in a
variety of technologies, e.g., catalysis, gas sensors,
electronic devices, etc. [1]. Investigating the electronic properties of insulating oxides using surface
analytical techniques is often complicated because
of sample charging. To overcome this problem,
oxide thin films can be synthesized on refractory
metal substrates. Extensive studies have been devoted to the preparation and characterization
*
Corresponding author. Tel.: +1-9798456822; fax: +19798450214.
E-mail address: [email protected] (D.W.
Goodman).
of oxide thin films prepared on refractory metals
[2–5].
Previous investigations have shown that certain
properties of thin films can differ from those of
well-ordered, bulk single crystals. An example of
the differences between thin film data and those
of the respective bulk material is the adsorption of
CO on MgO. The binding energy of CO on MgO
thin films grown on Mo(1 0 0) was estimated to
be 0.43–0.45 eV using temperature programmed
desorption (TPD) as well as Clausius–Clapyron
plots [6,7]. In contrast, early experiments using
MgO powders showed the binding energy of CO to
be 0.14–0.16 eV [8]. A recent investigation using a
vacuum-cleaved MgO(1 0 0) surface reports a binding energy for CO of 0.14 eV [9]. Recent theoretical investigations estimated the binding energy
of CO on MgO(1 0 0) to be 0.08–0.11 eV, consistent with the experimental values for MgO bulk
crystals, but contradictory to the results for the
0039-6028/02/$ - see front matter Ó 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 0 3 9 - 6 0 2 8 ( 0 2 ) 0 1 3 8 6 - 9
Y.D. Kim et al. / Surface Science 506 (2002) 228–234
thin films [10–13]. The higher binding energy of
CO on MgO thin films was attributed to CO adsorption at defect sites such as steps and corners,
i.e., the thin films are purported to contain higher
defect densities compared with bulk single crystals
[11,13].
Data from highly defective thin films can be
misleading since features arising from defects can
be misinterpreted as properties of the perfect oxide surface. Very recently thick (>30 monolayers (MLs)) MgO(1 0 0) films that contain a defect
density comparable to that of vacuum-cleaved
MgO(1 0 0) were synthesized on Mo(1 0 0) [14].
However, as shown in Fig. 1, surface charging is
observed during electron spectroscopic measurements on MgO films exceeding 30 ML in thickness. Surface charging can cause difficulty in
interpretation of data, e.g., X-ray photoelectron
spectra from charged samples exhibit shifts of
binding energies that are related to changes in the
chemical environment of the surface elements and/
or surface charging. An ideal MgO thin film must
have a relatively low concentration of defects in
order to serve as a proper model for a well-ordered
MgO single crystal yet thin enough to prevent
charging. Therefore, efforts to produce high quality MgO films are warranted.
In this work, MgO(1 0 0) films grown on Mo(1 0 0) under various conditions were characterized
Fig. 1. MIES spectra collected successively from bottom to top
from a MgO film with a thickness of 30 ML at room temperature. The MIES and UPS (not shown) spectra shift toward
higher binding energies as a function of time, indicating
charging of the MgO film.
229
using low energy electron diffraction (LEED) and
metastable impact electron spectroscopy/ultraviolet photoelectron spectroscopy (MIES/UPS) in
conjunction with CO adsorption experiments. A
well-ordered MgO(1 0 0) film could be prepared by
annealing the as-grown MgO(1 0 0) films with a
thickness of approximately 15 ML at 1150 K. CO
adsorption on well-ordered MgO(1 0 0) thin films
was not observed in agreement with data from
MgO single crystals [9]. No surface charging was
found during MIES/UPS measurements for the
MgO(1 0 0) thin films when the thickness of the
film was 15 ML or less.
2. Experimental
These experiments were carried out in an ultrahigh vacuum (UHV) system with a base pressure
of 1 1010 Torr. The UHV system consists of two
contiguous chambers, one equipped with an iongun for sputtering, LEED, and TPD, and a second
equipped with Auger electron spectroscopy (AES),
X-ray photoelectron spectroscopy (XPS), and
MIES/UPS experiments. MIES is a non-destructive and extremely surface sensitive technique since
the metastable helium atoms approach the surface
with thermal kinetic energy. For insulating materials, Auger de-excitation (AD) is the dominant
mechanism for electron emission in MIES. In AD,
the intensity of the ejected electrons as a function
of their kinetic energy corresponds directly with
the surface density of states (SDOS) for the topmost layer of the surface. More details can be
found in a recent review of the technique [15].
MIES/UPS spectra were measured simultaneously
using a cold-cathode discharge source [16,17] that
provides both ultraviolet photons and metastable
He23 S/21 S (E ¼ 19:8=20:6 eV) atoms with thermal
kinetic energy. Most of the metastable He atoms
are in the triplet state, and the contribution of the
singlet metastable He in MIES spectra is negligible. Metastable and photon contributions to the
signal were separated via differences in time-offlight using a mechanical chopper. MIES and UPS
spectra were taken with incident photon/metastable beams at 45° with respect to the surface normal
in a constant pass energy mode using a double
230
Y.D. Kim et al. / Surface Science 506 (2002) 228–234
pass cylindrical mirror analyzer (CMA). The energy denoted by EF in the spectra corresponds to
electrons emitted from the Fermi level of Mo(1 0 0)
substrate. In the following spectra, all binding
energies are referenced to EF .
The Mo(1 0 0) sample was cleaned by repetitive
heating to 2200 K, and the sample cleanliness
verified with AES. The LEED pattern of the clean
sample showed a (1 1)-periodicity with intense
spots.
3. Results and discussion
3.1. LEED and MIES results from various MgO(1 0 0) films
MgO films were grown on Mo(1 0 0) by deposition of Mg in an O2 background of 1 107 Torr
at a sample temperature of 600 K. The thickness of
the thin films was varied within the range of 2–15
ML. The film thickness was determined by careful
calibration of the Mg doser using the Auger break
point method. It is assumed that the sticking
probabilities of Mg on Mo(1 0 0) and on MgO(1 0 0) are identical. In all cases, AES did not show
a metallic Mg peak at 44 eV, whereas the Mg2þ
peak at 32 eV was distinctly visible [5] indicating that Mg was completely oxidized. The LEED
pattern of these thin films exhibited a (1 1)periodicity with diffuse spots and a relatively high
background intensity, indicating that the as-grown
MgO thin films are rough and poorly ordered.
These results are consistent with recent results that
showed a sharp (1 1)-LEED pattern from the asgrown MgO films only when the thickness of the
film was greater than 30 ML [14].
Previous LEED investigations found that subsequent annealing of the MgO thin films does not
lead to improvement in the LEED patterns [5]. In
fact, when the thickness of the MgO films was less
than 5 ML, the diffuse (1 1)-LEED pattern
remained unchanged upon annealing at 1150 K. In
contrast, a sharp (1 1)-LEED pattern could be
obtained by annealing a MgO thin film grown at
600 K with a thickness of 15 ML.
In Fig. 2, MIES data for as-grown MgO(1 0 0)
films taken before and after annealing at 1150 K
Fig. 2. MIES spectra acquired from a MgO film before and
after annealing (annealing temperature ¼ 1150 K). The thickness of the film was 15 ML.
are compared. The films were grown at 600 K
with a thickness of 15 ML. The full-width-halfmaximum (FWHM) of the O(2p) peak decreased
from 3.2 to 2.5 eV upon annealing. Narrowing of
the O(2p) band, indicating that the surface becomes more ordered upon annealing, is consistent
with the LEED results, i.e. the (1 1)-spots are
sharper after annealing. Our LEED and MIES
results agree with previous AFM investigations of
MgO single crystals where a high temperature
anneal (T > 1250 K) reduced the roughness of the
MgO single crystal surfaces significantly [18].
3.2. CO adsorption
3.2.1. CO adsorption on the as-grown MgO films
The MgO(1 0 0) film grown at a sample temperature of 600 K (thickness ¼15 ML) was exposed to CO at 90 K before annealing, and the
MIES/UPS spectra were collected as a function of
CO exposure. As mentioned above, this film exhibited a diffuse (1 1)-LEED pattern, and a relatively broad O(2p) feature in the MIES spectrum,
indicating that the surface is not well-ordered. Fig.
3a shows the MIES spectra collected as a function
of CO exposure, 3b shows the difference spectra,
and 3c shows the change of the O(2p) intensities
with increasing CO exposure.
Y.D. Kim et al. / Surface Science 506 (2002) 228–234
231
Fig. 3. (a) MIES spectra taken as a function of CO exposure from the as-grown MgO(1 0 0) film 15 ML in thickness. (b) Difference
spectra of (a). Before subtracting the MgO-spectrum from the CO/MgO spectrum, the MgO spectrum was attenuated taking into
account the CO-induced damping of the MgO features. (c) The change of the O(2p) intensities as a function of CO exposure. (T ¼ 90
K, P ðCOÞ ¼ 2 108 Torr.)
The intensity of the O(2p) peak decreases dramatically at the early stage of the CO exposure,
and begins to attenuate more slowly with increasing CO exposure. The intensity of the O(2p)
shows an attenuation of 35% upon CO exposures
of 10 L or more (Fig. 3c). At the beginning of the
CO exposure, additional peaks at 10, 12.4 and 14.6
eV develop in the MIES spectra. As the CO exposure increases, the peak at 12.4 eV shifts to 12.0
eV. Further CO exposure leads to a gradual shift
of the peaks at 12.0 and 14.6 eV to lower binding
energies. The attenuation of the O(2p) peak saturates at a CO exposure of 10 L, whereas the COinduced peaks continue to shift to lower binding
energies with higher CO exposures.
In Fig. 4, UPS spectra collected as a function of
CO coverage are shown. The overall changes of
the UPS spectra upon CO exposure are similar to
those of the MIES spectra in Fig. 3. The only
notable difference between the MIES and UPS
spectra is that the binding energies of the COinduced peaks in the UPS spectra are higher than
those in the MIES spectra by 1.2–1.3 eV. This
same phenomenon was also observed in the case of
D2 O adsorption on MgO(1 0 0)/Mo(1 0 0) [19].
The three CO-peaks at 10, 12.4, and 14.6 eV in
the MIES spectra in Fig. 3 (or the peaks at 11.2,
13.5 and 15.8 eV in the UPS spectra) can be assigned to the 5r, 1p, and 4r orbitals of CO. In the
gas phase, the separation between the 5r and 1p
232
Y.D. Kim et al. / Surface Science 506 (2002) 228–234
Fig. 4. (a) UPS spectra taken as a function of CO exposure from the as-grown MgO(1 0 0) film. These UPS spectra were collected
simultaneously with the MIES spectra in Fig. 3 from a film 15 ML in thickness. (b) Difference spectra of (a). Before subtracting the
MgO-spectrum from the CO/MgO spectra, the MgO spectrum was attenuated taking into account the CO-induced damping of the
MgO features. (T ¼ 90 K, P ðCOÞ ¼ 2 108 Torr).
levels is 2.9 eV, whereas on transition metal surfaces, the 1p and 5r levels typically overlap because of the shift of the 5r orbital [20]. On
MgO(1 0 0) surfaces, the 5r and 1p levels are
separated by 1.6 eV in both the MIES (Fig. 3) and
UPS spectra (Fig. 4). A separation of 1.6 eV
for the 1p and 4r peaks was also observed for
ZnO(1 0 1 0) [21]. The large separation between the
5r and 1p orbitals on both oxide surfaces is likely
a result of weaker interaction of the 5r level with
the oxide surfaces as compared to transition metal
surfaces.
The shifts of the CO-peaks with increasing CO
coverage are consistent with a strong lateral interaction between the CO molecules. Thus, it is
likely that the CO coverage at the adsorption site
or sites is relatively high at 90 K. However, the
O(2p) attenuation in the MIES spectra is only 35%
upon a 10 L or more exposure of CO at 90 K,
indicating that the overall coverage of CO on the
surface at 90 K is quite low. Note that the completion of the CO monolayer should result in a
100% attenuation of the O(2p) feature because
MIES is exclusively sensitive to the outermost
surface layer. From these results, it is reasonable
to conclude that CO adsorption takes place only
on a specific area of the surface, most probably
extended defect sites such as step edges and corners.
3.2.2. CO on the annealed MgO(1 0 0) surface
To investigate the interaction between CO and
the annealed MgO(1 0 0) surfaces (thickness 15
ML), MIES/UPS spectra were collected during CO
exposure (Fig. 5). No change of the MIES/UPS
spectra could be observed even with a CO background pressure of 1 106 Torr. Consequently,
no adsorption of CO takes place at 90 K on the
MgO(1 0 0) surface after the MgO(1 0 0) thin film
has been annealed at 1150 K. These results are in
agreement with a recent TPD study using a bulk
MgO single crystal where no CO TPD peaks were
found above 90 K [9]. Furthermore, the COinduced c(4 2)-LEED structure is observed on the
MgO(1 0 0) single crystal at 40 K, whereas above
55 K, no superstructure spots were observed, indicating that desorption of the first CO monolayer
takes place at this temperature [22].
Recently, we have found that the D2 O TPD
spectrum from an annealed MgO(1 0 0) thin film
with a thickness of 15 ML is identical to that
from a flat and well-ordered MgO(1 0 0) single
crystal [19]. Because water TPD spectra from
MgO(1 0 0) are very sensitive to the surface roughness [23,24], we suggest that the roughness of the
vacuum-annealed MgO(1 0 0) film is comparable
to that of well-ordered MgO(1 0 0) single crystals, which were characterized by AFM [18].
Therefore, the preparation method in this work
Y.D. Kim et al. / Surface Science 506 (2002) 228–234
233
Fig. 5. MIES/UPS spectra taken as a function of CO exposure from the annealed MgO(1 0 0) film (thickness ¼ 15 ML). T ¼ 90 K,
P ðCOÞ ¼ 1 106 Torr.
Fig. 6. MIES spectra collected successively from the MgO thin
film with a thickness of 15 ML at room temperature. No shift
of the MIES spectrum as a function of time was found, indicating no surface charging.
apparently produces well-ordered MgO(1 0 0) surfaces with chemical properties essentially identical
to those of in situ cleaved single crystals. Fig. 6
illustrates that as long as the thickness of the MgO
films is 15 ML or less, no shift of the MIES spectra
as a function of time is observed, indicating no
surface charging.
ML are highly defective. For MgO films approximately 15 ML in thickness, annealing leads to the
formation of a flat, well-ordered MgO(1 0 0) surface. CO adsorbs on the as-grown MgO(1 0 0) film,
whereas no CO is adsorbed on an annealed MgO
film. Therefore, CO selectively adsorbs on defect
sites on MgO(1 0 0) at 90 K. These results agree
with previous bulk MgO single crystal data [9] as
well as recent theoretical calculations [10–13]. In
addition, these results show that CO can be used as
a probe molecule to titrate defect sites on MgO.
We have demonstrated that the MgO films do not
charge as long as the thickness of the film is 15 ML
or less. Therefore, MgO films, properly prepared,
can provide useful models to explore the structural
and chemical properties of MgO single crystals.
Acknowledgements
Funding for this work was provided by the
Department of Energy, Office of Basic Energy
Sciences, Division of Chemical Sciences and the
Robert A. Welch Foundation.
References
4. Conclusions
By using LEED and MIES, we have shown that
as-grown MgO films with thicknesses below 15
[1] V.E. Henrich, P.A. Cox, The Surface Science of Metal
Oxides, Cambridge University Press, Cambridge, 1994.
[2] X.P. Xu, D.W. Goodman, Surf. Sci. 282 (1993) 323.
[3] P.J. Chen, D.W. Goodman, Surf. Sci 312 (1994) L767.
234
Y.D. Kim et al. / Surface Science 506 (2002) 228–234
[4] W.S. Oh, C. Xu, D.Y. Kim, D.W. Goodman, J. Vac. Sci.
Technol. A 15 (1997) 1710.
[5] M.C. Wu, J.S. Corneille, C.A. Estrada, J.W. He, D.W.
Goodman, Chem. Phys. Lett. 182 (1991) 472.
[6] J.W. He, J.S. Corneille, C.A. Estrada, M.C. Wu, D.W.
Goodman, J. Vac. Sci. Technol. A 10 (1992) 2248.
[7] J.W. He, C.A. Estrada, J.S. Corneille, M.C. Wu, D.W.
Goodman, Surf. Sci. 261 (1992) 164.
[8] S. Furuyama, H. Fujii, M. Kawamura, T. Morimoto, J.
Phys. Chem. 82 (1978) 1028.
[9] R. Wichtendahl, M. Rodriguez-Rodrigo, U. Hartel, H.
Kuhlenbeck, H.J. Freund, Phys. Status Solidi A 173 (1999)
93.
[10] K.M. Neyman, S.P. Ruzankin, N. Rosch, Chem. Phys.
Lett. 246 (1995) 546.
[11] J.A. Snyder, D.R. Alfonso, J.E. Jaffe, Z.J. Lin, A.C. Hess,
M. Gutowski, J. Phys. Chem. B 104 (2000) 4717.
[12] F. Illas, G. Pacchioni, A. Pelmenschikov, L.G.M. Pettersson, R. Dovesi, C. Pisani, K.M. Neyman, N. Rosch, Chem.
Phys. Lett. 306 (1999) 202.
[13] M.A. Nygren, L.G.M. Pettersson, J. Phys. Chem. 105
(1996) 9339.
[14] Z. Dohnalek, G.A. Kimmel, S.A. Joyce, P. Ayotte, R.S.
Smith, B.D. Kay, J. Phys. Chem. B. 105 (2001) 3747.
[15] Y. Harada, S. Masuda, H. Ozaki, Chem. Rev. 97 (1997)
1897.
[16] W. Mausfriedrichs, M. Wehrhahn, S. Dieckhoff, V.
Kempter, Surf. Sci. 237 (1990) 257.
[17] W. Mausfriedrichs, S. Dieckhoff, V. Kempter, Surf. Sci.
249 (1991) 149.
[18] S.S. Perry, H.I. Kim, S. Imaduddin, S.M. Lee, P.B. Merrill,
J. Vac. Sci. Technol. A 16 (1998) 3402.
[19] Y.D. Kim, R.M. Lynden-Bell, A. Alavi, J. Stultz, D.W.
Goodman, Chem. Phys. Lett. 352 (2002) 318.
[20] G. Ertl, J. Kuppers, Low Energy Electrons And Surface
Chemistry, VCH Verlagsgesellschaft Mbh, Weinheim, 1985.
[21] J. Lin, P. Jones, J. Guckert, E.I. Solomon, J. Am. Chem.
Soc. 113 (1991) 8312.
[22] P. Audibert, M. Sidoumou, J. Suzanne, Surf. Sci. 273
(1992) L467.
[23] S.I. Ahmed, S.S. Perry, O. El-Bjeirami, J. Phys. Chem. B.
104 (2000) 3343.
[24] Y.D. Kim, J. Stultz, D.W. Goodman, Langmuir, submitted for publication.