вход по аккаунту


Size-Dependent Optical Properties of MgO Nanocubes.

код для вставкиСкачать
DOI: 10.1002/anie.200500663
Size-Dependent Optical Properties of MgO
Slavica Stankic, Markus Mller, Oliver Diwald,
Martin Sterrer, Erich Knzinger,* and
Johannes Bernardi
Understanding the nature and chemical activity of welldefined surface structures on polycrystalline oxide systems
[*] Dipl.-Chem. S. Stankic, Dipl.-Ing. M. M!ller, Dr. O. Diwald,
Dr. M. Sterrer,+ Prof. Dr. E. Kn+zinger
Institute of Materials Chemistry
Vienna University of Technology
Veterin4rplatz 1/GA, 1210 Vienna (Austria)
Fax: (+ 43) 1-2507-73890
Dr. J. Bernardi
University Service Centre for Transmission Electron Microscopy
Vienna University of Technology
Wiedner Hauptstrasse 8–10/137, 1040 Vienna (Austria)
[+] Present address: Department of Chemical Physics
Fritz-Haber-Institute of the Max-Planck-Society
Faradayweg 4–6, 14195 Berlin (Germany)
[**] We thank T. Berger, K. Beck, P. V. Sushko, and A. L. Shluger for
fruitful discussions and advice and gratefully acknowledge the
financial support from the Austrian Fonds zur F+rderung der
wissenschaftlichen Forschung (FWF grants P14731-CHE and
P14730-CHE) and the Hochschuljubil4umsstiftung der Stadt Wien
Angew. Chem. Int. Ed. 2005, 44, 4917 –4920
requires atomic-scale characterization.[1, 2] However, the properties of surface ions on polycrystals usually vary over a wide
range, because different crystallographic faces are disposed to
form edges, steps, and point defects. This situation leads to a
substantial energetic heterogeneity in the surface properties.
Consequently, it is one of the major challenges in current
catalysis research to single out ions and point defects
associated with specific surface structures and to characterize
their role in the complex surface chemistry.[1–4] Magnesium
oxide (MgO) is a particularly well-suited model system for
such studies because of its straight-forward rock-salt crystal
structure, the large abundance of thermodynamically stable
(100) faces, and the purely ionic, nonconducting nature of the
compound. Low-coordinate surface anions, which are considered to be chemically reactive sites, give rise to specific
optical transitions in the UV light range, which can be
assigned to the formation of surface excitons.[2, 5] Their
investigation provides an opportunity to establish a correlation between topological surface features such as corners,
steps, and edges and the corresponding chemical and spectroscopic properties.[5–8] For the efficient investigation of
surface sites, nanocrystalline model systems (where the
number of surface ions is of the same order of magnitude as
the number of bulk ions) are required.[9–11]
Herein we report the production of nanometer-sized MgO
nanoparticles by chemical vapor deposition (CVD). They
attain an almost perfect cubic shape after thermal activation
at 1170 K under high-vacuum conditions (p < 5 6
10 6 mbar).[12] Depending on the CVD parameters they can
adopt different average sizes between 3 nm and 10 nm.
Preferential adjustment of the cube size is important since it
sets the relative concentrations of corner and edge anions and
thus determines the optical and chemical properties. Sizedependent trends in the relative intensities of absorption
bands and photoluminescence emission bands were observed
in the UV light range. A novel photoluminescence emission
band was observed for nanocubes with edge lengths smaller
than 9 nm and is attributed to the radiative deactivation of
surface excitons at corner anions of the MgO nanocube.
The MgO nanoparticles are prepared by CVD in a flowreactor system by the reaction of metal vapor with different
gaseous oxidizing agents. At pressures under 20 mbar, metal
atoms are evaporated from metal grains and transported by
argon at a fixed flow rate to the downstream end of a quartz
glass tube where the oxidant is introduced. There the metal
vapor is combusted either with dry air (sample A), O2 (sample B), or N2O (sample C). The resulting metal oxide nanocrystals are then deposited downstream in a stainless steel net
at room temperature.[13] Afterwards, the material is subjected
to thermal activation at 1170 K. Figure 1 shows typical TEM
images of MgO nanocube samples A, B, and C after such
treatment. The nature and concentration of the oxidizing
agent has a major influence on the nucleation process and
consequently affects the average size of the particles after
thermal activation. This feature is shown by the size
distribution plots in Figure 2.[14] The particles of sample A
are slightly less regular cubes and have an average edge length
of 10 nm, whereas samples B and C consist of 5 and 3 nm sized
nanocubes, respectively.
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 1. Typical TEM images of MgO nanocubes that result from CVD in conjunction with subsequent thermal activation at 1170 K and
p < 5 I 10 6 mbar. For samples A, B, and C different oxidants were used at the same flow rates for the CVD process, as outlined in the text.
whereas the 4.6 eV (l = 270 nm) absorption band is attributed
to the excitation of three-coordinate anions in corner
positions.[2, 5, 8] Larger cube sizes imply lower relative concentrations of corner anions with respect to edge anions. This
feature is exemplified in Table 1, where the calculated
Table 1: Calculated numbers of three- and four-coordinate surface
anions located in ideal MgO nanocube edges and corners as a
function of particle size.
Figure 2. Size distribution plots of CVD MgO nanocubes. About
200 particles were observed to plot each distribution.
The optical properties of the nanocubes were measured by
UV diffuse reflectance spectroscopy. The spectra, recorded in
the presence of 10 mbar of O2 to avoid photoluminescence
(Figure 3),[15] reveal two clear trends with decreasing average
cube size from sample A (10 nm) to sample C (3 nm):
1) The maximum of the high-energy absorption shifts from
220 nm (5.6 eV) to 230 nm (5.4 eV).
2) The absorption at 270 nm (4.6 eV), which is essentially
absent on sample A (10 nm), gains relative intensity.
The second observation is perfectly in line with previous
assignments of these optical features: with an energy of 5.4 eV
(l = 230 nm) four-coordinate anions in cube edges are excited
Figure 3. UV diffuse reflectance spectra of samples A, B, and C. For
the sake of clarity, the curves were shifted on the ordinate scale.
edge length of parti[nm]
3.6 I 1017
8.4 I 1016
1.0 I 1016
Number of
corner O2
Number of
edge O2
Ratio of corner
to edge O2
1.4 I 1018
3.4 I 1017
4 I 1016
3.5 I 1019
1.5 I 1019
3.7 I 1018
[a] In 100 mg of MgO powder.
numbers of surface O2 ions for the three particle sizes are
listed. For the calculation we used an Mg2+–O2 distance of
2.1 @ and assumed a perfect cubic shape, where the particle is
terminated by (100) planes, cube edges, and cube corners. For
a constant quantity of 100 mg, the total number of corner
anions ranges from 4 6 1016 for 10 nm cubes to 144 6 1016 ions
for 3 nm cubes and the ratio of corner anions to edge anions
increases from 0.01 to 0.04. This enhancement in the relative
concentration of corner anions is qualitatively reflected in the
relative growth of the 270 nm band (absorption on corners)
compared to the one at 240 nm (absorption on edges) shown
in Figure 3.
Luminescence spectra of the three types of nanocube
samples were recorded at 300 K and p < 10 6 mbar (Figure 4).
Depending on the excitation wavelength l = 240 nm (5.2 eV)
or l = 270 nm (4.6 eV) the spectra are characterized by broad
emission bands with maxima at l = 370 nm (3.4 eV) and l =
380 nm (3.3 eV), respectively. The full width at half maximum
(FWHM) for both bands is 0.8 eV. The photoluminescence
signal at 370 nm (excitation at 240 nm) is in general agreement with results from previous studies.[16–20] However, to our
knowledge, the feature which is centered at slightly higher
wavelength l = 380 nm (excitation at l = 270 nm) has not
been reported as a distinct band to date.[21] The intensity of
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 4917 –4920
Experimental Section
Figure 4. Photoluminescence emission spectra of samples A (10 nm),
B (5 nm), and C (3 nm). For comparison, the emission signals excited
at 240 nm (black curves) are set to the same nominal intensity as for
the bands following excitation at 270 nm (gray curves).
this band is significantly higher for samples B and C with short
cube edges than for sample A.[22] Excitation of corner anions
produces surface excitons at these sites which then deactivate
radiatively at 3.3 eV (l = 380 nm). This photoluminescence
effect is thus preferentially observed on small MgO cubes
(samples B and C; Figure 4) and can only be understood in
terms of the enhanced surface concentration of corner anions
(Table 1). The higher Stokes shift of the emission band at
370 nm (5.4 to 3.4 eV compared to 4.6 to 3.3 eV) is explained
by an energy transfer from the four-coordinate edge anion,
where excitation initially occurs, to a three-coordinate corner
anion acting as the emission site[6, 8, 16] On the basis of the
demonstrated size-dependence, we attribute the novel band
at 380 nm to a new electronic transition on MgO surfaces.
This interpretation is in line with a recent electron paramagnetic resonance (EPR) study where the formation of two
types of trapped hole centers, that is, O radicals on the MgO
surface was reported.[23] The fact that the local crystal-field
splittings of these two types of radicals are similar, connects
well to the two emission bands of similar energy observed in
this study. The linking of optical properties and the generation
of surface radicals suggests that on MgO nanostructures the
selective physical and chemical activation of well-defined
surface structures is feasible. Further work is needed to
understand the nature and mobility of surface excitons on
MgO nanostructures and to investigate how changes in the
concentration ratio of corner to edge ions affect surface
In conclusion, size control over MgO nanocubes has been
achieved through the CVD technique in conjunction with
subsequent thermal activation steps at p < 5 6 10 6 mbar. The
resulting adjustment of the ratio between corner and edge
ions is clearly reflected in the optical properties of the MgO
nanocubes as measured by UV diffuse reflectance and
photoluminescence spectroscopy. Adjustment of the average
particle size makes MgO nanocubes a powerful model system
for surface-science studies on dispersed materials. In addition,
it represents a promising building block element for the
construction of functional nano- and mesostructures.
Angew. Chem. Int. Ed. 2005, 44, 4917 –4920
The three types of MgO nanocube samples A, B, and C were obtained
by using dry air (N2/O2 = 4:1 (A), O2 (B), and N2O (C) as the
oxidizing agent in the CVD process.[13] Flow rates for Argon and
the oxidizing agent were 1220 and 82 s cm3 min 1, respectively. For
UV Diffuse reflectance and photoluminescence measurements the
sample material was transferred into a high-vacuum tight quartz cell
providing pressures better than p < 10 6 mbar. For each type of
measurement, thermal treatment was carried out to remove hydroxy
and carbonate groups from the surface (confirmed by IR spectroscopy). During this treatment the annealing rate was 5 K min 1 and the
pressure above the sample was < 10 5 mbar. The sample was first
heated to 873 K and treated with oxygen to remove organic
contaminants. The final annealing temperature of 1173 K was kept
for 3–10 h at pressures < 5 6 10 6 mbar. Photoluminescence measurements were carried out with a Perkin Elmer LS 50 B system equipped
with a pulsed Xe discharge lamp as the excitation light source. The
excitation light beam was passed through a UG 5 filter (Schott) in
conjunction with a 295 nm long pass filter. UV diffuse reflectance
measurements were performed on a Perkin Elmer Lambda 15
spectrometer with an integrating sphere. For TEM investigations a
holey carbon grid was immersed into the metal oxide powder samples
which had been previously investigated by spectroscopy. The TEM
measurements were carried out with a TECNAI F20 analytical
transmission electron microscope equipped with a field-emission
electron source and a S-Twin objective lens. Images were recorded
with a Gatan 794 MultiScan camera. No nitrogen impurities were
detected on all samples as measured with electron energy loss
spectrometry (EELS) using a Gatan Image Filter GIF 2001.
Received: February 22, 2005
Published online: July 6, 2005
Keywords: magnesium oxide · nanostructures ·
photoluminescence spectroscopy · transmission electron
microscopy · UV diffuse reflectance
[1] a) V. E. Henrich, P. A. Cox, The Surface Science of Metal Oxides,
Cambridge University Press, Cambridge, 1996; b) G. Pacchioni,
ChemPhysChem 2004, 5, 1041 – 1047; c) C. Barth, C. R. Henry,
Phys. Rev. Lett. 2003, 91, 196102; d) M. Chiesa, M. C. Paganini,
E. Giamello, C. Di Valentin, G. Pacchioni, Angew. Chem. 2003,
115, 1801 – 1803; Angew. Chem. Int. Ed. 2003, 42,1759 – 1761;
e) Z. Yan, S. Chinta, A. A. Mohamed, J. P. Fackler, Jr., D. W.
Goodman, J. Am. Chem. Soc. 2005, 127, 1604 – 1605.
[2] a) G. Spoto, E. N. Gribov, G. Ricchiardi, A. Damin, D. Scarano,
S. Bordiga, C. Lamberti, A. Zecchina, Prog. Surf. Sci. 2004, 76,
71 – 146; b) E. Garrone, A. Zecchina, F. S. Stone, Philos. Mag.
1980, 42, 683 – 703.
[3] H. KnKzinger, Science 2000, 287, 1407 – 1408.
[4] P. K. Stoimenov, V. Zaikovski, K. J. Klabunde, J. Am. Chem. Soc.
2003, 125, 12 907 – 12 913.
[5] P. V. Sushko, J. L. Gavartin, A. L. Shluger, J. Phys. Chem. B
2002, 106, 2269 – 2276.
[6] M. Sterrer, T. Berger, O. Diwald, E. KnKzinger, J. Am. Chem.
Soc. 2003, 125, 195 – 199.
[7] D. Ricci, C. Di Valentin, G. Pacchioni, P. V. Sushko, A. L.
Shluger, E. Giamello, J. Am. Chem. Soc. 2003, 125, 738 – 747.
[8] a) A. L. Shluger, P. V. Sushko, L. N. Kantorovich, Phys. Rev. B
1999, 59, 2417 – 2430; b) P. V. Sushko, A. L. Shluger, Surf. Sci.
1999, 421, L157 – L165.
[9] R. Kakkar, P. N. Kapoor, K. J. Klabunde, J. Phys. Chem. B 2004,
108, 18 140 – 18 148.
[10] R. M. Richards, A. M. Volodin, A. F. Bedilo, K. J. Klabunde,
Phys. Chem. Chem. Phys. 2003, 5, 4299 – 4305.
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
[11] Z. Dohnalek, G. A. Kimmel, D. E. McCready, J. S. Young, A.
Dohnalkova, R. S. Smith, B. D. Kay, J. Phys. Chem. B 2002, 106,
3526 – 3529.
[12] This treatment leads to entire sample dehydroxylation as judged
by IR spectroscopy.
[13] Further details of this procedure are described in: E. KnKzinger,
K.-H. Jacob, S. Singh, P. Hofmann, Surf. Sci. 1993, 290, 388 – 402.
[14] The underlying differences in the process of nucleation and
particle growth during CVD are not yet understood and in the
process of investigation.
[15] Gaseous O2 quenches luminescence and prevents artifacts that
originate from light emission in the optical absorption spectra.
This quenching effect, in turn, clearly demonstrates that UVlight absorption and emission exclusively stem from surface
[16] a) S. Coluccia, A. M. Deane, A. J. Tench, J. Chem. Soc. Faraday
Trans. 1 1978, 74, 2913 – 2922; b) S. Coluccia in Adsorption and
Catalysis on Oxide Surfaces (Eds.: M. Che, G. C. Bond),
Elsevier, Amsterdam, 1985, pp. 59 – 69.
[17] M. Anpo, M. Che, Adv. Catal. 2000, 44, 120 – 248.
[18] M. Anpo, Y. Yamada, Y. Kubokawa, S. Coluccia, A. Zecchina,
M. Che, J. Chem. Soc. Faraday Trans. 1 1988, 84, 751 – 764.
[19] S. G. MacLean, W. W. Duley, J. Phys. Chem. Solids 1984, 45,
227 – 235.
[20] M. L. Bailly, G. Costentin, H. Lauron-Pernot, J. M. Krafft, M.
Che, J. Phys. Chem. B 2005, 109, 2404 – 2413.
[21] Emission were also observed at higher wavelengths l > 450 nm
in the course of these studies on samples which had not been
entirely dehydroxylated.
[22] Intensity trends of the absorptions bands at 240 nm and 270 nm
in Figure 3 do not directly scale with those of the associated
emission bands in Figure 4. This fact is attributed to one or the
combination of the following reasons: a) the cross sections for
absorption (measured in the presence of O2) and emission
(measured at p < 10 6 mbar) are unknown and may be subject to
particle size, b) Rayleigh scattering into the interior of the
sample cell causes the loss of detectable light and has a distinct
dependence on particle size (~ d3) and frequency (~ n4). The
relative extent of scattering loss is expected to be different for
both measurements owing to inherent differences in the optical
set up. c) Photoluminescence spectroscopy is by far the more
sensitive technique.
[23] O. Diwald, M. Sterrer, E. KnKzinger, P. Sushko, A. Shluger, J.
Chem. Phys. 2002, 116, 1707 – 1712.
2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 4917 –4920
Без категории
Размер файла
128 Кб
properties, optical, size, dependence, nanocube, mgo
Пожаловаться на содержимое документа