Photostimulated generation of defects and surface reactions on a series of wide band gap metal-oxide solids

Article abstract of DOI:10.1021/jp990664z

The interconnection between photostimulated formation and destruction of defect centers (F- and V-type color centers) and surface photochemical reactions taking place on the surface has been examined for a series of 21 wide band gap metal-oxide specimens, which comprise insulators and semiconductors whose band-gap energies span the range from ca. 3 eV (ZnO, TiO₂) to ca. 11 eV (BeO). Photostimulated post-adsorption of O₂ was seen for 9 of the 21 metal oxides tested. Three of these specimens, namely scandia (Sc₂O₃), zirconia (ZrO₂), and normal spinel (MgAl₂O₄), were chosen for detailed study to establish that spectral sensitization of metal oxides by UV illumination is a generally occurring phenomenon which is carried over to photostimulated adsorption of molecules and to surface photochemical reactions. Results demonstrate that irradiation into the absorption bands of the color centers of the photocolored metal oxides leads to a red shift of the spectral limit of surface photoreactions. A simple mechanism is described quantitatively for the photocoloration and surface reactions. Quantum yields of photobleaching of color centers in vacuo and in the presence of H₂ and O₂ are reported for scandia and zirconia, together with the quantum yields of photoadsorption of H₂ and O₂ molecules.

Full text of DOI:10.1021/jp990664z

9190
J. Phys. Chem. B 1999, 103, 9190-9199
Photostimulated Generation of Defects and Surface Reactions on a Series of Wide Band
Gap Metal-Oxide Solids
†
‡
‡
,†
A. V. Emeline, G. V. Kataeva, V. K. Ryabchuk, and N. Serpone*
Department of Chemistry & Biochemistry, Concordia UniVersity, 1455 de MaisonneuVe BouleVard West,
Montr e´ al (Qu e´ bec), Canada H3G 1M8, and Department of Physics, UniVersity of St. Petersburg,
UlianoVskaia 1, St. Petersburg, 198904 Russia
ReceiVed: February 23, 1999; In Final Form: June 19, 1999
The interconnection between photostimulated formation and destruction of defect centers (F- and V-type
color centers) and surface photochemical reactions taking place on the surface has been examined for a series
of 21 wide band gap metal-oxide specimens, which comprise insulators and semiconductors whose band-gap
energies span the range from ca. 3 eV (ZnO, TiO
was seen for 9 of the 21 metal oxides tested. Three of these specimens, namely scandia (Sc
ZrO ), and normal spinel (MgAl ), were chosen for detailed study to establish that spectral sensitization
2
) to ca. 11 eV (BeO). Photostimulated post-adsorption of
O
2
2 3
O
), zirconia
(
2
2 4
O
of metal oxides by UV illumination is a generally occurring phenomenon which is carried over to
photostimulated adsorption of molecules and to surface photochemical reactions. Results demonstrate that
irradiation into the absorption bands of the color centers of the photocolored metal oxides leads to a red shift
of the spectral limit of surface photoreactions. A simple mechanism is described quantitatively for the
photocoloration and surface reactions. Quantum yields of photobleaching of color centers in vacuo and in the
presence of H
2
2
and O are reported for scandia and zirconia, together with the quantum yields of photoadsorption
of H and O
2
2
molecules.
Introduction
impact displacement energy threshold (e.g., as might occur when
specimens are irradiated with high-energy radiation) down to
the band-gap energies Ebg of the solids, and perhaps even lower.
For wide band gap ionic crystals, defect formation may occur
by two fundamental pathways as a result of excitation of the
Interaction of light with a solid/gas or a solid/liquid system
has attracted much worldwide attention1-6 in advancing fun-
damental scientific knowledge and numerous technological
applications: e.g., optical and electronic devices, environmental
remediation, and heterogeneous photochemistry and photoca-
talysis. Such interactions differ by which partner of the
heterogeneous system absorbs photons. Absorption by molecules
of the gas or liquid phase adsorbed on the solid’s surface leads
to several consequences, most notable being formation of excited
states of adsorbed molecules, followed by (i) luminescence and
energy and/or electron transfer along the layer of adsorbed
molecules, or to molecules in the gas or liquid phase or to the
bulk of the solid, and (ii) by secondary chemical processes. The
second possibility is light absorption by the solid producing its
own excited states: e.g. excitons, free charge carriers (electrons
and holes), phonons, and others. Energy or charge transfer
toward the surface may lead to surface chemical reactions with
adsorbed molecules or with molecules in the gas or liquid phase
diffusing toward the surface. Other decay channels for the solid’s
excited states are recombination of the charge carriers and the
excitons resulting in photon evolution (radiative recombination,
i.e., luminescence) or phonon emission (nonradiative recombi-
nation). Incomplete recombination of carriers and excitons leads
to radiation-stimulated formation of defects in semiconductors
and insulators.
7
electronic subsystem of crystals. The decay of excitons into
intrinsic ion vacancies and interstitials in regular lattices have
been well established for alkali metal halides and other
8-10
solids.
Two conditions favor this kind of defect formation:
(i) Eex > Ed, i.e., the exciton energy Eex must exceed the energy
Ed of defect (Fraenkel pair) formation (energy conservation law);
and (ii) τex > τvib, i.e., the lifetime of self-trapped excitons in
a given regular lattice site, τex, must be longer than the period
-12
-13
of lattice vibrations: τvib ∼ 10 -10
s (the momentum
11
conservation law). Formation of a Fraenkel pair defect as a
primary product of the photoreaction in a regular lattice is
equivalent to the photodissociation of free molecules. However,
the decay of excitons into Fraenkel pairs, well-known in alkali
metal halides, remains a contentious issue for wide band gap
7
metal oxides, primarily because of the failure of either one or
both of conditions (i) and (ii) above.
A complex interconnection between surface photochemical
processes and defect formation in solids is caused by common
stages of photoexcitation and the participation of photoexcited
electronic subsystems in both processes, which typically mani-
fest themselves as an influence of the photoadsorption of
molecules on the photocoloration of powdered solids.1
2-14
Defect formation in metal-oxide specimens occurs by excita-
tion of the electronic subsystem of the solids under irradiation
by light of moderate photon energies in the range below the
One result connected with photoinduced defects (color
centers) is the spectral sensitization of solids in connection with
the spectral limit of surface photoreactions. This spectral
sensitization was originally observed on alkali metal halides by
*
Author to whom correspondence should be addressed at Concordia
15
Ryabchuck and co-workers with respect to the photostimulated
University. Fax: 514-848-2868. E-mail: serpone@vax2.concordia.ca.
†
‡
adsorption of O2 and CO2, which manifests itself as a red shift
of the spectral limit of photoreactions.
Concordia University.
University of St. Petersburg.
1
0.1021/jp990664z CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/21/1999

Wide Band Gap Metal-Oxide Solids
J. Phys. Chem. B, Vol. 103, No. 43, 1999 9191
TABLE 1: Summary of Data from Preliminary Tests of Metal-Oxide Specimens Examined for the Photostimulated
Post-Adsorption of Oxygen
f
g
dP/dt (PhA)
(10 Pa s
dP/dt (PhSPA)
metal oxide
BeO
MgO
Al
MgAl
SiO
TiO
GeO
E
bg (eV) ∆Rvac (10-3
a
b
)
δRvac (10-4
c
)
δR 2d
O
(10-4
)
(δR/∆R)vace
-5
-1
)
(10 Pa s
-5
-1
)
PhSPA effect
h
10.6
8.7
9.5
8
10
3.2
15
0
0
0
0
0
0
0
0
0
0
no
no
no
no
no
no
no
no
32.6
4.2
12
0.30
8.6
8.5
10.2
4.4
40.2
24.2
9.9
15.2
18
90
53
30
0.1626
0.1905
2
O
3
8.0
8.0
1.0
O
2 4
2
1.0
-9.0
2.0
-5.0
-8.0
88
34
45
40
25
0.3333
-0.1047
0.0235
-0.0490
-0.1818
0.2189
0.1405
0.4546
0.2631
0.1389
2
-23
2.0
2.0
-19
88
22
34
41
2.0
3.0
101
120
0
8.0
81
90
3.0
84
2.8
12
2.0
1.0
1.0
8.0
2.0
3.0
92
37
0
8.0
155
150
62
81
30
320
138
93
49
83
2
SnO
ZnO
2
3.8
3.2
6.0
6.2
5.4
5
5.3
5
5.2
5.0
negative
no
Sc
2
O
3
Y
2
O
3
no
2
3
yes
yes
yes
no
yes
yes
yes
yes
yes
yes
3
3
3
3
36
68.9
71
37
30
80
106
50
80
110
173
75
149
60
90
120
174
0.2222
0.1539
0.0704
0.2162
0.3667
0.2978
ZrO
Y*ZrO
2
2
(5.5%)
Y*ZrO2 (8%)
Y*ZrO (15%)
2
HfO
2
5.5
58.1
a
Band-gap energy of the insulator or semiconductor specimen. b Extent of coloration of the specimen in vacuo. c Extent of photobleaching of a
d
e
colored specimen in vacuo. Extent of photobleaching in the presence of oxygen. Fraction of photobleached color centers after 200 s of visible
irradiation. Rate of photoadsorption of oxygen. Rate of photostimulated post-adsorption of oxygen. h Presence or absence of a photostimulated
f
g
post-adsorption effect in the case of oxygen.
The principal objectives of this study were 2-fold. First, we
wished to establish whether spectral sensitization of metal oxides
by UV pre-illumination is a more general phenomenon than
otherwise imagined, and whether it can be extended to photo-
stimulated adsorption of molecules and to photochemical
reactions. Second, we wished to confirm whether excitation of
metal oxides in the absorption bands of photoinduced color
centers is responsible for the red shifts of the spectral limits of
photoreactions as evidenced in the case of alkali metal halides.
In the present study we tested 21 metal oxides to ascertain the
effect of photoinduced spectral sensitization. Zirconia, scandia,
and spinel specimens were examined in some detail.
Metal Oxide Samples. Zirconia (ZrO2; Ebg ≈ 5.0 eV) is a
photoadsorbent and an active photocatalyst in accordance with
empirical correlations between the photochemical activity of
4
the metal oxide specimen and its band-gap energy Ebg. The
red limit of the photostimulated adsorption of O2, H2, CH4, and
1
2,13
C2H6 molecules on powdered ZrO2
is positioned in the
extrinsic absorption band region at about 3.0 eV (λ ) 410 nm).
The first stage of photooxidation of methane and ethane is
photostimulated adsorption followed by conversion to other
1
2
hydrocarbons. The quantum yield Φ of photooxidation of
16
propene over powdered zirconia is ca. 0.03; Φ of photostimu-
lated adsorption of O2 reaches 0.05 on irradiation of ZrO2 at
the red edge of the fundamental absorption band.17 The main
2
Experimental Section
types of preexisting intrinsic defects in ZrO are anion vacancies,
1
8
Va. Zirconia powder is photocolored under UV-excitation
owing to formation of stable defects (color centers) with trapped
The 21 powdered metal oxide samples (Table 1) examined
in preliminary tests were of “high purity grade”. Three of these
were studied in some detail. The powdered monoclinic form of
ZrO2 (IREA) was produced from zirconium chloroxide; specific
3
+
+
12,19
- •
electrons (Zr , F , F)
centers).
or with trapped holes (O -like V
1
2,20
Color centers localized at the surface serve as long-
2
lived adsorption centers (i.e., post-adsorption centers) for
acceptor (e.g., O2) and donor (e.g., H2 and CH4) molecules,
surface area (BET with nitrogen gas) of the sample was ∼7 m
-
1
g . Powdered Sc2O3 (IREA) and optical material grade spinel
MgAl2O4 were obtained from the Vavilov State Optics Institute;
2
1
respectively.
2
-1
2
-1
Scandium oxide (Sc O ) is a wide band gap insulator (E )
BET specific surface areas are ∼9 m g and ∼14 m g ,
respectively. Average crystal size was ca. 1 µm. These three
specific metal oxides were analyzed for content of impurities
2
3
bg
7
6
.0 eV) with a cubic crystal lattice belonging to the Th space
group; lattice constant, a ) 9.845 Å. Among the wide band
-
5
gap metal oxides, Sc O is not as frequently used in photocata-
in their preparations: analyses for ZrO2 gave 6 × 10 % Fe,
and less than 5 × 10 % for Mn, Cu, Co, Cr, and Ti; for Sc2O3,
analyses gave 1 × 10 % Cu, 2 × 10 % for Cr, Mn, and Ni,
× 10 % Yb, 2 × 10 % for V, Er, and Nd, 3 × 10 % Fe,
2
3
-5
lytic studies as zirconia. Nonetheless, it is an active metal oxide
2
2
-
5
-5
in water photolysis in a gas/solid system. Photoadsorption of
O , H , and CH , and methane photooxidation to CO and H O,
-
4
-4
-4
-4
1
2
2
4
2
2
-
3
along with its conversion to ethane, ethylene, and other
and 1 × 10 % Y; for MgAl2O4, analyses gave 5 × 10 % Cr,
14,23
-
3
-3
hydrocarbons have also been established.
Photo- and radia-
3
× 10 % Mn, and 5 × 10 % Ca and Fe.
tion-induced color centers of the F- and V-type display complex
Any adventitious organic impurity and molecules adsorbed
spectra in the near-UV and visible spectral regions.1
4,24,25
on the surface of the metal oxide specimens were removed by
thermal pretreatment (T ) 800 K) first in an O2 atmosphere (P
Spinel, MgAl2O4, is a complex wide band gap (Ebg ) 7.5-
2
6,27
)
100 Pa) and then in vacuo. Reproduction of the original state
9.0 eV
) insulator with a cubic crystal lattice and spinel-
of the samples between experiments was achieved by heating
the samples in oxygen for ca. 1 h. Any experimental error in
the spectral measurements caused by the nonreproducibility of
the original state of the sample does not exceed about 5%.
type structure Fd3m; lattice constant, a ) 8.11 Å. Cation
vacancies, Vc, are the predominant type of intrinsic defects. In
addition, synthetic spinels are partly (eca. 20%) inversed,
2
+
whereby a fraction of the Mg cations occupy the positions of

9192 J. Phys. Chem. B, Vol. 103, No. 43, 1999
Emeline et al.
Al3 ions and in turn the Mg positions are occupied by Al3+
+
2+
2
8
cations. Such defects serve as hole and electron traps,
respectively. Among the several UV- and radiation-induced
centers, F and V centers have been identified in MgAl2O4.
Photoadsorption of O2, H2, and CH4 and its influence on the
2
8-30
29,31
photocoloration of spinel MgAl2O4 have also been explored.
Procedures. Powdered specimens were contained in a quartz
2
cell (path length, 3 mm; illuminated area, 7 cm ) connected to
a high-vacuum setup equipped with an oil-free pumping system;
-
7
the ultimate gas pressure in the reaction cell was ca. 10 Pa.
3
The volume of the reactor occupied by the gas was 40 cm . A
-
1
Pirani-type manometer (sensitivity, 18 mV Pa for O2 and 24
-
1
mV Pa for H2) was used to measure gas pressures in kinetic
studies. Analyses of gas composition were carried out using a
mass spectrometer model MX-7301. Irradiation of the solid
samples was accomplished with a 120 W high-pressure mercury
lamp model DRK-120 (MELZ). Cutoff filters from a standard
set of colored glass filters (Vavilov SOI) were used to select
the spectral regions for sample excitation.
Figure 1. Kinetic curves as P versus time for the photostimulated post-
2 2 3 2 3 2 2 3
adsorption of O on: (1) Yb O ; (2) Gd O ; (3) HfO ; (4) Sm O ; and
(
5) Y-stabilized ZrO .
2
Diffuse reflectance spectra (vs BaSO4) were recorded with a
Specord M-40 (Karl Zeiss, Jena, Germany) spectrophotometer
equipped with an integrating sphere assembly; data processing
was carried out using the Data Handling I package. In
preliminary test experiments, a specially designed integrating
sphere interfaced to a computer was employed; it permitted
measurements of diffuse reflectance coefficients integrated in
the visible spectral range between 400 and 800 nm (powdered
samples); accuracy of the measurements, ( 0.0001. Unless noted
otherwise, the moveable high-vacuum setup/quartz cell system
was moved as a unit between the position of irradiation and
thermal treatment of the sample to a position for recording
spectra; both were fixed with high precision in appropriate
positions. Thermostimulated luminescence was recorded with
a KSVU-12 (LOMO) spectral setup adapted to record emission
from solid powders in vacuo and in a gaseous atmosphere. The
heating rate of the specimens in the linear heating regime was
(
5) The effect of photostimulated post-adsorption of oxygen
was considered established for irradiated samples when oxygen
pressure changes under irradiation with visible light (λ > 400
nm for ca. 600 s) exceeded 0.01 Pa and the rate of photoad-
sorption of oxygen was at least 10% higher than for non pre-
irradiated samples. The extent of photobleaching of the UV-
induced color δRvac () Rvis - Rhν) and δRO () Rvis - Rhν)
2
was also measured; Rvis is the integrated diffuse reflectance
coefficient after irradiation with visible light in vacuo and in
oxygen, respectively (Table 1).
Results
UV pre-irradiation of the samples decreases the integrated
diffuse reflectance coefficients Ro (∆R varies from 0.0003 for
SiO2 to 0.071 for the 5.5%Y-doped ZrO2 specimen) as
evidenced by formation of photoinduced color centers for most
of the metal oxides examined (Table 1). Detailed data concern-
ing the changes in the diffuse reflectance characteristics and
the influence of photoadsorption of gases on these features were
-
1
maintained at 0.3 K s .
Test Experiments. Preliminary studies were conducted on
2
1 wide band gap metal oxides (Table 1) to diagnose the effect
of photoinduced spectral sensitization for a simple surface
photochemical process, namely photostimulated adsorption of
oxygen. This set of experiments was carried out using the
following uniform procedure:
12
reported earlier by Ryabchuck and Burukina. Photobleaching
of UV-induced color centers by visible light is typical for most
of the metal oxides examined; relative values of δRvac/∆R ratios
range from 0.02 for GeO2 to 0.45 for La2O3. However, visible
light irradiation decreases the integrated diffuse reflectance for
TiO2, SnO2, and for ZnO (δRvac < 0). The effect of photo-
stimulated post-adsorption (PhSPA) was observed for La O ,
(1) The integrated diffuse reflectance coefficient Ro was
measured for each metal oxide specimen after sample pretreat-
ment (see above).
(
2) Samples were irradiated in vacuo for ca. 200 s using all
the wavelengths of light from the mercury lamp, followed by a
0-min pause to permit the phosphorescence emission (typical
2
3
Sm O , Yb O , Cd O , HfO , ZrO , and for ZrO stabilized in
the cubic modification with Y O . Photostimulated adsorption
2 3
2
3
2
3
2
3
2
2
2
1
of some oxides) to decay completely. The integrated diffuse
(PhA) of O by visible light was not observed for MgO, BeO,
2
reflectance coefficient Rhν was also measured to determine ∆R
Sc O , Y O , SiO , GeO , γ-Al O , and MgAl O , neither before
nor after UV pre-irradiation of the specimens in vacuo. No
2
3
2
3
2
2
2
3
2
4
)
{Rhν - Ro} so as to characterize the extent of photocoloration
column ∆Rvac; Table 1).
3) After addition of oxygen into the reactor (PO ) 1.5 Pa),
(
difference in rates of O photoadsorption was detected between
2
(
nonirradiated and UV-irradiated samples of TiO2, whereas UV
2
the system was exposed to the dark for about 10 min until post-
adsorption of oxygen (also typical of some metal oxides) was
complete. Subsequently, a given system was irradiated in the
visible region using a 400-nm cutoff filter and the kinetics of
photostimulated adsorption of oxygen determined (see Figure
irradiation of ZnO caused a negative effect, i.e., decreased the
rate of photoadsorption under visible light irradiation.
Thermostimulated adsorption of O2 takes place on UV-
colored ZrO , Sm O , and Yb O , but was not detected on
2
2
3
2
3
nonpreirradiated samples. Maximal rates of thermostimulated
1
).
2
O adsorption in the linear heating regime of the samples (see
(
4) A set of control experiments were also carried out for
Experimental Section) were obtained at 420, 600, and 630 K
for ZrO2, Sm2O3, and Yb2O3, respectively.
Photostimulated post-adsorption of O2 was not found for a
group of metal oxides (from BeO to SiO2, upper part in Table
nonirradiated samples to establish the existence and the extent
of photoadsorption of oxygen on a given metal oxide stimulated
by visible light (λ g 400 nm).

Wide Band Gap Metal-Oxide Solids
J. Phys. Chem. B, Vol. 103, No. 43, 1999 9193
2
Figure 2. Diffuse reflectance spectra of powdered ZrO : (1)
Figure 3. Dependencies of (1) the kinetic parameter ∆P
∞
, and of (2)
on ZrO on
nonirradiated R
o
; (2) irradiated in vacuo Rhν; (3) absorption spectrum
- Rhν of photoinduced color centers.
the initial rate of photostimulated post-adsorption of O
2
2
given as ∆R ) R
o
the extent of photoinduced coloration ∆R.
1
) whose band-gap energies Ebg are greater than 8 eV. Except
for MgO, all these specimens are poorly photocolored (∆R <
.015) by the UV light used (λ g 200 nm; Ehν e 6.2 eV). Only
color centers are formed in the bulk of ZrO2 by UV pre-
irradiation of the sample, the spectral limits of photostimulated
adsorption are red-shifted to 580 nm for H2 and CH4 and to
wavelengths above 800 nm for O2; these correspond to the
positions of the absorption bands of V- and F-type color centers,
respectively. Photostimulated adsorption of CH4 followed by
secondary chemical reactions leads to formation of higher
hydrocarbons: C2H6, C2H4, C3H8, along with CO2 and H2O.
The mass spectral analyses of the gas phase during the
photostimulated post-adsorption of CH4, and the temperature
desorption spectra recorded after irradiation of ZrO2, demon-
strate the evolution of C2H6 into the gas phase and formation
of C2H6, C2H4, and other reaction products on the specimen’s
surface. Identical reaction products were reported earlier when
photoadsorption of CH4 on ZrO2 and other wide band gap metal
0
intrinsic and extrinsic defects are photoexcited in these oxides.
On the other hand, the metal oxides from TiO2 to ZnO with
Ebg around 3 eV also displayed no photostimulated post-
adsorption of O2. The remaining oxides from Sc2O3 to the
Y-stabilized ZrO2 have Ebg around 5 eV; they show moderate
to strong coloration (∆R > 0.02) and photobleaching effects
(
δRvac/∆R > 0.15). Oxygen has an influence on the photo-
bleaching of these oxides (positive: increase of photobleaching
δRvac < δRO ; negative: δRvac > δRO ). However, there seem
to be no correlations between the rates of photostimulated post-
adsorption {dP/dt (PhSPA)} and the remaining parameters (e.g.,
∆
2
2
R) listed in Table 1.
2
3
Clearly, results from the preliminary studies show that UV-
oxides was examined under UV-light irradiation.
induced photocoloration of metal oxides is a necessary but not
a sufficient condition to observe either photostimulated or
thermostimulated post-adsorption of O2 on the metal oxide
surfaces. Except for ZnO, all the metal oxides tested can be
classed into two groups as regards the effect of photostimulated
post-adsorption of O2. Note that the thermostimulated post-
adsorption effect for O2 was also detected for metal oxides which
are otherwise also active in the photostimulated post-adsorption
process.
The kinetics of photostimulated photoadsorption of O2 and
H2 on ZrO2 follow the Solonitsyn approximation (eq 1):
t
kFt
kFt + 1
∆
P(t) ) ∆P∞
) ∆P∞
(1)
{
}
{
}
t + τ
where ∆P(t) describes the temporal change in gas pressure
(corresponds to the amount of photoadsorbed gas molecules)
during irradiation for a time t, ∆P∞ denotes the maximal
coverage of photoadsorbed molecules on the surface at some
given conditions when the time of irradiation approaches infinity
To better understand the mechanism and the reasons underly-
ing the photo- and thermo-stimulated processes on the surface
of UV-colored metal oxides, we selected ZrO2 (strong effect
of photostimulated post-adsorption of O2) from the first group
and Sc2O3 and MgAl2O4 (no effect found) from the second
group of metal oxides for a more detailed exploration of these
processes.
(
t f ∞), and τ is the time of half-decay of the pressure during
-1
the photoprocess; τ ) (kF) , and F is the photon flow (photons
cm s ). The initial rate of photoadsorption is described by
-2 -1
eq 2:
∆
P∞
dP
dt
)
) ∆P kF
(2)
Zirconia, ZrO2. The spectral limit of surface photochemical
processes of photoadsorption of H2, CH4, and O2, as well as
the photoinduced coloration (formation of defects) of ZrO2 is
positioned at about 410 nm. UV irradiation of powdered ZrO2
in vacuo with light of wavelength λ < 410 nm caused the
{
}tf0
∞
τ
Note that the linear dependence of the initial rates of
photostimulated post-adsorption on light intensity (photon flow)
was experimentally confirmed in special tests for all the metal
appearance of an additional broad absorption band (Figure 2).
oxides examined. The dependencies of ∆P and the initial rate
∞
As evidenced from the results of previous studies,1
3,19,20
this
of photoadsorption {dP/dt (PhA); Table 1} of O on powdered
2
band is formed by an overlap of absorption bands of electron
ZrO on the number of photoinduced color centers, represented
by ∆R, under photoexcitation with 436-nm light are illustrated
in Figure 3. The linear dependence of the initial rate on the
absorbance of the color centers ∆R infers that the quantum yield
2
+
3+
color centers F, F , and Zr with maxima at 620, 390, and
80 nm, respectively, and of hole color centers of the V-type
showing maxima at ca. 410 and 270 nm. When photoinduced
2

9194 J. Phys. Chem. B, Vol. 103, No. 43, 1999
Emeline et al.
Figure 4. Kinetic dependencies of the photobleaching of color centers
in ZrO by the actinic light at 436 nm (1) in vacuo, (2) in O , and (3)
in H
Figure 5. Absorption spectrum ∆R of photoinduced color centers in
powdered Sc O .
2 3
2
2
2
.
of photoadsorption of O2 (ΦO ) is independent of ∆R for the
2
whole range of coloration. Herein, we define1
4,24
the quantum
yield of photostimulated post-adsorption in relation to the rate
by which photons are absorbed by the UV-induced color centers
as
dN/dt
ΦO2
)
(3)
∆
R F
where the rate dN/dt ∼ dP/dt is expressed by the number of
adsorbed molecules per second. Under our experimental condi-
tions, the quantum yield of photostimulated post-adsorption of
O2 at λ ) 436 nm is ΦO2 ) 0.16 ( 0.02, whereas the
corresponding quantum yield for H2 at this wavelength is ΦH2
)
0.057 ( 0.007.
Irradiation of UV-colored ZrO2 in vacuo or in a gaseous
atmosphere with visible light causes photobleaching of the color
centers that is typical of many other metal oxides (see above).
The relevant kinetics in vacuo in O2 and in H2 are displayed in
Figure 4. The presence of O2 increases the rate of photobleach-
ing of electron color centers in ZrO2, whereas the presence of
H2 decreases the rate. Using the data for the calibration of the
absorbance ∆R of F-type color centers in ZrO2 reported by
Burukina and co-workers,17 we estimate the quantum yield of
photobleaching of F centers in vacuo, ΦF(vac), relative to the
number of photons absorbed by the color centers (see above)
to be ΦF(vac) ) 0.32 ( 0.03 at 436 nm, whereas in the presence
Figure 6. Kinetic dependencies of the photobleaching of color centers
in Sc by the actinic light at 436 nm (1) in vacuo and (2) in H
O
2 3
2
.
surface of powdered Sc2O3 is located at ca. 320 nm. UV-
irradiated powdered scandia shows a broad absorption band at
λmax 430 nm which consists of overlapping bands of hole color
centers and (probably also) electron F-type color centers with
λmax at 330 nm (Figure 5). The spectral limits of photostimulated
interaction of gases with the surface of the preirradiated sample
is red-shifted to 360 nm for O2 and to 560 nm for H2 and CH4.
The kinetics of photostimulated post-adsorption of H2 and O2
on excitation of the photoinduced color centers are described
by eq 1. The parameter ∆P∞ scales linearly with the extent of
coloration ∆R, whereas the dependence of the initial rate of
photoadsorption of H2 on ∆R tends to be sublinear, indicating
that the quantum yield of photostimulated adsorption of H2
of O2 and in the presence of H2 they are ΦF(O ) ) 0.52 ( 0.04
2
and ΦF(H ) ) 0.22 ( 0.04, respectively.
2
The linear heating regime of a UV-colored ZrO2 powdered
specimen in the presence of O2 leads to thermostimulated
adsorption of O2 with a maximum rate obtained at 420 K, which
also corresponds to the thermostimulated green luminescence
(TSL) maximum of ZrO2. The estimated activation energy of
1
4
decreases with increasing ∆R.
this TSL is Ea ) 1.1 ( 0.1 eV. No thermostimulated adsorption
was detected on noncolored samples.
Irradiation of UV-colored Sc2O3 with visible light leads to
photobleaching of the color centers, the rate of which increases
in the presence of H2 and does not change in the presence of
O2 relative to the rate of the photoprocess in vacuo (see Figure
6). The linear dependence between the concentration of adsorbed
H2 and the difference between the number of color centers
photobleached in H2 and in vacuo has been confirmed (see
Figure 7). Heating the UV-colored powdered scandia in the
linear regime causes adsorption of H2 which correlates with the
Scandium Oxide, Sc2O3. Scandium oxide is representative
of the second group of metal oxides which displayed no spectral
sensitization for the photoadsorption of O2 (see above). Results
of detailed studies on the interconnection between photoinduced
formation and destruction of defects with surface photochemical
14,24
processes on Sc2O3 specimens have been reported elsewhere.
The spectral limit of photoinduced defect formation in the
bulk and photostimulated adsorption of O2, H2, and CH4 on the

Wide Band Gap Metal-Oxide Solids
J. Phys. Chem. B, Vol. 103, No. 43, 1999 9195
Figure 9. Temperature dependence of the (1) thermoluminescence and
2 4
(2) thermostimulated post-adsorption of hydrogen on MgAl O .
Figure 7. Dependence of the amount of adsorbed H , ∆P, during the
2
photostimulated post-adsorption on the number of bleached color centers
δR.
of a spinel sample in vacuo increases the rate of photostimulated
adsorption of O2 by up to ca. 20%. Irradiation of UV-colored
samples with light in the range 300 nm < λ < 400 nm affects
photobleaching of hole color centers. A decrease in the
coloration of spinel was also observed during thermal treatment
of the colored sample in the linear heating regime. Concomi-
tantly, the thermostimulated luminescence and the thermostimu-
lated adsorption of H2 were detected for colored spinel (see
Figure 9); both dependencies correlate with each other. Neither
thermostimulated luminescence nor thermostimulated adsorption
were observed for noncolored spinel specimens.
Discussion
Most metal oxides belong to the class of photoresistant solids.
The only mechanism of defect formation caused by light in such
solids is photocarrier trapping by preexisting intrinsic and
extrinsic lattice defects to yield so-called color centers. An
example of this kind of defect formation is trapping of
photogenerated electrons by anion vacancies to give F-type color
centers. A thorough discussion of the mechanisms of processes
which lead to formation of defects is given elsewhere.7
,32-36
Figure 8. Absorption spectrum ∆R of photoinduced color centers in
2 4
powdered MgAl O .
Photoinduced centers differ from each other by the energy
distance, Ed, between the level of the trapped carrier within the
energy band gap of the insulator or semiconductor and the
bottom of the conduction band for electron centers, and the top
thermoannealing of hole color centers. No adsorption of H2
occurred on noncolored samples.14
Secondary reactions that follow the photostimulated adsorp-
tion of CH4 during photoexcitation of hole color centers with
visible light convert CH4 to higher hydrocarbons such as C2H6,
C2H4, and C3H8, among others.
Spinel, MgAl2O4. Spinel belongs to the second group of
metal oxides examined (Table 1). It displays no photochemical
activity toward O2 when spinel is irradiated with visible light,
neither before nor after UV pre-irradiation in vacuo. Nonethe-
3
5,37,38
of the valence band for hole centers.
From this point of
view, a given photoinduced center can be described as a shallow
center (or trap) if the energy of the trapped carrier is comparable
to the energy of thermal excitation of the crystal lattice (i.e., Ed
∼
kT), or as a deep energy center if Ed . kT. The difference
between the various types of photoinduced centers is determined
by the dominant decay path back to the initial state (without
trapped carrier).
less, spinel specimens are photoactive with respect to UV-
stimulated adsorption of H2, O2, CH4, and CO.31 The spectral
In general, any photoinduced center can lose a carrier as a
result of (i) thermal ionization with a thermal quenching
probability given by qth; (ii) recombination with a charge carrier
of the opposite sign for which the recombination probability is
qrec; (iii) photoionization caused by incident photons, qph; and
limit of these surface photochemical processes and photoinduced
formation of defects occurs at ca. 260 nm when the specimen
was not preirradiated. UV-irradiated powdered spinel (in vacuo)
displays a broad absorption band at λmax ∼ 390 nm, attributable
to formation of different photoinduced hole color centers (Figure
(iv) tunneling recombination of adjacent hole and electron
2
8,30
centers. The role of the latter process in heterogeneous systems
8
).
The spectral limit of photoadsorption of H2 and CH4 is
3
9-41
has been treated extensively elsewhere.
red-shifted from 260 nm to ca. 410 nm after UV-induced
coloration of the powdered spinel specimen. No red shift was
detected for photoadsorption of O2 because absorption by
electron color centers occurs at λ < 260 nm. Pre UV irradiation
For the simplest case of centers with one trapped carrier, the
kinetics of formation of centers typically follow an exponential
growth and approach the steady-state concentration of photo-

9196 J. Phys. Chem. B, Vol. 103, No. 43, 1999
Emeline et al.
induced defects Nd given by36
MgAl2O4 specimens; the results suggest the following simplified
mechanism:
p
N ) No
(4)
d
F + hν f V + e
(5a)
(5b)
(5c)
(5d)
{
}
a
p + q + q + qph
th
rec
V + hν f V + h
c
where No is the number of preexisting defects of a given type,
p is the probability of photocarrier trapping defined, as are the
corresponding quenching probabilities q, as the number of events
per unit time. For shallow traps, the thermal ionization usually
dominates, i.e., qth . qrec + qph. By contrast, thermal ionization
can be neglected for deep energy centers, i.e., qth , qrec + qph.
In the latter case, centers having a high ability to capture a carrier
of the opposite sign, qrec . p, can serve as effective recombina-
tion centers. These centers do not accumulate in considerable
quantity in illuminated crystals since Nd ) Nrec , No (eq 4). At
the same time, recombination centers are mainly responsible
for the steady-state concentrations of photogenerated carriers
in wide band gap solids because of the inefficient direct band-
to-band recombination compared with recombination involving
deep energy centers.35 Contrary to recombination centers, deep
energy centers (qrec ∼ p) accumulate in sufficient amount in
solids under illumination. Such defects are the color centers
because they usually have absorption bands located in the UV
and visible regions and determine the lasting stable color of
pre-illuminated, initially colorless wide band gap solids. Pho-
toionization of color centers by the actinic light may, to some
extent, reduce the color saturation level under illumination when
qrec ∼ qph.
F + heat f V + e
a
V + heat f V + h
c
V + e f F (+ hν)
(6a)
(6b)
a
V + h f V (+ hν)
c
V + e f V (+ hν)
(7a)
(7b)
c
F + h f V (+ hν)
a
-
S + e f S
(8a)
(8b)
+
S + h f S
where stages 5a, 5b and 5c, 5d represent the photo- and thermo-
ionization (photobleaching and thermoannealing), respectively,
of electron (F-type) and hole (V-type) color centers; stages 6a
and 6b describe trapping of electrons and holes, respectively,
by preexisting defects (e.g., anion Va and cation Vc vacancies)
to form electron (F) and hole (V) color centers; note that these
processes may be accompained by luminescence hν. Stages 7a
and 7b represent the recombination of free electrons and holes,
respectively, with trapped carriers of opposite sign localized at
UV-induced defects that can also yield luminescence; however,
recombination occurs primarily through recombination centers
rather than through color centers (see above). Finally, stages
Results of test experiments demonstrate that the red shift of
the spectral limits of photochemical processes on the surface
of UV-preirradiated metal oxides may indeed be a general
phenomenon. In fact, the photostimulated post-adsorption of O2
occurs for 9 of the 21 metal oxides tested. As evidenced by the
results, the necessary condition to observe this effect is
photoinduced formation of defects (color centers) during the
preliminary UV irradiation of the samples. All the metal oxides
investigated belong to the class of photoresistant solids, in which
trapping of photocarriers by preexisting surface and bulk defects
is the dominant mechanism of formation of color centers in these
solids. Evidently, photobleaching of UV-induced color centers
by visible light is identified with the effect of photostimulated
post-adsorption of O2. Notwithstanding that photobleaching is
typical of all the metal oxides tested, the photostimulated post-
adsorption event was detected only for 9 of the 21 samples.
Moreover, ZnO displays a negative effect, i.e., the rate of the
photoprocess decreases. To explain this observation, we note
that up to 100% of the photoinduced electron active centers on
the ZnO surface are stable at ambient temperatures with a
8
a and 8b represent the first steps of surface photochemical
reactions as a process of carrier trapping by either the surface
defects to form surface active centers, or by the preadsorbed
molecules to produce either anion radicals (electron trapping)
or cation radicals (hole trapping), respectively. We denote all
+
3+
types of electron color centers in metal oxides (F, F , and Zr )
as F-centers, whereas all the types of hole color centers (V,
-
2-
2+
-
V , V , and {[Mg ]Al O }) are described as V-centers.
Because of its acceptor character, O2 reacts predominantly
-
•
-•
with electrons to give such anion radicals as O2 and O ; the
-
•
adsorption complex O3 formed by the interaction of O2 with
-
•
a surface hole OS is unstable at ambient temperatures on most
metal oxides and has been detected only at low tempera-
2
3,42,44
tures.
Thus to observe photostimulated adsorption of O2,
free electrons must be generated which can be achieved by
visible light excitation of electron color centers. That is, UV-
induced coloration in the visible spectral region must be due to
electron color centers. This is the case for ZrO2 where UV
irradiation causes the appearance of a broad absorption band in
the visible region formed by an overlap of absorption bands of
42
lifetime as long as (at least) a few days. We infer that during
UV irradiation of ZnO in vacuo many of the surface active
centers are created in a manner analogous to when O2 is present.
Exposing a UV-pre-irradiated sample of ZnO to O2 (in the
dark) causes all the electron surface centers to be blocked by
adsorbed O2 molecules, thereby inhibiting additional electron
active centers from being created on the ZnO surface. This leads
to a decrease in the rate of photostimulated adsorption of O2
by visible light after UV pre-irradiation.43 However, the fraction
of long-lived surface active centers does not exceed ∼10% for
the majority of other metal oxides; UV pre-irradiation of these
metal oxides in vacuo does not decrease the rate of surface
photochemical processes. Reasons as to why photostimulated
post-adsorption of O2 was not universally observed for the metal
oxides of the second group necessitate certain considerations
that emerge from detailed studies on the ZrO2, Sc2O3, and
+
F and F centers (see Experimental Section). Excitation of these
centers with visible light generates free electrons (stage 5a)
which in turn produce electron active centers (stage 8a) that
react with O2 molecules. On the basis of this mechanism and
the steady-state approach for the concentration of carriers at
sufficiently high O2 pressures (the case in our experiments),the
expression for the initial rate of photoadsorption of O2 is given
4
5
by
d[S-]
k
k
8a [F] [S] F
5a
)
(9)
dt
k [V ] + k [V] + k [S]
6a
a
7a
8a

Wide Band Gap Metal-Oxide Solids
J. Phys. Chem. B, Vol. 103, No. 43, 1999 9197
where k5a, k6a, k7a, k8a are the rate constants of the corresponding
steps in the mechanism (5a-8b), and F is the photon flow of
irradiation. It is worth noting that the initial concentration of F
and V color centers and the concentration of anion vacancies,
and
k5b [F] F
[h] )
in hydrogen (17)
k [V ] + k [F] + k [S]
6
b
c
7b
8b
[Va], depend on the time of UV pre-irradiation and on the
relationship (eq 10):
-
+
We have neglected the concentrations of S and S centers
in vacuo since their lifetime during irradiation with visible light
[
V ] ) [V ] - [F]
(10)
-2
-2
a
ao
does not exceed ca. 10 s (i.e., τ e 10 s) and the fraction of
these centers produced in vacuo does not exceed ∼10% of those
where [Vao] is the concentration before UV-induced coloration
21
created in the presence of either O2 or H2. Combination of
of the metal oxide specimen. Considering that [V] ) (const)-
eqs 14-17 yields:
[F] from the charge conservation law (see discussions in refs
2
7 and 28), and assuming that k6a [Vao] . {k7a (const) - k6a}-
-
d[F]
d[F]
-d[S ]
dt
[F], eq 9 reduces to
-
-
)
)
(18)
(19)
( dt
)
vac
vac
( dt
)
O2
H2
d[S-]
dt
+
)
(const′) [F]F
(11)
d[F]
d[F]
-d[S ]
( dt
)
( dt
)
dt
where k5a k8a/{k6a[Va] + k8a[S]} ) (const′) such that the rate
scales linearly with the absorption (i.e., concentration) of F color
centers. (Note that [S] is constant since in the proposed
mechanism it is assumed that pre-irradiation does not affect this
concentration term).
so that
and
ΦF(O2) - ΦF(vac) ) ΦO2
(20)
(21)
To the extent that the rate of photon absorption is given by
ΦF(vac) - ΦF(H2) ) 2 ΦH2
Fabs ) σ [F]F
(12)
where σ is the cross-section of photon absorption by F centers,
the quantum yield of photoreaction will then be given by
which correspond to the experimental quantum yields of
photobleaching of F centers in vacuo (ΦF(vac) ) 0.32 ( 0.03),
in oxygen (ΦF(O ) ) 0.52 ( 0.04, ΦO ) 0.16 ( 0.02), and in
2
2
(
const′) [F] F
σ [F] F
hydrogen (ΦF(H ) ) 0.22 ( 0.03, 2ΦH ) 0.11 ( 0.02)
2
2
Φ )
) (const′′)
(13)
considering a dissociative adsorption for H2 on surface hole
centers to give twice the value of the quantum yield of
+
46
Thus, Φ is independent of the concentration of color centers,
photogeneration of S surface active centers (eq 21).
-
1
and since τ scales as F (from eq 1), the kinetic parameter
For Sc2O3 and MgAl2O4, representatives of the second group
of metal oxides, the UV-induced absorption in the visible
∆
∝
P∞ is proportional to the concentration of F centers, i.e., ∆P∞
[F] (eqs 2 and 11). The experimental results obtained for
14,25,28,30
spectral range is due to hole color centers.
Consequently,
ZrO2 are fully described by eqs 9, 11, and 13 (see Results and
Figure 3).
photoexcitation of these centers at λ > 400 nm causes the
photostimulated adsorption of donor molecules (H2 and CH4)
and does not lead to adsorption of acceptor molecules (oxygen).
We infer this to be the major cause for the lack of photostimu-
lated adsorption of O2 on other metal oxides of the second group.
Indeed, absorption bands in the visible spectral range for
The tail end of the complex absorption band of UV-induced
hole color centers in ZrO2 occurs in the visible spectral region
causing the spectral limit of the photostimulated adsorption of
H2 to be red-shifted to ca. 580 nm. Irradiation of UV-colored
ZrO2 by visible light at λ ) 436 nm leads to photoexcitation of
electron and hole color centers (stages 5a and 5b). The rate of
photobleaching of F centers can be written as:
4
7-49
50
47-49
14,25
28,30
MgO,
Al2O3, BeO,
Sc2O3,
and MgAl2O4
are
those of hole color centers; absorption bands of electron color
centers are seen in the UV spectral region. Consequently,
electron color centers are not photoexcited under visible light
irradiation conditions used herein (see Results), and photo-
stimulated post-adsorption of donor molecules is to be expected
for these metal oxides under photoexcitation with visible light.
Photostimulated post-adsorption of acceptor molecules, e.g. O2,
is anticipated for these metal oxides when illuminated with UV
light.
d[F]
)
-k [F] F + k [V ] [e] - k [F] [h] (14)
5a 6a a 7b
dt
with the steady-state concentrations of carriers given by
k [F] F
5a
[
e] )
in vacuo
(15a)
(15b)
The red shift of the spectral limit to 360 nm observed for the
photoadsorption of oxygen on Sc2O3 corresponds to the absorp-
tion region of photoinduced F centers (λmax ∼ 350 nm). In spinel,
the absorption bands of UV-induced F color centers (λmax, 240
and 260 nm) coincide with the spectral region of photoactive
absorption of nonirradiated samples (λ e 260 nm). There is no
red-shift of the spectral limit of the photoactivity of this metal
oxide within the experimental accuracy of our experiments,
owing to the restriction imposed by the sharp cutoff of the
optical filters used. However, photoinduced formation of
electron color centers does lead to an increase in the rate of
k [V ] + k [V]
6
a
a
7a
and
k [F] F
5b
[
h] )
in vacuo
k [V ] + k [F]
6
b
c
7b
whereas in the presence of gases, they are given by
k [F] F
5a
[e] )
in oxygen (16)
k [V ] + k [V] + k [S]
48
6
a
a
7a
8a
photoadsorption of O2. Kuznetsov et al have described a

9198 J. Phys. Chem. B, Vol. 103, No. 43, 1999
Emeline et al.
similar behavior for the photoadsorption of O2 on BeO, which
they associated with a redistribution of carriers between
preexisting F and F centers under UV excitation.
adsorption observed for the ZrO2/O2 and MgAl2O4/H2 systems
demonstrates the validity of the proposed mechanism described
+
52,53
by stages 5a to 8b. In fact, it was established earlier
that
Photoexcitation of hole color centers by visible light in Sc2O3
and MgAl2O4 (stage 5b) leads to photostimulated post-adsorp-
tion of donor molecules (H2, CH4, and CO) and to the
photobleaching of coloration. By analogy with the effect of
photostimulated post-adsorption of O2 on ZrO2 on the rate of
photobleaching of F color centers (see above), we can write a
similar expression to reflect the influence of photostimulated
post-adsorption of donor molecules on the rate of photobleaching
of hole color centers in Sc2O3 and MgAl2O4; thus,
the thermostimulated luminescence of ZrO2 originates from the
thermoionization of F-type color centers (stage 5c), that is, of
anion vacancies with trapped electrons. In our experiments, this
thermostimulated luminescence correlates with the adsorption
of O2. In spinel, the thermostimulated luminescence emanates
from ionization of V-type hole color centers (stage 5d) followed
by retrapping or recombination of the free holes with the
54
produced cation vacancies (stage 6b). Appearance of free holes
on the metal-oxide surface induces the thermostimulated post-
adsorption of H2 (Figure 9 and Results). Thermostimulated
adsorption of H2 on Sc2O3 also results from the thermodestruc-
d[V]
d[V]
d[S+]
dt
-
)
(22)
14
( dt
)
( dt
)
tion of hole color centers. By analogy, we infer that the
vac
H2
thermostimulated adsorption of O2 on the Sm2O3 and Yb2O3
surfaces (Table 1) is a result of the thermoannealing of electron
color centers in these metal oxides.
or after integration
+
∆
[V] ) [S ]
(23)
Concluding Remarks
where ∆[V] denotes the difference between the number of hole
The data reported here emphasize the notion that the influence
of spectral sensitization on surface photochemical processes by
preliminary UV irradiation of the metal oxide specimens is
encountered not only for the photostimulated adsorption of
molecules, but also, taking CH4 as an example, in the multistep
chemical reaction of CH4 oxidation and its conversion to higher
hydrocarbons. The results also illustrate the relevance of the
present studies in solving the problem of spectral sensitization
of wide band gap solids in heterogeneous photocatalytic
applications.
color centers destroyed in the presence of H2 and in vacuo, and
+
51
[
S ] is the number of adsorbed molecules or species. Equations
2
2 and 23 are in accord with experimental results obtained in
this study for Sc2O3 (see Results and Figures 6 and 7). Taking
into account that one molecule of H2 occupies one surface hole
+
14
center, S , on the surface of scandia, the experimental linear
correlation between δR and ∆P of H2 (Figure 7) and with eq
2
3 we estimate the number of hole color centers (V) produced
in our experiments by UV irradiation of Sc2O3 to be [V] )
1
7
15
(
0.9 × 10 )∆R ) 6.1 × 10 centers after UV coloration (∆R
51
was taken from Figure 5 at λ ) 430 nm). If we further assume
that during the UV irradiation only 5 to 10 layers of the
Acknowledgment. We are grateful to the Natural Science
and Engineering Research Council (N.S.E.R.C.) of Canada for
support. A.V.E. thanks N.S.E.R.C. for a North Atlantic Treaty
Organization Science Fellowship for 1997 and 1998. We also
thank Dr. A. A. Basov, Dr. G. K. Kuzmin, and N. S. Andreev
of the University of St. Petersburg for designing the setup used
in the measurement of the total reflectance.
-4
microcrystals, average crystal size ca. 10 cm, become colored
because of strong intrinsic absorption, the concentration of
V-type photoinduced color centers in Sc2O3 are ca. (1-3) ×
18
-3
1
0
cm which is the typical defect concentration in com-
mercial solid specimens. Knowledge of the number of hole color
centers in colored Sc2O3 now affords an estimate of the quantum
yield of photobleaching of these centers: Φ ) 0.022 ( 0.002
in vacuo and 0.046 ( 0.003 in a H2 or CO atmosphere. The
quantum yield of photostimulated adsorption of H2 and CO on
colored scandia is 0.023 ( 0.002,14 so that the quantum yield
of photobleaching of hole color centers in H2 (or CO) reflects
the sum of the quantum yield of photobleaching of coloration
in vacuo and the quantum yield of surface photochemical
processes. It is noteworthy that two V-type color centers in
Sc2O3 are photobleached on adsorption of one H2 molecule on
the surface (ratio of quantum yields, 0.046/0.023 ) 2; e.g.,
References and Notes
(
1) Photocatalysis. - Fundamentals and Applications; Serpone, N.;
Pelizzetti, E., Eds.; John Wiley & Sons: New York, 1989.
(
2) Semiconductor Nanoclusters - Physical, Chemical, and Catalytic
Aspects, Vol. 103 (Studies in Surface Science and Catalysis); Kamat, P.
V.; Meisel, D., Eds.; Elsevier: Amsterdam, 1997.
(3) Miller, R. J. D.; McLendon, G. L.; Nozik, A. J.; Schmickler, W.;
Willig, F. Surface Electron Transfer Processes; VCH Publishers: New
York, 1995.
(4) Kisel’ov, V. Th.; Krilov, B. O. Electronic Phenomena in Adsorption
and Catalysis at Semicondactors and Insulators; Nauka: Moscow, Russia,
1
979.
5) Baru, V. G.; Volkenstein, Th. The Effect of Irradiation on Surface
Properties of Semiconductors; Nauka: Moscow, Russia, 1978.
6) Photocatalytic Transformation of Solar Energy. Heterogeneous,
+
dissociation of H2 into two hydrogen atoms to yield 2 H ),
(
whereas the photostimulated post-adsorption of an acceptor O2
molecule on the ZrO2 surface implicates three F centers
destroyed (quantum yield ratio, 0.52/0.16 ≈ 3 for oxygen to
(
Homogeneous and Organized Molecular Structures; Zamaraev, K. I.;
Parmon, V. N., Eds. Nauka: Novosibirsk, Russia, 1991.
(7) Luschik, Ch. B.; Luschik, A. Ch. Decay of Electronic Excitation
with Formation of Defects in Solids; Nauka: Novosibirsk, Russia, 1989.
-
•
-•
yield three species, namely one O2 and two O ions).
Destruction of color centers can also be achieved by ther-
moionization (stages 5c and 5d) yielding free charge carriers.
Trapping and recombination of these carriers is accompanied
by (thermostimulated) luminescence (stages 6a, 6b, 7a, and 7b).
At the same time, the carriers can migrate to the solid’s surface
and take part in surface chemical processes (stages 8a and 8b).
Obviously, thermoannealing of hole color centers would lead
to adsorption of donor molecules, whereas thermoionization of
electron color centers would lead to surface reactions with
acceptor molecules. Experimental correlation between the
thermostimulated luminescence and the thermostimulated gas
(
8) Hersh, H. N. Phys. ReV. 1966, 148, 928.
(9) Pooley, D. Proc. Phys. Soc. 1966, 87, 257.
(
10) Vitols, I. K. IzV. Akad. Nauk. USSR, Ser. Fiz. 1966, 30, 564.
(11) Bradford, J. N.; Williams, R. T.; Faust, W. L. Phys. Lett. 1975,
5, 300.
3
(12) Ryabchuk, V. K.; Burukina, G. V. Fiz. Khim. 1991, 65, 1621.
(13) Emeline, A. V.; Kataeva, G. V.; Litke, A. S.; Rudakova, A. V.;
Ryabchuk, V. K.; Serpone, N. Langmuir 1998, 14, 5011.
(14) Emeline, A. V.; Petrova, S. V.; Ryabchuk, V. K.; Serpone, N. Chem.
Mater. 1998, 10, 3484.
(15) Ryabchuk, V. K.; Basov, L. L.; Solonitzyn, Yu. P. Usp. Photoniki,
Iss. 0.7.; Vilesov, Th. I., Ed.; LGU (Leningrad State University), 1980, p
3.

Wide Band Gap Metal-Oxide Solids
J. Phys. Chem. B, Vol. 103, No. 43, 1999 9199
(
16) Pichat, P.; Herrmann, J.-M.; Disdier, J.; Mozzanega, M.-N. J. Phys.
Chem. 1979, 83, 3122.
17) Burukina, G. V.; Vitkovsky, G. E.; Ryabchuk, V. K. Vestn. LGU
; LGU (Leningrad State University), 1990, 25, 93.
18) Strekalovsky, V. N.; Polejaev, Yu. M.; Palguev, S. V. Oxides with
Extrinsic Disorder; Nauka: Moscow, Russia; 1987.
solely by their lifetime. In most metal oxides the lifetime is rather short,
ca. 1-10 ms, such that once irradiation is stopped there will no longer be
any surface active centers after a few tens of milliseconds because of their
decay. In the case of ZnO, however, the lifetime of the active centers is at
least a few days (perhaps because the energy levels of these centers are
sufficiently deep that no thermal ionization takes place and because the
probability of recombination through these centers is very low). Conse-
quently, UV irradiation of ZnO in vacuo generates a maximal number of
active centers under our conditions, and once irradiation is stopped
introduction of O2 (in the dark) leads to reaction with these centers and
blocks these surface sites. If the specimen is now irradiated again, no new
active sites can be generated and the reaction rate for the photostimulated
adsorption of O2 will consequently decrease sharply. Clearly the factor that
contrasts ZnO with other metal oxides of Table 1 is the lifetime of the
surface active centers produced by UV irradiation.
(
4
(
(
(
19) Mikhailov, M. M.; Kuznetsov, N. Y. Inorg. Mater. 1988, 24, 656.
20) Mikhailov, M. M.; Dvoretsky, M. I.;. Kuznetsov, N. Y. Inorg.
Mater. 1984, 20, 449.
21) Emeline, A. V.; Rudakova, A. V.; Ryabchuk, V. K.; Serpone, N.
J. Phys. Chem. B 1998, 102, 10906.
22) Basov, L. L.; Efimov, Yu. P.; Solonitsyn, Yu. P. Uspechi Fotoniki,
Iss. 4; Vilesov Th. I., Ed.; LGU (Leningrad State University), 1974, 12.
23) Kotova, O. B. Ph.D. Thesis, Leningrad State University, Leningrad,
Soviet Union, 1986.
24) Emeline, A. V.; Lobyntseva, E. V.; Ryabchuk, V. K.; Serpone, N.
J. Phys. Chem. B 1999, 103, 1325.
25) Abdurazakov, A.; Antonov, V. A.; Arsen’ev, P. A.; Bagdasarov,
H. S.; Kevorkov A. N. Inorg. Mater. 1991, 27, 795.
(
(
(
(
(44) Che, M.; Tench, A. J. AdV. Catal. 1982, 31, 77.
(45) The concentration of F color centers in eq 9 has a distinct meaning
depending on whether it occurs in the numerator or in the denominator
because of the dual role of F centers. In the numerator, [F] denotes the
concentration of centers that absorb light (stage 5a), whereas [F] (or [V] )
const [F]) in the denominator represents the concentration of F color centers
that act as recombination centers. Consequently, under our proposed
conditions, observation of a linear dependence of the reaction rate on the
concentration of color centers infers that the efficiency of recombination
k7a through color centers must be comparable to the efficiency described
by k6a for electron trapping by anion vacancies in order to keep the
concentration of color centers relatively low.
(
(
26) Woosley, J. D.; Wood, C. Phys. ReV. B, 1980, 22, 1066.
(27) Savihina, T. I.; Merilloo, I. A. Trudi Inst-ta Fiziki AN ESSR, Tartu;
Ed. Luschik, Ch. B., 1979, 49, 146.
(
28) Cain, L. S.; Pogatshnik, G. I.; Chen, Y. Phys. ReV. B 1988, 37,
2
645.
(
(
29) Emeline, A. V.; Ryabchuk, V. K. Fiz. Khim. 1998, 72, 305.
30) White, G. S.; Jones, R. V.; Crawford, J. H., Jr. J. Appl. Phys. 1982,
5
3, 265.
(
(
31) Emeline, A. V.; Ryabchuk, V. K. Fiz. Khim. 1997, 71, 2085.
32) Schulman, J. H.; Compton, W. D. Color Centers in Solids.
(46) (a) Solonitzyn, Yu. P.; Kuzmin, G. N.; Shurigin, A. L.; Yurkin, V.
M. Kinet. Katal. 1976, 17, 1267. (b) Basov, L. L.; Kuzmin, G. N.;
Prudnikov, I. M.; Solonitzyn, Yu. P. Usp. Photoniki, Iss. 6; Vilesov, Th. I.,
Ed., LGU (Leningrad State University), 1977, pp 82-120. (c) Andreev, N.
S.; Koltel’nikov, V. A. Kinet. Katal. 1974, 15, 612. (d) Kondo, J.; Sakata,
J.; Domen, K.; Marya, K.; Onisi, T. J. Chem. Soc., Faraday Trans. I 1990,
82, 397.
(47) Schrimer, O. F. J. Phys. C: Solid State Phys. 1978, 11, L65.
(48) Kuznetsov, V. N.; Skaletskaya, T. K.; Lisachenko, A. A. Fiz. Khim.
1980, 54, 2596.
Pergamon Press: Oxford-London-New York-Paris, 1962.
(
33) Williams, R. T. Semiconductors Insulators 1978, 3, 251.
(34) “Defects and Impurity Centers in Ionic Crystals: Optical and
Magnetic Properties,” Part 1. Special Issue, J. Phys. Chem. Solids 1990,
1, N. 7.
35) Bonch-Bruevich, V. L.; Kalashnikov, S. G. Physics of Semiconduc-
tors; Nauka: Moscow, Russia, 1990.
36) Siline, A. R.; Trukchin, A. N. Point Defects and Elementary
Excitation in Crystalline and Glass SiO2; Zinatne Publishing House: Riga,
985.
37) Pankove, J. I. Optical Processes in Semiconductors; Dover Publica-
tions: NewYork, 1971.
38) Bube, R. H. PhotoconductiVity of Solids; John Wiley & Sons: New
York, 1960.
5
(
(
(49) Henderson, B.; Wertz, J. E. Defects in Alkaline Earth Oxides with
Application to Radiation Damage and Catalysis; Taylor & Francis: London,
1977.
1
(
(50) Crawford, J. H. Semiconductors Insulators, 1983, 5, 599.
(51) In eq 23, it should be noted that ∆[V] is proportional to δR and
(
+
that [S ], which reflects the number of V centers, is proportional to ∆P.
(
39) Aristov, Yu. I.; Parmon, V. N.; Zamaraev, K. I. Khim. Fiz. 1982,
The number of V centers given in the text was calculated using the following
procedure: from the slope of the relationhip between ∆P and δR in Figure
7 we first estimate the number of molecules adsorbed using the ideal gas
law and then we weight it by the extent of coloration ∆R from Figure 5 of
the UV-pre-irradiated specimen under the conditions of the experiment.
(52) Llopis, J. Phys. Status Solidi A 1990, 119, 661.
9
, 1233.
(
40) Serpone, N.; Khairutdinov, R. F. In Semiconductor Nanoclusters
-
Physical, Chemical, and Catalytic Aspects; Kamat, P.V.; Meisel, D.; Eds.;
Stud. Surf. Sci. Catal. 1997, 103, 417.
(
(
41) Serpone, N.; Khairutdinov, R. F. Prog. React. Kinet. 1996, 21, 1.
42) Solonitzyn, Yu. P.; Prudnikov, I. M.; Yurkin, V. A. Fiz. Khim.
(53) Arsenev, P. A.; Bagdasarov, Kh. S.; Niklas, A.; Ryazantsev. A. D.
Phys. Status Solidi A 1980, 62, 395.
(54) Lorincz, A.; Puma, M.; James, F. J.; Crawford J. H. J. Appl. Phys.
1982, 53, 927.
1
980, 57, 2028.
43) Irradiation of a metal-oxide specimen produces active centers at
the surface; if done in vacuo, the concentration of these centers is determined
(