Gas-Phase Hydroxyl Radical Generation by the Surface Reactions over Basic Metal Oxides

Article abstract of DOI:10.1021/jp9731314

Hydroxyl radical desorption in the heterogeneous catalytic reactions of water or hydrogen with oxygen was examined by laser-induced fluorescence spectroscopy. The catalytic activities of Pt, Al2O3, and basic metal oxides (MgO, CaO, SrO, BaO) supported on Al2O3 were studied in the pressure range 0.1-10 Torr and the temperature range 1100-1300 K. In the case of OH generation from water, the catalytic activities of Pt and all oxides except Al2O3 were very high and the OH concentration reached the equilibrium value within a residence time of 4 ms. In the case of hydrogen oxidation, differences in the catalytic behavior were clearly observed. The surface reaction mechanisms and the effects of Al2O3 and MgO on gas-phase ignition are discussed.

Full text of DOI:10.1021/jp9731314

J. Phys. Chem. B 1998, 102, 3185-3191
3185
Gas-Phase Hydroxyl Radical Generation by the Surface Reactions over Basic Metal Oxides
Suguru Noda,*,† Masateru Nishioka,† Azuchi Harano,‡ and Masayoshi Sadakata†
Department of Chemical System Engineering, Graduate School of Engineering, The UniVersity of Tokyo,
7-3-1, Hongo, Bunkyo-ku, Tokyo 113-8656, Japan, and Department of Biological and Chemical Engineering,
Faculty of Engineering, Gunma UniVersity, 1-5-1 Tenjin-Cho, Kiryu, Gunma-ken, 376-8515, Japan
ReceiVed: September 26, 1997; In Final Form: December 31, 1997
Hydroxyl radical desorption in the heterogeneous catalytic reactions of water or hydrogen with oxygen was
examined by laser-induced fluorescence spectroscopy. The catalytic activities of Pt, Al2O3, and basic metal
oxides (MgO, CaO, SrO, BaO) supported on Al2O3 were studied in the pressure range 0.1-10 Torr and the
temperature range 1100-1300 K. In the case of OH generation from water, the catalytic activities of Pt and
all oxides except Al2O3 were very high and the OH concentration reached the equilibrium value within a
residence time of 4 ms. In the case of hydrogen oxidation, differences in the catalytic behavior were clearly
observed. The surface reaction mechanisms and the effects of Al2O3 and MgO on gas-phase ignition are
discussed.
Introduction
Gas-phase OH radicals are strong oxidizing agents and known
to be important chain carriers in combustion, so they are
expected to act as promoters in some reactions such as catalytic
combustion.1 To date, much research has been performed on
the OH desorption in heterogeneous catalyzed reactions over
the surfaces of noble metals. Ljungstro¨m et al. studied the
catalytic activities of several metals, and they found that the
order of the catalytic activities of the metals was Pt > Pd > Rh
> Ir > Ni.2 Pt is the most frequently studied catalyst, and the
kinetics of catalyzed reactions on its surface were explained by
a simulation incorporating a series of reactions among the neutral
species.3-6 On the other hand, only a few studies have been
performed on the OH desorption from metal oxides. The
reactions of methane or water with oxygen over basic lanthanide
oxides were examined by Lunsford’s group.7,8 The more basic
oxide proved to have the higher activity (La2O3 > Nd2O3 >
Sm2O3 > MgO > Yb2O3 > CeO2) for OH generation, and
reactions containing ionic species were proposed.
Thermodynamically, there are two strategies to produce OH
by heterogeneous catalytic reaction. One is to produce OH
below the thermodynamic equilibrium concentration from
species that have low chemical potential only by applying heat.
The other is to produce OH at a high concentration by applying
heat to species that have high chemical potential. In this work,
OH generation was examined in the catalytic reactions of water
(which has low chemical potential) or hydrogen (which has high
chemical potential) with oxygen by laser-induced fluorescence
(LIF) spectroscopy. We examine the catalytic properties of Pt,
basic metal oxides (MgO, CaO, SrO, BaO), and Al2O3 and
discuss the mechanisms governing OH generation.
Figure 1. Schematic diagram of the LIF system.
(rhodamine 590, pumped by a Q-switched Nd:YAG laser) was
used for OH-LIF. The OH was excited by the A2Σ+(V′ ) 1)
r X2Π(V ) 0) Q1(4) transition at 282.44 nm. The change of
the amplitude of this transition relative to the total amplitude is
insignificant in the temperature range of our experiments.8 The
fluorescence of the A2Σ+(V′ ) 1) f X2Π(V′′ ) 1) transition
around 314 nm was collected on a photomultiplier tube equipped
with two quartz glass lenses, a band-pass filter (313 nm,
bandwidth of 10 nm), and a slit set vertically with respect to
the laser beam. The signal was amplified, accumulated by a
boxcar integrator, and transferred to a computer. The time decay
of the LIF was also observed by a digital oscilloscope to estimate
the collisional quenching.
OH Measurement under Thermodynamic Equilibrium
Conditions. First, we measured the OH concentration under
thermodynamic equilibrium conditions to derive the absolute
OH concentration from the LIF signal. OH was generated by
heating a gas mixture of H2O, O2, and Ar in the reaction cell.
The cell was composed of a quartz glass tube with a heating
line, a thermocouple, and fire bricks as shown in Figure 2. Pure
oxygen, argon, and water-saturated argon were mixed at
atmospheric pressure. The gas mixture was supplied to the
quartz glass tube and pumped out by a rotary pump with a liquid
Experimental Section
LIF System. The LIF measurement system is shown in
Figure 1. Briefly, the second harmonic of the dye laser
* To whom correspondence should be addressed. Telephone: +81-3-
5802-8996. Fax: +81-3-5802-2997. E-mail: snoda@env-lab.t.u-
tokyo.ac.jp.
† The University of Tokyo.
‡ Gunma University.
S1089-5647(97)03131-3 CCC: $15.00 © 1998 American Chemical Society
Published on Web 04/02/1998

3186 J. Phys. Chem. B, Vol. 102, No. 17, 1998
Noda et al.
Figure 2. Top view of the experimental apparatus for OH measurement
under the thermal equilibrium conditions.
Figure 3. Side view of the flow tube reactor for the catalytic OH
generation.
nitrogen trap. Thermodynamic equilibrium conditions were
obtained when the valves on both sides of the tube were closed.
OH was excited by the laser, and the fluorescence was observed
through the window in the firebrick. From the lifetime of the
OH-LIF, the quenching rate constant of each gas was obtained.
These rate constants were used to derive the absolute concentra-
tion from the LIF signals in this and the following experiments.
Catalytic Generation of OH over the Surfaces. The flow
tube reactor for the catalytic generation of OH is shown in
Figure 3. It was composed of two quartz glass tubes and a SiC
ceramic heater covered with three alumina tubes. A K-type
thermocouple was put on the surface of the middle alumina tube
and covered with alumina paste, and the catalyst was supported
on it. For the basic metal oxides, a specific amount (3 mL) of
an aqueous solution of nitrate (0.1 M) was painted on the middle
alumina tube and dried in the atmosphere at about 420 K. Then
the nitrates were heated and kept at 1300 K for an hour under
a 20% O2/Ar flow at 6 Torr where they were decomposed into
oxides. In the case of Pt, platinum paste was painted and
calcined at 1300 K for an hour in the atmosphere. Reactant
gas was supplied to the reactor and pumped out by a rotary
pump with a liquid nitrogen trap. The desorbed OH was
measured 2 mm above the catalyst. Standard conditions for
the experiment were 1200 K, 6.0 Torr, and a 4 ms residence
time.
Figure 4. (a, top) Time profile of the OH-LIF signal. OH was
generated in the cell shown in Figure 2 at 1200 K. (b, middle) Effect
of the total pressure on the LIF lifetime at 1200 K. (c, bottom) Effect
of the total pressure on the OH concentration under thermal equilibrium
conditions at 1200 K.
radicals. From the work of Schofield and Steinberg,17 the
relative concentration can be expressed as8
A. Rose´n’s group5,6 pointed out that the profile of the OH-
LIF intensity differed largely from the profile of the desorption
rate, when they changed the total pressure by changing the input
mass flow of the reactant. In their experiments, the reactant
was supplied to the catalyst surface vertically in the stagnation
point flow geometry, so the OH concentration was largely
influenced by the mass transport of OH by the flow vertical to
the surface. Since the reactant was supplied parallel to the
surface in our experiments, the mass transport of OH vertical
to the surface by the flow was negligible. The residence time
of the reactant is kept more constant with a constant pumping
speed than with a constant mass flow of the reactant. Therefore,
we adjusted the total pressure by changing the input mass flow
with a constant pumping speed.
n
∑
OH,i
Aij
[OH]′rel
)
I (1)
nOH,i
A +
A + k + k [M ]
im d q q
∑
∑
ij
where [OH]′rel is the relative concentration of OH, ∑nOH,i is
the total number of OH, nOH,i is the number of OH in the ith
state, Aij is the Einstein A coefficient for the observed emission,
∑Aim is the sum of the Einstein A coefficient for the unobserved
emission, kd is the rate constant for predissociation, ∑kq[Mq] is
the sum of quenching rate constants times the concentration of
the quenching agents, and I is the OH-LIF intensity normalized
by the laser power. The value of ∑nOH,i/nOH,i remains almost
constant under our experimental conditions, and Aij is also
constant. The value of Aij + ∑Aim is equal to the inverse of
the lifetime (τ0 ) 0.75µs) of OH fluorescence without quench-
ing, which can be found in the work of J. Brzozowski et al.16
The value of kd equals zero here. So the equation is simplified
to
Results
Derivation of the Absolute Concentration from the OH-
LIF Signals. The quenching effect of the reactant gas should
be considered when we derive the relative concentration of OH

Hydroxyl Radical Generation
J. Phys. Chem. B, Vol. 102, No. 17, 1998 3187
TABLE 1: Quenching Rate Constants of OH (A2Σ+, W ) 1,
N ) 4, F1) at 1200 K (units of cm3 molecules-1 s-1
)
this expt
in ref 8
kH2O
kO2
kAr
6.2 × 10-10
1.6 × 10-10
5.2 × 10-12
(3.9-7.4) × 10-10
(1.1-2.2) × 10-10
kHe
5.6 × 10-14
I
[OH]rel
)
(2)
1
+
k [M ]
q q
∑
τ0
where [OH]rel is also the relative concentration of OH. The
dominator is equal to the inverse of the LIF lifetime observed
in the experiments.
Figure 4a shows the time profile of the OH-LIF signal
obtained by the oscilloscope. The LIF lifetime was derived from
the exponential fitting of the time profile. We obtained the
relation between the lifetime and the gas composition by
changing the total pressure or the partial pressure of each
reactant. Figure 4b shows the effect of the total pressure on
the LIF lifetime. The straight line is the result of fitting the
plots with the following equation:
k x
∑
q
q
1
1
)
+
P
(3)
τ
τ0
RT
where τ is the observed lifetime, kqxq is the quenching rate
constant times the molar ratio of each gas, P is the total pressure,
R is the gas constant, and T is temperature. The relations
between the partial pressure of each gas and the lifetime were
also analyzed by the following equation:
k P
∑
r
r
kq
1
1
r*q
)
+
+
Pq
(4)
τ
τ0
RT
RT
Figure 5. (a, top) Arrhenius plot of the OH concentration generated
in the catalytic reaction of H2O and O2 at 6.0 Torr. (b, middle) Effect
of the total pressure on the OH concentration generated in the catalytic
reaction of H2O and O2 at 1200 K. The data over Al2O3 were obtained
at 1300 K. (c, bottom) Effect of the reactant molar ratio on the OH
concentration generated in the H2O and O2 reaction over CaO at 1200
K and 6.0 Torr.
where Pq and Pr are the partial pressures of gases q and r,
respectively. From the results of the fitting, simultaneous
equations for the quenching rate constants were obtained, and
the rate constants shown in Table 1 were derived. The values
agreed well with those in ref 8.
Figure 4c shows the effect of the total pressure on the relative
OH concentrations. The OH-LIF signal normalized by the
laser power (2) was corrected by the procedure described above,
and the relative OH concentration (b) was obtained. The
relative OH concentration was also obtained by another method
(O). The OH-LIF intensity is proportional to the Einstein A
coefficient times the number of excited OH molecules. Since
there is no quenching effect immediately after the laser
excitation, the peak height normalized by the laser power in
the LIF time profiles in Figure 4a is proportional to the OH
concentration.
The results from both methods (b and O) agreed well with
the theoretical curve (-), proving that the relative concentrations
were successfully derived. The absolute OH concentrations in
other experiments can be obtained from the relation between
the measurements under equilibrium and the thermodynamic
calculations.
Catalytic Generation of OH over the Surfaces. Since the
experiments were performed under steady-state conditions,
adsorption on the surface did not change the composition of
the elements in the gas phase. We calculate the gas-phase OH
concentration under equilibrium conditions and compare it with
the measured OH concentration in the following sections.
OH Generation in the Reaction of H2O and O2. Figure
5a is the Arrhenius plot of the OH concentration generated in
the reaction of H2O and O2 over the catalysts. Al2O3 showed
the lowest catalytic activity. The OH concentration and the
apparent activation energy over the other catalysts were almost
the same as those under the equilibrium conditions. Figure 5b
shows the effect of the total pressure on the OH concentration.
The OH concentration changed almost linearly with the total
pressure over all the catalysts. Figure 5c shows the effect of
the molar ratios of H2O and of O2 on the OH generation over
CaO. Both molar ratios had a positive effect on the OH
generation, but H2O had a larger effect. The other alkaline earth
metal oxides had almost the same properties. Over Pt, O2 had
a smaller effect on OH concentration than the metal oxides.
Over MgO, both of the gases had almost the same effect. The
characteristic constants of the OH generation for each catalyst
are shown in Table 2a.

3188 J. Phys. Chem. B, Vol. 102, No. 17, 1998
Noda et al.
TABLE 2: Characteristics of OH Generationa
(a) In Catalytic Reaction of H2O and O2
[OH]b
Eac
H2Od
O2
d
equilibrium
Pt
Al2O3
MgO
CaO
SrO
BaO
29
15
2.1
13
17
16
28
153
152
167
144
140
145
143
0.5
0.25
0.11
0.20e
0.32
0.19
0.18
0.19
0.45
0.37e
0.37
0.45
0.40
0.47
(b) In Catalytic Reaction of H2 and O2
[OH]b
Eac
H2
O2
d
d
equilibrium
Pt
Al2O3
MgO
CaO
SrO
BaO
29
56
ignition
0.5
3
2
5
153
238
0.5
0.13
0.25
0.54
210
198
194
182
f
f
f
0.41
0.49
0.47
a Standard experimental conditions: 1200 K, 6 Torr, 4% H2O or
H2, 20% O2, 76% Ar. b OH concentration under the conditions listed
above [1010 molecules/cm3]. c Apparent activation energy [kJ/mol].d The
order of OH concentration dependence on the partial pressure of each
gas. e At 1300 K. f There was an optimum H2 molar ratio for OH
generation.
metal oxides showed similar properties. Over Pt, both of the
gases had a positive effect with O2 having the larger effect.
The characteristic constants of the OH generation for each
catalyst are shown in Table 2b.
Estimation of the OH Generation in the Gas-Phase
Reaction. To estimate the OH generation by the gas-phase
reactions, a kinetic simulation using CHEMKIN9 code was
carried out. The following reactions were included in the code.
H2 + O2 ) 2OH
OH + H2 ) H2O + H
O + OH ) O2 + H
O + H2 ) OH + H
H + O2 + M ) HO2 + M
OH + HO2 ) H2O + O2
H + HO2 ) 2OH
(I)
(II)
(III)
Figure 6. (a, top) Arrhenius plot of the OH concentration generated
in the catalytic reaction of H2 and O2 at 6.0 Torr. (b, middle) Effect of
the total pressure on the OH concentration generated in the catalytic
reaction of H2 and O2 at 1200 K. (c, bottom) Effect of the reactant
molar ratio on the OH concentration generated in the H2 and O2 reaction
over CaO at 1200 K and 6.0 Torr.
(IV)
(V)
(VI)
(VII)
(VIII)
(IX)
(X)
OH Generation in the Reaction of H2 and O2. Figure 6a
is the Arrhenius plot of the OH concentration generated in the
hydrogen oxidation over the catalysts. The OH concentration
showed the Arrhenius type dependence over all the catalysts
except Al2O3. Pt had the highest catalytic activity, and the OH
concentration exceeded the thermodynamic equilibrium value.
The alkaline earth metal oxides (e.g., CaO, SrO, BaO) had the
intermediate activity, and MgO had the lowest activity. As for
Al2O3, the OH concentration changed sharply around 1220 K
and exceeded that over Pt at higher temperatures. Figure 6b
shows the effect of the total pressure on the OH concentration.
The OH concentration changed almost linearly with pressure
only over Pt. Over alkaline earth metal oxides, OH concentra-
tion was constant from 2 to 8 Torr but started to increase above
8 Torr. As for Al2O3, OH was hardly observed below 6 Torr
but rapidly increased above 6 Torr. Figure 6c shows the effect
of the molar ratios of H2 and O2 on the OH generation over
CaO. O2 had a positive effect on the OH generation, but there
was an optimum molar ratio for H2. The other alkaline earth
O + HO2 ) O2 + OH
2OH ) O + H2O
2H + M ) H2 + M
2H + H2 ) 2H2
(XI)
(XII)
(XIII)
2H + H2O ) H2 + H2O
H + OH + M ) H2O + M
H + O + M ) OH + M
2O + M ) O2 + M
(XIV)
(XV)

Hydroxyl Radical Generation
H + HO2 ) H2 + O2
J. Phys. Chem. B, Vol. 102, No. 17, 1998 3189
(XVI)
(XVII)
(XVIII)
(XIX)
2HO2 ) H2O2 + O2
H2O2 + M ) 2OH + M
H2O2 + H ) HO2 + H2
H2O2 + OH ) H2O + HO2
(XX)
Parts a and b of Figure 7 show the time profiles of the OH
concentration in the reaction of H2O and O2 and in the reaction
of H2 and O2, respectively. In the reaction of H2O and O2 the
OH concentration increased slowly, reaching 90% of the
equilibrium value after 900 s. To estimate the effect of the
surface-generated OH, 1 ppm of OH was added to the initial
gas composition (broken curve). The addition promoted the
OH generation by 35 s, but there was no difference in shape
between the two curves.
In the reaction of H2 and O2 the OH concentration increased
rapidly, reaching the equilibrium value after a few milliseconds
and 300 times that after 10 ms. The OH addition promoted
OH generation for 10 ms but had little effect after 10 ms.
Discussion
Figure 7. (a, top) Calculated results of the change of OH concentration
with time generated in the gas-phase reaction of H2O and O2 at 1200
K and 6.0 Torr. (b, bottom) Calculated results of the change of OH
concentration with time for the gas-phase reaction of H2 and O2 at 1200
K and 6.0 Torr.
From the experimental results stated above, the catalysts can
be classified into four groups as shown in Table 3. The
following discussion of the reaction mechanism proceeds with
this classification in mind.
OH Generation in the Reaction of H2O and O2. In the
OH generation from H2O, all the catalysts except Al2O3 had
high catalytic activity, and the OH concentration was almost
the same as the equilibrium value. The activation energy was
similar to the calculated equilibrium value, and the dependence
on the molar ratio of the reactants also agreed well with the
equilibrium value particularly for the alkaline earth metal oxides
(CaO, SrO, BaO). So we conclude that the OH generation rate
was so fast that the OH concentration both on the surface and
in the gas phase almost reached equilibrium within a residence
time of 4 ms for all the catalysts except Al2O3.
OH Generation in the Reaction of H2 and O2. In this
reaction, the catalysts showed quite different characteristics from
each other. OH concentration exceeded the equilibrium con-
centration in the reaction over Pt. It is known that the OH
desorbed from Pt can induce the gas-phase radical chain reaction
through the mechanism of catalytic ignition,1 which explains
the OH concentration over Pt. For the other catalysts, the OH
concentration was below the equilibrium. We discuss the
reaction mechanisms next.
Rate-Determining Step of the OH Generation in the
Reaction of H2 and O2 over Basic Metal Oxides. In this
reaction, the OH generation was less than that in the reaction
of H2O and O2, in which the OH concentration reached
equilibrium. Since the inlet gases were supplied in almost the
same elementary ratio in the two reactions (H/O was 2/10 and
2/11, respectively), the OH concentration should be the same
in the two reactions under thermodynamic equilibrium condi-
tions. It is clear that the OH generation did not reach
equilibrium, so there exists a rate-determining step in the
reaction of H2 and O2. In this reaction system, three steps
(adsorption, surface reactions, and desorption) should be
considered to determine the rate-determining step. Since the
experimental results were obtained under steady-state conditions,
all the steps except the rate-determining one are at equilibrium.
TABLE 3: Classification of Catalysts for OH Generation
catalyst
H2 + O2
high
middle
H2O + O2
i
Pt
high
high
high
low
ii
iii
iv
CaO, SrO, BaO
MgO
Al2O3
low
low ignitiona
a OH concentration rapidly increased with temperature increase by
gas-phase ignition.
If the desorption was the rate-determining step, the other steps
should be at equilibrium. As stated above, the inlet gases were
supplied in almost the same elementary ratio as in the reaction
of H2O and O2, and the composition of the surface species
should be also the same. Since the OH desorption rate is
determined only by the catalyst and the composition of the
surface species, the desorption rate should also be the same.
This conflicts with the experimental result, so desorption cannot
be the rate-determining step. This conclusion is contrary to
previous research, in which the desorption process controlled
the OH generation rate over noble metals under high vacuum
conditions.5
Next we consider adsorption. In the case of the alkaline earth
metal oxides (CaO, SrO, BaO), the OH generation did not
depend on the total pressure over a wide range (2-8 Torr). This
typically occurs when the adsorption of the reactants reaches
equilibrium at high coverage. If adsorption is the rate-
determining step, the coverage should be lower and depend on
the total pressure more strongly than when adsorption is at
equilibrium. Thus, adsorption is probably at equilibrium, and
the total reaction rate was most likely determined by surface
reactions.
Mechanism of the Surface Reactions of H2 and O2 over
Basic Metal Oxides. In the surface oxidation reactions over
metal oxides, ionic species can play important roles. The
adsorption of O2 on basic metal oxides has been studied by the

3190 J. Phys. Chem. B, Vol. 102, No. 17, 1998
Noda et al.
ESR technique, and the reactions of surface oxygen anion
species were well investigated especially on MgO at room
temperature or lower temperatures.18 Hewett et al.8 studied the
reaction of H2O and O2 over lanthanide metal oxides under
nearly the same conditions as our experiments, and they
proposed the following reaction mechanism:
From the discussion above, both routes for the OH generation
seem to be possible. The dominant route cannot be determined,
however, because the OH generation from the reaction of H2
and O2 is always accompanied by the generation of H2O.
Effect of the MgO and Al2O3 Surfaces on the Gas-Phase
Ignition. Both MgO and Al2O3 had little catalytic activity in
the reaction of H2 and O2 below 1200 K, but they showed quite
different behaviors at higher temperatures. The OH concentra-
tion over MgO remained low but that over Al2O3 increased
rapidly with a temperature increase. A rapid increase of OH
over Al2O3 was also observed with a total pressure increase.
The surface reaction mechanism can hardly explain these
discontinuities. Since these phenomena typically appear in the
gas-phase chain reaction, we concluded that the reaction
conditions were close to the limit of inflammability and that
gas-phase ignition occurred. The difference between MgO and
Al2O3 can be explained by the consumption rate of OH on the
surface as follows. Above 1200 K, OH is produced by the gas-
phase reaction of H2 and O2. The produced OH diffuses to the
solid surface when the OH concentration near the surface is
low. Since the reversible reaction of H2O and O2 to form OH
is fast on the MgO surface, the sticking coefficient of OH on
MgO is large. OH adsorbs on the surface and is consumed
rapidly through recombination or reaction with surface species
(H2 or H ions), so the gas-phase ignition is suppressed. Al2O3
has so little OH generation activity that OH is hardly consumed
on its surface, and the gas-phase ignition occurs. Hence, the
consumption rate of OH on the surface may be one of the factors
controlling the gas-phase ignition.
-
H2O + OS- T OH + OHS
(5)
2-
2OHS- T H2O + 0 + OS
(6)
(7)
O2 T 2OS
-
OS + 0 + OS2- T 2OS
(8)
where “0” refers to an oxygen vacancy and the subscript “S”
refers to a surface species.
Next, we consider the adsorption and the reactions of H2.
MgO proved to have a high catalytic activity in the H2-D2
exchange reaction after neutron or ultraviolet irradiation,11,12
as well as after evacuation at high temperature.13 MgO and
the alkaline earth metal oxides were also found to show high
catalytic activity for the hydrogenation of olefins after evacu-
ation at around 1100 K.14,15 The strong dependence of catalytic
activity on the evacuation temperature indicated that the surface
structures as well as the desorption of surface species were
important. It is believed that H2 heterolytically dissociates to
produce H+ and H- ions on the O2- site and Mg2+ site,
respectively. In our experiment, the OH generation was carried
out under the same conditions as in these previous studies
(1100-1300 K under vacuum), so the adsorption of H2 should
proceed by the same mechanism.
Conclusions
Basic metal oxides proved to be good catalysts for rapid OH
generation from water but not for the OH generation from
hydrogen.
H2 and H2O both can act as hydrogen sources in OH
generation. MgO showed little activity in our experiment, so
it is difficult to discuss the reaction mechanism. For the other
oxides (CaO, SrO, BaO), there was an optimum concentration
of H2 for the OH generation. This phenomenon can be
explained as follows.
In the OH generation from H2O and O2, basic metal oxides
(MgO, CaO, SrO, BaO) showed high catalytic activity, and the
OH concentration over the oxides almost reached the thermo-
dynamic equilibrium value within a residence time of 4 ms.
In the OH generation from H2 and O2, the alkaline earth metal
oxides (CaO, SrO, BaO) showed catalytic activities that were
less than that of Pt. The other oxides (MgO and Al2O3) showed
the lowest activities. The rate-determining step for OH genera-
tion over the alkaline earth metal oxides was probably the
surface reaction process under our experimental conditions. The
OH concentration showed a large change with changes in
temperature or pressure only over Al2O3. This phenomenon
was attributed to gas-phase ignition, and the consumption rate
of OH at the surface is proposed to be one of the factors
controlling this process.
For H2, the following reactions are assumed.
-
H2 dissociation: H2 f HS+ + HS
(9)
(10)
(11)
(12)
OH generation: HS+ + OS- f OHS
OH consumption: HS- + OS f OHS
-
HS+ + OS2- f OHS
-
followed by
-
OHS- + OH f H2O + OS
(5′)
Acknowledgment. We thank professor Mitsuo Koshi of the
University of Tokyo for his kind advice regarding the LIF
measurement and the computer calculations. We also thank Dr.
Toshihide Baba of Tokyo Institute of Technology and Dr.
Takehiko Sasaki of the University of Tokyo for their fruitful
comments on the surface reaction mechanisms. Moreover we
thank professor Tatsuya Okubo of the University of Tokyo for
his advice on our overall research.
At low H2 pressure, an increase in the molar ratio of H2 has a
positive effect on the OH generation by accelerating reaction
10. However, at high H2 pressure, reaction 11 or 12 can be
important, resulting in the consumption of surface OH by
reaction 5′.
OH generation from H2O is considered to proceed by
reactions 5-8. At low H2 pressure, H2 contributes positively
to the generation of H2O, and therefore, the OH generation is
promoted by reaction 5. At high H2 pressure, inhibition of
reaction 5 by the direct reaction of H2 with O- could become
important, and this reaction would lead to the suppression of
OH generation.
References and Notes
(1) Pfefferle, L. D.; Griffin, T. A.; Winter, M.; Crosley, D. R.; Dyer,
M. J. Combust. Flame 1989, 76, 325.
(2) Ljungstro¨m, S.; Hall, J.; Kasemo, B.; Rose´n, A.; Wahnstro¨m, T. J.
Catal. 1987, 107, 548.

Hydroxyl Radical Generation
J. Phys. Chem. B, Vol. 102, No. 17, 1998 3191
(3) Hsu, D. S. Y.; Hoffbauer, M A.; Lin, M. C. Surf. Sci. 1987, 184,
25.
(4) Williams, W. R.; Marks, C. M.; Schmidt, L. D. J. Phys. Chem.
1992, 96, 5922.
(5) Gudmundson, F.; Fridell, E.; Rose´n, A.; Kasemo, B. J. Phys. Chem.
1993, 97, 12828.
(6) Fridell, E.; Rose´n, A.; Kasemo, B. Langmuir 1994, 10, 699.
(7) Anderson, L. C.; Xu, M.; Mooney, C. E.; Rosynek, M. P.; Lunsford,
J. H. J. Am. Chem. Soc. 1993, 115, 6322.
(8) Hewett, K. B.; Anderson, L. C.; Rosynek, M. P.; Lunsford, J. H.
J. Am. Chem. Soc. 1996, 118, 6992.
(9) Gordon, S.; McBride, B. J. Report NASA SP-273; NASA:
Washington, DC, 1971.
(10) Kee, Robert J.; Rupley, F. M.; Miller, J. A. Sandia National
Laboratories Report SAND89-8009; Sandia National Laboratories: Albu-
querque, NM, 1989.
(11) Lunsford, J. H.; Leland, T. W. J. Phys. Chem. 1962, 66, 2591.
(12) Lunsford, J. H. J. Phys. Chem. 1964, 68, 2312.
(13) Boudart, M.; Delbouille, A.; Derouane, E. G.; Indovina, V.; Walters,
A. B. J. Am. Chem. Soc. 1972, 94, 6622.
(14) Tanaka, Y.; Hattori, H.; Tanabe, K. Chem. Lett. 1976, 37.
(15) Hattori, H.; Tanaka, Y.; Tanabe, K. Chem. Lett. 1975, 659.
(16) Brzozowski, J.; Erman, P.; Lyyra, M. Phys. Scr. 1978, 17, 507.
(17) Schofield, K.; Steinberg, M. Opt. Eng. 1981, 20, 501.
(18) Lunsford, J. H. CRC Crit. ReV. Solid State Sci. 1976, 6, 337.