Selective CO oxidation on a Ru/Al2O3 catalyst in the surface ignition regime: 1. Fine purification of hydrogen-containing gases

Article abstract of DOI:10.1007/s10975-008-1011-8

Selective CO oxidation in a mixture simulating the methanol steam reforming product with an air admixture was studied over Ru/Al₂O₃ catalysts in a quasi-adiabatic reactor. On-line monitoring of the gas temperature in the catalyst bed and of the residual CO concentration at different reaction conditions made it possible to observe the ignition and quenching of the catalyst surface, including transitional regimes. A sharp decrease in the residual CO concentration takes place when the reaction passes to the ignition regime. The evolution of the temperature distribution in the catalyst bed in the ignition regime and the specific features of the steady-state and transitional regimes are considered, including the effect of the sample history. In selective CO oxidation and in H₂ oxidation in the absence of CO, the catalyst is deactivated slowly because of ruthenium oxidation. In both reactions, the deactivated catalyst can be reactivated by short-term treatment with hydrogen. A 0.1% Ru/Al₂O₃ catalyst is suggested. In the surface ignition regime, this catalyst can reduce the residual CO concentration from 0.8 vol % to 10-15 ppm at O₂/CO = 1 even in the presence of H₂O and CO₂ (up to ～20 vol %) at a volumetric flow rate of ～100 1 (g Cat)^-1 h^-1, which is one magnitude higher than the flow rates reported for this process in the literature.

Full text of DOI:10.1007/s10975-008-1011-8

ISSN 0023-1584, Kinetics and Catalysis, 2008, Vol. 49, No. 1, pp. 92–102. © MAIK “Nauka /Interperiodica” (Russia), 2008.
Original Russian Text © A.Ya. Rozovskii, M.A. Kipnis, E.A. Volnina, P.V. Samokhin, G.I. Lin, 2008, published in Kinetika i Kataliz, 2008, Vol. 49, No. 1, pp. 99–109.
Selective CO Oxidation on a Ru/Al2O3 Catalyst
in the Surface Ignition Regime:
1. Fine Purification of Hydrogen-Containing Gases1
A. Ya. Rozovskii, M. A. Kipnis, E. A. Volnina, P. V. Samokhin, and G. I. Lin
Topchiev Institute of Petrochemical Synthesis, Russian Academy of Sciences, 119991 Russia
e-mail: rozovsk@ips.ac.ru
Received January 31, 2006; in final form, June 21, 2007
Abstract—Selective CO oxidation in a mixture simulating the methanol steam reforming product with an air
admixture was studied over Ru/Al2O3 catalysts in a quasi-adiabatic reactor. On-line monitoring of the gas tem-
perature in the catalyst bed and of the residual CO concentration at different reaction conditions made it possi-
ble to observe the ignition and quenching of the catalyst surface, including transitional regimes. A sharp
decrease in the residual CO concentration takes place when the reaction passes to the ignition regime. The evo-
lution of the temperature distribution in the catalyst bed in the ignition regime and the specific features of the
steady-state and transitional regimes are considered, including the effect of the sample history. In selective CO
oxidation and in H2 oxidation in the absence of CO, the catalyst is deactivated slowly because of ruthenium
oxidation. In both reactions, the deactivated catalyst can be reactivated by short-term treatment with hydrogen.
A 0.1% Ru/Al2O3 catalyst is suggested. In the surface ignition regime, this catalyst can reduce the residual CO
concentration from 0.8 vol % to 10–15 ppm at O2/CO = 1 even in the presence of H2O and CO2 (up to
~20 vol %) at a volumetric flow rate of ~100 l (g Cat)–1 h–1, which is one magnitude higher than the flow rates
reported for this process in the literature.
DOI: 10.1134/S0023158408010114
1
In recent decades, there have been extensive studies
on the catalytic removal of CO from hydrogen-contain-
ing gases (selective CO oxidation). This interest is due
to the development of technologies using H2 in electric-
ity generation (fuel processors), separation processes
using platinum metal membranes, etc. As a rule, fuel
processors require that CO in the hydrogen-containing
gas be reduced to 10–100 ppm. A promising solution to
this problem is use of catalysts based on noble metals,
in particular, Ru-containing catalysts [1–7].
Since oxygen is also spent on hydrogen oxidation,
its proportion in the gas mixture should be above the
stoichiometric proportion. The normal reactant ratio
seems to be O2/CO = 1. In fact, the desired reduction of
the CO content in the case of ruthenium catalysts is
usually achieved at O2/CO > 1 [2–6, 8, 10–12], at a low
or zero water content [1, 6, 13], and at moderate GHSV
values (no higher than 15000 h–1) [2–4, 7, 12]. At a low
CO content of hydrogen (0.6 vol %), in the absence of
water and CO2, and a low GHSV of 7500 h–1, the
desired purification efficiency can be achieved at
O2/CO = 1 [6]. However, if the CO concentration in the
initial mixture is raised from 0.6 to 1%, it will be nec-
essary to increase the O2 content from 0.6 to 1.5 vol %
in order to reduce the residual CO content to the desired
level of 10 ppm [6].
A possible source of hydrogen is methanol [1, 2, 7,
8]. Methanol steam reforming yields a gas containing
hydrogen, ~1% CO, ~20% water, and CO2. This gas is
purified from CO by oxidizing the latter into CO2. In
order to achieve a low residual CO level at high flow
rates, CO is removed in two steps (or more) [8–10] with
oxygen introduction between the stages or the O2/CO
ratio is raised. However, two-step purification compli-
cates the overall process (since it requires CO and O2
control between the stages, an extra air feed, and a
larger volume of the CO converter). Raising the O2/CO
ratio leads to extra oxygen consumption for hydrogen
oxidation and, hence, a lower capacity of the fuel pro-
cessor.
In vehicular applications of fuel cells, the catalyst
has to operate at rather high gas flow rates. To ensure an
electric power of 10 kW, the methanol steam reforming
products must be passed through the fuel processor at a
rate of 200 l/min [8]. If the desired power is one mag-
nitude higher, which is the case for modern automo-
biles, the required flow rate will be ~100 m3/h. At
GHSV values usual for catalytic processes (~10000 h–1
and below), CO removal from this amount of the gas
1
The article includes materials from the authors’ report at the II
Russian Conference on Current Problems in Petroleum Chemis-
try, Ufa, October 11–13, 2005.
92

SELECTIVE CO OXIDATION ON A Ru/Al2O3 CATALYST
93
(‡)
240
220
200
180
160
140
120
100
10000
8000
6000
4000
4
3
2
2000
0
1
180
220
260
300
Time, min
340
380
420
10000
8000
6000
225
200
175
150
125
100
75
(b)
1
4000
2000
0
2
100
110
120
130
140
150
Furnace temperature, °C
Fig. 1. Selective CO oxidation on the 1% Ru/Al O catalyst. Fragment of a record of the process made while heating the surface in
2
3
steps: (a) dynamics of (1) the residual CO concentration, (2) the furnace temperature, (3) the gas temperature at the catalyst bed
entrance, and (4) the gas temperature at the bed exit; (b) (1) residual CO concentration and (2) gas temperature at the bed exit versus
the furnace temperature. Feed composition (vol %): CO, 0.75; O , 0.75; H , 57; CO , 18, H O, 20; N , balance. The gas flow rate
2
2
2
2
2
–1 –1
is 87 l (g Cat) h .
will require a catalyst volume of ~10 l. Since almost achieved a high-purity product at O2/CO = 1 and flow
complete CO removal is required, the real catalyst vol- rates of about 100 l (g Cat)–1 h–1.
ume will be much larger.
Thus, a challenging problem is to design a highly
active and selective catalyst capable of removing CO at
high GSHV values and in the presence of H2O and CO2.
EXPERIMENTAL
Catalyst Synthesis
The catalyst support was commercial γ-Al2O3
extrudate (Ryazan Refinery, A-64k brand, specific
surface area of 200 m2/g, size fraction of 0.200–
0.315 mm after crushing). Before loading the active
component, the support was calcined in air at 500°C
for 2 h. The starting ruthenium compound was
Ru(OH)Cl3 (see the caption to Fig. 5). Catalysts were
In earlier studies [14, 15], we demonstrated by the
example of a modified platinum catalyst that the purifi-
cation efficiency can be significantly raised by conduct-
ing selective CO oxidation in the catalyst surface ignition
regime (the specific features of chemical reactions in this
regime are considered in earlier monographs [16, 17]).
Here, we report our study of selective CO oxidation prepared by impregnating alumina with a ruthenium
on ruthenium catalysts. By combining an active cata- salt solution followed by drying at 120°C for 6 h.
lyst with a favorable macrokinetic regime, we have After being charged into the reactor (without an inert
KINETICS AND CATALYSIS Vol. 49 No. 1 2008

94
ROZOVSKII et al.
60
50
40
30
20
10
0
150
140
130
120
110
100
2
1
175
225
275
325
375
Time, min
Fig. 2. Catalyst bed heating dynamics in selective CO oxidation on the 1% Ru/Al O catalyst: (1) difference between the gas tem-
2
3
peratures at the bed exit and entrance and (2) furnace temperature. The process conditions are the same as in Fig. 1.
component), the catalyst was reduced in flowing H2 added to this mixture as the oxidizer so that the initial
(3.5 l/h) for 2.5 h at 350 or 400°C.
CO/O2 ratio was 1 : 1. Furthermore, a mixture of H2,
O2, and N2 was used to study the specific features of H2
oxidation.
IR gas analyzers were calibrated and tested against
standard gas mixtures from the Balashikha Oxygen
Works, including mixtures containing less than
100 ppm CO.
Gas Mixtures
The gas mixtures to be used in experiments were
prepared by the partial pressure method from H2 (RF
Specifications 6-20-00209585-26-07), CO2 (USSR
State Standard GOST 8050-85), N2 (USSR Specifica-
tions TU 6-21-39-79, grade B), O2 (USSR State Stan-
dard GOST 5583-78, grade 1), and CO obtained by for-
mic acid decomposition. Vapor–gas mixtures used in
selective CO oxidation simulated the methanol steam
reforming product (H2, ~55–60 vol %; CO, ~0.8 vol %;
CO2, ~20 vol %; H2O, ~20 vol %). Oxygen was usually
Catalytic Tests
Catalytic activity was studied using a modified vari-
ant of the quartz flow reactor described in [15].
As compared to its previous version [15], the reactor
had a sealed-in porous quartz partition for supporting
the catalyst bed and a bypass line with two three-way
valves for bypassing the gas when necessary.
Selectivity
0.9
CO, ppm
6000
Experiments were performed at atmospheric pres-
sure. The initial gas mixture was fed into the reactor
through a unit allowing the gas flow rate to be measured
and regulated. The flow rate was varied in the range of
60–300 ml/min. The flow rate of the dry gas was mea-
sured with an IRG-1000 flowmeter with a scale interval
of 1 ml/min.
The gas mixture resulting from the reaction was
dried at 0°C and was divided into two streams with
metering valves. One stream was analyzed chromato-
graphically for CO, CH4, and O2 (LKhM 80 chromato-
graph, thermal-conductivity detector, zeolite 13X col-
umn, O2 detection limit of ~20 ppm, CH4 detection
limit of ~40 ppm, CO detection limit of ~100 ppm). The
second stream was directed to a BINOS 100 dual-chan-
nel IR gas analyzer to quantify CO (measurement range
of 0–9999 ppm) and CO2 (measurement range of
0−25%, relative error no greater than 5%). The error of
CO determination in the gas was mainly determined by
0.8
0.7
0.6
0.5
0.4
4500
3000
1500
1
2
0
140
160
180
200
220
240
260
Furnace temperature, °C
Fig. 3. Temperature dependence of the reaction selectivity
for CO oxidation in an isothermal reactor on the 0.1%
Ru/Al O catalyst: (1) oxygen consumption selectivity and
2
3
(2) residual CO concentration. The feed composition is the
same as in Fig. 1. The gas flow rate is 90 l (g Cat) h .
–1 –1
KINETICS AND CATALYSIS Vol. 49 No. 1 2008

SELECTIVE CO OXIDATION ON A Ru/Al2O3 CATALYST
95
230
210
190
170
150
130
110
90
700
600
500
400
300
200
100
0
5
4
3
2
1
70
50
90
100
110
120
130
140
150
Furnace temperature, °C
Fig. 4. Selective CO oxidation on the 1% Ru/Al O catalyst in the catalyst surface ignition regime during the stepwise cooling of
2
3
the furnace: (1–3) outlet O , CH , and CO concentrations, respectively; (4, 5) gas temperatures at the bed exit and entrance, respec-
2
4
–1 –1
tively. The feed composition is the same as in Fig. 1. The gas flow rate is 91 l (g Cat) h .
CO, ppm
1200
2
1000
3
800
1
3
600
400
200
0
2
4
4
80
100
120
140
160
180
200
220
240
Furnace temperature, °C
Fig. 5. Residual CO concentration in selective CO oxidation on (2–4) 0.1% Ru/Al O low-percentage catalysts as compared to (1)
2
3
1% Ru/Al O in the catalyst surface ignition regime during the stepwise cooling of the surface. The feed composition is the same
2
3
–1 –1
as in Fig. 1. The gas flow rate is 91 l (g Cat) h .
the error in the composition of the calibration mixtures temperature. The gas temperature in the catalyst bed
and by the baseline drift. At low CO concentrations was measured with two thermocouples placed at the
(~20 ppm), it was ≤2 ppm.
Current values of the CO concentration (in 1 ppm
increments) and CO2 concentration (in 0.1% incre-
ments) were displayed on the IR analyzer. The analog
bed entrance and exit, each at a distance of 1–2 mm
from the bed surface. The bed thickness was 1.2–
1.4 cm, and the catalyst weight was ~0.2 g.
When there was no reaction, the steady-state readings
signal from the CO channel was directed to a five-chan- from the thermocouples differ insignificantly (by a few
nel measuring controller (designed by A.P. Manyakin). tenths of a degree Celsius). At the furnace temperatures
This device was also used to monitor the furnace and 110°C and 250°C, they differ from the reading from the
catalyst temperatures ( 0.1 K) and control the furnace control thermocouple by 2.5 and 7°C, respectively.
KINETICS AND CATALYSIS Vol. 49 No. 1 2008

96
ROZOVSKII et al.
A special-purpose program (written by A.S. Korot- rate and, accordingly, the heat evolution rate are the
kov) allowed on-line computer monitoring of the highest. In this case, it is obvious that, under preignition
parameters being measured.
conditions, the maximum temperature (hot spot) will
occur at the downstream end of the catalyst bed. It is
also obvious that the critical ignition conditions will
primarily be established around the hot spot to enhance
the CO conversion and heat evolution. Owing to the
longitudinal transfer of the extra heat and the corre-
sponding perturbation, the ignition regime gradually
propagates in the bed as long as the perturbation inten-
sity is sufficient for satisfying the critical ignition con-
ditions for a given cross section. Note that the rate of
this process depends not only on the rate of heat trans-
fer through the reactor walls, but also on the mass flow
rate of the gas mixture in the reactor, because part of the
extra heat transferred to a reactor cross section is spent
on the heating of the gas coming to this cross section.
Thus, the ignition of the catalyst surface begins at
the bed exit and the ignition zone gradually expands
toward the upstream end of the bed. As the ignition
zone propagates to the upstream end, the CO and O2
conversions increase. Accordingly, the CO and O2 con-
centrations at the bed exit decrease to slow down the
oxidation reaction and to reduce the heat evolution due
to these reactions at the bed exit. The hot spot thus
shifts toward the bed entrance.
In a steady state, nearly all of the oxygen around the
hot spot is reacted. Therefore, moving further along the
reactor axis, the gas mixture only cools down because
of heat transfer through the reactor walls.
These features determine the dynamics of the longi-
tudinal temperature distribution in the catalyst bed: in
an unsteady state, a dramatic surge of the temperature
difference (Tout – Tin) is observed immediately after
ignition. After that, the temperature difference gradu-
ally decreases to the value characteristic of the steady-
state regime. The transitional period in our experiment
was ~7 min long.
Prior to experiments, the reactor was tested for tight-
ness by pressurization and was then heated in an H2
flow (3.5–4.0 l/h) to the preset temperature. Next, water
was fed into a preheated evaporator and hydrogen was
replaced with the reaction mixture by switching the
flows with a Swagelok five-way valve. After the exper-
iment, the reactor was purged with hydrogen and the
catalyst was left under a hydrogen pressure.
RESULTS AND DISCUSSION
Specific Features of the Reaction under Ignition
Conditions: Transitional Regimes
The CO and H2 oxidation reactions are very exo-
thermic (∆H = –283.0 and –241.8 kJ/mol, respectively
12]). In a quasi-adiabatic reactor, these reactions
readily pass to a catalyst surface ignition macrokinetic
regime [16, 17]. This transition usually takes place
abruptly. It can be observed as a dramatic rise of the gas
temperature in the catalyst bed and as a decrease in the
residual CO content of the gas mixture.
Figure 1a shows fragments of the residual CO, fur-
nace temperature, and the entrance and exit gas temper-
atures for the 1% Ru/Al2O3 catalyst bed (stepwise heat-
ing in 5 K increments, 10- to 20-min-long arrests at
each temperature point). Figure 1b plots the residual
CO content and the gas temperature at the bed exit ver-
sus the furnace temperature.
As is clear from these plots, raising the furnace tem-
perature up to 139°C causes a temperature rise in the
catalyst bed and a decrease in the residual CO content.
Raising the furnace temperature from 139 to 144°C
causes the ignition of the catalyst surface. This is man-
ifested as a dramatic increase of the catalyst bed tem-
perature and as a sharp decrease in the residual concen-
trations of CO (down to ~490 ppm) and O2 (according
to chromatographic data, the oxygen conversion is as
high as 99.6%). The gas temperature at the exit of the
catalyst bed under the ignition conditions was 215°C.
Let us consider the catalyst heating dynamics in this
experiment in greater detail (Fig. 2). As the furnace
temperature is raised in steps from 115 to 139°C, the
difference between the bed exit and entrance tempera-
tures increases monotonically from 1 to 7.8 K, stabiliz-
ing at each step. By contrast, raising the temperature by
another 5 K, from 139 to 144°C, causes the catalyst sur-
face to ignite. The difference between the bed exit and
entrance temperatures is as large as 57.5 K at the igni-
tion point and stabilizes at 31.8 K in the steady-state
region.
The reverse behavior will be observed in surface
quenching. For example, if quenching is initiated by
furnace cooling or by catalyst deactivation, the hot spot
will move toward the bed exit, followed by the ignition
zone boundary moving in the same direction.
Effect of Temperature on the Oxidation Selectivity
The effect of temperature on the selectivity of oxy-
gen consumption in CO oxidation in a steady-state
regime (0.1% Ru/Al2O3, sample 2; see caption to Fig. 5)
was studied in an isothermal stainless steel flow reactor
using a KL-3D setup (designed by the Design Office of
the Zelinskii Institute of Organic Chemistry, Russian
Academy of Sciences). The catalyst batch was diluted
The presence of a sharp peak in the heating dynam- with a tenfold volume of quartz (0.25–0.315 mm size
ics curve (Fig. 2) deserves special consideration. For fraction) and was placed into the annular space between
very exothermic reactions with a positive reaction the reactor walls and the thermocouple well. The resid-
order, the heat-induced ignition of the catalyst surface ual CO and O2 concentrations were determined chro-
should take place at low conversions, when the reaction matographically.
KINETICS AND CATALYSIS Vol. 49 No. 1 2008

SELECTIVE CO OXIDATION ON A Ru/Al2O3 CATALYST
97
The results obtained in this experiment are pre- hydrogenating a CO + CO2 mixture. It was found that
sented in Fig. 3. As the temperature is raised, the oxida- the reaction between hydrogen and CO2 begins only
tion selectivity decreases from a near-stoichiometric after the disappearance of CO. When the gas tempera-
value to 0.5 (Fig. 3, curve 1), forming a plateau, and ture in the catalyst bed is <200°C, no methane is
then to lower values. The plateau at ~190–230°C corre- detected among the reaction products.
sponds to the so-called “window” in which the lowest
residual CO concentration is observed. As the tempera-
ture is further raised, the residual CO concentration
increases at a rather high, progressively increasing rate
and, accordingly, the oxygen consumption selectivity
decreases.
The longitudinal temperature profile in the catalyst
bed for the experiment whose results are presented in
Fig. 4 is different from the profile shown in Fig. 1: the
hot spot is shifted toward the bed entrance, so that the
gas temperature near the entrance cross section exceeds
the gas temperature at the bed exit until the catalyst sur-
face is quenched.
Holding the catalyst bed at 104°ë for 70 min shifted
the hot spot toward the bed exit. Further cooling of the
furnace caused a marked decrease in the gas tempera-
ture in the catalyst bed: the quenching of the catalyst
surface set in.
Initiation of the Reaction
above the Critical Ignition Temperature
The simplest way of establishing the ignition regime
is by admitting the reaction mixture into a reactor pre-
heated to a temperature well above the critical ignition
temperature of the catalyst surface (Tcr). In this case, the
ignition regime will be established immediately after
the admission of the feed. The larger the difference T –
Tcr, the shorter the time required for the establishment
of a steady-state regime.
In the ignition regime, the furnace temperature may
be reduced well below the critical ignition temperature
(down to the critical surface quenching temperature)
without breaking down this regime. This provides a
means of enhancing the oxygen consumption selectivity.
Effect of the Reaction “Start” Temperature
on the Macrokinetic Situation in the Catalyst Bed
Operated in the Ignition Regime
As follows from the above, the axial temperature
distribution in a flow reactor for selective CO oxidation
depends on how the reaction was initiated, namely, by
starting it directly above the critical catalyst ignition
temperature (Tcr) or by gradually bringing it to this tem-
perature “from below.”
Figure 4 illustrates the dynamics of the component
concentrations and gas temperature at the entrance and
exit of the 1% Ru/Al2O3 bed during a stepwise decrease
of the furnace temperature with 10- to 20-min-long
temperature arrests. In this experiment, the reactor was
preheated to 144°ë in flowing H2, and the hydrogen
flow was then replaced with a feed flow. This caused the
ignition of the catalyst surface. The residual CO con-
centration was ~470 ppm, and the O2 conversion was
~99.9%. The highest gas temperature in the steady-
state surface ignition regime was observed near the bed
entrance and was 220°ë (versus 197°ë at the bed exit).
We have already considered the establishment of the
ignition regime in a flow reactor by gradual heating.
This regime usually “nucleates” in the vicinity of a hot
spot (which usually occurs near the bed exit because of
the low conversion) and then propagates toward the bed
entrance owing to longitudinal heat transfer, gradually
occupying all or part of the catalyst bed.
When the system is brought to the ignition regime
“from above” (that is, from a temperature far exceeding
Tcr, as in the experiment illustrated in Fig. 4), this
regime is rapidly established in the upstream part of the
bed. Subsequently reducing the temperature will not
quench the catalyst surface in the upstream layers until
the critical quenching temperature (Tquench) is reached.
Next, we began to cool the furnace, making temper-
ature arrests at 134, 124, 114, and 104°ë.
As the furnace was cooled, the residual CO concen- Since the quenching temperature is well below Tcr (the
tration decreased to ~100 ppm. Oxygen was consumed difference is tens of degrees), there is some interval
almost completely throughout the temperature range Tquench < T < Tcr in which different regimes can simulta-
until the quenching of the surface, which begins neously exist (or, as is probably more correct, coexist)
≥104°ë. Thus, under conditions of the experiment con- under the same conditions. Indeed, the surface ignition
sidered, temperature reduction in the ignition regime is regime established at a higher temperature will persist
favorable for selective CO oxidation. This is in agree- in the upstream part of the bed if the temperature is
ment with the data obtained for an isothermal reactor decreased within this interval. At the same time, if the
using a low-percentage catalyst (Fig. 3).
system is brought to ignition from lower temperatures,
it is possible that the ignition regime will not be estab-
lished in the upstream part of the bed because of the
insufficient longitudinal heat flux in the direction oppo-
site to the gas flow direction.
When the gas temperature at the exit of the catalyst
bed is >200°C, the desired oxidation reaction is accom-
panied by methane formation (Fig. 4). On the ruthe-
nium catalysts, methane results from CO hydrogena-
tion. As the temperature is raised, the contribution from
As a consequence, under given external conditions,
this reaction to the overall CO conversion increases. the system in which the ignition regime was established
The hydrogenation of CO, not CO2, was proved by at a high temperature and in which this temperature was
KINETICS AND CATALYSIS Vol. 49 No. 1 2008

98
ROZOVSKII et al.
then reduced may be nonidentical to the system in
Thus, we have got another means of controlling the
which the ignition regime was established by a gradual selectivity of the process in the catalyst surface ignition
temperature rise. Here, we deal with one more memory regime.
effect in heterogeneous catalysis, which, however,
arises from pure macrokinetic features of the system.
Stability of the Ruthenium Catalysts
The enhanced selectivity effect in CO oxidation in
Effect of the Properties of the Catalyst
the catalyst surface ignition regime is rather stable.
on the CO Oxidation Selectivity
Nevertheless, it is known from the literature that ruthe-
nium catalysts undergo comparatively rapid deactiva-
tion in CO oxidation. In CO + O2 + He mixtures
(CO/O2 = 0.5, 2, 4) reacting on a Ru/SiO2 catalyst
heated in steps to 140°C, the CO conversion at each
temperature point decreased rather soon (within a few
tens of minutes); furthermore, the CO conversion was
lower during stepwise cooling than during stepwise
heating [19].According to in situ IR spectroscopic data,
the intensities of the absorption bands due to CO lin-
early adsorbed on metallic ruthenium (2010 and
2030 cm–1) and on oxidized ruthenium (2080, 2130,
and 2135 cm–1) increased monotonically at 100°C.
Raising the temperature caused a weakening of all the
absorption bands. A possible cause of catalyst deactiva-
tion is that ruthenium oxides reacting sluggishly with
CO form on the catalyst surface under the oxidation
conditions [19].
Along with 1% Ru/Al2O3, a series of low-percent-
age catalysts containing 0.1 wt % Ru was prepared and
tested. Under the conditions of selective CO oxidation,
we observed surface ignition for all catalyst samples.
Figure 5 plots the residual CO content versus the fur-
nace temperature in the catalyst surface ignition regime
2
for three samples prepared by different methods. For
comparison, we present data for the 1% Ru catalyst.
Equal weights of the four samples were tested. In all
runs, the sample was heated in flowing hydrogen to a
temperature exceeding the critical ignition temperature
and the feed was then admitted.
As is clear from the plots shown in Fig. 5, the low-
percentage catalysts are much less active than the 1%
Ru catalyst. A given CO conversion is achieved on the
former at temperatures ~50 K higher than on the latter.
Oxygen is converted almost completely in the ignition
regime. The reaction over the low-percentage catalysts
yields no methane.
The effect of the oxidation temperature on the selec-
tivity of the reaction in the ignition regime is the same
as is described above: reducing the temperature causes
a decrease in the residual CO concentration until the
quenching of the surface. Therefore, the lower the sur-
face quenching temperature for a given catalyst, the
lower the residual CO level attainable by oxidation. The
surface quenching temperature in the low-percentage
catalyst series varies in a wide range (Fig. 5), and it is,
therefore, possible to achieve much lower residual CO
levels than are attainable with the 1% Ru catalyst.
Note that, in the catalyst surface ignition regime, the
reaction is controlled by external diffusion. It might be
expected that, in this case, the nature of the catalyst will
have no effect on the reaction kinetics and even selec-
tivity because the hydrogen concentrations at the cata-
lyst surface and in the flow core are approximately
equal. In fact, the catalyst composition effect is quite
evident (Fig. 5). As follows from the above discussion,
this effect is mediated by the effect of the catalyst com-
position on the surface quenching temperature. In turn,
the latter effect makes it possible to carry out the oxida-
tion process at a lower temperature within the ignition
limits and thus enhance the selectivity of the reaction.
The Ru/MgO catalyst also undergoes rapid deacti-
vation during CO oxidation in the absence of H2 [20].
At 100°C, the CO conversion on a prereduced Ru/MgO
sample decreased from 85 to <10% within 15–20 min.
In the authors’ opinion, the deactivation of this catalyst
is due to ruthenium oxidation yielding an oxide phase.
In the above experiments on CO oxidation in excess
hydrogen in the presence of H2O and CO2 in the cata-
lyst surface ignition regime at comparatively high tem-
peratures, we did not directly observed any decline of
catalytic activity. However, as was indicated by indirect
data, it did take place, though at a low rate.
Figure 6 illustrates the evolution of the residual CO
concentration and of the gas temperatures at the cata-
lyst (0.1% Ru/Al2O3) bed entrance and exit (sample 2
in Fig. 5). After the replacement of hydrogen with the
reaction mixture (57th minute) and the ignition of the
catalyst surface, the residual CO concentration does not
exceed 15 ppm and even shows a decreasing trend (10–
11 ppm at the end of the run) in spite of the high flow
rate, the presence of considerable amounts of H2O and
CO2, and a comparatively low O2/CO ratio. Methane
was not detected chromatographically in this experi-
ment.
At the same time, the run of the temperature curves,
specifically, the slow decrease in time of the gas tem-
perature near the catalyst bed entrance and the corre-
sponding increase of the gas temperature at the bed exit
(Fig. 6), indicates that the hot spot moves toward the
catalyst bed exit. This means that the activity of the cat-
alyst decreases slowly with time on stream (see above).
It would be expected that the gas temperature will then
2
Samples 1, 3, and 4 were synthesized by impregnating alumina
with an aqueous solution of the ruthenium salt. For samples 1 and
3, the solution was maintained at pH 10. Sample 2 was synthe-
sized by vacuum impregnation with a toluene solution of a metal
complex prepared from the original ruthenium salt and triocty-
lamine [18].
KINETICS AND CATALYSIS Vol. 49 No. 1 2008

SELECTIVE CO OXIDATION ON A Ru/Al2O3 CATALYST
99
40
30
20
10
200
175
150
125
100
75
4
3
2
1
0
100
200
300
400
500
Time, min
Fig. 6. Dynamics of the (1) residual CO concentration, (2) furnace temperature, (3) gas temperature at the bed exit, and (4) gas tem-
perature at the bed entrance for selective CO oxidation on a 0.1% Ru/Al O low-percentage catalyst (sample 2) in the surface ignition
2
3
–1 –1
regime. Feed composition (vol %): CO, 0.77; O , 0.77; H , 57; CO , 18; H O, 21; N , balance. The gas flow rate is 90 l (g Cat) h .
2
2
2
2
2
H is replaced with the feed in the 57th minute.
2
175
150
125
100
75
9000
7500
6000
4500
3000
1500
4
3
4
3
2
2
1
1
50
0
100 150 200 250 300 350 400 450 500 550
Time, min
Fig. 7. Ignition, quenching, and activation of the 1% Ru/Al O catalyst. In the 68th minute, humid H is replaced with the feed
2
3
2
mixture (for the mixture composition, see Fig. 1). From the 424th till the 446th minute and from the 452nd till the 469th minute,
the catalyst is activated with humid hydrogen containing ~21 vol % water (see the arrows). (1) Residual CO concentration, (2) fur-
–1 –1
nace temperature, (3) bed entrance temperature, and (4) bed exit temperature. The gas flow rate is 106 l (g Cat) h .
decrease at the bed exit and this will finally cause the tivation). During the reaction conducted at this low fur-
quenching of the surface and, accordingly, a dramatic nace temperature, the gas temperature at the bed
decrease in the gas temperature and conversion.
entrance gradually decreased. The gas temperature at
the bed exit first increased slightly (in agreement with
the shift of the hot spot) and then began to decrease.
The runs of both curves reflect the slow deactivation of
the catalyst. After ~400-min-long operation of the cat-
alyst, the critical surface quenching temperature was
reached and the reaction regime was changed: the gas
temperature in the catalyst bed decreased down to the
The results of an experiment illustrating the dis-
placement of the hot spot in the 1% Ru/Al2O3 bed and
the quenching of the surface are presented in Fig. 7.
After the replacement of H2 with the working mix-
ture (68th minute), the surface ignited. Next, the fur-
nace temperature was reduced from 109 to 94°ë and
was then maintained constant (as follows from Fig. 6, furnace temperature and the CO conversion fell dra-
reducing the temperature is favorable for catalyst deac- matically. After being exposed to humid hydrogen for
KINETICS AND CATALYSIS Vol. 49 No. 1 2008

100
ROZOVSKII et al.
200
180
160
140
120
100
80
100
80
60
40
20
4
3
2
4
3
1
1
290
60
140
190
240
340
Time, min
390
440
490
Fig. 8. H oxidation on the 1% Ru/Al O catalyst. The arrows point to the instants H is replaced with the reaction mixture or vice
2
2
3
2
versa. (1) Furnace temperature, (2) bed entrance temperature, (3) bed exit temperature, and (4) O conversion. Feed composition
2
–1 –1
(vol %): O , 098; H , 60; N , balance. The gas flow rate is 54 l (g Cat) h .
2
2
2
20 min starting at the 424th minute, the catalyst par-
Holding the catalyst at the furnace temperature of
tially regained its activity. After additional 17-min-long 85°C for ~200 min caused a gradual decrease in the
treatment starting at the 452nd minute, the catalyst sur- oxygen conversion to 31% and, accordingly, a decrease
face ignited on replacing hydrogen with the reaction in the catalyst bed temperature. For example, the gas
mixture and the initial catalytic activity was completely temperature at the bed exit is 101.2°C in the 231st
minute and 96.9°C in the 426th minute. In the ~430th
minute, the reaction mixture was replaced with hydro-
gen and the catalyst was held in flowing H2 for ~8 min.
Next, hydrogen was again replaced with the reaction
mixture. This replacement of the hydrogen flow with
the reaction mixture flow at a fixed furnace temperature
caused the catalyst surface to ignite. The ignition event
manifested itself as a dramatic increase of the O2 con-
version (to 99.7%) and of the gas temperature in the
catalyst bed (Fig. 8).
Thus, short-term exposure to hydrogen activates the
catalyst. Since there was no CO in the system in this
experiment, the deactivation of the catalyst in the oxi-
dation of CO and/or hydrogen can be due to some oxy-
gen species tightly bound to the active sites.
For better illustration of the macrokinetic features of
the system, we present, in Fig. 9, the complete record of
experimental data whose fragment is shown in Fig. 1.
In this experiment, the reactor was heated to 109°C
under flowing hydrogen, the admission of water was
switched on, and the H2 flow was replaced with a feed
(CO + O2 + H2 + N2) flow (47th minute). After the
admission of the feed, we observed the ignition of the
catalyst surface, which manifested itself as a rapid heat-
ing of the reaction mixture (the exit and entrance tem-
restored. Thus, hydrogen serves as a reactivating agent.
These results are not in conflict with the conclusion
that the deactivation of the ruthenium catalyst is caused
by some form of tightly bound oxygen appearing on the
surface [19, 20]. Exposure to hydrogen brings the deac-
tivated sites back into their active state. However, it is
not impossible that the deactivation of the catalyst is
due to the formation of tightly bound CO species on
active sites and that these species are removed upon
exposure to hydrogen or upon the admission of the
feed. If this were the case, by eliminating CO from the
reaction mixture, we could eliminate the cause of deac-
tivation.
Figure 8 displays the results of the experiments in
which hydrogen was oxidized in the absence of CO.
The catalyst was heated in hydrogen to 130°ë, cooled
to 70°ë, and purged with helium for 15 min. Next, the
helium flow was replaced with the feed flow (145th
minute; feed composition (vol %): O2, 0.98; H2, 60; N2
as the balance gas) and the furnace temperature was
raised in steps to 85°C. After the admission of the reac-
tion mixture, the catalyst underwent heating so that the
gas temperature at the bed exit was higher than the gas
temperature at the bed entrance. As the furnace temper-
ature was raised, the bed temperature grew and the oxy- peratures were 200 and 131°C, respectively) and an
gen conversion increased from 25.6% at a furnace tem- equally rapid decrease of the residual CO concentration
perature of 70°C (149th minute) to 40% at a furnace (down to ~130 ppm). Therefore, Tcr < 109°C. The igni-
temperature of 85°C (231st minute; the furnace temper- tion regime persists as the furnace temperature is
ature was raised to 85°C in the 205th minute).
reduced to 99°C. Further reduction of the temperature
KINETICS AND CATALYSIS Vol. 49 No. 1 2008

SELECTIVE CO OXIDATION ON A Ru/Al2O3 CATALYST
101
225
200
175
150
125
100
75
10000
8000
6000
4000
2000
4
4
3
3
2
3
2
1
50
1
25
0
100
200
300
400
500
Time, min
Fig. 9. Selective CO oxidation on the 1% Ru/Al O catalyst: complete record of the experimental data presented in part in Fig. 1.
2
3
In the 48th minute, H is replaced with the reaction mixture. Between the 48th and the 407th minutes, the gas flow rate is
2
–1 –1
–1 –1
87 l (g Cat) h ; later, 17.5 l (g Cat) h . (1) Residual CO concentration, (2) furnace temperature, (3) bed entrance temperature,
and (4) bed exit temperature.
to 89°C caused the quenching of the surface (119th
ACKNOWLEDGMENTS
minute).
This work was supported by the Russian Foundation
for Basic Research (grant no. 06-03-32848).
Subsequent heating of the furnace to 109°C did not
cause the ignition of the catalyst surface (Fig. 9). Igni-
tion was observed only at a furnace temperature of
144°C (380th minute). Since all oxidation conditions
were identical, this dramatic change in the critical igni-
tion temperature can be due solely to a change in the
activity of the catalyst. Thus, the activity of the ruthe-
nium catalyst depends strongly on the pretreatment
procedure.
REFERENCES
1. Edwards, N., Ellis, S.R., Frost, J.C., Golunski, S.E., van
Keulen, A.N.J., Lindewald, N.G., and Reinkingh, J.G.,
J. Power Sources, 1998, vol. 71, p. 123.
2. US Patent 6576208, 2003.
3. Echigo, M. and Tabata, T., Appl. Catal., A, 2003,
vol. 251, p. 157.
Decreasing the gas flow rate from 87 to
17.5 l (g Cat)–1 h–1 (407th minute) reduces the residual
CO concentration in the catalyst surface ignition
regime to ~20 ppm. A further decrease in the CO con-
centration can be achieved by reducing the furnace tem-
perature without going beyond the ignition limits. This
procedure (steps at 149, 141, and 137°C) allowed the
residual CO concentration to be reduced to 12 ppm.
4. Snytnikov, P.V., Extended Abstract of Cand. Sci. (Chem.)
Dissertation, Novosibirsk: Inst. of Catalysis, 2004.
5. Eur. Patent 1485202, 2003.
6. Snytnikov, P.V., Sobyanin, V.A., Belyaev, V.D., Tsyrulni-
kov, P.G., Shitova, N.B., and Shlyapin, D.A., Appl.
Catal., A, 2003, vol. 239, p. 149.
7. Kawatsu, S., J. Power Sources, 1998, vol. 71, p. 150.
Thus, by establishing the catalyst surface ignition
regime in some way, and by varying the gas flow rate
and the furnace temperature within the ignition limits,
it is possible to improve the performance of the selec-
tive CO oxidation process. The combination of the
advantages of the low-percentage ruthenium catalyst
and the selectivity gain provided by the catalyst surface
ignition regime enabled us to remove CO almost com-
pletely (down to a level of 10–15 ppm; see Fig. 6) at gas
flow rates of about 100 l (g Cat)–1 h–1 in the presence of
H2O and CO2 (~20 vol % each).
8. Dudfield, C.D., Chen, R., and Adcock, P.L., J. Power
Sources, 2000, vol. 86, p. 214.
9. Kipnis, M.A., Volnina, E.A., Samokhin, P.V., Lin, G.I.,
and Rozovskii, A.Ya., Tezisy dokl. 1 Vseross. konf.
“Khimiya dlya avtomobil’nogo transporta” (Proc. 1st
Russian Conf. on Chemistry for Automotive Applica-
tions), Novosibirsk, 2004, pp. 199.
10. Dudfield, C.D., Chen, R., and Adcock, P.L., J. Power
Sources, 2000, vol. 85, p. 237.
11. Rosso, I., Galletti, C., Saracco, G., Garrone, E., and
Specchia, V., Appl. Catal., B, 2004, vol. 48, p. 195.
KINETICS AND CATALYSIS Vol. 49 No. 1 2008

102
ROZOVSKII et al.
12. Worner, A., Friedrich, C., and Tamme, R., Appl. Catal., 17. Rozovskii, A.Ya., Kinetika topokhimicheskikh reaktsii
A, 2003, vol. 245, p. 1.
(Kinetics of Topochemical Reactions), Moscow: Nauka,
1980.
13. Han, Y.-F., Kahlich, M.J., Kinne, M., and Behm, R.J.,
Phys. Chem. Chem. Phys., 2002, vol. 4, p. 389.
18. Zhilyaeva, N.A., Volnina, E.A., Shuikina, L.P., and
Frolov, V.M., Neftekhimiya, 2000, vol. 40, no. 6, p. 422
[Pet. Chem. (Engl. Transl.), vol. 40, no. 6, p. 383].
14. Rozovskii, A.Ya., Kipnis, M.A., Volnina, E.A., Lin, G.I.,
and Samokhin, P.V., Kinet. Katal., 2004, vol. 45, no. 4,
p. 654 [Kinet. Catal. (Engl. Transl.), vol. 45, no. 4,
p. 618].
19. Kiss, J.T. and Gonzalez, R.D., J. Phys. Chem., 1984,
vol. 88, p. 892.
15. Rozovskii, A.Ya., Kipnis, M.A., Volnina, E.A., Lin, G.I.,
and Samokhin, P.V., Kinet. Katal. (in press).
20. Amann, J., Crihan, D., Knapp, M., Lundgren, E., Loffler, E.,
Muhler, M., Narknede, V., Over, H., Schmid, M., Seit-
sonen, A.P., and Varga, P., Angew. Chem., Int. Ed. Engl.,
2005, vol. 44, p. 917.
16. Frank-Kamenetskii, D.A., Diffuziya i teploperedacha v
khimicheskoi kinetike (Diffusion and Heat Transfer in
Chemical Kinetics), Moscow: Nauka, 1967.
KINETICS AND CATALYSIS Vol. 49 No. 1 2008