Note: Descriptions are shown in the official language in which they were submitted.
110~292
: `'
; This invention relates to a catalytic process for use in
the purification of air and other ga~eous media. More particularly, it
relates to the catalytic removal of one or more oxides of nitrogen,
especially in the presence of excess oxygen, and also to the catalytic
removal of carbon monoxide and organic compounds such as the lower hydro-
carbons.
The exhaust gases from an internal combustion engine include
- oxygen, carbon monoxide, unburnt hydrocarbons and nitrogen oxides. If the
engine is fuelled by petrol, the oxygen is generally present in a stoichio-
metrically lean concentration, that is, the atmosphere is a "net reducing
atmosphere". Under such conditions, the nltrogen oxides may be catalytically
reduced in one catalyst bed and the carbon monoxide and hydrocarbons may
be mixed with additional oxygen and catalytically oxidised in a second
catalyst bed. If, on the other hand, the engine is fuelled by diesel,
there is a stoichiometric excess of oxygen in the exhaust gases and the
nitrogen oxides may only be catalytically reduced after adding an excess
of a reducing fuel. The exhaust of a petrol~fuelled internal combustion
engine will also be oxygen-rich under certain operating conditions.
In the manufacture of nitric acid, tail gases contain
appreciable quantities of N0 and other oxides of nitrogen which it is
desirable to remove before venting
.
29Z
to atmospllcre.
Accordillg to one feature of the invention, a process
for the cat;alytic removal of oxides of nitrogen present in
a gas comprises contacting the gas at elevated temperature
with a catalyst comprising one or more metals from the
platinum group, gold and silver, supported on or associa-ted
with a refractory oxide.
By "the platinum group" we mean platinum, rhodium,
iridium, osmium, palladium and ruthenium, and of these we
particularly prefer to use platinum, rhodium or iridium.
The metals may be alloyed together, or may be present as
a mixture or alloy with one or more base metals or one
or more precious metals, other than those of the platinum
group.
The refractory oxide may be an oxide or a mixture
of oxides of elements of Groups IIA, IIB or IVB of the
Periodic Table, or of the transition metals. Such elements
are, for example, beryllium, magnesium, calcium, scandium,
yttrium, titanium, zirconium, aluminium, silicon or
germanium. Additionally, such compounds as the aluminos-
ilicates may be used. We prefer to use either silica or
alumina, or mixtures or compositions c,ontaining silica and
alumina, or an aluminosilicate.
A process according to the present invention will
remove oxides of nitrogen from a gas under either reducing
or oxidising conditions Under oxidising conditions, removal
may occur either by decomposition or by selective reduction.
Such a process represents a - -
--3--
Z92
considerable advance in the technology of purification of petrol
and diesel exhaust gases and nitric acid tail gases.
- According to a further feature of the invention, a process
for the catalytic removal of oxides of nitrogen and the catalytic
oxidation of carbon monoxide and hydrocarbons present in a gas
comprises contacting the gas at elevated temperature with a catalyst
comprising one or more metals of the platinum group supported on
or associated with a refractory oxide, where "platinum group" and
"refractory oxide"- have meanings as defined above. In its most
important embodiment, the present invention provides a process for
the catalytic reduction under oxidizing conditions for one or
more oxides of nitrogen present in the exhaust from an internal
combustion engine or from an industrial plant gas which comprises
contacting the gas under oxidizing condi*ions at an elevated tem-
perature with a metal catalyst comprising at least one metal
selected from the group consisting of platinum, rhodium, and
palladium metals supported on a particulate silica support, said
particulate oxide support being itself deposited on a catalyst
substrate.
Optionally, the catalyst may be associated with one
or more other catalytic metals or metal oxides to assist in pro-
moting the oxidation reaction.
Under certain circumstances, the catalytic decomposition
of nitrogen oxides may lead to the formation of nitrous oxide. A
further option, therefore, is that the catalyst be associated with
one or more catalytic metals or oxides to promote the reduction or
decomposition of nitrous oxide. We have found that oxides of one
or more of lanthanum, cerium and titanium are particularly useful
- 4 -
~.'
292
for this purpose.
By "elevated temperature" we mean a temperature sufficient
for catalytic decomposition of a significant quantity of the oxide
of nitrogen etc. present or catalytic oxidation of a significant
quantity of the organic compound
- 4a -.
~, ~
-
'
29Z
or carbon monoxide present to occur as a result of contact with a catalyst
such as described above. In general; the elevated temperature would not
exceed about 500-600 C. and, depending upon the exact nature of the catalyst,
the optimum temperature could be in the region of 300~450 C.
The refractory oxide on which the catalytic metal or mixture
of metals is deposited or with which it is associated, may be in the form
of pellets, granules or powders but is preferably in the form of fine grains,
the individual grains being typically slightly coarser than particles of
powder.
An example of a suitable refractory oxide is Davison 70
(T.M.) silica.
The catalyst is optionally ultimately supported on a substrate
known in the art. Examples of preferred substrates which may be used are
porous refractory ceramic honeycombs or modules formed from corrugated
metallic sheets rolled up to form a cellular structure. Other supports may
be used, however, forlexample particulate solids sueh as granules, pellets,
shapes and extrudates which can be constructed of ceramic or metallic
refractory materials. Alternatively, the catalyst may be mixed with a
catalytically inert refractory material, for example in the form of granules,
as an extender or diluent.
The ultimate supporting structure may have a first coating,
layer or deposit of a refractory oxide
- 5 .
,.~
29Z
which may be the same as or different from the refractory oxide on which
the catalytic metal is supported. This first refractory oxide is deposited
on the ultimate support, preferably in the form of a film, and the film may
have a thickness of 0.0004 to 0.001 inches.
Such oxides are calcined and preferably have a porous structure
which has a relatively large internal pore volume and total surface area.
Such oxides are referred to as "active", that is, catalytically active,
refractory oxides.
Refractory oxides suitable for this purpose may be selected
from the group consisting of silica, silumina, silica/alumina, titania,
zirconia, magnesia, quartz and proprietary molecular sieve zeolites.
Catalysts for use in the process of the invention may be
deposited directly on the support without the intervention of an intermediate
refractory oxide layer. Such an arrangement includes a catalyst comprising
a platinum group metal supported on or associated with, for example, silica
and deposited on a ceramic or metallic ultimate support, so that, although
no separate intermediate refractory oxide layer is included, the refractory
oxide of the catalyst acts in lieu of such an intermediate layer.
' , .
.
Z9Z
In addition to Davison 70 silica a range of colloidal
suspensions sold under the trade mark SYTON by Monsanto may be
used. For example, "Syton X30" which has a silica concentration
of approximately 30%, has been used for applying silica to both
ceramic and metallic monoliths prior to applying the precious metal.
The general procedure of preparation of such catalysts
is to dip the monolith directly into the "Syton" solution, blow
off the excess from the channels and surface of the monolith, dry
and fire, and then to apply the precious metal by dipping into an
aqueous solution of precious metal salt followed by drying and
reducing in a stream of H2/N2. Metal supports (e.g. Fecralloy
(T.M.) should be pre-oxidised by heating for 1 hour at 130C.
This has the effect of improving corrosion resistance to the gases
with which the catalyst will come into contact and also improving
the adhesion of the silica layer. Also, since metal supports,
unlike ceramic supports, are not porous, two immersions in the
silica sol, each followed by drying etc, are necessary to build
up the desired coating of silica.
In detail, one method of applying a ca alyst according
to the invention on a ceramic support will now be described.
A ceramic honeycomb monolith (Corning M20 (T.M.)~; 300
cells per square inch) made from cordierite and being in the shape
of a cylinder 2 inches in diameter and 3 inches in length and
therefore having a volume of 9.42 cu. ins. was immersed in 500 ml
of Syton X30, the silica therein having a surface area in excess
of 200 m2g-1, for 2 minutes. After withdrawal, the excess silica
sol was first shaken off
s~
.
29Z
and then blown off at an air pressure of 15 p.s.i. The weight of the monolith
had increased from 92 to 108 g. The ~oated monolith was then dried at about
100C in a convection current of hot air flowing at>300 ft.min. 1 and was
then roasted in a static kiln furnace at 500 C for ~ hour. The resulting
coated monolith weighed 97.2 g, hence the weight of silica was 5.2 g which
represented a loading of 0.55 g.in 3.
The film thickness represented by this loading falls within
the range previously mentioned namely 0.0004 - 0.001 inch.
To deposit the precious metal on the silica, the following
method was used:
The water absorption of the silica-coated monolith was
measured and the amount of precious metal xequired to give a loading of
30 g.ft 3 of monolith was calculated. ~n aqueous solution of precious
metal salt, for example, RhC13.xH20, at a concentration of 14 g.l 1 of Rh
metal was made up and the silica-coated monolith immersed for 2 minutes in
the solution, removed, allowed to drain and the excess salt solution blown
off. The process was repeated, if and as necessary, until the predetermined
weight, that is, of silica-coated monolith carrying the required quantity
of solution corresponding to a metal loadlng of 30 g.ft 3, was achieved.
The process was completed by hot air drying and reduction in a tube furnace
for one hour under an atmosphere of H2/N2 containing 5% H2 at 450 C (for
rhodium) and 220C (for iridium).
Methods of deposition of oxide are described in
- 8 -
Z92
our British Patent Specification No. 1,390,182. Descriptions
of typical ceramic supports and useful metal supports which may
be used are given in British Patent Specification No. 882,484
~Corning) and in our published DOS No. 2,450,664.
The following descriptions, examples, tables and
figures are illustrative of the invention.
Preparation of the Catalysts
- Catalysts comprising pla*inium, rhodium and iridium
supported on Davison 70 silica were prepared by impregnation of
the silica with a solution of the relevant chloride or acid
chloride of the required strength and volume. The impregnated
silica was then dried in vacuo (0.1 torr) at room temperature
and reduced in a stream of hydrogen at 350C. (for platinum)
and 450C. (for rhodium and iridium).
The following silica-supported catalyts were pre-
pared using the above method.
TABLE 1
` ' , , ,
. ~
Catalyst A Rh/SiO2 6.14 x 10 2At~ (approx. 0.1 w/w%)
Catalyst BRh/SiO2 6.14 x 10 At% (approx. 0.1 w/w%)
Catalyst CPt/Sio2 6.14 x 10 At% (0.2 w/w%)
Catalyst D`Pt/Sio2 6.14 x 10 At% (0.02 w/w%)
Catalyst EIr/SiO2 6.14 x 10 At~ (approx. 0.02 w/w%)
. .
g _
llC3~Z92
I
.
In the above table, atomlc percentages are quoted as atoms
of the metal calculated as a percentage of the number of moles of silica.
In a similar way, weight percentages are expressed as weight of metal to
! weight of silica.
Experimental Test Rig
The experimental runs, results of which are given below, were
¦ carried out in a silica reactor comprising two concentric tubes mounted
¦ vertically. The outer tube had an o.d. of about l~" and the inner tube
had an o.d. of about 3/8". The catalyst was carried on a bed of silica
wool in the inner tube and the reaction gas stream was passed down the
outer tube and up the inner. The gases were passed through a pre-heating
zone upstream of the catalyst. Reactor gas inlet temperature was measured
by means of a chrome-aluminel thermocouple located in a well below the
catalyst. The reactor was surrounded by a heater equipped to give variable
heating rates. The inlet and outlet concentration of gases was monitored
using a BOC Luminox Analyser and Perkin-Elmer ~17 gas chromatograph.
Experimental Results~and Discussion
Two principal types of experiment were conducted on the
catalysts. These were:
(a) a profile of activity in removal of oxides of
nitrogen versus inlet temperature at a fixed ratio
of oxidising to reducing species in the inlet gas
stream; and
(b) an isothermal profile of activity against ratio at a
temperature chosen from the result of experiment type (i).
The ratio (R) of oxidising to reducing species in the inlet
gas stream is calculated according to the equation
- lU ~
`'';~'
29Z
R 2 2 x %2 ~ ~ NOx
.
. ' ~ CO
~ Generally, the particular oxide of nitrogen (NOx) in the
~, equation is nitric oxide, NO.
In the experiments described below, the weight of catalyst
tested was generally lg. In some experiments, smaller weights were tested,
typically 0.005 g., but in all experiments using less than 1 g. of catalyst,
pure silica was added as an extender to the catalyst to give a total weight
of 1 g. in order to prevent bed fluidisation.
Space velocity was calculated on the basis of catalyst weight
exclusive of the weight of added pure silica, if present, and is quoted in
units of hr 1
Experimental runs were conducted as in the following table
(Table 2) in which the "Catalyst" designation refers to Table 1, the
"Experimental type" refers either to (a) or (b) as above, and the "R or
temperature" refers to the fixed ratio R (for experimental type ~a)) or to
the isothermal temperature (for experimental type (b)).
-- 11 --
1:
Z9Z
,................. ... _ .. . .. ~
PS E~ _ _u~ _~ u) ., . o o ~ ~
X ~
t~ ~ ~1 O oNO o o o oO o o o o o
,, 0~_1 C~ O O C~ `1' 0 0 ~ I V
O ~ N u ) ~ 0~ ~ V
~!~ ~ L~
u nl~0 0 G t~ O 0 ~ V
V ~ _
¢ ¢ ¢
._ . . Cd
Z~ ~
~ 1~ ~ O ~ ~`I ~ .-1
1~ . _ ~ 1~
~` 12;
11(3~Z9~
Run l shows the actiVity of a ~h/SiO2 catalyst at an R value
of 1.9 (i.e. oxidising conditions). NOx decomposition showed a maximum of
95% at about 230C. and the "half peak activity range" (that is, the active
temperature range at half the maximum activity) was 220C. Nitrous oxide
formation (as a percentage of the total NOx input) reaches a maximum just
above the NOx decomposition threshold temperature and decreases steadily to
zero at about 450C.
Run 2 was carried out under similar conditions to Run 1
using the same catalyst, except that the actual catalyst weight was 0.005g.
instead of lg. ~he effect of this was to raise the temperature of maximum
activity to 400 C. At this temperature, decomposition of NOx is still 75%
despite the small amount of catalyst and the correspondingly high space
velocity compared with Run 1. Formation of N20 did not rise above 18% and,
at the temperature of maximum decomposition of NOx, was already dropping
rapidly.
Run 3 was carried out using the same catalyst as Run 2 but
under isothermal conditions, a temperature near the maximum activity of NOx
decomposition from Run 2 being selected (395 C.). Maximum NOx decomposition
(and minimum N20 formation) occurred at about the stoichiomet~ic point (R=l)
and the value of R at 50% of the maximum NOx decomposition was 2.8.
Run 4 used ig. of a Rh/SiO2 catalyst of lower metal loading
compared to Runs 1, 2 and 3. Diluting the metal loading results in a lower
maximum of NOx decomposition at a higher temperature, and also a lower half-
peak activity of 160C.
Run 5 - the isothermal plot using the catalyst of Run 4 -
showed that the catalyst was less resistant to oxygen poisoning between R
values of about 1 and 5 and that production of N20 was higher than with the
- 13 -
I` 110~29Z
similar catalyst of higher metal loading. The value of R at 50% of maximum
NOx decomposition was 2.5.
Run 6 used lg. of a Pt/SiO2 catalyst and is comparable to
Run 1 using a similarly loaded Rh/SiO2 catalyst. The results of Run 6
indicated that the maximum activity in NOx decomposition for Pt/SiO2 was
only 50% compared with 95% for Rh/S102. However, this poor result does not
necessarily reflect accurately the true activity of thls catalyst since little
is known as yet about the relatlve dispersions of the metals on silica. The
half~peak activity or Pt/SiO2 was 200C. compared to 220C. for Rh/SiO2.
Run 7 used 0.005g. of the same catalyst and is therefore
comparable to Run 2. As in Run 2, lowering the quantity of catalyst increased
the temperature of maximum activity in NOx decomposition but the maximum
decomposition was only 17%.
Run 8 was carried out under isothermal conditions at a tempera-
ture of 400C. Compared to Run 3 tusing Rh/SiO2 catalyst) the resistance
to oxygen poisoning was poor. The value of R at 50% of maximum NOx decompo-
I sition was 1.25.
Run 9 was carried out using lg. of a Pt/SiO2 catalyst havinga lower metal loading than that used in Runs 7 and 8. Run 9 therefore compares
with Run 4 for Rh/SiO2 and the results show that Pt/SiO2 is inferior to
Rh/SiO2 in respect of NOx decomposition and about the same for half peak
activity.
Run 10 was carried out under isothermal conditions at a
temperature of 360C. Again, comparing the results with Run 5, Pt/SiO2 shows
less resistance to oxygen poisoning than does Rh/SiO2. The value of R at 50%
of maximum NOx decomposition was 1.7.
Run 11 used an Ir/SiO2 catalyst and compareswith Run 4 ~Rh/SiO2)
- 14 -
292
and Run 9 (Pt/SiO2). Maximum Nox decompositlon was 65% at 345 C. and at this
temperature production of N20 was vir-tually zero. Half peak activity was
190C
Run 12 was a re-run of Run 11, using the identical catalyst,
and shows the remarkable effect of heating to 600C. in Run 11. Maximum Nox
decomposition was 90% at a temperature of 350 C. and production of N20 was
still extremely low, peaking at 2%. The half peak range was comparatively
sharp at 100C.
Run 13, the isothermal experiment on Ir/SiO2 at 329C., shows
that this catalyst is considerably more resistant to oxygen poisoning than
either Rh/SiO2 (Run 5) or Pt/SiO2 (Run 10). The value of R at 50% of maximum
NOx decomposition was 5. N20 production is consistently either very low or
zero throughout the entire range of R.
In conclusion, it may readily be seen that these catalysts
show remarkable results for the decompositlon of NOx under oxidising conditions.Furthermore, all of them showed high efficiency in the oxidation of carbon
monoxide and it is anticipated that these catalyst, either alone or in asso-
ciation with known oxidation catalyst, for example alloys of the precious
metals such as Pt/Rh, will successfully oxidise the lower hydrocarbons typically- 20 present in the exhaust gas stream from an internal combustion engine.
Rh/SiO2 and Pt/SiO2 tended under certain conditions to cause
the formation of nitrous oxide. We have found that certain additives may be
incorporated or admixed with the catalysts to catalyse the reduction or de-
composition of nitrous oxide. Examples of such additives are oxides of
lanthanum, cerium and titanium.
- 15 -
