Note: Descriptions are shown in the official language in which they were submitted.
1 155~27
The present invention relates to novel catalyst struc-
tures prepared with use of glass fiber products.
Granular catalysts 5 to 10 mm in size have been generally
used for vapor-phase catalytic reaction systems. Such catalysts
have been found empirically favorable with respect to minimizing
pressure loss of the reactant gas in the catalyst layer and
minimizing clogging of the catalyst layer, and further from the
viewpoint of economy. On the other hand, however, such granular
catalysts do not permit gaseous reactants to diffuse into the
catalyst effectively, that is, fail to permit a high rate of mass
transfer. A great difference therefore occurs in the reactant
concentration within the grains or pellets, with the result that
the catalyst is low in the effectiveness factor defined by
Thiele et al. Thus there is the likelihood that almost all the
charge of catalyst is unable to function substantially effectively.
The low catalyst effectiveness factor poses a serious problem with
expensive noble metal catalysts and is very unfavorable
A ~,~
1 ~55827
economically even in th. case oL re~Lallvely incxpensive
catalysts o~ metallic oxides ln resp(-ct oi` the pressure
loss of l~eactant gaC;eS~ ~he size o~ the reactor, -the
~electivi-ty oL react;ion, etc.
Generally catalysts for use in indus-tries must
f`ulfill the Lollowing requirements:
1) High activity per unit weight o~ the catalys-t.
2) High activity per unit vol~ne of the reactor.
3) ~alall pressure loss of the reactant gases in the
catalyst layer.
4) High overall strength enabling -the catalys-t to fully
vli-thstand the impact of charging.
5) High surface strength against -the external forces
to be exerted on the catalyst during use.-/
~) Reduced variations in activi-ty despite the lapse of
time.
7) ~ow CoGt.
The true activity of a catalyst, free fro~
rnass transfer resistance, per unit weight thereof is
dependent on the composition of the catalyst, the structure
of the crystals thereof, etc. and is inherent in the
catalyst, but the actual activity varies with the grain
eiæe, the pore structure of the grain, the flow rate, i.e.
linear velocity, of the reactant gas, etc. Generally
the activity increases with decreasing grain size,
1 155827
ncrea~in~ Pore size, incre~sin~ l)ore vollime an~i
nc~ea~in~ L`low rate of` the react~lrlt ~;ac,. Nevertheless,
the~e l~lctors contribu-ting to the lncreElse ol'' ac-tivi-ty
are entirely in conI'lict with the require~ments in respect
of` the reac-tant gas pressure lo~,s and catalyst strength.
Accordingly it is extremely dilficult -to l`ull'ill all the
f`oregoing requirements l) to 7).
In order to satisfy -t'le above requirements l)
to 7), several catalysts OI the honeycomb tvpe have been
~eveloped in recent years. These catalysts include those
prepared by forming a paste from a catalytic component
and a binder, extruding the paste into a honeycomb body
and baking the body under suitable condi-tions, and those
prepared by making a honeycomb s-tructure from ceramics
having no catalytic ac-tivity and depositing a cataly-tic
component on the surface of the structure with a binder.
Ylith the former catalysts, there is the need to form a
honeycomb body with a large wall thicknese~ and to obtain
a compacted baked 'body which is less amenable to the
diffusion of reactant gases into the catalyst, in order
to give the desired strength to the catalyst. It is
therefore impossible to afford an improved catalyst
effectiveness f'actor. Thus difficulties are encountered
in improving both strength and activity at the same time.
With the latter oase, the to~gh ceramics horeycomb
1 1S5827
structure has ideal overall strength, fulfilling the requirement
4), but the use of the binder reduces the inherent activity of the
catalyst. Further the layer of deposited catalytic component,
which is made very thin to assure high surface strength 5) and
high activity per unit weight 1), reduces the activity per unit
volume of the reactor 2) and impairs the stability of reaction.
SUMMARY OF THE INVENTION
The main object of the present invention, which has been
accomplished in view of the foregoing problems, is to provide novel
catalyst structures better adapted to meet the foregoing require-
ments 1) to 7) for industrial catalysts.
The invention provides a catalyst structure comprising a
shape retaining metal core having a desired shape, a catalyst
particle holding layer attached to the core and made of a glass
fiber product having a plurality of filament-to-filament inter-
stices, and catalyst particles held in the interstices of the hold-
ing layer.
The invention will further be described, by way of
example, only, with reference to the accompanying drawings,
wherein:-
~ - 4 -
.~,
1 155827
BRIF,i~ S~RIP~ N ~ Hl, ~RA~'IlNG',
!~'igs. 1 ~ln,d 2 are sectional views showing
sandwich-type c~-talys-t c,truc-turec,;
l~'igs. 3 and 4 arC,- front views partl,y in section
and showing col~nnar catalyst structures;
l~'ig. 5 is a gr.lph showing the relation between
the concentra-tion of slurry and t;he amount of' ~articles
held by glass clo-th;
~ ig. 6 is a view in vertical section showing a
reactor;
E'ig. 7 is a graph showing reactivi-ties and
selectivities at varying reaction temperatures; and
Fig. 8 is a pers?ective view showing a reactor.
DESCRIPTION 0~' THE PRE~'ERRED E~IBOD~IENl'S
The ~hape retaining core gives the desired
overall ~trength to the catalyst structure. The core is
in the shape of a flat plate or cylinder or is otherwise
shaped suitably in accordance with the shape of the
reactor. The core is made generally of steel.
IJseful glass fiber products include nonv~oven
fabric, paper, rnat, cloth and roving cloth of glass fibers.
Throu~r,hout the specification and the appended claims, these
terms rnean the follovling.
The glass nonwoven fabric or glass paper is a
planar web of glass filaments which are bonded to one
1 1S5827
another ~Jit~l an adhesive. 'l`he web is nonvloverl f~bric
when long t'lber~ are used, or is paper ~hen short i`ibers
are used.
The glass mat is ~repared by forming a wad
of specified -thickness i`rom rovings Or glas~ f`ilaments
anci rll2,h-ing -the wad into a mat.
The gla5S cloth is woven of' yarns made by
twisting from a multiplicity of glass filaments about
5 to 15 ~ in diameter. The weaving method is plain
weave~ twill weave, satin weave or the like.
The glass roving cloth is prepared by forming
rovings from long glass iibers and ~eevin~ the rovings
into a cloth without twisting.
All of' these glass f'iber products ~ e
i'ilament-to-filament interstices which hold particles
therein and permit dif'fusion of gaseous reactants there-
through. Thus these glass f'iber products are useful as
a component of the catalyst structure of the invention.
Since the glass nonwoven fabric, paper and mat
each have a large number of f'ilament-to-filament
interstices, and further because component filaments
thereof are easily shiftable relative to one another when
subjected to an external ~'orce, they hold a large amount
of' cataly~t particles therein. Accordingly these non-
~loven fabric, paper and mat are well suited as ~ass f'iber
1 1 55827
~r;duct3 L'or formin~ pa~-tlc~le holdin~~L-~!er~-.
i~`ibers are incor~orclted into t;he ~ C; cloth
with increacecl tension and wi-th a reduced number of
filament--to-filarnent interstice6, whi1e the component
1'i1aments thereof' are lescs shiftable relative to one
another, therefore hold a smal]er amount of catalyst
par~icles therebetween, but will not permit the release
of such p~rticles therefrom when subjected, for exam~le,
to vibration. Accordingly the glass cloth is well suited
as a glass fiber product for forming the layer for
preventing release of particles.
The glass roving cloth as it is is similar to
usual cloth, but when the roving cloth is subjected to
a needling, brushing or like process to nap one side
thereof, the napped surface is given suitable properties
to hold ~articles. Accordingly when glass roving cloth
having a napped surf'ace is fitted to the core with the
napped surface inside, the roving cloth serves as a
glass fiber product having the functions of both the
particle holding layer and the release preventing layer.
I]seful catalyst particles are those not larger
than 100 mesh. Especially particles in the range of 1 to
20 ~ are most preferable since they are easy to hold but
diff'icult to release. Particles srnaller than 1 ~ in size
are ~omewhat easily releasable but are still satisfacto-
1 ~55827
rily IsabLe t~n(ler conc.l~;ions ~'ree ol' st;rong n~echanicalvibrstion. Although par~lc]es exceed:Lng 20 ~ in size
have ~il`f`icul-ty in entering ~'ilarnent-to-:t`il.ament inter-
~tices of ',he ]ayer, a sufficient amount of` such particles
can be held in the la,~er if` the layer is pressed with
ruhber rollers a~ i~creased nu~lber of -times as wi]l be
described later.
To cause a glass fiber product to hold catalyst
particles ? the Produc~ is immersed in a slurry comprlsing
water or a suitable dispersing medium and catalyst
particles and having a viscosity oI` 100 to l~00 cps. It
is deirable to press the glass fiber product wi-th
rubber rollers in the slurry to force the slurry into
the filament-to-filament in-terstices and the~Je,by cause
the product to hold -the par-ticles effectively To
squeeze an excess of the slurry from the glass fiber
product, the product is then passed between rubber rollers
under a pressure of 5 to 30 kg per unit length (m) of the
rollers. The wet product is thereafter fitted to a
shape retaining core and fixed thereto with fasteners.
The glass f'iber product thus fitted to the core is dried
and, when desired, baked. The fiber product may be dried
first and then attached to the core. In this case the
product will release some amount of dust during handling,
so that care should be taken to assure a good work
1 155827
enviror~lent.
In this way, a calaLyct structure of' t,his
invention ic~ obt;ained.
~ l'he cat21yst structures o~ this in~rention can
be in various forms. Llligs. 1 to 4 sho~ t;~pic21 examples.
l'he catalyst structure shown in ~`ig. 1 is of the candwich
type. This drawing shows a metal core 1 in the f'orm of
a flat plate, a pair of planar particle holding layers
2 fitted to the opposite sides of the core 1, and fasteners
3 for fixing the layers 2 to the core 1. The catalyst
structure shown in ~ig. 2 further includes ~lanar layers
4 for preventing release of catalyst particles which
layers are fitted to -the outer sides of the pair of
particle holding layers. Fig. 3 shows a columnar catalyst
structure comprising a cylindrical metal core 11, a
particle holding layer 12 f'i-tted around the core 11, and
fasteners 13 for fixing -the layer 12 to the core 11.
Fig. 4 shows a catalyst structure further comprising a
release preventing layer 14 covering the particle holding
layer 12.
The catalyst structures thus constructed
according to the invention have the f'ollowing advantages.
The core, which is made of metal, gives high
mechanical strength to the structure.
The particle holding layer, which is a glass
_g _
1 155827
fiber ~?roduc-l, ellec-tively hol~ a large amount of`
cataly~!t particles therein without using any binder,
permits efficient cll1`fusion of reactant gases through
the interior of` the str~cture Wit~! a reduced pressure
loss and a3~ure~ efiective contact between -the gases
an~ the catalyst, enablin~ the catalyst to exhibit
exceedingly high activity.
The materi~ls of the structure are all
inexpensive and easily available.
Accordingly the Gatalyst structures of the
invention are ideal and fulfill all the f`oregoing
requirements l) to 7).
Reference ~xample l
! Different kinds of particulate materials were
checked for differences in the amount of particles held
by glass cloth.
A glass cloth was prepared which was 0.2 mm
in thickness, 200 g/m2 in weight and l9 yarns/25 mm in
yarn density and which was woven of yarns each having
one twist/25 mm and composed of 800 glass filaments
14 ~m in diameter. Surries of varying concentrations
were prepared by adding water to 200- to 500-mesh
particles of r-alumina. A piece of the glass cloth was
immersed in each of the slurries and pressed with rollers
~wice in the slurry. The piece o~ glass cloth thus
--10--
1 155827
im~regn~ted with the ~ url~y ~.as dried ~t 100 C for l
'rlour and th~n baked or f'i~r?(l at 4()0 C f'or 3 hours.
The amount ol parlicles held in each of' the cloth pieces
thus treated ~as measured.
The same proced~lre as above was repeated
cxcept that in place of the Y-alumina particles, silica-
alurnina partlcles or magnesia ~articles, the same as the
r-alumina particles in size, were used. The ~mount of
particles held in each of the resulting clo-th pieces was
measured.
The relation between the concentration of slurry
and the amount of' particles held in the glass cloth was
deterrnined f'or each kind of the particles. ~'ig. 5 is a
graph showing the results. The graph reveals that the
arnount of each kind of particles increases with the
increase of the slurry concentration.
Reference ExamPle 2
Diff'erent kinds of glass fiber products were
caused to hold particles therein and checked for differences
in the amount of the particles.
The same glasæ cloth as used in Reference
Example 1 was used as a glass fiber product. Water was
added to 200- to 500-rnesh magnesia particles to prepare
a slurry havin~ a concentration of 25~ by weight. In the
sarne manner as in Reference Example l, the glass cloth
1 155827
was caused -t,o hol(l magrlesia part,ic~es, an~l the ~mount of
particlec held therein was meas~red.
The C,eme procedure ac, above was repeated with
the exceo-tion oL using ~lass non~oven fabric, mat or
roving cloth na-pped on one side ~by needling with
108 needles/cm2) instead of the glass cloth. The amount
of magnesia particles held in each of the resulting glass
fiber products was measured.
Table 1 shows the results.
rl'able 1
Glass fiber product Particles
K d Weight* Amount Amount held/weight
1n (kg/m ) (kg/m3) of fiber product
-
~loth 1000 550 0.55
Nonwoven 600 520 0.87
Mat 80 650 8.1
Napped roving 470 620 1.32
cloth
* Y~eight of product per apparent unit volume of
the product.
** Weight of particles per apparent unit volume of
the fiber product.
Example 1
~he same glass cloth as used in Reference
Example 1 was caused to hold magnesia particles in the
-12-
- 1 155827
same manner as in ~`~eference Exam~11e 2, and ~Nas then
attached to -the opposite ~ides o~` a core o~ stainle~s
steel pla-te, 35 mm x 50 mm x 1.5 mm, whereby a cat~lyst
struc-~ure A of the sandwich type sho~n in l~'ig. 1 was
prepared. Catalys-t structures B, C and D ~ere prepared
in the same manner as above except tha-t glass lionwoven
fabric, glass mat or glass roving cloth naped on one
side was used for the particle holding layers in place
of the glass cloth. The roving cloth, the same as -the
one used in Reference Example 2, was attached to the
steel plate with the napped surface inside.
Example 2
~ lass nonwoven fabric, mat and cloth were caused
to hold magnesia particles therein in -the same manner as
in Reference Example 2. In the same manner as in
Example 1, the glass nonwoven fabric was attached to a
steel panel serving as -the core to provide particle
holding layers therei)n. The glass cloth having magnesia
particles held therein as above was placed over the
holding layers to provide particle release preventing
layers thereon, whereby a catalyst structure E of the
~ame construction as shown in Fig. 2 was prepared.
Further a catalyst structure F was prepared in the same
manner a~ above except that the above glass mat having
magnesia particles held therein was used in place of the
1155827
glass nonllJo~en f`abrlc to nrovide particle holding layers.
'['~ble ,' shows thc thicknesses and weights of
the layers lncluded in the catalyst s-tructures prepared
in the foregoing exa~ples, and also the Qmounts of
magnesia held ln -the layers.
Tab]e 2
Cata- Holding la~er Preventing ~yer Thick- Amount
lyst ness of of` mag-
struc- ~'iber Weight ~iber Weight layer* nesia
ture product (g~m2) product (g/m2) (mm) (g/m2)
A Cloth 200 - - 0.2 llO
Non-
B woven 300 - - 5 260
fabric
C hiat 400 - - 3 1900
! Napped J
roving 570 - - 1.2 - 260
cloth
Non-
E woven 300 Cloth 200 0.8 380
fabric
F ~lat 400 Cloth 200 3.1 1900
* The thickness of the particle holding layer or
the combined thickness of the holding layer and
the particle release preventing layer.
Reference Example 3
The catalyst structures A to ~` were tested for
abra~ion resistance.
~ he catalyst structure A was fixed to the center
-14-
- 1155827
of a 30-mesh metal screen having a diameter of 300 mm,
200 g of alumina ba'lls with a diame-ter of' 5 mm were -~laced
into the screen, and the screen was vibrated on an
automa-tic screening device (290 vibraiions/min ~ith an
amplitude of 30 mm) f'or 4 hours. The amount of catalyst
released was measured l hour, 2 hours and 4 hours af'ter
the start of the vibration to determine the amount of'
particles held in the struc-ture. The same p-rocedure as
above was repeated for the catalyst structures B to P'.
Table 3 shows the results.
Table 3
Catalyst Amount of magnesia held (g/m )
structure Initial In 1 hour In 2 hours In 4 hours
_
A 110 90 86 84
B 260 180 llO 40
1900 llO0 620 120
D 260 235 228 225
E 380 345 330 321
1900 1550 1420 1400
~ able 3 reveals that all the structures have
great ability to hold particles.
Reference Example 4
The catalyst structures A to F were tested f'or
re~i~tance to vibrations.
1 155827
l'he catalyst structure A was fixed to a 6-
mesh metal ~creen and ~ubjected to vibrations for 2 hours
by a 50-H~ vibrator mounted on the structure A. The
amount of` particles released was measured 0.5 hour, 1
hour and 2 hours after the start of the vibration to
determine the amount of particles held in the structure.
I~he same procedure as above was repeated for the catalyst
structures ~ to ~. Table 4 shows the results.
Table 4
Catalyst Amount of magnesia held (g/m2~
structure Initial In 0.5 hour In 1 hour In 2 hours
A 115 108 105 100
B 260 202 145 J~
C 1830 1620 1400 -~ 1250
D 255 250 248 245
E 390 350 325 318
1900 1750 1720 1700
Table 4 reveals that all the catalysb
structures have high resistance to vibrations.
Example 3
A 300 g quantity of a-alumina particles not
larger than 200 mesh in size were placed into a solution
o~ 11.8 g of a~monium metavanadate, 3.8 g of ammonium
molybdate tetrahydrate and 8 g of oxalic acid in 130 g
-16-
1 155827
ol water. 'l~he mixture was eva~orated to dryness wi-th
stirring in a vacuum at a temperature of up to 60 C
and was iurther fired at 400 C for 3 hours. The fired
product was pulverized to obtain 200- to 500-mesh
catalyst particles, to which water was aclded to ~prepare
a ~lurry having a concentration of 25~ by weight.
A strip of glass roving cloth napped on one
side and having a thickness of 1.2 mm, wid-th of 20 mm
and weight of 570 g/m2 was immersed in the slurry and
pressed with rollers twice in the slurry. The strip
thus impregnated with the slurry was fitted, with the
nap inside, around a mild steel cylinder 18 mm in
diameter and 80 mm in length and fastened thereto at
its upper and lower portions. The strip was then dried
at 100 ~ for 1 hour and thereafter fired at 400 C for
1 hour, whereby a columnar catalyst structure was
prepared which had the construction shown in Fig. 3, with
the roving cloth covering measuring 60 mm in length. The
amount of particles held in the cloth layer was 240 g/m2.
Activit~ test
~ he columnar catalyst structure (having about
0.9 g of catalyst particles held therein) indicated at
22 and prepared in Example 3 was placed into a large-
diameter portion 21, 23 mm in inside diameter, of a
reactor of quartz as seen in ~ig. 6. The reactor was
1 ~55827
controlled to a predetermined reaction temper~lture. A
benzene-~ir gaseous mixture (with a benzene concentration
o~ ~.8 to o.gqO) was led in~o the reactor throu~h a
small-diameter portion 23 at a flow rate of 60 N~/h,
and the resulting outflow was collected from the upper
end of the large-diameter portion 21 to measure the
amount of maleic anhydr~ide formed by the oxida~ion of
benzene. The same procedure as above was repeated at
vary1ng reaction temperatures. The same procedure as
above was further repeated at varying reaction tempera-
tures with thè exceptlon o~ using a benzene-air mixture
having an increased benzene concentr~ation of 1.2 to 1.3%.
~or comparison, 0.9 g of 6- -to 8-mesh
I a-alumina 300 grains were admixed with 9 g o~ - to 8-
mesh ~ilica glass grains to obtain a catalyst ln the
~orm of~small pellets, which~was tested for activity
under the same conditions as above. ~ig. 6 further
shows a heater 24, a molten salt bath 25 and a the~rmo-
couple 26.
Reactivities and selectivities were calculated
~rom the above measurements at each of the reaction
temperatures acoording to the following equations.
Reactivity - (1 ~ Benzene cConncnn of ¢htargeW)x 100
-18-
I
1 1558~7
~Mount Or maleic anhydride
~electivity = lor~med (mol/h) x lO0
Amount o~ benzene
reac-ted (mol/h)
The resul-ts are shown in l~'ig. 7, which reveals
that the columnar catalyst structure of Example 3 is much
more excellent than the pellet cat~lyst in both reactivity
and selectivity.
Example 4
To a slurry composed of 30 parts by weight of
anatase-type TiO2 particles not larger than 200 mesh and
70 parts by weight of water made acidic with sulfuric
acid and having a pH of 2 to 3 was added ammonium meta-
vanadate in an amount corresponding to 0.05 in atomicratio relative to the TiO2. The mixture was stirred at
room temperature for 5 hours to dissolve -the ammonium
metavanadate in the slurry and cause the TiO2 to adsorb the
vanadate radical, affording a slurry of TiO2 retaining
vanadium therein. Glass roving cloth weighing 300 g/m2
was immersed in the slurry and pressed with rollers
twice in the slurry. ~he cloth thus impregnated with
the ~lurry was attached to the opposite sides of a
stainless steel plate, 50 mm x 29 mm x 0.3 mm, then dried
at 150 C for 1 hour and thereafter fired at 400 C for
3 hour~, whereby a catalyst structure G of the sandwich
type shown in Fig. l was prepared.
--19--
1 155827
Cata~yst structures H to N were prepared in
the same manner as above wi-th the exception of' using the
glass fiber products shown in Table 5 in place o~ the
glass roving cloth. The napped roving clo-th was used
with the nap inside.
Table 5
~atalyst Glass fiber productNumber ~ 2*
G Xoving cloth (300 ~/m ) O
H Napped roving cloth (300 g/m2) 162
I Napped roving cloth (300 g/m2) 216
J Roving cloth (570 g/m2) 0
K Napped roving cloth (570 g/m2~ 108
L Napped roving clo-th (570 g/m2) .J 216
1~ Nonwoven fabric, (300 g/m )
1 0.8 mm in thickness
N ~loth, 0.2-mm-thick (200 g/m2)
* The number of needles used for napping.
Comparison Example
One hundred parts by weight of TiO2 powder
(not larger than 44 ~ in particle size, and 150 m2/g in
~urface area) and 100 parts by weight of colloidal silica
(containing about 2010 by weight of SiO2) serving as a
binder were thoroughly mixed together to obtain a slurry.
The slurry was applied to the opposite sides of 18-mesh
-20-
1 155827
metal net-ting (30 mm x 50 mm in size an(l macle of ~1rires,
0.5 mm in diameter, of SUS 304 steel), then dried et
100 C ~`or 1 hour and thereafter fired at 400 ~ for 3
hours, whereby a 'liO2 carrier was ~`ormed on the netting
in the ~orm of a plate abou-t 0.8 mm in thickness. ri`he
netting was immersed in a 2N oxalic acid solut~ion of
ammonium metavanadate (1.0 mole/liter) at room temperature
for 30 minutes, then withdrawn from the solution, there-
after dried at 100 ~ for 1 hour and further fired at
about 400 ~ for 3 hours, affording a catalyst 0 having
the catalytic component suPported on the metal ne-tting.
Activity test
Into a tubular stainless steel reactor 32 having
a reactant gas flow channel 31, 6 mm x 30 mm, were placed
three catalyst structures 33 as arranged in a row
longitudinally of the channel and supported by metal nets
34 a~ shown in ~'ig. 7. A test reaction gas having the
composition shown in Table 6 vras passed through the
reactor 32 at a flow rate of13 N~/min to test the catalyst
for activity to reduce N0 with NH3.
1 155827
l'able ~
Component Proportion (by volume)
N0 Abou-t 150 ppm
NH3 About 150 ppm
H20 10~/f
C o 2 1 05o
2 5%
N2 Balance
In this way, the catalyst struct~lres G to N
and catalyst 0 were te~ted for activity at va.rying
reaction temperatures to calculate values K defined by:
K = -(AV) x ~n(l-x) l~
where: AV = ~'low rate of reaction gas (Nm3jh)
Geometric surface area of cataly~t (m )
x - N0 reactivity
'The relation between K and the amount of TiO2 supported
was established. Table 7 shows the results.
-22-
1 155827
Table 7
t,atalystAmount of K
structureT .' 2 200 C 250 C 300 C 350
G 62 10.3 29.8 47.8 69.0
H 255 30.7 66.4 98.9 132
I 3t)5 32.1 ~7-9 100 137
J 7'0 10.9 28.2 41.4 54.1
K 362 35.1 67.0 103 141
L 361 36.7 73.4 122 172
120 17.5 42.2 70.0 99.5
N 65 9.8 24.9 36.2 47-3
0 400 12.3 26.3 40.9 49.4
Table 7 reveals that the catalyst structures
G to N of Example 4 have higher K values per unit amount
of TiO2 and higher activity than the catalyst 0 of
~omparison Example.
