Special Suppplemellt to Chemical Engineering Science, vol. 8. 1!)5 ·~.

Oxidations carried out by means of vanadium oxide catalysts

P. MARS and D. W. VAN KREVELEN

Staatsmljnen in Limburg, Central Laboratory, Geleen, Netherlands

Summary-All technically interesting reactions carried out with vanadium oxide catalysts
arc marked by their highly exothermic character, whieh forms an impediment to the Investi-
gation of the kineties of these processes. In the present study use was made of a fluid bed, in
which the temperature is uniform. The oxidation of the following substances: benzene,
toluene, naphthalene, and anthracene has been studied,
The partial pressures of the reacting substances were varied to the greatest possible extent.
Both reaction components appeared to influence the reaction rate. A formula depicting this
influence is derived. This formula may be interpreted by assuming two succes sive reactions,
namely the reaction between the aromatic and the oxygen on the surface, and the re-oxldatlon
of the partly reduced surface by means of oxygen.
The formula may be reduced to an equation by which also the data on the oxidation of sul-
phur dioxide by means of vanadium oxide catalysts found in the literature are well described.
Using kinetic data it is possible to determine the optimum temperature distribution in a fixed
bed reactor used for the oxidation of sulphur dioxide and to make calculations of the ratio be-
tween the amounts of catalyst to be used in the various stages of a multiple-stage reactor. The
results of these calculations have been compared with practical experience.

Resume-Toutes les reactions qui sont importantes du point de vue technique et sont effectuees
a l'nide de cntnlyseurs d'oxyde de vanadium se distinguent par leur caractere tres exothermique
qui entrnve I'exumen ·d e In cinetique de ees processus. C'est pourquoi on utilise, dans l'etude
presentee, un lit Iluidille dans lequcl Ia temperature est uniforme.
Dans cet examen, on n etudle l'oxydation des substances sulvantes : benzene, toluene, naph-
tnlenc ct anthracene. Les pressions partielles des substances de reaction sont vnrlees dans In
mesure du possible. II paraissait que les concentrations des deux reactants influencent In vitessc
de reaction. La formule indiquant eette influence a etc etublle, Cette forrnule s'Interprete en
admcttant deux reactions sueeessives , savoir Ill. rl!action entre I'arornatc et l'oxyg~ne ~ In surface
et In reoxydation, a l'aide d'oxygenc, de In surface partiellement reduitc,
Sous certaines conditions, ccttc formulc peut Nre transformec a une equation indiqunnt les
donnees, trouvecs dons In litternture, de l'oxydation d 'oxydc sulfurcux a I'aidc de eutnlyseurs
d'oxyde de vanadium.
Les donnees cinetiques pcrmettent de determiner la distribution de In temperature optimum
dans un rCucteur a lit fixe utilise pour I'oxydntlon d'oxyde sulfureux et de ealculer Ie rapport
entre Ies quantites de cntalyseurs a utiliser dans un reacteur a multiples etages, Les resultats
de ces calculs sont compares a l'expcrience pratique.

Zusnmmenfnssung-Siimtliche in tcclmischer Hinsicht intcressenerregendc Reaktionen, die

mit Ililfe von Vnnadiumoxydc-Kntalysatoren durehgefUhrt werd en, untcrseheiden sich durch
ihren stark exothermen Charakter, der die Untersuehung dcr Kinetik dieser Prozesse ersehwert,
Dies ist der Grund dafiir, das s bel der betrellenden Untersuehung cine Wirbelschicht ungcwandt
wird, in dem cine uniforme Temperatur vorherrscht. Es wurde die Oxydation nnchstehender
Stoffe untcrsueht: Denzol, Toluol, Naphthnlln und Anthrazen. Die Partialdrlieke der
Reaktionsstoffe wurdcn zwischen moglichst weiten Grenzen va riicrt. Die Konzcntmtionen bcider
Reaktionskomponenten beeinlliissen die Renktlonsgeschwlndigkclt. Die kinetisehe Gleiehung
liisst sieh Intcrprcticrcn, indem man zwci aufeinanderfoigende Renktionen unnimmt, nlimlieh die
Heaktion zwischen der aromatischen Verbindung uml dem SauerstoU uuf der Oberlliiche und die
Oxydatlon dcr tcilwcise reduzierten Obcrfliiche mittcls Sauerstoff.
 CONFERENCE ON OXIDATION PROCESSES

Unter Umstanden lasst sich dicse Formcl in eine G1ciclmng umwnndcln, in der auch die
Litcraturunterlagen tiber die Oxydatlon von SehweCeldioxyd mittels Vllnlldiumo),,-yd-Klltlllysa.
toren, gut zum Ausdruck gelangen.
An Hand der Kinetik ist es mogllch die optimale Temperaturvertcilung in eincm Festbctt-
Rcnktor, der zur Oxydation von Schwefcldioxyd angcwandt wird, zu ermittcln und das VcrhaItnis
zwischen den in eincm mehrstullgcn Reaktor angewandten Katlllysatormengen zu bercchren,
Die Ergebnisse dieser Bcrechnungen sind mit den Erfahrungen nus der Praxis vcrglichcn norden.

INTRODUCTION reduces the catalyst, which Was oxidizedby oxygen

In various manufacturing processes vanadium from the air. The oxidation reaction preper can
catalysts play an important part. In particular, be carried out without oxygen being present: by
the oxidation of sulphur dioxide with air (the heating, say, anthracene with nnadium perr-
.. contact process "), the oxidation of naphthalene oxide in the absence of air a good-yield of the
to phthalic anhydride and, to a less degree, the oxidation product (anthraquinone) is obtained. In
oxidation of benzene to maleic anhydride over this proccss the catalyst is reduced to vanadium
this type of catalyst arc widely applied in tech- tetroxide (SENSlllIAN [1]). However, a catalyst
nical processes. In an analogous manner toluene which is in interaction with an oxidizable sub-
can be oxidised to benzaldehyde and benzoic acid, stance and air will also be partly in the tetravalent
anthracene to anthraquinone, methyl alcohol to state. This was found by W. SCHREIDER [2], who
formaldehyde, whereas from aliphatic hydro- weighed a catalyst actively engaged in the oxida-
carbons a large number of oxidation products tion of naphthalene. Previously, J. 1\L WEISS [8]
(aldehydes, ketones and acids) can be derived. had already arrived at the same 'conclusion by
Precisely those reactions which arc of great determining tetravalent vanadium in the catalyst
importance in chemical industry arc strongly in the benzene oxidation.
exothermic. This has been a difficulty both for By comparing the activities of one catalyst in
the technical realization and for the studying of various reactions, additional information about
the kinetics of these reactions. the reaction mechanism may be obtained. From
Because of the fact that oxidation by means an investigation made by E." B. MAXTED [4] it
of vanadium oxide catalysts finds much applica- appeared that markedly differing oxidations, e.g,
tion in the authors' organisation, it was considered that of naphthalene to phthalic anhydride, that
important to investigate the technological bases of benzaldehyde to benzoic acid and that of
of these processes. For this investigation the toluene to benzoic acid, invariably show a good
catalytic oxidation of aromatic hydrocarbons was yield at about the same temperature. The fact
chosen as the starting point, as this oxidation that the operating tcmperature of a catalyst
proceeds as an irreversible reaction, so that there proves to be independent of the substance to be
is no need to take into account the influence of oxidized, justifies the presumption that the rate
the equilibrium position. The conclusions, how- is mainly determined by the reoxidation of the
ever, have a much wider scope and in principle catalyst.
apply also to reversible reactions. By way of P. H. CALDERDANK [5] carried out experiments
example we shall consider the contact process for on the oxidation of naphthalene and concluded
the production of sulphuric acid. With the aid that, within a wide temperature range, the
of the reaction kinetics further information on the .. chemosorption of oxygen by the catalyst"
design and performance of the apparatus will be determined the gross reaction rate.
derived.
Experimental Part
1. OXIDATION OF AnOMATIC COMPOUNDS ON As the kinetic data on the reactions taking place
VANADIUM OXIDE CATALYSTS on this type of catalyst were incomplete and, in
The reaction may be viewed as an oxidation- part, unsatisfactory, we decided to perform an
reduction process: the substance to be oxidized investigation into this matter in our laboratory.

42
 P. MARS and D. W. VAN KREVELEN : Oxidations carried out by means of vanadium oxide catalysts

Because of the above-mentioned difficulties a fluid water are formed. The organic oxidation products
bed was applied. With this method, the tem- obtained from toluene are benzaldehyde and
perature differences in the catalyst bed are benzoic acid. Naphthalene yields naphthaquinone
negligible, while the temperature level can be and phthalic anhydride, while from anthracene
adjusted in a simple manner. the oxidation products anthraquinone and phtha-
lic anhydride are obtained. The reaction pro-
Apparatus ducts (together with the non-converted aromatic
A diagram of the apparatus used is given in substance) were condensed by cooling them to
Fig. 1. In a .. Pyrex" tube with an internal - BO°C (in the case of benzene - 90°C), and
diameter of 43 mm, the catalyst powder (par- subsequently analysed. These analyses were

Figure I Diagram of apparatus

Voporizer Converter Condenser Absorption trcm gas meter

ticle diameter between 200 and BOO JL) on a carried out in the following way: of the aromatics
sintered glass plate is fluidized by means of a the refractive indices were determined in n-
mixture of air and reactant. Heating is effected hexane; the quinones were analysed by reduc-
by means of an external heating wire, the tem- tion with titanium trichloride, and benzcldehyde
perature being kept constant by a relay system. by the hydroxylamine reaction; the acids were
The variations in temperature remain within 0'6°C titrated with lye. The amount of CO and CO2
A study was made of the oxidations of benzene, was measured by converting the CO into CO2
toluene, naphthalene and anthracene. When with hopealite, binding the CO2 to soda lime, and
benzene is oxidized, benzoquinone, maleic anhy- weighing.
dride and carbon dioxide, carbon monoxide and In the experiments, the partial pressures of the

4B
 C O·N F ·E n ENe EON O'x I D A T ION P no C E SSE S

figure 2 Conversion as aIunctlcn of partial pressure of the The dynamic state of the fluid bed
reactant.
The gas velocity in these experiments was kept
4.01r---------;--r--.-----,---,~___,
constant, because the conversion changed in a
IO:S ·'llcl/rnlft 0 col
manner difficult to describe if variations occurred
t in the gas vclocity or in the amount of catalyst.
301+---,----r::+-1~-1--+-_+-___Ir--
Therefore, efforts were made to keep the dynamic
state of thc catalyst particles circulating in the
bcd constant throughout the experiments, If, in
a fluid bcd, the linear gas velocity increases or
the amount of catalyst decreases, the number of
bubbles passing through the bed is greatly
increased. As the interaction between bubbles
and catalyst becomes much less intensive the
reaction rate measured experimentally falls off.
J. W. ASKINS et ai. [6] performed measurements
on a large catalyst bed to examine the composition
of the gas at various heights in thc bed. The
a 2 4 12 14 16 diffcrenccs in concentration measured wcrc only
small. These investigators concluded that a fluid
aromatic hydrocarbons and of the oxygcn were bed reactor must be considered an intensive
varied, while the other variunts: splice velocity mixer.
and temperature, were kept constant. The par- As the dynamic situation in a large catalyst
tial oxygen pressure was varied between 80 and bed may diffcr considerably from the situation
760 mm Hg, the partial pressure of the aromatic in a small bcd, it seemed possible that the mixing
reactant between 1 and 30 mm Hg, Typical con- in a fluid bed of smaller diameter is not so
version curves arc shown in Fig. 2 and 3. thorough and that in elaborating the kinetics of
Both partial pressures proved to influence the reactions carried out in such ri. bed the usual
gross reaction rate. mathematical methods might be followed.
Figure 3 Conversion as a function of the partial pressure of This problem of the mixing of the gas in a
oxygen. fluid bed is of importance in the interpretation
120
of our measurements only in that wc want to
10' "",'/m,n 0 col know whether the partial pressures/of the com-
0
100 t l- ponents vary with the height of the bed or arc
V 0 kept constant owing to the mixing. For this

eo
noplltholon. 33~
V 0
reason we measured the partial pressure of naph-
thalene as a function of the bed height under

~
the same conditions as were applied for the
kinetic measurements, By means of a sampling
60 tube a small amount of gas was taken from the

V bed and after washing with lye to absorb the

phthalic anhydride and naphthaquinone formed,

vet »-: . .
4a / the naphthalene concentration was determined
\ bonnno 317 ·c with an ultraviolet absorption meter.
The results of .thcse measurements arc shown
2a
/ /1 I I in Fig. '1-. From this figure it can be seen that

~
I I
anthracene Z70 ·C the concentration in the bed gradually decreases
I I, I ~- in the direction of the flow, from which it follows
1--' Pe2 I"nITIHg

a 200 400 600 eoo that the entering and outgoing gil" streams are
 P. MARS and D. 'V. VAN KREVELEN : Oxidations carried put b)' means of vanadium oxide catalysts

not thoroughly mixed. For this reason we lattice two kinds of oxygen ions may be distin-
consider the reactor to be an integrating converter guished. The greater part of the oxygen is
in the circumstances applied in our experiments. present ill about the same plane as the vanadium
ions. Two-fifths of the ions, h-owever, are arranged
16
H e,~ht In the bed
in planes parallel to and alternating with the
I , In em first [9] (sec Fig. 5). We suppose the latter
14

12
IFigure 4 Pcrtml pressure af naPhtholene'l-
measured m a fluidized bed Figure 5 Projection of the structure of V20S on (001)
a ir
bed
to

6
"

~\ \ -,
aor
L

4 \ 0, bed

,'\ \
2 '.
-,
0025 0030
<, r-...,
0035 0040
-
P naphthalene mm Hg

0025
~
0030 0035 0040

Discussion of Results. Kinet ic Basic Equation

In what way should the influences of the partial
pressures of the aromatic hydrocarbon and of the
oxygen on the reaction rate be interpreted? Bystrom, Wi/Mimi and Brotzen
Acto Chem,ca Scandmavica
Calculations of the mass transfer in a fluid bed, !!.1I26 (19S0)
using the known relations, show that in the border-
line case, where the external diffusion of the
aromatic component would determine the gross oxygen ions to interact with aromatic molecules
reaction rate, the number of molecules reacting in at the surface. The rate of oxidation will then
a given period surpasses the number actually depend on the concentration of these oxygen ions
reacting in the experiments by at least 100 times. (which will in the following be called the degree
Neither docs the diffusion in the pores of the of coverage, by the oxygen, and referred to by B),
catalyst restrict the reaction rate, as appeared and on the amount of aromatic molecules adsorbed
from an application of the formulae developed on or coming down upon the surface from the
by TIIIELE [7] and WAGNER [8]. gaseous phase. The degree of occupation will
If adsorption of the substance to be oxidized automatically be fixed at a certain value, depend-
on the reduced part of the surface should be a ing on the rate of the reaction of the oxygen on
condition of the reaction, there would be a the surface with the substance to be oxidized and
maximum rate at certain partial aromatic and the rate of re-oxidation of the reduced surface.
oxygen pressures. As such a maximum rate was This is in agreement with the experiment: if
never found we imagine the reaction to take place the reaction is carried out at high partial pressure
between th e aromatic compound and the oxygen of the oxygen and/or low partial pressure of the
present on the underlying catalyst surface. aromatic. the colour of the catalyst wiII be yellow-
The structure of crystalline vanadium pentoxide ish-brown, whereas at lower partial oxygen"
induces us to attribute a catalytic action to certain pressure and/or high partial pressure of the sub-
oxygen ions. For, in the vanadium pentoxide stance to be oxidized {provided the reaction
 CONFERENCE O.N OXIDATION PROCESSES

temperature be not too high) the colour changes to area not covered by oxygen. This can be expressed.
greenish-blue and on analysis a high concen- in the following formula :
tration of tetravalent vanadium is found. . k'2
noI = • Pon• . (1 - 0) (2)
The two reactions can be formulated:
I. aromatic compound + oxidized catalyst ~ If, for the oxidation of one aromatic molecule, fJ
oxidation products + reduced catalyst molecules of oxygen are required, the following
+
II. reduced catalyst oxygen - - + formula also holds :
oxidized catalyst . 1 .
nH=~nOI (8)
If, for a simple approximation of these reaction
rates, we assume them to be of the first order so that for the steady state we may write:
with respect to the partial pressures of the
substances to be oxidized, it follows that the fJ •k l • Pn- 0 = k2 • po~' (1 - 0) (4)

8 FIgure 6 OXidation role of ncohmclene

/.
os 0 Iuncton of lhe porbol
V
.:
pressures of lhe reectonts

6
/
I

V·/
V. 109 Po temp ' 337 'C
p

t •
4

2 ./
.>
/.;::
,
log
POz

POz-P(Po-P)

109 ~
o
01 02 03
- P

04 05

rate of the oxidation of the aromatic compound .For the degree of occupation by the oxygen it
may be expressed in the following formula: follows that:
0= k 2 .Po~ (5)
(1)
fJ .kl • Pn + k 2 ·Po"•
supposing the reaction rate to be proportional By combining formulae (1) and (5) we find for
to the degree of occupation on the surface of the the reaction rate :
catalyst. The rate of re-oxidation of the surface . 1
reduced in the first reaction may be assumed to tt n =- - - - - - -
be proportional to a certain power of the partial _1_+ fJ (6)
pressure of the oxygen and to the catalyst surface k l ·Pn k2 • Po~

46
 P. l\lARS and D. W. VAN KREVELEN: Oxidations carried out by means of vanadium oxide catalysts

If this formula is to be tested experimentally it Figurt! 8 Oxidation rote of benzene as a function at the
partial pressures of the reactants
must first be integrated. In the special case of
n = 1, we find the following relation:

-!. = .!:.. . In (Pn)o +.!.. In (Po,)o (7)

1.3

12
log paz
Paz· B [Pbo-Pb!
1*-
»; Pb0
V6
where (PO)linnl can be calculated as :
(Pn) final k2 (PO.)final
II
t 109_
Pb
, /
(Po)nnni = (Po)o - f3 (PR. - PRllnal) 10 0
/
and V 6 = space velocity.
09
0
/
377 ·C /337 ·C
Paz 08
,g PO z ' ,8 [PAo'PAl 1/

/-/ *
PAo 0.7

-
I 109PA"

H Figure 7 O,idomn rate of anthracene

as a functIon of the porncl
II /"
-: 0.6

0
0

*;-
pressures of the reactants
V< - 05

V
V- o. 4
-: /
• Vtemp 270-C
O.3
/. ./
p .* ~/
I

-:
~ O.2 nO
•*
&
"",*
~ - I
PA o
Vs · log llA
o.I~-
V - Vs
\
lo,/ho
Pb
z 4 6 B 10 IZ 14 ZO
100 200 300 400 500

In Figs. 6, 7 and 8 it is shown that the results check this point, various aromatic substances
of our experiments, both with the oxidation of were oxidized on a Porter camlyst** at the same
benzene, naphthalene and anthracene, arc in good temperature, while by variation of the partial
agreement with this formula.
For the activation energy in the re-oxidation
reaction (naphthalene oxidation) of the catalyst --l Figurt: 9 OXidation of naphthalene I
used in these experiments* we found 22 kcalyrnol ~
k2 as a fun chon _of temperature I
(sec Fig. 9).
An activation energy for the reaction constant ~
of the surface reaction between the oxygen and
the naphthalene nucleus cannot be given on
k2
1
r-,
"-
account of the spread in the results. e.
...0 <,
If the reactions occurring in the oxidation of
various substances always show one common
step: the re-oxidation of the reduced catalyst, k 2
<,
must bc equal for the oxidations of the various
substances at the same temperature. In order to
..... +.,0 3 <,
1.6 1.7 1.8 1.9

* Composition: 0% by weight V:P4' 21% by weight ** Containing 0% by weight V20 4• 2'0% Ly weight
Hil0.j,; carrier: silica gel. MoOs and 0'03% }'20/l. the carrier being corundum;

47
 CONFERENCE ON OXIDATION PROCESSER

pressure of the reactants the reaction constants The kinetics of the benzene reaction can also
were determined. Identical k 2 values were found, be described in this way. (See PIGULEWSKI [10]).
in conformity with our expectations (see Fig. 10).
Only the k 2 value for toluene showed a deviation, 2. OXIDATION OF SULPHUR DIOXIDE ON
about the cause of which we are not clear. VANADIUlII OXIDE CATALYSTS
The numerical values of the reaction constants Now the question arises as to whether a kinetic
are such that under technical conditions the formulation of the oxidation-reduction mechanism
influence of the oxygen pressure will be greater can throw light upon the numerous descriptions
than that of the partial pressure of the aromatic of the rate of the S02 oxidation reaction found in
compound. literature.
30
(P0 2 )O'

I;!'
10
lo~ ~
02 e 2
.10
101o~ PRo
25 f- I PRe

20
I
IFigure 10 Oxidation of aromatics
on porter catalyst
I
I
'j
/
15 /

10 ~
~oq
e
~
',!>r:J
•
/
I
q;if

5 / ./
0

'l 1'@
/
/ • - Il
S
I
PRo
10Iog _ _
PRe

5 10 15 20 25 30

Thus it may be understood why the most differ- Also in this process we have a great heat
ent substances arc oxidized on V20 S in the same release during the reaction:
temperature range (MAXTED [4]) and further, how S02 +
!02 --~ S03 23 k.eaIjmol +
CALDERDANK [5] could find the initial rate of However, this reaction differs from the oxidation
oxidation of naphthalene to be proportional to of aromatic compounds in that it is reversible.
(Po )0'96. For, at constant k 1 and PR' formula (6) The following formula describes the equilibrium
can: for a wide range of Po ,, very well be approxi- 2
Pso, (
mated with the following formula: K = 2 pressures in atm) (8)
Pso, . Po,
i~R = const, (po)m
K
where m < 1.
101
og = 'D880
i' - D'34

48.
 P. l\IARS and D. W. VAN KREVELEN : Oxidations carried out by means of vanadium oxide catalysts
630r-- --,

600

-r>.
Figure 1/ Ternperolure distnbunon In a
P 570 two-stage 502 conve rter
~
2
e
~ 540
E
~

I 510

480

450

420 i ("51 converl~r

-> ---- secondconverter

-
-
He'Qhl in the bed
-----.i
In practice the gas is usually eonvertcd for those cases in which SOa is present we must
adiabatically in scvcrallayers of catalyst. Between take into account the fact that this reaction
the catalyst beds there arc heat exchangers, by product strongly reduces the reaction rate. This
means of which the temperature distribution can mllY be due to the affinity of the SOa to the
be controlled. Fig. 11 shows this d istribution in non-oxidized part of the surface.
such a reactor, 'I'he catalyst bed may, e.g., be This implies that, besides the two reactions
distributed over two converters. In thc first the mentioned in Purt 1 :
gas is converted rapidly but very incompletely k1
owing to the unfavourable value of the equilibrium S02 + Cat - O~ SOa + Cat
constant i the second converter is operated at a and
lower tempcrature level, so as to make the Figure 12 Oxidat ion of S02 on pure V2 Os
reaction as nearly complete us possible. All the Kriehevskaycl (Iii. II)
kinetic formulae given for this process in litera- 300
ture show that the reaction product S03 strongly PSO
z
T l f e.veat.
reduces the reaction rate. /
For this reason we will first consider the
reaction rate in the border-line case, when the t /
formation of SOa is still negligible, 200 / 0-
JE. L. KlllCHEVSKAYA [11] carried out experi-
0
V
ments with pure vanadium pentoxide under such
conditions that the external and internal diffusion u
/
V
did not influence the reaction rate. When extra-
polating her experimental results to infinite space 100 ~
velocity (contact time = 0), the influence of a g'
small rise in temperature caused by the heat of o 0/
reaction is eliminated at the same time. The
experimental results are expressed by formula (6)
with n = ! (see Fig. 12). V - PS0
- - a2i m
vPOa
Vz

o 02 0.4 0.6
In setting up a formula for the reaction rate
-- I I
 CONFERENCE ON OXIDATION PROCESSES

k2 where
Cat +02~ Cat-O
Teq = ( --
pso. ) equilibrium
there is a. third reaction : PSo.
ks
SOs + C a t - S02 + Cat - ° or:
.
nso. =
k n
. PO.
C1eq - ot kl1 n
rOt I
=----,-
- ot'
(I8a)
From the strong influence exerted by SOs on ot • CXeq oteq ot
the reaction rate we may conclude that this third
reaction proceeds at a high rate. where oteq is the degree of conversion at equili-
Therefore we must examine the equilibrium : brium, and ot' = degree of conversion as compared
S02 + Cat - °( ) SOs + Cat to that at equilibrium = .!:..... After integration
oteq
which, apparently, is rather far to the left; for we find:
this equilibrium the following formula applies:
Pso • (1 -0)
K* = -....:.:...---- !. . {po.)n = C1 . {In _1__ ot'} (14)
Pso .0 V. (Pso.)o eq 1 - a.'
•
from which it follows that:
where V. = space velocity.
o =-= Pso. (10)
These formulae will now be compared with the
K* ·Pso • +Pso• data given in literature.
F. W. KUSTER [12] carried out experiments in
Since for the gross reaction process the rate- which a mixture of sulphur dioxide and oxygen
determining reaction is the re-oxidation of the was led over a vanadium oxide catalyst on
catalyst: asbestos. The degree of conversion was measured
no, = .
k2 .p~ . (1 - 0)
(as we found also for the oxidation of hydro-
(11) at several gas velocities. From Table 1 it appears
that formula (13) gives a satisfactory description
of the experimental results.
.
carbons). And since, moreover, nso.
follows from (10) and (ll), that:
= ,2no it
Table 1.
n Sulphur diOJ:ick oxidation (F. W. KilsTER [12]).
* Po, · Pso,
nso = 2 k 2 • K . K* (12)
• Pso, + Pso. Time Conversion
sec. % k'(1)
Now K* is small (as appears from the strong
retardation effected by SOs) so that the following ll'O 57·6 2'82
equation will approximately apply if the degrees 21'4 72'{) 3·41
of conversion arc not too low: 26·5 77·1 3·35
.
Po. ·Pso.
37·7
38'0
83'7
84·1
3·30
3·25
'liso = k
• 52'1 ar-e 3·12
Pso•
where k = 2k2 • K*. (1) Sinccpo and (Pso )0 were not varied. the combination
H at the same time the equilibrium conversion • •
k (po.ln/(pso.lo is given as k',
of S02' 02 and SOs is taken into consideration
(and hence also the decomposition of the oxidized n. NEUMANN and his collaborators [13] investi-
catalyst) it follows that: gated the contact sulphuric acid process, especially
with the purpose to identify the intermediate
Po, ( pso.) products formed on the catalyst. For a check of a
nso = k • - Pso - - (18)
• Pso, • Teq kinetic formula NEUMANN'S investigation into the

50
 P. MARS and D. W. VAN KREVELEN : Oxidations carried out by means of vanadium oxide catalysts

rate of conversion of S02 over silver vanadate as formula (14). Also here a satisfactory agreement
a catalyst can be applied. The conversion per- is found.
centage was determined at various gas velocities. In the experiments performed by E. L.
Table 2 lists the results of his measurements KRlCIIEVSKAYA [11] gas composition and contact
between 325 and 550°C, together with the time were varied. The oxygen content of the
reaction constants calculated with the aid of gas ranged from 5'0 to 89'7%, the sulphur dioxide
content from 3'1 to 40'1%. The catalyst used
Table 2. Sulphur dioxide oxidation on silver vanadate.
Experiments of n. NEUMANN [13]
was pure V20'" From Fig. 18 it is seen that the
experimental results are described by equation
Temp. Gas Conversion k'll) k' (12), if n = 0'5(2). From the values found for
°C. velocity % average K*, it appears that equation (13) is, indeed, a
good approximation at high conversion.
825 150 82·5 0·095 0·107 Previously, G. K. BORESKOV and F. I.
300 73'0 0-120
500 59·0 0'107
SOKOLOVJ\ had investigated the kinetics of this
process on a technical catalyst [14]. They found
875 150 92·0 0·172 0·188 that the reaction rate for conversions between
800 84·5 0·220 45 and 60 % were adequately represented by the
500 73'5 0'212 formula
750 57'0 0'147

425 150 96-5 0'291 0·315

d
Pso. =k. Po• (Pso.) 0·8

300
500
750
90·5
82·0
72·0
0·324
0-880
dt .PSO .
0'800 In a later publication G. K. BORESKOV [15]
450 150 97·0 0·333
mentions the possibility of the internal diffusion
0'366
800 92'0 0·370 interfering 'with the kinetic measurements in their
500 84·0 0·392 experiments.
750 76·0 0'370 P. H. CALDERBANK'S experiments [16] were
performed with a technical catalyst containing,
465 150 97'0 0·400 0·889
300 92·5 0·398
besides about 10% V 20 ", some percents ofK2S04 •
500 85·0 0'363 This investigator determined the reaction rate as _
750 78·5 0'392 a function of the gas velocity at five temperature
levels and three different gas compositions. How-
475 150 96·5 0·412 0'419 ever, two of the gas mixtures chosen deviate from
300 92·5 0'417
the usual compositions in that they are deficient
500 85·5 0·412
750 79·0 0·430 in oxygen. On this account it is not possible
to check by this relation the influence of the
500 150 95'0 0·418 0·43(1 partial pressure of the oxygen. As the gas
300 91·0 0'410 velocity was varied over a sufficiently wide range
500 85·5 0'428
formula (12) can be checked (sec Fig. 14).
750 81·0 0·495
(Formula 12 is used here for the same reason as
525 300 90'0 0·483 0'500 in checking the results of B. NEUMANN.) This
500 85·0 0·500 formula provcs to describe the results with
750 80·5 0-517 satisfactory accuracy.
550 300 87·0 (0'444) 0·535 (2) Since in these experiments the conversions are small and
500 84·0 0-527 therefore SOs-eoncentration is low in comparison with the
750 79·5 0-545 S02-concentration, the term K*pso. may not be neglected;
since further the conversions are Car from equilibrium,
(1) See remark In Table 1. formula 12 can be applied in its integrated form.

51
 CONFERENCE ON OXIDATION PROCESSES

10
,-«
In-!-
.
6
-" ...In

-I
I
1-« Figure /3 Oxidation of sulphur dioxide
on pure V20S
V
Je. L Knchevskoyo (ht. II) /'/

6
V
V
4 »: •

2 /
~
•
•
....
• _. !
(Paz)·l

0 /
V - (Pso.)o
-O(+ln ..L
1-«

20 40 60 80 100 120 140 160 180 200

Flguie 14 OXidation of sulphur dioxide on technical calolysl the transport of the molecules taking part in the
PH Cold'fbonk (lit 16) reaction) may, however, influence the gross
20 reaction rate.
In-'-,-0< I, For the calculation of the "external mass
18 -cw.+ln~ transfer" - i.e, the transfer from the bulk of the
,-«
\6
I /
410'~~00~ /
/
390~/
gas to the external surface of the catalyst grain-
the relationships arc well known. It is also

14 . . . possible to estimate in a simple manner the

influence of the" internal transport," viz. of the

/
/ /
./ / rate of diffusion in the catalyst pores on the
overall reaction rate for the- special ease of a

JO / /
·V
/ first-order reaction, by means of the formulae
elaborated by TIIIELE [7] and WAGNEU [8].
./ / J However, in case of a reaction such as the one
8

/
/i ~
discussed here, these formulae cannot be applied,
as exactness would here lead to very complicated
formulae [17].
•
4I • ·1 ./
/ A simple consideration of the diffusion in a
pore of a catalyst grain enables us to estimate
/ '/ / roughly the alteration of the kinetic equation as
-
PO,' 1
2
V ( PS02)0 a result of this internal mass transfer.
o1/
VII
5 15 20 25 30
-CI(+ln-L-

35
,-"
40
Taking a pore (which, for the sake of simplicity,
is imagined to be cylindrical) and considering only
10
the mean reaction and diffusion rates, wc may
'I'his agreement between the data found in write (see Fig. 15) :
literature and our kinetic equation occasioned us feed rate owing to diffusion = reaction rate
to take this equation as the starting point of our
further discussion of the process.

INTERNAL DIFFUSION

With a catalyst used under conditions such as

occur in the technical process, mass transfer (i.e.

52
 P. MARS and D. W. VAN KREVELEN : Oxidations carried out by means or vanadium oxide catalysts

Figure /5 Diagram of dIffusion in pore EXAlIlI}iATlON OF A TECHNICAL CONVERTER FOIl

THE 1IIANU}'ACTURE OF SOa
We will now compare the conclusions drawn from
P
theory with the contact process as it is carried
out in practice.
I paz
One of the most commonly used cquipmcnts
II comprises a combination of catalyst beds in which

---
the reaction proceeds adiabatically, with heat
l---. Pso 3
PSO z exchangers interposed; Fig. 17 is a diagram of
such an equipment. By means of the heat
1!2rp
am I

so that for the maximum depth of penetration X m

the following formula holds :

X'" JD.rIe
m
p (pso.)o
(Po.)'" IX • IXeq

from which it follows that:

1/so. '" r'J Ju»; (Po.)" (Pso.)o IX' C~~') (I5)

By integration we now find : Fir;ure /7 Twa sto)ie S02 converter

_
k*
2
J__'_.- = J-;:_ -
(Po )"
(Pso.)o
t
IXeq Clteq
tanh? JIi" _
iXeq
(I6)
with monlle-eooling

exchangers the temperature distribution in the

Figure /6 Temperature distribution as a function of conversion reactor may be controlled. Fig. 11 shows a
typical temperature curve. In converter 1 the
Flprrpturr ~
I
op',ml,lm~
gas is converted at high temperature, rapidly, but
1 owing to the unfavourable position of theequili-
I\. 0
<, brium, incompletely. After heat exchange with

l><~ the gases flowing towards the converter has taken

~

t
artua,) V \
place, the degree of conversion is raised as far as
possible in converter 2, which operates at lower
temperature. Converter 1 contains three layers
i\ of catalyst, converter 2 comprises 'five catalyst
beds. The catalyst is in the form of kieselguhr
V

V
/ "\ I~
blocks of 0·8 X 1·1 X 1·1 em, impregnated with
potassium-eontaining vanadium oxide.'
From our kinetic equation we will now draw
conclusions relating to the following points: .
1. The temperature distribution in the reactors;

1 01 az 0.3 04 0-5 06
-'"
0 .7 08
0
/0
1
I\'
09 10
2.
3.
4.
The
The
The
distribution of the catalyst;
size of the catalyst grains;
composition of the reaction gas.

58
 CONFERENCE ON OXIDATION PROCESSES

Ad 1. Table 3.
The optimum temperature distribution can be
calculated in a simple manner if the rate equation Temp. k** hr-1 Effectivity of k*y(po.)n. (Pso.)o
°C. catalyst % atm. hr-l
and the apparent activation energy of the reaction
is known. As the activity of a catalyst may 445 2000 45 164
decrease during the operation we derived the rate 570 9140 20 400
constants at different temperatures from tech-
nical data on a catalyst which had been in use
for some years. The rate of conversion can be Table 8. It appears that both in the first and in
calculated from the temperature differcnce across the second converter the reactor rate is strongly
the catalyst bed, since the reaction proceeds reduced by the rate of diffusion in the pores: the
adiabatically within the bed. From an analysis effectiveness factor of the catalyst grains in the
we know the concentration of the sulphur dioxide layers we examined did not exceed 45% and 20%
before the reaction and the degrees of conversion for the two converters.
after the reactions in converters I and 2. This implies that we cannot use formula (14)
By applying the well-known relation between but that we have to use formula (15) tocalculate
reaction conditions and mass transfer it was found the reaction rates and the activation energy. The
that only in the first layer of converter I docs the apparent activation energy proved to be about
diffusion to the catalyst tablet-surface slightly 9 kcal /mol. For the re-oxidation reaction of a
reduce the reaction rate. This very small influence vanadium oxide catalyst not retarded by diffusion
of the external diffusion will be neglected. we found 22 kcal/mol (see page 7). Conformable
It is quite different, however, where thc to theory, it appears that the activation energy
influence of the internal diffusion is concerned. decreases by about a factor 2 if there is internal
In order to estimate roughly the influence of diffusion. This result is important: the fact
the internal diffusion, we used a mean degree that the rate of the sulphur dioxide oxidation is
of conversion a in the examined layer, so that largely dependent on internal diffusion in the
formula (18a) passes into a formula, which des- catalyst pores, involves that it would be incorrect
cribes an apparent first-order reaction : to carry out calculations using literature data on
the activation energy.
For every conversion (rJ.) there is a temperature
where at which the reaction rate reaches its maximum,
k n as both k* and Teq (and hence cxeq ) are dependent
k** =_ . Po. on the temperature.
rJ. • rJ.eq
By differentiation of formula (15) the optimum
Now the formulae of TUIENLE [7] and WAGNER temperature and the maximum reaction rate can
[8] can be applied as a first approximation. be determined as a function of the conversion ;
The diffusion coefficients of 502 - 503 in the in Fig. 16 Toplimum =f(cx) is given. It appears
(mean) gas mixture were calculated using GILLI- that there is a considerable deviation from the
LAND'S formula [18]. The ratio of the cffcctive actual temperature distribution.
diffusion coefficient to the diffusion coefficient The question arises as to whether this deviation
proper was determined in the manner described is attended by a great drop in potential product-:
by C. BOKIIOVEN and J. HOOGSCIIAGEN [19]. For ion. An approximative calculation shows that I

this calculation the diffusion in the pores was under optimum temperature conditions the pro-
assumed to proceed according to the laws govern- duction may be appreciably higher than under
ing the normal gas diffusion, while the influence the actual conditions. It should be remarked
of the transport along the catalyst surface was that the temperature levels of the two converters
left out of consideration. are already so adjusted that the maximum pro-
The results of our calculations are given in duction under the actual conditions is attainer-
 P. MARS and D. W. VAN KREVELEN : Oxidations carried out by means of vanadium oxide catalysts

It appears that it is important to design the reaction. This would result in a rise of the
conversion apparatus in such a way that the temperature at which the reaction rate
optimum temperature curve can be followed, reaches its maximum for a given value of
although the construction will undoubtedly be the conversion.
more complicated. (c) The pressure drop of the catalyst beds would
increase, involving a rise in energy costs.
A.d 2.
Under Ad 1 we estimated the optimum temper- Ad. 4.
ature distribution. Even, if for technical reasons or Depending on the amount of air used in the
for reason s of heat economy, the optimum temper- combustion of the sulphur or the roasting of the
ature curve is not followed in most of the practical pyrite, a gas rich in oxygen and poor in sulphur
applications, it remains a useful guide for the dioxide or a gas with a lower oxygen content
temperature level to be applied in the second and containing more sulphur dioxide can be
converter, while, moreover, it may be used for obtained.
checking the performance of the two converters. It follows from our considerations on the
For, from the optimum temperature distribution kinetics of this process and on the internal
we can calculate the proportion between the diffusion in the catalyst tablets that the maximum
contact times to be applied for attaining a given production (P max' to be obtained by following
degree of conversion. Then we know the ratio the optimum temperature curve) can be calculated
between the amounts of catalyst to be used and as:
we are able to check if the reaction conditions in P max = (V.)opt • Vcat • ex • (% 502)0
.the two converters are well adapted to each other. The optimum space velocity to be applied is
A calculation showed that under optimum proportional to
conditions the ratio of the amounts of catalyst
necessary to obtain a conversion of from 0 to VP8, (Pso)o
81 % [converter 1) and a conversion of from 81 % which reaches its maximum value at certain O2
to 97% (converter 2) should be 23: 77. The and 502 percentages. Table 4 shows the corres-
actual ratio between the amounts of catalyst in ponding values of these percentages, besides the
the two converters is 27: 73. Of course, the corresponding values of
significance of this agreement should not be
overestimated. It may be concluded, howcver,- V(% 02)" (% 502)0
that t he distribution of the catalyst is rather for 1t =1 and 1t = i.
impervious to the temperature distribution.
Table 4.
Ad.8. Influence of the 02 and 802 concentrations on the Teaction
It appears that rather large catalyst tablets rate.
arc not very effective. Table 3 shows that only
45% and 20% of the total catalyst surface parti- %°2 %5°2 V%02. %502 V(%02)1. %802
cipate in the reaction at tempcraturcs of 445 and 5'2 12 7·0 5 ·2
570°C, respectively. 6 ·6 11 8·5 5'3
However, reduction of the tablet size has 7'8 10 8'8 5'3
several consequences: 0'0 o e-o 5'2
10'6 8 o-s 5 ·1
(a) As thc mean diffusion path becomes smaller, IH) 7 {j'1 4·9
the effectiveness of the catalyst will increase, 13'0 6 8'8 4·7

(b) In consequence hereof the diffusion will exert

less influence on the reaction rate, involving a A weak maximum is found at 8 % 502 and 11 %
rise of the apparent activation energy of the 502' respectively. As, however, the decrease of

55
 CONFERENCE ON OXIDATION PROC ESSES

thc reaction rate at increasing sulphur dioxide 1. Tube converters;

content and decreasing oxygen content is only 2. Tray converters;
smull , it remains ad vantageous to operate at the 3. Combinations of 1 and 2.
highest sulphur dioxide content possible, pro- Examples of these converters are shown in
vided that the temperature level is under control. Fig. 18. As it is difficult to control the tempera-
ture distribution in such converters, and, moreover,
strong influences are exerted by the sulphur
-~
, , ,,

50 2
Figure 19a J.G. Farben converter
(1938)

. .°2 - ~

cold gases

air air
...,
= heal exchanger

~ J
i:=1 - 50 3
50 3 Figure 19b Bachelard converter
Figure 18b Schroeder converter (1939)
IGm,IIn 9°, Fiv 48')
50 2

oir - -

50 3 - 1 - - - - - ' Figure 19c Chemico converter

(1950)
Agure 18 C Tentelew-converter
co 1900
dioxide content of the gas and by the gas
velocity, the apparatus according to Fig. 19a
8. CONSIDERATION OF THE TRENDS IN TECHNICAL
appeared to be an improvement: In this apparatus
REACTION DESIGN
internal and external heat exchange arc combined.
Finally, we shall examine some aspects of the When improvements in the roasting technique
development of the technical oxidation process made it possible to obtain cheap gases of high
relating to the points discu ssed in this paper. sulphur dioxide content, a method of cooling by
"I'here are several ways of designing the con- supplying cold gases was soon invented (BACHE-
verters. Three types of converters have been LARD, 1980, sec Fig. lOb). A parallel improvement
known for a long time : was the application of converters having 3 or ~

56
 P. MAns and D. W. VAN KnEIELEN : Oxidations carried out by means of vanadium oxide catalysts

instead of 2 stages, between which the gases are 11 = reaction rate

cooled by means of heat exchangers. Also, in P = partial pressure
rp = pore radius
this manner a better approximation to the
optimum temperature curve can be obtained r eq = Pso,
- - a t equiilib .
I rium . I I prevai'1'mg Po
Wit
Pso, ,
(Simon-Carves Monsanto). One of the latest
t = time of reaction
developments, the Chemico wet sulphuric acid Xm = maximum depth of penetration
process, is carried out with a concentrated gas K = gas equilibrium constant
(12% S02)' which makes it possible to apply K* = gas-catalyst equilibrium constant
direct air cooling (Fig. 1ge). R = organic reaetant(s)
Thus we see a trend in the converter design to Vent = catalyst volume
bring the temperature distribution in the con-
Va = space velocity
IX = conversion (fraction converted)
verters nearer to the optimum (i.e., as far as
economical considerations will allow). In the one a.' =~
IXeq
case this is attained by applying several converters f1 = number of oxygen molecules needed to convert one
in series, in the other by direct cooling of the gases. molecule of aromatic compound into the oxidation
The effect of the size of the catalyst tablets products found experimentally.
has been examined in the I.G. laboratories. It
£) = degree of coverage of the catalyst surface by the
oxygen
-appeared that reduction of the grain size, while "eq = equilibrium
aiming at the same degree of conversion, made it "0 = value of the quantity concerned at IX = 0
possible to increase the production capacity [20]. ". = value of the quantity concerned at the end of the
For this reason in the Lurgi convertor 6 mm dia, converter
cylinders arc applied in the first few catalyst
layers (which layers arc operated at the highest REFERENCES
temperatures). [1] SENSIMAN, C. E.; Ind. Eng. Chern. 1923 15 521. [2]
In the technical oxidation of aromatic com- Scrmnmnn, 'V.; in KROGER, M" Grcnzflachen-Katalyse;
pounds the fluid bed technique has come to play Leipzig 1033, p, 203. [3] WEISS, J. M., DOWNS, C. R. and
an increasingly large part. There is an important BURNS, R.l\I.; Ind. Eng. Chern. 1923 15905. [4] MAXTED,
advantage in the uniformity of the temperature E. B.; J. Soc, Chern. Ind. 1028 47 101 T. [5] CALDERDANK,
in a fluid bcd, which, together with the pos- P. II.; The Indust. Chemist, July 1052, r- 201. [0] ASKINS,
sibility of applying higher concentrations of the J. W.; Chern. Eng. Progr. 1051 47401. [7] TIIIELE, E. W.;
organic compound, is a strong point in favour of Ind. Eng. Chern. 103031 010. [8] WAGNER, C.; Z. Phys,
the application of a fluid bed. Chern. 1M3 A 193 1. [9] KETELAAR, J. J.; Nature 1936
It is obvious that this is not the case for the 139 310.; Chern. Wbld. 1936 33 51.; BYSTROM, A., WIL-
S02 oxidation: owing to the homogeneity of the I1ELMI, K. and BROTZEN, 0.; Acta Chemica Seandinavica
temperature in a fluid bed a great number of 1050 4 1119. [10] PIGULEVSKI, V. V.; Chemical Abstracts
reactors would have to be arranged in series in 1037 31 5315. ell] !{mcIIEVSKAYA, JE. L.; J. Phys,
order that the optimum temperature course might Chern. (USSR) 1947 21 287. [12] KUSTER, F. W.; Z. Anorg.
be better approximated. Thus, for this particular Chern. 1904 42 453. [13] NEUMANN, B.; Z. Elektroeh.
192834000. [14] BORESIWV, G. IC, and Sorcor.ova, F. I.;
case the fluid bed technique loses much of its
J. Chern. Ind. (USSR) 1037 14 1241. [15] BORESKOV,
attractiveness.
G. K.; J. Phys. Chern. (USSR) 194519535. [16] CALDER-
NOl\~ENCLATUItE nANK, P. H.; J. Appl. Chern. 10522482. (17] BORESKOV,
G. K. and SLlNJKO, 1\1. G.; J, Phys. Chern. (USSR) 1952
k = reaction rate constant 26 235.8. [18] GILLILAND, E. R.; Ind. Eng. Chern. 1034
k* = } apparent reaction rate constant 26 683. [W] BOKIIOVEN, C. and HOOGSCUAGEN, J.; J.
k** = Chern. Phys. 1953 21 159. [20] BIOS Final Report 1032.

57
 CONFEnENCE ON OXIDATION PROCESSES

DISCUSSION
Prof. Joe M. Smith: (Purdue University, Lafayette, A strong homogenization of the gas phase is to be
lndiana) expected if the particles also adsorb great amounts ofthe
The authors arc to be commended for attempting to reaction components and hence transport them through
bring some unification into kinetic phenomena." Their the bed.
2. The mixing in the gas phase of a fluidized bed is oC
scheme of choosing a particular class of reaction and
specific catalyst appears to be a fruitful way to approach interest to us in this connection only in so far as we are
necessitated to m ake a choice between two extreme
this problem.
In connection with the paper proper, would the authors situations. From our measurements of the partlal pres sures _
care to comment on the following three points : of one reaction component at different heights in the' bed
1 . It was interesting to note that the measured partial it is sufficiently clear, in our opinion, that n good approxl-
pressure profiles in the fluidized bed were not flat. In mntion can be made by considering the b ed to be . an
contrast the temperature is generally considered to he integrating reactor.
uniform in such an apparatus and the authors utilized this 3. Naturally, agreement between a kinetic formula and
simplification in interpreting their kinetic data. This experimental data L~ not a proof of the correctness of .the
situation would suggest that the transport mechanisms assumed reaction mechanism. In our investigation a
for heat and mass are not the same, or at least individual mechanism made acceptable by experiments (Refs. 1; 2 .
me chanisms do not h ave thesame importance "inheat and and 3) was taken as the basis for a consideration oC the
mass transfer. 1<; this due presumably, to the ability of kinetics. The agreement between"the kinetic formula and
the solid particles to carry energy but not significant the experimental data makes it superfluous to consider
quantities of mass from the surrounding fluid? other mechanisms and may be held to support the assumed
2. On the basis of the measured partial pressure pro- mechanism.
files, the authors con sidered that the Iluidized bed did not
behave as a stirred tank reactor, but instead integrated the Dr. H. Jockush: (Farbwerke Ilocchst A.G., Frankfurt a.
proposed rate equation. 'fhis procedure assumes that Muirl)
there is no mixing in tile direction of flow, which is some- I should like to ask whether the authors accounted for
what in contradiction to experimental results of GILLILAND the formation of radlcals in the course of the reaction. It
and DANCKWERTS. 'Vould it be possible that some longi- is known that maleic acid is formed also from paraffinic
tudinal mixing did occur in the fluidized bed and that a hydrocarbons. The formation oC maleic anhydride is
precise description of the reactor would be som ewhere in certainly the function of the catalyst surface. This explains
between the stirred tank type, where mixing is complete, the fact, that no intermediates arc formed. It also explains,
and the narrow tubular form where mixing is negligible? why crotonic aldehyde is oxidized to maleic annydride in
Perllaps tile shnpc of the partial pre ssure profiles led the the gas phase whereus in the liquid phase crotonic acid is
authors to conclude that the no-mixing assumption wus formed and it seems to be impossible to surmount the
the closest to the actual process. barrier even by applying high temperature and pressure.
3. One of the discouraging features of heterogeneous
catalysis is that frequently several different mechanisms The Author:
can be found which describe kinetic data equally well. A
knowledge of the intermediate products is helpful in such In contrast with the oxidation in the liquid phase
circumstances, although their identlflcution is not always mentioned by you, catalytic oxidation reactions in the gas
possible. In this connection a mechanism based upon the phase do not provide certainty that the reaction proceeds
three steps, adsorption of oxygen on reduced catalyst via radicals.
surface, reaction between oxidized catalyst and hydro-
carbon in the gas phase, and desorption of the oxidiation H. A. Cheney: (Sllell Deo, Co., Emeryville, Cal.)
product, leads to a kinetic equation similar to the one
proposed . By assuming different individual steps to be The que stioner would like to be informed on the way
controlling and the others to be close to equilibrium con- in which tile catalyst was made. He further asks whether
ditions, it is possible for the resultant rate equation to there were any mechanical difficulties and whether
indicate, or not indicate, a maximum rate at certain attrition or .. sticking" were serious.
partial pressures of hydrocarbon and oxygen. 'Vould the
authors like to comment on other mechanisms they may The Author:
have considered? For our investigations wc used three catalysts :
(a) The" I.G. Farben catalyst" for thc preparation
The Author: of phthalic anhydride, which was prepared by impreg-
1. Indeed I In our opinion the transport mechanisms nation oC potassium aulphatc-i-contnlnlng silica gel with
of heat and mnss arc not the same in the fluidized bed, a vanadyl sulphate solution. (Fiat Final Report No. 649).
'Vhen dealing with ma ss transfer to the catalyst surface (b) The .. Porter catalyst," prepared by drenching
in a fluidized bed no viewpoints differing from those held corundum in a solution of vanadium, molybdenum and
in connection with fixed bed transfer present themselves. phosphorus oxides in hydrochloric acid. (Am. pat. spec.
In a fluidized b ed, however, the catalyst particles serve no. 2,204,130).
us heat accumulators in the removal of heat; these (c) A trade-mark catalyst for the oxidation of sulphur
particles nrc cooled as they circulate along the reactor wall. dioxide containing 7·7% V20jl and 10'0% K20 (carrier:
kieselguhr).

58
 P. MARS and D. 'V. VAN KREV£LEN : Oxidations carried out by means of vanadium oxide catalysts

Attrition of the catalyst was serious in the tirst week successful operation of a miniature converter in which
only. After that, the amount of mnterial entrained bv the such a liquid was circulated over a packed tower.
gas from the bed was less than 1 % of the catalyst weight I think any theory of the mechanism of potassium
in a week. The amount of gas we used for fluidizing the sulphate promoted vanadium sulphuric acid catalysts
bed was so large that also the particles immediately over must be made consistent with the presence of a catalytic-
the grid were caused to I;ll0ve through the bed. ally active liquid phase in a large proportion of the catalyst
in the conversion system.
Dr. C. P. Van Dijk: (Kon./ShellLab., Amsterdam)
Temperature on the catalyst surface at the spot where
the reaction takes place, will be higher than the tempera- The Author:
ture measured. In our opinion, the fact that part of the catalyst used
The dirrerenee will of course increase with higher heat in the 802 oxidation is present on the carrier in the liquid
of reaction, poorer heat transfer of the catalyst material phase (the exact amount depending on reaction conditions,
and greater speed of reaction. especially on temperature, degree of oxidation and potas-
Has any indication of such a difference been obtained sium content) has no influence on the consideration of the
and what is the influence of such a difference on the reaction mechanism and the kinetics of the reaction. The
considerations given? phenomena described may equally well occur on a liquid
The Author : surface,
From measurements made by means of a thermocouple
enclosed in a glass tube, wnich was inserted in the bed at Dr. G. C. A. Schult: (Kon./Shell Lab., Amsterdam)
different heights, it appeared that the temperature is •The conclusion of the analysis concerning the reaction
practically the same in all parts of the bed. mechanism of the oxidation of aromatics over V205,
The temperature of the particles themselves is very arrived at by the authors, is equivalent to the assumption
dilflcult to measure. Calculation of the temperature of a fast reaction between" oxidized catalyst" and hydro-
dirrerenee between the catalyst particles and the surround- carbon, followed by a much slower one between .. reduced
ing gas by the generally known heat transfer relations catalyst." and oxygen. If this is accepted, the catalyst
showed this difference to be always less than 1°C. would be predominantly in the .. reduced" state during
catalysls, This, however, implies a situation which is
C. F. P. Bevlnllton : (Imp. Smelting Corp., Avonmouth) extremely favourable for partial oxidation, while it is a
The essential components of all modern commercial wellknown fact that in V205-catalyscd oxidation often
vanadium sulphuric acid catalysts are potassium sulphate, considerable total oxidation occurs.
vanadium pentoxide and silica. In the presence of sulphur Arc the authors aware of n possibility of including these
trioxide under operating conditions the potassium sulphate effects in their meehnnism? Or docs the .. total" oxi-
is partly converted to the pyrosulphate-K2&.lOr-which dation occur as a homogeneous reaction subsequent to
is molten at temperatures above about 400°C and is able the catalytic oxidation?
to dissolve pentavalent and tetravalent vanadium com-
pounds. There is evidence that in much of the catalyst in The Author:
a sulphur dioxide converter a liquid phase is present. It is
possible to prepare this liquid by treating a mixture of From our measurements we concluded that the situation
K~04 and V201S with gas mixtures containing appropriate on the surface of a .. reduced catalyst" is not parttcularly
proportions of 802, 803 and 02 at temperatures in the favourable for obtaining high yields of phthalic nnhydrlde :
range 450·GOO°C, and to usc this liquid as a catalyst. The the highest yields were obtained under conditions pro-
Danish workers TOPsoE and NIFLSEN* have reported the moting strong oxidation of the catalyst. 'Vllether radical
reactions play a part in this connection is not known to
* Trans. Danish Akad. Techn. Sci. No.1 (1048). us.

50