The acrolein and acrylonitrile synthesis over a bismuth
molybdate catalyst : kinetics and mechanism
Citation for published version (APA):
Lankhuijzen, S. P. (1979). The acrolein and acrylonitrile synthesis over a bismuth molybdate catalyst : kinetics and mechanism. Technische Hogeschool Eindhoven. https://doi.org/10.6100/IR1200
DOI:
10.6100/IR1200
Document status and date: Published: 01/01/1979
Document Version:
Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)
Please check the document version of this publication:
• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website.
• The final author version and the galley proof are versions of the publication after peer review.
• The final published version features the final layout of the paper including the volume, issue and page numbers.
Link to publication
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal.
If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:
www.tue.nl/taverne Take down policy
If you believe that this document breaches copyright please contact us at: openaccess@tue.nl
'
THE ACROLEIN AND ACRYLONITRILE SYNTHESIS
OVER A BISMUTH MOLYBDATE CATALYST
Kinetics and mechanism
PROEFSCHRIFT
TER VERKRIJGING VAN DE GRAAD VAN DOCTOR IN DE TECHNISCHE WETENSCHAPPEN AAN DE TECHNISCHE HOGESCHOOL EINDHOVEN, OP GEZAG VAN DE RECTOR
MAGNIFICUS, .PROF. DR. P. VAN DER LEEDEN, VOOR EEN COMMISSIE AANGEWEZEN DOOR HET COLLEGE VAN DEKANEN IN HET OPENBAAR TE· VERDEDIGEN OP
VRIJDAG 22 JUNI 1979 TE 16.00 UUR
DOOR
SIMON PIETER LANKHUIJZEN
Dit proefschrift is goedgekeurd door de promotoren:
Prof. drs. H.
s.
van der Baan, le promotorAan Anny Aan Jannelies Gerdiene Joanne Machteld Han
CONTENTS 1. Introduction 1.1. General 1.2. Acrylonitrile manufacture 1 2
1.3. The mechanism of the oxidation and arnmoxidation of
propene 3
1.4. Aim and outline of the present investigation 4
2. Literature
2.1. Introduction 2.2. Kinetics
2.3. Adsorption of reactants and products 2.4. Hydrocarbon surface intermediates
2.5. Nitrogen containing surface intermediates origi-nating from ammonia
2.6. ~ole of oxygen
2.7. Catalyst 2.8. Models
3. Apparatus and analysis 3.1. Introduction
3.2. The flow reactor system 3.2.1. Analysis
3.3. The thermobalance
3.4. The pulse reactor system 3.4.1. Analysis
3.5. Safety
3. 5 .1. Toxicity
3.5.2. Flammability and explosive ranges 4. The catalyst 4.1. Introduction 7 8 9 12 13 14 15 16 21 23 27 32 33 34 35 35 37 39
4.2. The structure of the bismuth molybdate catalyst 40
4.3. Catalyst preparation 43
5. Experimental methods
5.2. The rate of reaction 48
5.3. Factors governing the reactor behavioUr 49
5.3.1. Plug flow 49
5.3.2. Temperature gradients 51
5.3.3. Catalyst dilution 53
5.3.4. Mass and heat transfer 54
5.3.5. Pressure drop 57
5.4. Data handling and analysis of errors 57
6. Results and discussion
6.1. Introduction 61
6.2. Flow experiments 62
6.2.1. Introduction 62
6.2.2. Preliminary experiments 62
6.2.3. The oxidation of propene to acrolein 64
6.2.3.1. Experiments at 673 K 64
6.2.3.2. Experiments at other temperatures 68
6.2.4. The oxidation of ammonia to nitrogen 69
6.2.5. The ammoxidation of acrolein to acrylonitrile 70
6.2.5.1. Preliminary experiments 71
6.2.5.2. Experiments at 673 K 72
6.2.6. The ammoxidation of propene to acrylonitrile 74 6.2.6.1. Introduction and preliminary
ex-periments 74
6.2.6.2. Experiments at 673 K 76
6.2.6.3. Experiments at other temperatures 79
6.2.6.4. Experiments at non-initial
con-ditions 80
6.3. Thermobalance experiments 84
6.3.1. Introduction 84
6.3.2. Preliminary experiments 85
6.3.3. Reduction of the catalyst with propene 85
6.3.4. Reduction of the catalyst with hydrogen 90
6.3.5. Reoxidation of a reduced catalyst 93
6.3.6. The reduction of the catalyst with mixtures of propene, nitrogen and small quantities
6.4. Pulse experiments
6.4.1. Introduction 100
6.4.2. Preliminary experiments 101 6.4.3. Experiments with propene-helium mixtures 103 6.4.4. Experiments with mixtures of propene and
oxygen 105
6.4.5. Experiments with mixtures of ammonia and
helium 107
6.4.6. Experiments with mixtures of propene,
ammo-nia and helium 107
7. Final discussion 7.1. Introduction
7.2. The mechanism of the catalytic reaction
113 113 7.3. Adsorption and adsorption sites 115 7.3.1. The adsorption of ammonia 115 7.3.2. The adsorption of propene 117 7.3.3. The adsorption of acrolein 120 7.3.4. The role of the catalyst in the activation
of molecular oxygen
7.4. The formation of the main products 7.4.1. The formation of acrolein 7.4.2. The formation of nitrogen 7.4.3. The formation of acrylonitrile 7.4.4. The formation of water
7.5. The kinetic model
7.6. Selectivities in the acrylonitrile synthesis reac-tion List of symbols summary Samenvatting Dankwoord 121 125 125 128 129 130 130 134 139 143 147 150
CHAPTER 1
INTRODUCTION
1.1. General
Since acrolein was discovered in 1843 by Redtenbacher (1) and acrylonitrile was synthesized fifty years later by Moureu (2) these compounds remained laboratory chemi-cals until the development of some large scale polymeri-zation processes. Preparative methods were replaced by commercial catalytic processes and expensive chemicals by cheaper raw materials to obtain these important unsatura-ted intermediates in chemical industry.
As a result of the development of certain heterogeneous catalytic processes acrylonitrile is obtained nowadays from propene, ammonia and air instead of acetylene and hydrogen cyanide. Contrary to the latter the former process is based on catalytic oxidation, which is now a major tool for the incorporation of carbonyl and nitrile groups in hydrocarbons. The necessity of a catalytic oxidation
re-'
action follows from thermodynamic calculations. The com-plete oxidation of propene and ammonia to carbon dioxide nitrogen and water at temperatures of interest is prefer-red to the formation of acrylonitrile or acrolein.
An intensive research effort has been directed towards the development of selective catalysts for the partial oxidation of hydrocarbons to a number of products. This has led to the development of a number of selective oxi-dation catalysts made from transition metal oxides.
A major event in the history of oxidation catalysis was the discovery of bismuth molybdate (3,4,5,6) as a selec-tive catalyst for the partial oxidation of propene and also in an one-step operation for the ammoxidation of pro-pene. As other metal oxide combinations this catalyst is
very versatile, being effective in a number of processes, e.g. the oxidative dehydrogenation of n•butenes and me-thyl-butene to butadiene and isoprene and the formation of aromatic carbonyls and nitriles.
Although in the 1960's the catalytic ammoxidation of propene over bismuth molybdate became a commercial process, the basic understanding of the role of the catalyst lagged far behind the technological knowledge of the process. Many research studies have been carried out to
characte-rize the excellent catalytic properties of bismuth molyb-date based catalysts. These studies have increased the un-derstanding of oxidation and.ammoxidation kinetics and mechanisms, but the controlling parameters are not yet completely understood.
Two basic principles however have become clear. Firstly the oxygen atoms for the selective reaction come from the catalyst. Secondly the metal oxide combinations must be able to transfer oxygen by a redox reaction. The latter requirement explains the suitability of transition metal oxides in selective oxidation. The catalyst surface is alternately in an oxidized and reduced state. Anion va-cancies play an important role.
1.2. Acrylonitrile manufacture
The impressive growth of the acrylonitrile production (average annual growth between 16 and 20% from 1960-1975) has resulted from the manufacturing technology that could use cheaper raw materials at high selectivity. Older com-mercial processes were based on the reaction of acetylene with hydrogen cyanide or the reaction of ethylene oxide with hydrogen cyanide followed by dehydration. Both pro-cesses required more expensive starting materials and more
*
extensive safety measures than the modern (SOHIO) process
which is based on the direct ammoxidation of propene, according to
The other advantages of this process are the high selecti-vity and the long lifetime and high actiselecti-vity of the cata-lyst.
After the use of a bismuth molybdate catalyst promoted
by phosphorus (composition 50% Bi
9 P Mo12
o
52/50% Si02)and an uranium-antimony oxide catalyst (composition usb3
o
10> nowadays the socalled multicomponent molybdates {MCM) catalyst (SOHIO 41) is the most important catalyst (7,8). It is composed of a variety of elements like nickel, co-balt, iron, manganese, potassium, phosphorus but always contains bismuth and molybdenum. Today SOHIO's processaccounts for over 95% of the world's installed capacity
of 2.4 106 metric tons/year.
Some of the older ammoxidation processes used multitu-bular fixed bed reactors but all major modern processes use fluidized bed reactors. The advantages of the fluid bed are a better temperature control and a removal of the limitations on propene and ammonia concentrations due to the explosivity of the reactor feed (9), The process ope-rates at 1-3 bar and 673-773 K. The main feature of the process is the high conversion obtained on a once-through basis in the fluid bed, thanks to the high selectivity of the catalyst.
Over the past two decades the rapidly expanding market for acrylonitrile has shifted more and more from the acry-lonitrile elastomers (NBR) to the acrylic fibers and re-sins (ABS and SAN). High impact resistance, low porosity
acrylic copolymers with over 75% acrylonitrile are a
re-cent development in the manufacture of bottles and con-tainers. In the near future the demand for polyacrylamide
may capture a good deal of the acrylonitrile market (10).
1.3. The mechanism of the oxidation and ammoxidation of propene
During the past ten years the commercial production of acrylonitrile has been attended with an increasing number of investigations dealing with the elucidation of the
mechanism of the ammoxidation reaction. Several reviews have appeared about reaction mech~nisms of olefin oxida-tion { 11-18) •
In literature relatively little attention has been paid to the mechanism of the ammoxidation reaction itself. On the other hand research has been focussed mainly on the behaviour of the catalyst under various conditions that differ considerably from commercial ones.
Generally a relatively simple test reaction is used for that purpose. It has been stated that in commercial multi-component catalysts bismuth molybdate performs the cataly-tic role. We decided to investigate the ammoxidation reac-tion itself in order to obtain further knowledge of the re-action kinetics and insight in the mechanism of the acry-lonitrile synthesis.
1.4. Aim and outline of the present investigation
It is the aim of this investigation to derive a reac-tion model based on kinetic results for the catalytic ammoxidation of propene. Unsupported y-bismuth molybdate is used as a catalyst.
In chapter 2 a survey is given of the literature with respect to the subject of this investigation.
The apparatus and the methods of analysis for studying the kinetics are described in chapter 3.
Chapter 4 deals with the preparation and the properties of the catalyst.
Chapter 5 is devoted to the experimental methods
applied in the kinetic studies of heterogeneous catalysis. Chapter 6 deals with the kinetic experiments carried out in the different reactors. The kinetics of the propene oxidation and ammoxidation reactions are studied to de-termine the conditions for the selective reaction proce-dures. The significance of acrolein in the reaction model is investigated by means of its ammoxidation to acryloni-trile. Special attention is paid to the oxidation of ammo-nia to nitrogen.
Thermobalance and pulse reactor experiments are per-formed to investigate the reaction of propene and ammonia with the catalyst in the absence of oxygen in the gas phase. In this way the role of gasphase and catalyst oxygen is studied during the catalytic reaction.
Finally in chapter 7 the reaction model for the
diffe-rent oxidation and ~oxidation reactions based on the
experimental results is given. In a final discussion a mechanistic model is proposed which may contribute to the understanding of the catalytic activity of bismuth molyb-date.
For the explanation of symbols, abbreviations and sub-scripts see List of Symbols.
References
1. Redtenbacher, J., Anw. Chern. Liebigs 47, 114 (1843)
2. Moureu,
c.,
Bull. Soc. Chim. Fr. ~ (3) 424 (1893)3. Idol, J.D. (Standard Oil Co.),
u.s.
Pat. 2.904.580(Sept. 15, 1959)
4. Callahan, J.L., Foreman, R.W., Veatch, F. (Standard
Oil Co.)
u.s.
Pat. 3.044.966 (July 17, 1962)5. Veatch, F., Callahan, J.L., Idol, J.D., Milberger, E.C., Chern. Engng. Progr. 56 (10) 65 (1960)
6. Callahan, J.L., Grasselli, R.K., Milberger, E.C.,
'
Strecker, H.A., Ind. Engng. Chern. Prod. Res. Dev. ~.
134 (1970)
7. Krabetz, R., Chern. Irig. Techn. 46, 1029 (1974) 8. Wolfs, M.W.J., Thesis Eindhoven (1974)
9. Anon, Hydroc. Proc. 56 (11) 124 (1977)
10. Pujado, P.R., Vora, B.V., Krueding, A.P., Hydroc. Proc. 56 (5) 169 (1977)
11. Sachtler, W.M.H., Catal. Rev.
!•
27 (1970)13. Hucknall, D.J., Selective oxidation of hydrocarbons.
Acad. Press, London (1974)
14. Skarchenko, V.K., Russ. Chem. Rev.~~ 731 (1977)
15. Schuit, G.C.A., J. Less. Com. Met: 36, 329 · (1974)
16. Gates, B.C., Katzer, J.R., Schuit, G.C.A., Chemistry of Catalytic Processes, Ch. IV, McGraw Hill N.Y. (1979)
17. van der Wiele, K., van den Berg, P.J., in Bamford
C.H. and Tipper C.H.F. (Eds.) Comprehensive Chemical
Kinetic~ Vol. 20 Complex Catalytic Processes, Chapter
21 12~ Elsevier Publ. Cy Amsterdam 1978
CHAPTER 2
LITERATURE
2.1. Introduction
During the past two decades more than 400 papers and reviews have been published about the selective oxidation of olefins in general and the ammoxidation of propene over bismuth molybdates or bismuth molybdate containing catalysts in particular.
In this chapter we will give a brief literature survey to situate the subject of our investigation. It is not our aim however to add a new comprehensive review of the li-terature to the excellent ones that have appeared already
(1,2,3).
Catalytic oxidation reactions can be explained accor-ding to two different mechanisms, viz.
a) the reduction-oxidation mechanism, proposed by Mars and van Krevelen (4) operating in the higher tempe-rature range;
b) the associative mechanism set up by Roiter (5) at lower temperatures.
In the redox mechanism two separate steps are distinguish-ed: in the first step the hydrocarbon is oxidized with lattice oxygen whereas in the second step the reduced oxide is reoxidized by oxygen of the gasphase. In the associative mechanism a reaction between adsorbed oxygen species and the hydrocarbon occurs. Evidence for the two
mechanisms is obtained from i~otopic exchange experiments,
as has been pointed out by Boreskov (6) and Winter (7,8) and from catalyst reduction experiments carried out by Batist et al (9) and Sachtler et al (10).
Bismuth molybdate catalysts show a high activity in com-bination with a good selectivity both in the oxidation and
in the ammoxidation of propene. During the oxidation of propene besides acrolein only small quantities of carbon dioxide, carbon monoxide, acetaldehyde and formaldehyde are formed. Acrylonitrile is the main product of the propene ammoxidation. Other products are acetonitrile, hydrogen cyanide, carbon dioxide and carbon monoxide whereas acrolein is only a trace product. The stoichiometric equations are
(r 2.1) (r 2.2)
2.2. Kinetics
Broadly there is a great similarity in the overall features of the oxidation and ammoxidation of propene over bismuth molybdate catalysts: the rates of oxidatiop and ammoxidation are both first order with respect to propene and zero order with respect to oxygen. The rate of
ammoxi-dation is zero order in ammonia (11,12).
Activation enthalpies for the formation of acrolein and acrylonitrile show a considerable spread mainly caused by the different catalysts and temperature ranges as can be
seen in table 2.1.
Eact (kJ/mol) Bi/Mo
T(K) Ref. C3H40 C3H3N mol mol-l 84 38 2/1 670 (13) 54
-
1/1 650-825 (14) 121-
1/1 670-730 (15) 159 71 1/1 650-750 (16) 71-
.74/1 700-775. (17) 104 104 .74/1 670 (13) 84 75 BigPMo12o
52/Si02 670-700 (12)Table 2.1. Activation enthalpies (kJ mol-l) for the forma~
tion of acrolein and acrylonitrile over diffe-rent bismuth molybdate catalysts.
The rate of ammoxidation of acrolein, according to the stoichiometric equation
(r 2.3)
is first order with respect to acrolein and zero order
both in ammonia and oxygen (12,13). The activation enthalpy
is 29 kJ mol-l (12).
The rate of oxidation of ammonia over bismuth molybdate according to the stoichiometric equation
+ (r 2 .4)
is first order with respect to ammonia and zero order with
respect to oxygen. The activation enthalpy is 155 kJ mol-l
( 18) •
If we compare the rate constant data presented by Callahan et al (12) with those of Cathala et al (19) carried out with slightly different catalysts it becomes clear that the rate of propene ammoxidation at 700 K is higher than the rate of propene oxidation. Callahan (12) found the rate of acrolein ammoxidation at least twice as high as the rate of propene ammoxidation. Contrary to Shelstad et al (20), Callahan et al (12) conclude that acrylonitrile is formed largely by a mechanism not in-volving acrolein as a vapour phase intermediate.
2. 3. AdsorE,t'i,on of reactants an:d ·products
The adsorption of the reactants and products of the ammoxidation of propene on the catalyst has been studied by Matsuura et al (18,21,22), who investigated not only the adsorption behaviour of a fully oxidized but also that
of a partly reduced catalyst and of Bi2o3 and Moo3•
Mat-suura linked the adsorption data obtained at low pressures and at temperatures between 325 and 475 K to the perform-ance of some oxidation catalysts at atmospheric pressure
reaction mechanism. He distinguishes between two types of adsorption viz. the socalled A-type and the B-type adsorp-tion.
The A-type is an activated, strong and slow adsorption, observed for butadiene, acrolein and ammonia on oxidized Bi 2Mo06 and for acrolein on Bi2o 3• All adsorptions are of the dual site type except the butadiene adsorption which is a single site type. Enthalpies of adsorption are between 88
-1 .
and 100 kJ mol • Prereduction of the catalyst linearly
de-creases the number of A-sites, so an A-site contains an oxygen ion (OA). To allow for the two types of adsorption
and for a similar adsorption of acrolein on Bi2o3 it is
assumed that the A-site contains two anion vacancies (VBi) located at two Bi-ions next to the oxygen ion OA. So the A-site is VBiOAVBi'
The B-type adsorption is a weak and fast adsorption observed for butadiene, acrolein, olefins and ammonia.
This type of adsorption occurs on Bi2Mo06 as well as on
Moo 3 , but not on Bi2o
3
~ On Bi 2Moo6 all B-type adsorptionsare of the dual site kind, except the ammonia adsorption.
Enthalpies of adsorption are in the range of 25-50 kJ mol-l.
Previous reduction does not remove B .. sites, provided the
reduction temperature does not exceed 673 K. At tempera~
ture above 673 K Batist et al (23) found a rapid reduction of the catalyst by butene-1 at degrees of reduction less than 8.3% and without loss of activity after reoxidation. According to Matsuura (21) the reoxidation above 673 K is first order in oxygen with an activation enthalpy of 72 kJ mol- 1 •
The removal of B-sites, mentioned by Matsuura is pro-bably connected with some rearrangement in the solid viz. the formation of metallic Bi in a separate phase. This
phenomenon has been mentioned by Batist et al '(24,25).
B-sites are claimed to be combinations of an anion vacancy (VM
0 ) and two oxygen ions (OB). So the B-site is OBVMooB.
The A-type adsorption of ammonia and acrolein is strong and the enthalpies of adsorption are so high that desorption can only occur at reaction tP;mperatures. Adsorption of.
oxygen on non reduced catalysts does not occur. However,
the catalyst shows some reversible dissociation when
.
the gas phase oxygen partial pressure is lower than the equilibrium oxygen pressure (i.e. p
02eq ( 673 K) = 1.3
10-8 bar). According to Matsuura (21) the adsorption of oxygen on partially reduced catalysts at room temperature is small, rapid and independent of the degree of reduction and does not lead to complete reoxidation.
Between 373 and 673 K the rate of reoxidation is zero
order in oxygen. The enthalpy of activation is 113 kJ mol-1,
a value found also by Batist et al (9) for the reoxidation
of reduced Bi 2Mo 2
o
9• Above 673 K the rate of reoxidationbecomes first order with respect to oxygen and has an enthalpy of activation of- about 72 kJ mol-l, depending on the degree of reduction. According to Matsuura the acro-lein adsorption occurs on both A and B-sites. The strong
and slow adsorption on site A, also observed on Bi2
o
3 is a dual site adsorption. This acrolein adsorption fits the adsorption model proposed by Sachtler et al (26}. This dual site adsorption must influence the dual site adsorp-tion of propene on site B, which needs also OB' This could be verified experimentally as the weak propene adsorptiqn decreased after a pretreatment of the catalyst with acro-lein.
The adsorption behaviour of ammonia on an oxidized cata-lyst is very complicated. Matsuura (18) concludes that the strong dual site adsorption is connected with the A-site, with the donation of a proton to OB of the B-site. The ammonia adsorption on partially reduced catalysts is con-nected with a reduced A-site.
The weak and dual site adsorption of propene on site B decreases after a pretreatment of the catalyst with
ammo-nia. Experimental data of the strong adso~ption of ammonia
on Moo3 and Bi2o 3 are lacking because of the nitrogen for-matio.n already occuring at low temperatures. Kfivanek et al (27) calculated the enthalpy of adsorption of propene under reaction conditions at 440 K on bismuth molybdate to be
2.4. Hydrocarbon surface intermediates
By the use of isotopic labels i t is established by Sachtler et al (10), McCain et al (28) and Adams et al
(29,30) that the oxidation of propene over bismuth molyb-date proceeds via the formation of the allylic interme-diate which is negatively charged. According to Schuit (2) the proton is donated to an 02- ion at the surface, expe-rimentally confirmed by Beres et al (31), and the carbanion is bonded to a metal ion at an anion vacancy. This mecha-nism resembles that taking place during the chemisorption of benzaldehyde on a Sno2
-v
2o5 catalyst, studied bySachtler (32).
Recent molecular orbital calculations by Haber et al (33) carried out for different transition-metal cations support the postulate that the IT-bonding electrons are transferred from the allylic intermediate to the Mo6+ ion. The Mo 6+ is reduced to Mo 5+ or Mo4+ and the positive char-ge on the c
3H5+-ion is concentrated on the terminal C-atoms in a symmetrical distribution. After the transfer of
2-electrons the allylic intermediate is cr-bonded to an 0 as was confirmed by Kondo et al (34). Dozono et al (35) studied the ammoxidation of 3-13
c
propene at 450°C in the presence of bismuth molybdate. Half of 3-13c in the acry-lonitrile was found to be in the CN-group. This points to a symmetrical intermediate also in the acrylonitrile syn-thesis. The appearance of 13c
in both the methyl- and the cyanogroup of acetonitrile, although not completely dis-tributed (60/40 respectively) can only result from bond rupture in the allylic intermediate rather than from the breakage of a C=
C bond in propene, acrylonitrile or acrolein.Further dehydrogenation must lead to a c
3H4 -interme-diate and proton donation to another o 2- ion. Adams et al
(29,30) suggested that the allylic intermediate undergoes this hydrogen abstraction before the incorporation of oxygen which has been experimentally confirmed by means of kinetic isotope effect measurements. Cathala et al (19)
connected this step with a parallel bond rupture which gives rise to degradation products. Daniel and Keulks (36) reported at 725 K an enhanced conversion of propene in a reactor having a large post-catalytic volume. It appeared that a surface-initiated homogeneous gas phase reaction caused the formation of side products. Without the post-catalytic volume this formation disappeared, Recently Kobayashi et al (37} have studied the mechanism of the oxidation of propene by applying a transient response me-thod. It was found that a stable surface intermediate exists which can be formed either from propene or from acrolein.
Further dehydrogenation of the c3H4 intermediate is highly
unlikely. In the case of ammoxidation Cathala et al (19) supposed that dehydrogenation occurs after the formation of allylidene-imine (C
3H4NH). This was also suggested by
Grassel+i et al (38) for the ammoxidation catalysed by Usb 3o 10 • 1
2.5. Nitrogen containin<J surface intermediates ori<Jinatin<J from ammonia
The NH2-intermediate follows from the adsorption expe-riments of Matsuura {18). Ammonia is dissociatively ad-sorbed, according to Matsuura donating a proton to an oxygen ion of the B-site. Ammonia adsorption on a reduced catalyst is supposed to occur preferentially on the anion vacancy left after reduction. Matsuura (18) and Cathala et al (19) drew for mechanistic reasons a parallel between the dehydrogenation of the allylic intermediate and the amide group and supposed the formation of allylidene-imine, synthesized by Bogdanovic et al (39), which proba-bly has adsorption properties comparable with acrolein and butadiene. Germain et al (40) classified the oxides that catalyse the oxidation of ammonia and postulated that the imine-intermediate is a substitute for the double bonded
oxygen ion. He classified Moo3 and not Bi2o
3 among the
As mentioned already for the
c
3H4 intermediate further dehydrogenation of the imine is supposed to be very unlikely.2.6. Role of o~gen
It is generally assumed that the
o
2- ion on the surfaceof the oxide catalyst is responsible for the oxidation of
the hydrocarbon.
Reoxidation by gas phase oxygen leads to the formation of
o
2- but needs four electrons for every oxygen molecule,as follows from the equation
o
2 + 4 e + 2c
{r 2.5)Gates et al (2) suggest a more stepwise donation of electrons, viz. the formation of some intermediate oxygen
- 2- ~
species e.g.
o
2 1o
2 and o at lower temperatures.
In that region the Mars van Krevelen mechanism does not apply as was indicated by Boreskov et al (41) and Sancier et al (42). The evidence of these intermediates is esta-blished by ESR spectroscopy (41). Van Hooff (44) suggested that these intermediates lead to chain reactions. Haber
(45) assumed the oxygen intermediates to be electrophilic reagents and the oxidizing species in the total oxidation of hydrocarbons, whereas lattice oxygen ions are nucleo-philic reagents with non oxidizing properties. Van Dillen
(46) investigated the existence of these species extensi-vely.
I 18 16
By means of 0 - 0 exchange, however, it is
esta-blished by Keulks (47) and Wragg et al (48) that bismuth
molybdate catalysts do not exchange with
o
2 at.temperatu-.res below 773 K in the absence of an oxidation reaction.
. 18
Keulks (47) suggested from experiments with
o
2 gas phase
oxygen and Bi2Mo 16
o
6 that during the oxidation of propeneat 698 K the oxygen of about 500 layers participated in
the reaction and that these layers were oxidized by a rapid diffusion of oxygen from the bulk of the catalyst
rather than by gas phase oxygen. However the gas phase oxy-gen was gradually incorporated in the product. An imme-diate incorporation would be expected if the reaction with catalyst oxygen was confined to the surface layer only on which gas phase oxygen would be chemisorbed. Wragg et al
(48) with experiments at 748-773 K came to the same
con-clusion. As also 180 is gradually incorporated in the
carbon dioxide Keulks assumed that the selective and com-plete oxidation of propene occurs at the same site.
Pendleton et al (49) studied the reaction between
pro-pene and 18
o
2 over bismuth molybdate between 623 and 673 K.They showed the incorporation of lattice oxygen into the acrolein, whereas oxygen for the carbon dioxide formation in that temperature region comes from both the gas phase
and the lattice. Keulks et al (50} however in a later
in-vestigation at 703 K concluded that there is no
distinc-tion be~ween the lattice oxygen incorporated into carbon
dioxide and into acrolein.
Sancier et al (42) determined the relative contribution of sorbed and lattice oxygen during propene oxidation over
silica supported bismuth molybdate between 590 and 670 K in
a pulse reactor and concluded that above 623 K lattice oxygen becomes more important whereas below 623 K the mo-bility of lattice oxygen is low and adsorbed oxygen takes over the role. Recently van Oeffelen (51) found a rapid increase of the electrical conductivity during the
re-duction of Bi2Mo1•02
o
6•06 with propene at 673 K. Heascribed this phenomenom to the formation of bismuth metal particles on the surface. Similar evidence was also
ob-tained by Peacock et al (52). E.s.r.-signals due to Mo5+
were detected when the catalyst was exposed to propene
but these signals were absent when oxygen was added (53).
Sancier et al (54) and Burlamacchi et al (55) obtained the same results.
2.7. Catalyst
excellent catalytic properties of bismuth molybdate and the nature of the active phase have been made by Schuit, Ba-tist and coworkers (2,9,56,58).
It would carry us too far to give a literature survey about the structure of the active catalyst. We refer to the recent review of Gates et al (2) and to chapter 4. 2.8. Models
Some authors have proposed models for the reaction me-chanism of the oxidation or ammoxidation of propene. These models are summarized in table 2.2 without detailed infor-mation. In chapter 7 these models will be discussed.
REACTANTS/INTERMEDIATES C3H6 C3H5 is is C3H4 is NH3 is NH 2 is NH is
c
3a
4o
is END PRODUCTS C~H3NI H~O ~s ~s REOXIDATION site during to . reaction re- 0-!absorbed formed formedladsorbedlformedlformedlformed formed! formedwith with oxi- transfer dize from to on I on on VM.o VMo4 VMo VBi VBi 2-Mo04 VMo..,.VBiiVBi VMo~VBiiVBi VMo VBi VBi 2-Mo04 VMo? VMo VBi 2-Mo04 on VBi VBi on on VBi VBi VBi VBi with OBi OBi 0 Mo? 0 Mo OBi 0Mo 2-0Mo04 OBi OBi OBi OBi OBi VBi VM 04 Mo 4->-Bi VBi Bi+Mo VBi IBi+Mo VMo IMo+Bi 0Moo 4 2-IVBi Bi->-Mo0 3 REF. * (18) (2) (53) (59) (64) (60) (61) ( 6 5) ( 62) ( 63)
Table 2.2. Different models for the reaction mechanisms of the oxidation and ammoxidation of propene.
*
1. Hucknall, D.J., Selective oxidation of hydrocarbons, Road Press London (1974)
2. Gates, B.C., Katzer, ~.R., Schuit, G.C.A., Chemistry
of Catalytic Processes, McGraw Hill, Ch. 4 (1979) 3. Vander Wiele, K., van den Berg, P.J., in Bamford C.H.
and Tipper C.F.H. (Eds.), Comprehensive Chemical Ki-netics, Vol. 20 Complex Catalytic Processes, Chapter 2, 123, Elsevier Publ. Cy. Amsterdam (1978)
4. Mars, J., van Krevelen,
o.w.,
Chem. Eng. Sci. Suppl.l,
41 (1954)5. Roiter, V.A., Kin. i . Kat • .!_, 63 (1960)
6. Boreskov, G.K., Adv. Cat, 15, 285 (1964) 7. Winter,
8, Winter,
9. Batist, G,C.A. I
E.R.S., Adv. Cat. 10, 196 (1958) E.R.S., J, Chem. Soc. A, 479 (1968)
Ph,A., Kapteijns, C.J., Lippens, B.C., Schuit,
J. Catal.
z,
33 (1967)10. Sachtler, W,M,H., Rec. Trav. Chim. 82, 243 (1963) Sachtler, W.M,H., de Boer, N.H., Proc. 3rd Int. Congr. Catal. Amsterdam 1964, Vol. I, 252, NH Publ. Co. Am-sterdam ( 1965)
11. Adams, C.R., Voge, H,H., Morgan, C.Z., Armstrong, W.E.,
J. Catal.
l•
379 (1964)12. Callahan, J.L., Grasselli, R.K., Milberger, E.C.,
Strecker, H.A., Ind. Engng. Chem. Prod. Res. Dev.
1•
134 (1970)
13. Wragg, R.D., Ashmore, P.G., Hockey, J.A., J. Catal.
ll·
293 (1973)14. Gorshkov, A.P., Kolchin, I.K., Gribov, J.M., Margolis,
L.Ya,, Kin. i. Kat.
2•
1086 (1968)15. Keulks, G.W., Rosynek, M.P., Daniel,
c.,
Ind. Engng.Chem. Prod. Res. Dev.
1Q,
138 (1971)16. Cathala, M., Germain, J.E., Bull, Soc. Chim. Fr. 2167, 2174 (1971)
17. Peacock, J.M., Parker, A,J,, Ashmore, P.G., Hockey,
J .A., J. Catal,
g,
398 (1969)19. Cathala, M., Germain, J.E., Bull. Soc. Chim. Fr. 4114 (1970)
20. Shelstad, K.A., Chong, T.C., Can. J. Chern. Engng. 47, 597 (1969)
21. Matsuura, I., Schuit, G.C.A,, J. Catal. ~, 19 (1971)
22. Matsuura, I., Schuit, G.C,A., J. Catal. ~, 314 (1972)
23. Batist, Ph.A., Prette, H.J., Schuit, G.C.A,, J. Catal.
ll·
267 ( 1969)24. Batist; Ph.A., Bouwens, J.F.H., Schuit, G.C.A., J. Catal. 25, 1 ( 1972)
25. Batist, Ph.A., Lankhuijzen, S.P., J. Catal. 28, 496 (1973)
26. Sachtler, W.M.H., Dorgelo, G.J.H., Fahrenfort, J., Voorhoeve, R.J.H., Proc. 4th Int. Congr. Catal. (1968)
(1), 454 (1971)
27. Krivanek, M., Jiru, P.,
z.
phys. Chemie, Leipzig 256,(1) 153 (1975)
28. McCain,
c.c.,
Gough, G, 1 Godin, G.W., Nature, Lond.198, 989 (1963)
29. Adams, C.R., Jennings, T.J., J. Catal. ~, 63 (1963)
30. Adams, C.R., Jennings, T.J., J. Catal.
l•
549 (1964)31. Beres, J., Bruckman, K., Haber, J,, Janas, J., Bull.
Acad. Pol. Sci. Ser. Sc. Chim. 20, (8) 813 (1972)
32. Sachtler, W.M.H., Catal. Rev.
!
(1) 27 (1970}33. Haber, J., Sochacka, M., Grzybowska, B., Golzbiewski,
A., J. Mol. Catal.
l•
35 (1975)34. Kondo, T., Saito,
s.,
Tamaru, K., J, Am. Chern. Soc.2§_, 6857 (1974)
35. Dozono, T., Thomas, D.W., Wise,
H.,
J. Chern. Soc. Far.Transa~t. I 69, 620 (1973)
36. Daniel, C., Keulks, G.W., J. Catal. ~, 529 {1972)
37. Kobayashi, M. , Futaya, R. , J. Catal. 56, 73 ( 1979) 38. Grasselli, R.K., Suresh, D.D., J. Catal. 25, 273 (1972) 39. Bogdanovic, B., Velie, M., Angew. Chern. 79, 818 (1967) 40. Germain, J.E., Perez, R., Bull. Soc. Chim. Fr. 2042
(1972)
41. Boreskov, G.K., 2nd Jap. Sov. Catal. Sem. Tokyo (1973) 42. Sancier, K,M., Wentreck, P.R., Wise, H., J. Catal. 39,
43. Lunsford, J.H., Catal. Rev.
l•
135 (1973) 44. van Hooff, J.H.C., Thesis, Eindhoven (1968) 45. Haber, J., 4th Roermond Conf. on Catal. (1978) 46. van Dillen, A.J., Thesis, Utrecht (1977)47. Keulks, G,W,, J. Catal. ~' 232 (1970)
48. Wragg, R.D., Ashmore, P.G., Hockey, J.A., J. Catal. 22, 19 (1971)
49. Pendleton, P., Taylor, D01 J. Chem. Soc. Far. Trans.--I
72, 1114 (1976)
50. Keulks, G.W., Krenzke, L.D., Proc. 6th Int. Congr.
Catal. ~' 806 (1977)
51. van Oeffelen, D.A.G., Thesis, Eindhoven (1978) 52. Peacock, J.M., Parker, A.J., Ashmore, P.G., Hockey,
J.A., J. Catal. ~' 387 (1969)
53. Peacock, J.M., Sharp, M.J., Parker, A.J., Ashmore,
P.G., Hockey, J.A., J. Catal. ~· 379 (1969)
54. Sancier, K.M., Dozono, T., Wise, H., J .• Catal.
ll'
270 ( 1971)
55. Burlamacchi, L., Martini, G., Ferroni, E., J. Chem.
Soc. Far. Trans. I ~' 1586 (1972)
56. Bleijenberg, A.C.A.M., Lippens, B.C., Schuit, G.C.A.,
J. Catal.
!1
481 (1965)57. Batist, Ph.A., Lippens, B.C., Schuit; G.C.A., J. Catal.
1·
55 (1966)58. Batist, Ph.A., der Kinderen, A., Leeuwenburgh, Y.,
Metz, F., J. Catal. 12, 45 (1968)
59. Haber, J., Grzybowska, B., J. Catal. 28, 489 (1973)
60. Otsubo, T., Miura, H,, Morikawa,
Y.,
Shirasaki, T., J.Catal. ~· 240 (1975)
61. Miura, H., Otsubo, T., Shirasaki, T., Morikawa, Y,,
J. Catal. 56, 84 (1979)
62. Trifiro, F., Kubelkova, L., Pasquon, I., J. Catal. ~'
121 (1970)
63. Sleight, A.W., Adv. Mat, Catal. (eds. J.J. Burton,
R.L. Garten) Acad. Press N.Y. (1976)
64. Grzybowska, B., Haber, J., Janas, J., J. Catal. ~'
150 (1977)
65. Dadyburjor, D.B., Ruckenstein, E., J. Phys. Chem. 82, 1563 ( 1978)
CHAPTER 3
APPARATUS AND ANALYSIS
3-. 1. Introduction
It is generally accepted that the catalytic activity of bismuth molybdate is closely related to its oxidizing properties. In the absence of molecular oxygen for short periods the catalytic activity and selectivity in the oxidation and ammoxidation of propene are not affected
i.e. a reduction-oxidation mechanism is operative.
To study the behaviour of bismuth molybdate under sta-tionary and non-stasta-tionary conditions three different techniques have been used.
A. Reaction kinetics in a stationary state as carried out in different plug flow fixed bed reactors, operating under differential as well as under
inte-gral conditions~
B. The behaviour of bismuth molybdate as an oxidant and the reoxidation of partially reduced bismuth molyb-date are studied in a thermobalance, acting as a semi-batch reactor.
c.
Additional information about the behaviour of thecatalyst under non-stationary conditions at a low degree of reduction is gained with a pulse reactor system.
In order to obtain reliable data the experiments have to meet certain requirements, such as:
- the experimental variables (temperature, flow and reactant inlet concentrations) have to be measured
and controlled accurately~
- the concentration and temperature differences be-tween the bulk gas phase and the catalyst surface should be as small as possible1
- the chemical analysis has to provide for a mass ba-lance over the whole range of experimental concentra-tions;
- isothermicity has to be pursued as much as possible; - as the residence time distribution of the reaction
mixture generally has an effect on the conversion level and on the selectivity of the reaction and more-over strongly depends upon the applied technique and on the experimental variables this distribution should be minimized and properly determined.
All reactors are connected to an on-line gaschromato-graphic analysis system for the determination of the reaction components. However, since. such a GLC-analysis takes at least 15 minutes and has only a moderate sensiti-vity i t is less suitable for the examination of non-sta-tionary processes in which rapid change of the reaction rate occurs.
To gather information about the rate of oxygen deple-tion of the oxidant, the thermobalance in combinadeple-tion with a GLC apparatus with a flame ionization detector is suit-able because i t gives additional information about the weight of the oxidant. Moreover this apparatus is useful for the study of the catalyst reduction and for the oxi~
dation of previously reduced samples. However the thermo~
balance has the drawback that the flow around the catalyst is poorly defined and one has to keep in mind that the concentrations at the catalyst surface can differ consi-derably from those in the bulk gasphase.
As our thermobalance is not resistant to ammonia vapour the ammoxidation reactions could not be studied in this apparatus. Additional information about these reactions and about the behaviour of the catalyst has been obtained with a pulse reactor. The pulse reactor is a good instru-ment to detect small changes in the catalyst properties but, unless the concentrations of the pulse in the reactor and its residence time are carefully studied the kinetic information leaves much to desire.
The conversion level at which one performs the kinetic experiments with the various techniques is a compromise between the low level necessary for the study of the ki-netics at differential conditions and the higher level re-quired for reliable analytical data.
As we deal with moderately or strongly exothermic re-actions, the kinetic data can be affected by non-isother-mic conditions in the fixed bed reactors. We have re-pressed the axial and radial temperature gradients by means of the dilution of the catalyst with silicon carbide
that has good heat conducting properties (A
673 K
=
105J s- 1 m- 1 K- 1 ) (1). Although the commercial operation for
the acrylonitrile production takes place in a fluid bed we have not used such reactors because of the unclear flow pattern and the attrition of our unsupported catalyst.
3. 2. The flow reactor system
The flow reactor system used for the kinetic experi-ments described in section 6.2.4 to 6.2.6 is shown in fi-gure 3.1. It consists of
NH3 C3H6 M He
Figure 3.1 Flow reactor system.
r--Jf'---r;orm111'yncAL SYSTEM
(SEE FIG. 3.3)
1. VAN OYCK MIXER
2. VAPORISER
3. CIRCULATION PUMP
4. REACTOR
(SEE ALSO FIG. 3.2) 5. OXYGEN ANALYSER
SEl SELECTION VALVE
FEED/PRODUCT
M ARTIFICIAL AIR
a) a gas mixing part in which carefu~~y controlled flows
of propene, ammonia, artificial air (20% vol
o
2, 80% vol He) and helium can be mixed in the desired
compo-sitions. For the experiments involving a liquid reac-tant (acrolein) and for the determination of the sub-stance specific correction factors of the liquids in the analysis of the feed and the product composition helium can be passed through a double-walled thermo-stated vaporizer filled with the pure component in question. The desired partial pressure of the reactant can be established by adjusting and controlling the temperature of the vaporizer. It has been ascertained
that the rising heli~ bubbles were completely
satura-ted with vapour. We used the Fourier-number as a mea-sure for the saturation of the dispersed phase
Dt
Fo
=
r2 (3.J.)with D is the molecular diffusion coef.ficient (m2 s- 1 ), t i s the residence time of the bubble in the liquid (s),
r is the radius of the bubble (m). We found Fo > 4,
whereas already at Fo
=
.5 for Biot numbers >> 10 (noconcentration gradient in the continuous phase), the concentration distribution over the bubble is practi-cally constant (2). Moreover we analysed the vapour gaschromatographically at varying liquid levels in the vaporizer and we found a constant vapour concentration. b) a tubular fixed bed reaator, which is made of AISI 321
stainless steel. Three reactors have been used for the various reactions as can be seen in table 3.1.
Reactor B is shown in figure 3.2. An aluminium jacket has been cast around the reactor tube to improve the temperature profile in the reactor. This aluminium jacket is divided in three sections that are indepen-dently heated. The temperature is measured at eight places, three in the catalyst bed and five in the
siliconcarbide bed under and above the catalyst section. The temperature is controlled at the three sections within 1 K with Eurotherm thyristor controllers.
Reactor Catalyst SiC bed
Reaction type dia. length weight dia. weight height
cat.
lliiii lliiii g lliiii bed g lliiii
P-+ACO A 6 90 .6 • 5- • 85 1.7 0/ 53/ 0 a) jACO-+ACN B 20 340 1.5 1.0-1.2 25 50/ 50/120 INH3-+N2 B 20 340 7.5 1. 0-1.2 43 50/100/120 P-+ACN B 20 340 7.5 1. 0-1.2 43 50/100/120 IAco-+ACN c 11 110 .5 1.0-1.2 7 10/ 50/ 10
Table 3.1. Flow reactors: dimensions and fillings. a) 50/50/120 means 50 mm SiC, 50 mm diluted catalyst,
120 mm SiC.
Under stationary reaction conditions the maximum axial temperature differences over the whole reactor at comparable temperature and flow,were as shown in table 3.2.
Reactor Reaction t.Ta (K)
A P-+ACO 'V3
B ACO-+ACN 2
B P-+ACN 2
c
ACO-+ACN 4Table 3.2. Axial temperature differences in the reactor heart line.
These axial temperature differences are mainly due to heat conduction to the colder inlet and outlet lines of the reactor,
II
I
~i
f;. ~r
¥ ~ ~ ~ ~"'
i~ fl-1-=~~~~~~:~~=--l ·~H~.11
rg~~·i "'== • '--'-.--J.I
~~"-'-'--·
J
hfJ~(V/tf(l( .IVSI ZZt {lllf¥,y·~s¥MMJrl«. Figure 3.2 Flowreactor B.Radial temperature profiles were measured in the cata-lyst section of reactor B and C during the ammoxidation of acrolein when the greatest differences could occur
and a temperature difference of not more than 1 K was found in the radial direction.
c) an analysis system.
The feed or the product stream is introduced by means of sampling valves in the analysis system, which will be dealt with in the next section. The feed and product lines are heated electrically and the tempera-ture of these lines is controlled at about 425 K to prevent the condensation of water and hydrocarbons and the polymerization of acrolein and acrylonitrile. 3.2.1. Anal:y:sis
All flow reactors are equipped with an on-line gas chroma-tograph. With this apparatus we can determine quantitatively the components oxygen nitrogen carbon monoxide carbon dioxide ammonia water formaldehyde acetaldehyde acetonitrile acrolein propene acrylonitrile
During the catalytic oxidation of propene we used at fLrst only one GLC-apparatus with katharometer detection
(3). For the separation of the components the column tem-perature had to be programmed in that case from 338 to 433 K with 12 K min- 1 . With the introduction of ammonia for the ammoxidation experiments however the reproducibi-lity of the temperature programmed analysis decreased. Crozat and Germain (4) analysed ammonia and water on two
columns, i.e. on Porapak Q at 360 K one peak for NHj+H
2
o
was obtained, whereas on a PEG column at the sametempera-ture an inaccurate H2
o
determination was carried out.With the introduction of two GLC's at constant tempera-ture (5) i.e. one for the analysis of the low boiling com-ponents and the other with a flame ionization detector for the analysis of the combustible components we took advan-tage of the better separation of the low boiling
compo-VENT
PRODUCT FEED
PRE SURE STABILIZER
VENT
He
He
Figure 3.3 Scheme of the analytical system.
nents at a constant low column temperature, Moreover we could perform a greater number of analyses in a given time. The system with two GLC's, schematically shown in
figure 3. 3, consists of a 4-way Whity-valve (S.E 1) for the
selection of the feed or product stream1 two 8-way Be.cker gas sampling valves S1 and S2 for the sampling of the gas stream and an 8-way Becker valve (SE 2) for the selection
of the columns during the analysis on GLC 1. The sampling
Samples containing formaldehyde, propene, acetaldehyde, acetonitrile, acrolein and acrylonitrile are analysed on the first gas chromatograph GLC-1, a Philips Pye series 104 gas chromatograph with flame ionization detector. The se-cond gas chromatograph GLC-2, a Philips Pye series 44 with katharometer detector is used for the analysis of oxygen, nitrogen and carbon monoxide by means of the separation on a Molsieve l3X column and for the analysis of carbon di-oxide, ammonia, water and propene on a Porapak Q4 column. By means of a selection valve SE 2 the components separa-ted on the Porapak Q4 column are detecsepara-ted in channel num-ber 1, whereas then the carrier gas passes through channel number 2. The components separated on the Molsieve 13X
Figure 3.4 Chroma-togram of an analysis on GLC-1. a"' Figure 3.5 Chro-matogram of an ana-lysis with Molsieve 13X on GLC-2,ch.2. 0 :rf'' Figure 3.6 Chroma-togram of an analysis with Porapak Q on GLC-2 ,ch .1.
column are detected in channel number 2, whereas the carrier gas passes through channel number 1. As the pro-pene peak is found in the chromatograms obtained with GLC-1 as well as with GLC-2 a quantitative analysis of all the components is feasible. For the prevention of a reac-tion between ammonia and acrolein in the Porapak Q4 column of GLC-1 this column is preceded by a small column, filled
with docosanoic acid (c21H
43cooH, melting point 353 K)
which adsorbes ammonia completely. The only drawback is the periodic regeneratio.n that is required for the Pora-pak Q4 column of GLC-2.
The analysing conditions are summarized in table 3.3, whereas the chromatograms are shown in figure 3.4, 3.5 and 3.6.
GLC-1: Philips Pye 104, temperature 523 K with flame ionization detector. Hydrogen 30 cm3 min- 1 • Air 50 cm3 min- 1 •
CoZumn: Porapak Q4, 50-80 mesh
length: 3,5 m, i.d. 2 mm temperature: 423 K
carrier gas flow: 25 cm3 min- 1 He analysing time: 3 minutes
column material: glass
GLC-2: Philips Pye 44, temperatu~e 523 K with katharo-meter detector. Bridge current 150 m A.
CoZumna: a) Molsieve 13X, particle size .5-.7 mm
length: 2 m, i.d. 4 mm temperature: 298 K
carrier gas flow: 25 cm3 min- 1 He b) Porapak Q4, 50-80 mesh
length: 2.75 m, i.d. 2 mm temperature: 333 K
carrier gas flow: 25 cm3 min- 1 He/NH 3 total analysing time: 15 minutes Table 3.3. Analysing conditions for the flow reactor
At low mole fractions the peak area of a component in . a chromatogram is proportional to its mole fraction and the quantitative analysis of the diluted product mixture can be carried out using the relation
(3.2)
with XA mole fraction of component A
XC H : mole fraction of propene A 3 6. peak area of component A A • AC H : peak area of propene
f 3 6. substance specific correction factor of A •
component A
This equation is based on the assumption that fc H = 1.
In order to determine the f-values of the variou~ 6
components, propene-helium gasmixtures of different com-positions are obtained with two plunger pumps (type
Wosthoff) and analysed on the two gas chromatographs. The f-values of the gaseous components are obtained in the same way by mixing with propene/helium gasmixtures in the range of the experimental mole fractions. The f-values of the liquid components can be determined with the thermostated vaporizer already mentioned in section 3.2. The f-value of ammonia is obtained by means of titration. Peak areas are determined with Infotronics model CRS 208 electronic integrators. The reliability of the f-values was checked periodically, because of the continuous ageing of the columns. Although temperature programming to 423 K had a favourable effect on the lifetime of the Porapak Q4 co-lumn the isothermal method is preferred, as was already stated above.
The slight increase in the number of moles as a result of the oxidation and ammoxidation of propene can be ne-glected, especially because the reacting gas mixtures con-tain at least 80% vol He.
For the stability and activity of the catalyst the gas mixtures must contain oxygen and therefore we analysed the oxygen content of the product continuously by means of a Servomex oxygen analyser.
3.3. The thermoba:la:nce
The experiments described in chapter 6, section'6.3 are carried out in a Dupont series 900/950 thermobalance. This apparatus is shown in f i,gure 3. 7. Mixtures of ni troqen and
~---~--·vm
"z AA Cff; "z
1. Quartz glass furnace tube 8. BTS catalyst
2. Furnace 9. Mol sieve
3. Sample holder 10. van Oyck mixer
4. Thermocouple 11. Thermos tate 5. Balance housing
s.
Sampling valve6. Photo voltaic cells AA. Artificial air 7. Counter weights
Figure 3.7 Flow diagram thermobalance.
propene or nitrogen and hydrogen prepared as usual and care-fully freed from oxygen over a bed of 120 gram of reduced
*
BTS catalyst and dried with 30 gram molsieve, are
intra-* BTS stands for the reduced BASF R3-ll catalyst (30% wt
duced into the sample chamber of the thermobalance. This sample chamber consists of a quartz tube with i.d. 2.1 em heated by an electric furnace. The chamber is at atmos-pheric pressure. The experiments are carried out under isothermal conditions. The temperature is measured with a chromel-alumel thermocouple placed just above the 12.5 x 8.4 x 1.20 mm quartz glass sample bucket usually con-taining 75 mg of the oxidant sample. To avoid the presence of reactants in the part of the balance where the weight changes are recorded with a photoelectric cell, this side of the system is continuously purged with nitrogen.
The accuracy of the temperature measurement is
±
1 K.The sensitivity of the thermobalance is .01 mg, which corresponds to an error in the degree of reduction of bis-muth molybdate of .OS%. Before the reduction experiment is carried out the thermobalance is carefully freed from oxygen by means of flushing the system for 15 minutes at room temperature with pure and dry nitrogen. Subsequently the balance is flushed with nitrogen at reaction tempera-ture for one hour. We did not observe a weight
loss larger than .01 mg during this conditioning.period. The effluent of the reduction experiment with propene containing gas mixtures is analysed by means of gas chroma-tography as described in section 3.2.1 for the combustible components.
3.4. The pulse reactor system
As shown in figure 3.8 a constant flow of helium passes through the pulse reactor into an Hewlett Packard 5700 A gas chromatograph with katharometer detector and further into a flame ionization detector. A pulse of a gasmixture containing the reactants for the oxidation and the ammoxi-dation reactions is injected closely before the pulse reactor inlet. After the reaction in the catalyst bed the pulse
is subsequently analysed. The carrier gas is carefully freed from oxygen and dried as described in section 3.3.
r - - - I V E M T
1. Pulse reactor 2. Furnace
3. BTS - catalyst 4. Molsieve 5. van Oyck mixer 6. Thermostate S.· Samling valve V. Switching valve AA. Artificial air Figure 3.8 Pulse reactor system.
Th~ pulse reactor D is a micro reactor, inner diameter
5 mm, length 14.6 mm made of AISI 316 stainless steel. 100 mg Bismuth molybdate, particle size .3-.5 mm is placed· between two plugs of quartz wool. The pulse reactor is
heated by means of an electric furnace. The temperature is
continuously recorded with a chromel-alumel thermocouple in the midale of the fixed bed. The temperature of the furnace is controlled with an Eurotherm thyristor
con-troller. The pulse volume is .155 cm3 NTP. The pressure
in the reactor is 2.5 bar and the carrier gas flow is
18 cm3 min-1 NTP. By means of an 8-way
Becker sampling valve
s,
which is switched pneumatically,a pulse is introduced in the line to the reactor. 3.4.1. Analysis
The analysis of the pulse after reaction differs from that of the flow reactor effluent because of the small sample quantity, the maximum admissible pressure and the analysis
time. The separation of the components is obtained with a
temperature programmed Porapak Q4 column and the detection occurs with a Hewlett Packard 5700 A katharometer. As the
quantities of the combustible p~oducts are very small the
ionization detector of a Philips Pye 104 GLC for the determina-tion of the combustible components. The analysing
condi-tions are summarized in table 3.4.
Column: Porapak Q4, 80-100 mesh length: 2.8 m i.d. 2 mm
temperature programs and analysing times: a) p~opene/helium
333-393 K with 2 K min- 1 analysing time: 50 minutes b) p~opene/oreygen/helium
16 minutes on 333 K, 333-393 K with 16 K min- 1 8 minutes on 393 K
analysing time: 40 minutes c) ammonia/helium
333 K constant temperature analysing time: 10 minutes
d) p~opene/ammonia/helium
16 minutes on 333 K, 333-423 K with 16 K min- 1 16 minutes on 423 K
analysing time: 38 minutes
Katha~omete~-deteato~: HP 5700 A, temperature 523 K bridge current 150 m A
Flame ionization-deteato~: Philips Pye 104, temperature
523 K
Hydrogen: 30 cm3 min- 1 Air: 50 cm3 min- 1
Table 3.4. Analysing conditions for the pulse reactor system.
3.5. Safety
3. 5. 1. Toxicity
As acrolein and acrylonitrile are highly toxic sub-stances (6), all experiments are carried out in a hood
with adequate exhaust ventilation. Due to its extreme lachrymatory effect (the smelling limit is • 2 to • 4 ppm
(7)) acrolein serves as its own warning agent. It affects particularly the membranes of the eyes and respiratory tract.
Acrylonitrile closely resembles hydrogen cyanide in its toKic action. By inhibiting the respiratory enzymes of tissue it renders the tissue cells incapable of oxygen usage. In table 3.5 the Treshold Limit Values (time
* **
weighted average) (TLC-TWA) and the LCLo values are
given.
TLV - TWA (8) LCLo (9)
ppm mg m- 3 ppm
c
3H40 • 1 .25 150/10 min (inhalation human)4. Qa)
c
3H3N 9.0 600/4 hrs (inhalation cat)a) DuPont de Nemours has reduced the TLV-value to 2.0
ppm (I 0)
Table 3.5. Treshold Limit Values for acrolein and acrylo-nitrile.
_We calculated the mean concentration of acrolein in the hood when condensation and subsequent destruction would have been omitted as .19 mg m- 3 • For acrylonitrile a value of .18 mg m- 3 would have applied. These values are ·smaller than the adopted TWA-values.
*
**
TLV-TWA
=
"the time weighted average concentration fora normal 8 hour workday to which all workers may be repeatedly exposed, day after day, without adverse effect" (8).
LCLo
= "the lowest lethal concentration of a
sub-stance in air, which has been reported to have caused death for a given period of ex-posure" (11).
3.5.2. Flammability and explosive ranges
Most of the reaction components are flammable and have explosive properties over wide ranges when mixed with air, as can be seen in table 3.6. Propene is the main hazardous substance, whereas acrolein and acrylonitrile are Class I flammable liquids (6).
Flash point Ignition temp. Explosive range
K K vol %
C3H6 165 770 2 -11
NH 3 924 15 - 28
c
3H40 255 551 2.8 - 31c
3H4N 273 754 3.1 - 17Table 3.6. Flammability and explosive ranges of the main reaction components (6).
Therefore a flame extinguisher is included in the feed line to the reactors. Strong ventilation is required.
References
1. Y.S. Touloukian ed., Thermophysical properties of High Temperature Solid Materials Vol, 5, 125 (1967) The McMillan Cy., New York
2. H.A.C. Thijssen, Masstransfer Processes, Lecture-notes 6.605, 10,25 (1973) University of Technology, Eind-hoven
3. Verhaar L.A.Th., Lankhuijzen S.P., J, Chrom, Sci.~,
4. Crozat, M., Germain, J.E., Bull. Soc. Chim. Fr. 3526 (1972)
5. A.P.B. Sommen, Int. Report TC (1975) University of Technology, Eindhoven
6. Sax, N.I., Dangerous Properties of Industrial Mate-rials, 4th Ed., van Nostrand Reinhold Cy., N.Y. (1975) 7. Hommel, G., Handbuch der gefahrlichen Guter, 2 Aufl.,
Springer Verlag, Berlin (1973)
8. Association of American Governmental Industrial Hy-gienists, Index TLV, Am. Ind. Hyg. Ass. Journ. 37, 721 (1976)
9. The International Technical Information Institute; Toxic and Hazardous Industrial Safety Manual, Tokyo
(1977)
10. Anon, Chern. Weekblad
1.1
(22) 1 (1977)11. Registry of Toxic Effects of Chemical Substances,
u.s.
Dept. of Health, Education and Welfare, NIOSH (1977)
