The catalytic reduction of nitrate and nitric oxide to hydroxylamine : kinetics and mechanism

The catalytic reduction of nitrate and nitric oxide to

hydroxylamine : kinetics and mechanism

Citation for published version (APA):

vd Moesdijk, C. G. M. (1979). The catalytic reduction of nitrate and nitric oxide to hydroxylamine : kinetics and mechanism. Technische Hogeschool Eindhoven. https://doi.org/10.6100/IR22369

DOI:

10.6100/IR22369

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

providing details and we will investigate your claim.

THE CATALYTIC REDUCTION OF NITRATE

AND NITRIC OXIDE TO HYDROXYLAMINE:

KINETICS AND MECHANISM

PROEFSCHRIFT

TER VERKRIJGING VAN DE GRAAD VAN DOCTOR IN DE TECHNISCHE WETENSCHAPPEN AAN DE TECH-NISCHE HOGESCHOOL EINDHOVEN, OP GEZAG VAN DE RECTOR MAGNIFICUS, PROF. IR. J. ERKELENS, VOOR EEN COMMISSIE AANGEWEZEN DOOR HET COLLEGE VAN DEKANEN IN HETOPENBAAR TE VER-DEDIGEN OP VRIJDAG 12 OKT. 1979 TE 16.00 UUR

DOOR

CORNELIS GERARDUS MARIA VAN DE MOESDIJK

GEBOREN TE LOSSER

DIT PROEFSCHRIFT IS GOEDGEKEURD DOOR DE PROMOTOREN PROF.DB. G.C.A. SCHUIT

EN IR. J. GEUS

Aan Trees,

DIT PROEFSCHRIFT WERD BEWERKT IN HET CENTRAAL LABORATORIUM VAN DE NAAMLOZE VENNOOTSCHAP DSM

CONTENTS

1. PRODUCTION OF CAPROLACTAM Summary

1.1. The importance of the inorganic chemistry of nitrogen 1.2. Historical review of the caprolactam industry

1.3. Nylon 6 in the fibre industry 1.4. Description of industrial processes 1. 5. References

2. APPARATUS, ANALYSIS, EXPERIMENTAL PROCEDURES AND TECHNIQUES Summary

2.1. Introduction

2.2. Purification of phosphoric acid 2.3. Preparation of hydroxylamine phosphate 2.4. Preparation of buffer solutions

2.5. Catalyst pretreatment 2.6, Analysis procedures 2,7. Activation procedures

2.8. Experimental procedure and apparatus 2.9. Evaluation and calculation

3. NITRATE REDUCTION 9 9 11 24 25

47

53 53 53 54 55 55 56 57 58 61 SumMary 69 3.1. Literature survey 69 3.2. Preliminary experiments 76

3,3, Features of activation by non-precious metal compounds 79

3.4. The system (Pd + Pt)/C 102

3.5. Influence of reaction conditions on reduction rate 109 3.6. Mass transport, influence of stirring rate

and catalyst particle size distribution 3.7. References

130 140

4. THE REDUCTION OF NITRIC OXIDE Summary

4. 1. Literature survey 4.2. Introduction 4.3. Activation of Pt/C

4.4. Influence of reaction conditions 4.5. Pd/C and Pd + Pt/C catalysts 4.6. References

5. THE HYDROGENATION OF HYDROXYLAMINE

Summary

5.1. Literature survey 5.2. Stoichiometry

5. 3. Influence of reaction conditions 5.4. Influence of the catalyst system 5.5. References

6. THE DISPROPORTIONATION OF HYDROXYLAMII~E

Summary

6. 1. Literature

6.2. Introduction and stoichiometry of the react ion 6.3. Influence of reaction conditions

6.4. Influence of catalyst composition 6 .5. References

7.

THE CATALYTIC REACTION OF HYDROXYLAMINE WITH NO Summary 7. 1. Literature survey 7.2. Introduction 7.3. Stoichiometry 7.4. Influence of reaction 7.5. Influence of catalyst 7.6. Tracer experiments

7.7.

References conditions composition 143 143 145 147 152 167 170 173 173 176 177 185 190 191 191 193 196 203 207 209 209 211 212 214 223 225 227

8. THE REDUCTION OF N

20 TO N2 Summary

8. 1. Introduction

8.2. Influence of reaction conditions 8.3. Influence of the catalyst system 8.4. Simultaneous reduction of nitrate

9. r,ENERAL DISCUSSION Summary

and N₂0

9.1. The catalyst in nitrate and nitric oxide reduction 9.2. The reaction medium

9.3. Kinetics and mechanism 9.4. Proposed mechanism 9. 5. References

APPENDIX

10. THE STRUCTURE AND TEXTURE OF ACTIVATED CARBON CATALYSTS Summary

10.1. Activated carbon

10.2. Particle size distribution

10.3. BET surface area and pore size distribution 10.4. SEM examination of activated carbon

10.5. Metal surface area 10.6. References SUMMARY SAMENVATTING DANKWOORD CURRICULUt1 VITAE 229 229 231 232 232 237 237 244 248 254 255 257 257 259 260 261 262 268 271 273 275 277 7

1. PRODUCTION OF CAPROLACTAM

This thesis deals with the kinetics and mechanism of hydroxylamine syn-thesis by the catalytic reduction of nitrate and nitric oxide. Because

of the industrial importance hydroxylamine as a intermediate in

synthesis a literatur'e svJVey is given in 1 on the

of ca;prolaetam. discussion of the

the chemistry of nitrogen, a historical review is trial routes to ca;prolactam and its intermediates as

the ind:'us-several

30 years. Special reference is made to the

at DSM, these the most familiar to the author. Next,

the role of nylon 6 in the industry and the consumer products based

on nylon 6 are described. A detailed description the different

commerc-ial routes to the f::'wo intermediates, and

hydroxyl-and an overall description other commercial routes

these intermediates. Chapter with a table the

companies caprolactam, sites,

and technologies utilized.

1.1. The of the chemistry of nitrogen

Nitrogen is one of the fundamental elements necessary for life,

Soon after its discovery it was recognized as an important factor in plant growth. lt is almost entirely due to the achievements of Haber and Bosch, who developed a suitable process for the synthesis of ammonia, that the

industrial chemistry of nitrogen has developed into a field of such enor-mous importance. Today, a wide variety of commodities based on ammonia are produced on a large industrial scale. Much of the product1_s signifi-cance is of course due to the fact that ammonia and its oxidation product nitrate are natural links in the biochemical nitrogen cycle. Without the addition of N-fertilizers in modern agriculture it would be impossible to produce the large quantities of food necessary to feed the present world population. Although the nitrogen cycle has been recognized for a

long time, the mechanism of individual steps is still not yet fully under-stood, especially with respect to the separate catalytic steps and

mediates involved. On the oxidation of ammonia and the bacterial reduction of nitrate in the soil there exists a vast body of 1 iterature. The same holds for the catalytic oxidation of ammonia to nitrous gases, but again the exact mechanism of the reaction is still unknown.

Studies on the electrochemical reduction of nitrate, nitric oxide and nitrite and the electrochemical oxidation of hydroxylamine and nitrite have been under way for more than a century. A considerable amount of progress has been made by electrochemists during the last ten years, following the development of the rotating disc electrode, which has proved to be a powerful device in electrochemical studies.

Another area of work deserving attention is the catalytic removal of NO •

X

Owing to the gradually more stringent air pollution legislation imposed in most of the industrialized countries, an enormous number of publica-tions dealing with this subject have appeared in the last decade.

The above survey of the relevance of nitrogen and its compounds, while in-dicating the areas of major importance, is of course by no means complete. Countless other examples could be given, such as the reduction of nitrate to nitrite or ammonia, which has for a long time been used as an analyt-ical tool to determine the nitrate content in various types of media; the corrosion problems and uses of nitrate in metallurgy; and the problems with batteries containing nitrate.

In all these studies, by scientists from a wide range of disciplines, the intricacy of the nitrogen system is obvious, but it is interesting to note that these studies also point to the same type of reaction paths and in-termediates. One of the key intermediates frequently mentioned is hydro-xylamine. As discussed in the general introduction, this compound also happens to be one of the major precursors in the industrial manufacture of caprolactam, the monomer of nylon 6.

World caprolactam production today totals more than 2 mill ion tonnes per year, a clear indication of the importance of hydroxylamine as an indus-trial commodity. This is one of the main reasons for the development, in the last 30 years, of a number of industrial processes in which nitrogen-oxygen compounds derived from ammonia are reduced to form hydroxylamine.

In these, mainly catalytical, processes the same types of reaction and by-product are encountered as in the analytical, biochemical and electro-chemical studies mentioned above.

1, 2. HISTORICAL REVIEW OF THE CAPROLACTAM INDUSTRY

1. 2.1. Developments WorZd War Two

The synthesis of f -caprolactam - a molecule consisting of a saturated ring of six C-atoms and one N-atom, with a double-bonded oxygen atom on the C-atom adjacent to the N-atom- was first reported by S. Gabriel 1 in 1899. The starting material used by Gabriel was E-aminocaproic acid.

In 1900, 0. Wallach 2 reported the synthesis of caprolactam from pimel-ic acid. Later on he was able to prepare caprolactam by heating cyclohex-anone oxime (prepared from phenol) in sulphuric acid, a reaction that had been familiar since 1886 and was known as Beckmann1_s 3 _{rearrangement. Low} yields of caprolactam were found in addition to a polymeric tar product. In January 1938 (almost 40 years after the discovery of caprolactam) Schlack 4 of I.G. Farben produced a spinnable polymer from caprolactam, which was given the name 11_Perlon".

Prior to this, Carothers 5 • 7 of Du Pant had succeeded in producing from adipic acid and hexarethylene diamine a polymer which he called "Nylon".

In the same paper Carothers also describes polymerisation of various he-terocyclic compounds, bein9 unsuccessful, however, with piperidone and pyrrolidone, He concluded that caprolactam could not be polymerized, even though he had already published a paper 6 on a product consisting of cap-rolactam entities the year before. Du Pont developed a commercial process for "Nylon", while I.G. Farl:ien started to develop their "Perlon" process. In .1938- 1939 an extensive exchange of know-how took place between I.G. Farben and Du Pont.

I.G. Farben acquired rights on Du Pont1_{s patent for preparation and} po-lymerisation of "Nylon" 8•9• Du Pant also licensed the process in Europe to I.C.I., Montecatini and Rhodiaceta.

In Japan, Toyo Rayon obtained samples of 11_{Nylon", and after comparative} analysis with "Perlon", the firm decided to start production of the lat-ter, A pilot plant with a capacity of 5 kg/day was started up in 1942 10

I.G. Farben started development of a continuous caprolactam process

simultaneously at two sites. The process was based on phenol, the route fol· Jowed by 0. Wallach in 1900. A research group led by H. Hopff set up the process at the former BASF Ludwigshafen site, while a team headed by

J. Giesen worked out a process at the Leuna works. Both groups developed a continuous process starting from phenol, which was hydrogenated in the

1 iquid phase at 150 °C and 25 bar on a nickel catalyst to produce cyclo-hexanol. The cyclohexanol was distilled twice and dehydrogenated to cyclohexanone in the gas phase at 400 °C on a zinc/iron oxide catalyst, Hydroxylamine was produced according to Raschig1_{s method from sodium}

nitrite at Ludwigshafen and from ammonium nitrite at the Leuna works. The better lactam quality and yields achieved at Ludwigshafen were attributed to the use of NaOH for neutralisation, as opposed to the NH

3-water used at the Leuna plant. There were, furthermore, differences in the equip-ment and method used for the Beckmann rearrangeequip-ment.

In 1943, both groups were operating small plants with a capacity of 150 tonnes a month. At the same time Schlack had developed the polymerisation and spinning of caprolactam on the basis of Du Pont1_{s technology, at}

Ber-1 in. Spinning plants were sited at Landsberg (Premnitz) and the Schwarza works in 1943. The product was used mainly for aircraft tyres and para-chutes. A more detailed description is given by Achilladelis

S,9,

In 1945, I.G. Farben was placed under Allied control, and was shortly afterwards split into BASF, Hoechst, Bayer and the Leuna works in Eastern Germany. The I.G. Farben process was freed from patent restrictions in most AI lied countries and the process was published after the war. BASF restarted caprolactam production after the war, as did the Leuna works. Hopff stayed at BASF, Schlack at Hoechst; Giesen first remained a while with Snia Viscosa before moving to Bayer and, later on, with eo-workers Dr. Zorn and Dr. Kahr from the Leuna works to the Emser plant (lnventa). Dr. Kahr later

(1959)

joined BASF 9• A scheme of the I.G. Farben route to caprolactam is given below.

Oleum NaOH/NH40H

1, 2, 2, Developments after World War Two at Dutch States Mines 12

After the war, DSM obtained the permission of the Dutch Goverment to apply the l.G. Farben patents, which had been freed from patent restrictions as part of war reparations. Know-how was also obtained from AKU, who had al-ready investigated 11_Nylon11 _{and "Perlon" on a preparative scale before the}

war. Due to the fact that AKU had obtained rights to Du Pont1_{s patents on} polymerisation and spinning and also, some time later, those of Rhodiaceta, DSM only developed a caprolactam process.

13 16

From 1946 to 1948, a research group headed by R. Zeegers ' made a study of caprolactam synthesis on laboratory scale with reference to the CIOS reports, the I.G. Farben patents and the know-how of AKU. In 1948-1949 a caprolactam pilot plant was constructed by DSM, while polymerisation and spinning facilities were completed by AKU. After investigation of several possibilities (phenol/cyclohexane/aniline) a phenol-based process was cho-sen. A 1 icense was obtained from Tyrer for a phenol plant based on benzene, sulphuric acid and calcium hydroxide via the potassium salt of benzene sulphonic acid. A 6000 tonnes/year plant was started up in January 1953, but was closed down again in October 1955 owing to economic and technical problems.

In 1964 a plant using a new process based on toluene was started up at Rotterdam by Chemische lndustrie Rijnmond, CIR in a (50-50) joint venture with DOW, and based on DOW know-how. (CIR was later to become a 100% daughter of DSM,) The process has since then been improved considera-bly as a result of research by DSM, and has been expanded in three steps to 130,000 tonnes/year,

A second phenol plant (100,000 tonnes/year) was started up in December of last year (1978) 17• Part of the phenol necessary for caprolactam production has throughout the years been purchased outside DSM. At the beginning of 1950, a semi-commercial plant for caprolactam was started up with a decision being made in the same year for construction of a 3600 tonnes/year plant. The semi-commercial plant produced lOO ton-nes of caprolactam in 1951, this total being increased to 500 tonton-nes in 1952. In 1950 a bid was made to obtain Leuna know-how on the dehydrogenation of cyclohexanol to cyclohexanone from Holzverzuckerungs AG (later Emser-werke and lnventa) at Zurich (Dr. Zorn), without result, however.

how on this part of the process was oBtained from Montecatini some months later. The commercial plant incorporating DSM developments on hydroxyl-amine preparation, oxime rearrangement and lactam purification was built

in 1951 - 1952 13,16; 18,21.

Around the same time extensive research was started on the purification of crude lactam by a group consisting of Boon, Kampschmidt, Zeegers and Soe-terbroek, and later on Ottenheym, with the results being incorporated in the actual production plant by Kretzers. Continuous research on hydroxyl-amine preparation, lactam purification, oximation and rearrangement sub-sequently improved the process and the caprolactam quality considerably22

•3

2• Owing to a growing demand for caprolactam, the capacity of the plant was gradually increased from year to year.

The 2000 tonnes of caprolactam produced in 1953 was raised to 3000 tonnes in 1954, 6000 tonnes in 1958 and 9000 tonnes in 1959.

From 1954 onwards, pure white crystal! ine ammonium sulphate (A.S.) was produced as a by-product in about five times the amount of caprolactam, and found extensive application as a fertilizer. In 1953 research was also started on the oxidation of cyclohexane in air, in view of the long-range forecasts for caprolactam demand. In October 1956 a continuous pilot plant was commissioned, headed by Steeman as research leader.

In 1959 research on the cyclohexane oxidation and a plant design had been finished, but a decision on a cyclohexane-based plant was postponed tem-porarily

33 •34•

The oxidation of cyclohexane with boric acid was also stu-died by Steeman and eo-workers 35,37.

The boric acid economy in such a plant was felt to be too complicated, however, and further research was stopped. In 1964 the first (25,000 mtpa) cyclohexane oxidation plant was started up at Geleen; this subsequently became standard in nearly every Stamicaroon caprolactam plant. The cyclo-hexane is obtained by gas-phase hydrogenation of benzene, for which a

DSM process had also been developed 38• The catalyst for this process, Pt/Al

203' is more expensive than the Ni-catalyst used by IFP, but has an extremely long lifetime ( > 10 years) and is not very sensitive to catalyst poisons like sulphur.

In .1959 i111portant improve111ents were 111ade on the dehydrogenation of hexanol to cyclohexanone, and the direct hydrogenation of phenol to cyclo-hexanone was patented by Phielix

39,

40 •

Later on, an even better catalyst for the dehydrogenation of cyclohexanol to cyclohexanone was found. In the second half of the fifties the BASF

pat-ents of Dr. K. Jockers c.s. on the reduction of NO were published,

Three patents on the catalytic reduction of nitric acid were also

publish-. 42 44

ed by E. Marhofer and eo-workers of Spencer Chemicals Co. ' • Marhofer reduced nitric acid on precious metal catalysts in small-scale autoclave experiments. Yields of 20- 65% a~e mentioned in these patents. A license was obtained from Spencer C~, but the research carried out at DSM on the

basis of these patents from 1957 - 1959 (by Hoogendoorn, de Rooij and Muytjens) on continuous bench scale was not succesful. Prohibitive factors were: 1. accelerated deterioration of the catalyst due to the solubility of the precious metal in nitric acid, making it necessary to keep the

cata-lyst under hydrogen, 2. rapid deactivation of the catacata-lyst, and 3. problems with hydroxylamine nitrate, which can decompose vigorously under certain conditions. Research was therefore stopped in 1960.

With demand for caprolactam rising fast after 1959 (18,000 tonnes in 1961 and 35,000 tonnes in 1963) and prospects even better for the years

follow-ing, there was a growing concern as to whether fertilizer markets could be maintained for the increasingly large amounts of by-product ammonium sulphate.

Research was therefore restarted on a small scale in the second half of 1963, aiming at a hydroxylamine process yielding little on no ammonium sulphate by-product. This research was initiated by de Rooij and his eo-worker Elmendorp, joined later By Aggenbach.

What they were aimin!,!J at, an ammoniumsulphate-free process, was a near con-tradiction, for all hydroxylamine preparation methods known are performed at low pH valJles at which the hydroxylamine salt is stable, whereas the oximation reaction- an equiliBrium reaction- can only be completed at pH values around 4.5. Moreover, the pH drops as a function of the hydroxyl-amine salt conversion due to the release of the free acid. In spite of these apparent impediments de Rooij continued on his brilliant idea to perform the oximation with hydroxylamine salts in a highly buffered me-dium. If the oximation could be carried out in the same solution in which the hydroxylamine is synthesized, it vmuld no longer be necessary for the hydroxylamine salt to be isolated or concentrated. The concept of the recycle process was thus born 45,47

At first a route was investigated producing a nitrite ester from nitrosyl sulphuric acid and an alcohol, with subsequent catalytic reduction of the ester to hydroxylamine and the original alcohol

48 •49•

This route was soon abandoned.

Patent restrictions led to hydroxylamine formation by NO reduction in a buffered liquor also being abandoned, although patents have been granted on a NO process. Research efforts were subsequently directed to the formation of hydroxylamine by nitrate reduction in a phosphoric acid buf-fered liquor, for which a 1 icence under the Spencer patents had already been obtained before 1960, De Rooij showed that the low yields of only 20 - 65 % mentioned in the Spencer patents could be raised to above 80 %. A Pd/C catalyst was found to function optimally, The highest time-volume yields were attained with a relatively high catalyst metal content. This contrasts with the Pt catalyst for the NO reduction, which gives the high-est selectjyities at metal loadings of 1 -

3

%platinum, Carbon and gra-phite turned out to be the most inert and stable carrier materials. The com-plete process was elaborated on a continuous bench scale.

The hydroxylamine-containing I iquor obtained by the reduction of nitrate in a phosphate medium was able to be used for the oximation step with excellent yields (~8% conversion, lOO% selectivity),

The main problem was the phosphoric acid quality to be used in the pro-cess. lt was found that while some phosphoric acids (of A.R. quality!) activated the catalyst durrng the first days, they also led to rapid de-activation of the catalyst after some time. Analysis of the phosphoric acid, in particular, showed diverging levels of impurities, but could not predict the activati-ng or deactivating effect of the phosphoric acid.

Investigations were started by Mars and eo-worker Gorgels to solve this problem. Mars afterwards became a professor at the Technical University of Twente, and this study was continued by Duyverman, who investigated the activating and deactivating behaviour of nearly all the elements of the periodic system

so.

Duyverman also developed a purification method 59 for phosphoric acid.

A pilot plant was built in 1966 and brought on stream in December of that year, headed by Damme, who was later on joined by van Goolen. The pilot plant ran to the end of 1969, with van Goolen acting as research leader for part of this period~

The pilot plant was a stirred reactor with separate filtration through filter candles (70 I liquor}. The circulation of liquor over the filtra-tion secfiltra-tion was maintained by the Mammoth pump principle. The capacity of the pilot plant was about 3000 tonnes of oxime per year. In this

re-51-58 search period, several methods for optimizing the recycle process ' 60

-63were investigated by de Rooij on bench scale. These were then put to the test in the pilot plant by Damme and van Goolen. The most efficient of which was utilized in the commercial plant. The new process became known as the Hydroxylamine Phosphate Oxime (HPO) process, later on also referred to as the DSM Low Sulphate Process (LSP).

Aided by the rules developed by van Dierendonck 64 and model experiments performed in a small bubble column, Van Goolen was able to enlarge the 30 l stirred hydrogenation reactor directly to a bubble column of about 60m3 • This enormous scaling-up factor of 1000 had not before 65 been achieved.

When the decision on the first commercial plant (Nypro UK, Fl ixborough) was taken, relatively I ittle kn0w-how existed on the fundamentals of the process, a situation which, it was felt, had to be improved. A small-scale 5-1 autoclave (3 l liquor} provided with a filtering system and working con-tinuously and fully automatically under reaction conditions identical to those in the commercial plant was therefore built to study the nitrate reduction. This apparatus was started up and run by a small group led by the author from 1970-1974; the equipment was also used for trouble-shooting and catalyst testing for the commercial plants. The research on the

kinet-ics of the nitrate reduction was carried out in batch experiments on la-boratory scale in the period after 1970, with the author as research leader. A selection of these experiments is reported in this thesis. In the period

1970-1973 the more active Pd-Pt catalyst 66 was also discovered, fol !owing Duyverman's idea to investigate in more detail combinations of precious metals, in analogy to those described in known patents on other processes. Duyverman had left before this, and did not have the opportunity to inves-tigate the systems himself.

The DSM caprolactam processes have been licensed widely by Stamicarbon, the engineering and licensing subsidary of DSM. In 1964 the first contracts were signed. Two caprolactam plants (50,000 mtpa), based on cyclohexane

oxidation and Raschig's route, were sold to Techmashimport in Russia, to be sited in Grodno and Kemerovo. In 1965 the contract was signed for a 20.000 mtpa plant in Augusta (USA) for Nipro lnc, DSM participating with P.P.G. for this project. A similar plant was constructed for Nypro (UK) at Fl ixborough in 1967-1968; DSM, (Fisons Ltd), and the National Coal Board

(NCB) are the participants in Nypro (UK). Around 1971 a 16,500 mtpa plant was built for Monomeros Columbia at Barranquilla in Columbia and at the same time a 40,000 mtpa plant in Salamanca (Mexico) for Univex S.A. In both plants DSM has a minor shareholding. Another conventional (Raschig) plant of 33,000 mtpa capacity was started up some years later in Ulsan (S. Korea) for the Hankook Caprolactam Corporation.

In the meantime, the first HPO plant had been started up in May 1972 for Ube Kosan Sakai in Japan, followed shortly by the Nypro {UK) plant at Flix-borough and the Nipro (USA) plant at Augusta, USA. All three plants had a capacity of 70,000 mtpa. In 1977 the 50,000 mtpa plant at Kaohsiung in Ch.ung Tai was started up, and subsequently a second plant of the same

capac-ity at Toufen in 1978. A 35,000 mtpa plant was also commissioned in 1978 for Nitrocarbono at Cama~ari-Bahia in Brazil.

These plants are based on cyclohexane oxidation and Stamicarbon1_{s Low}

Sulphate (HPO) Process. The caprolactam facilities at Geleen (NL) were ex-panded in 1976 with a 70,000 mtpa (HPO) plant, whilst the small cyclohexane oxidation plant was closed. All caprolactam production at Geleen is phenol/ HPO-based with the being Raschig route for hydroxylamine. After the disaster at Flixborough 190 in 1974 in the cyclohexane oxidation section the plant was reconstructed, and will be started up again in 1979 at the original 70,000 mtpa capacity. Production will be based on phenol hydrogenation and HPO. The most recent contract was signed in 1978, for a new 100,000 mtpa caprolactam plant (cyclohexane oxidation/HPO) for the Hyundai

Corporat-ion to be constructed at Yeo Cheon (S. Korea).

l 1~- 'd 8-10 67-86

1.2.3. Deve opments after tr~ war outs~ e DSM ~

8?-95

1o2.J.1. Eastern Europe

The Leuna works was situated in the Russian sector of Eastern Germany and was in fact completely cut off from the Western part of I.G. Farben. Production at the Leuna plant was slowly restarted after the war. During the war the caprolactam was produced batchwise from step to step, but a

pilot plant for a continuous process was built in 1945. This pilot plant was, however, never started up, and when production was restarted after the war it was with the discontinuous process. Developments were slowed down considerably, probably due to the fact that all leading scientists (Giesen, Kahr, Zorn) had left the Leuna works 89• lt was not before 1950 that the production level was up to 1000 mtpa again. Capacity was gradually increased to about 9000 mtpa in 1959.

Between 1960 and 1968 the process was improved and changed from a discon-tinuous to a condiscon-tinuous process. The capacity was Increased from 13,000 mtpa In 1963 to 25,000 in 1968 and finally to 40,000 in 1975.

Although some research has been carried out on the cyclohexane oxidation, the process is still based on phenol obtained from coal and hydroxylamine according to the old Raschig process (absorption of nitrous gases in am-monia 87-9°). In other Eastern Europe countries (e.g. Russia) western tech-nologies have mainly been used for caprolactam production (lnventa, BASF, Snia Viscosa, OSM and Zimmer). Zaklady Azotowe, Poland's state-owned caprolactam producer, recently announced a cyclohexane oxidation process -the Cyclopol process- in a specially designed multistage reactor. A commercial plant of 25,000 mtpa capacity was built in 1974 in Tarnow. A 50,000 mtpa plant has been constructed in Pulaway, while a 80,000 mtpa plant based on this process is planned in Czechoslovakia. Cyclohexane is obtained via vapour-phase hydrogenation of benzene. The cyclohexane is conventionally oxidized with a Co catalyst, the purification of the oxi-dation oil differing from known cyclohexane oxioxi-dations in the neutrali-sation and saponification steps 91 •92• Hydroxylamine is obtained via the Raschig route. A plant producing hydroxylamine by NO reduction according to lnventa's technology has been under construction since 1974; as of 1978, however, start-up had not yet been reported.

1,2.3.2.

Toyo Rayon (Toray) started research in 1941 and. a pilot plant was operated in 1945. Full production of caprolactam and nylon-6 was started in 1951 based on I.G. Farben and Du Pont technology. In 1956 Ube Kosan joined the producers, also using the I,G. Farben process. In 1961 Ube changed from phenol to cyclohexane based on lnventa1_{s technology, This process has been}

improved by Ube and is also licensed by them. Ube started the first 70,000 mtpa HPO plant in 1972, using DSM technology. After 10 years of research

Toyo Rayon developed photonitrosation of cyclohexane (PNC) into a process first commercialized in 1962 at Nagoya at a capacity of 15,000 mtpa. Toyo Rayon is still the only company operating its process 94-102•

In 1965 a new plant based on BASF technology (NO reduction/cyclohexane oxidation) was started by Sumitomo (Japan Lactam). At about the same time Mitsubishi started the production of caprolactam based on cyclohexane oxidation and Raschig1_{s route to hydro~ylamine based on lnventa}1_s technol-ogy 105• In 1969 Mitsubishi changed over to the boric acid cyclohexane oxidation developed by Scientific Design. Although Toa Gosei 104 has an-nounced a new process no commercial plants based on this process have been reported.

1.2.3.3. USA

Allied Chemicals started the production of caprolactam in 1954, based on I.G. Farben technology. The product ion level in 1958 was about 20,000 .ntpa, being increased in 1960 to 45,000 mtpa. 30,000 mtpa was added in 1961. Allied improved the I.G. Farben process on several points. The ammonium nitrite solution required for the hydroxylamine production was prepared from ammonium carbonate, phenol was hydrogenated directly to cyclohex-anone in the liquid phase on a Pd/C catalyst and the catalyst for

the dehydrogenation of cyclohexanol was improved (Zn/Ca). Further, the coal-extracted phenol was replaced by cumene-phenol, of which Allied is now one of the largest producers.

Du Pont commercialized a 20,000 mtpa plant in 1961 based on the nitrosa-tion of cyclohexane. In 1969 this plant was closed down for safety reasons and because of problems with caprolactam quality 93.

Union Carbide developed a process based on the oxidation of cyclohexanon to caprolactone and lactamization with NH

3• This process was commercialized in 1966 on a 20,000 mtpa scale, but was shut down again in 1971.

Dow-Badische - a joint venture between BASF and Dow - started the produc-tion of caprolactam in Freeport (Texas) in 1962 (15,000 mtpa) based on the technology of BASF (NO reduction/cyclohexane oxidation). The plant was ex-panded to 80,000 mtpa in 1968 and later on to the current 160,000 mtpa.

PPG and DSM started operation of a 25,000 mtpa plant at Nipro Incorporat-ed's Augusta site in 1966 (Raschig/cyclohexane), and in January 1977 an-other 70,000 mtpa extension (HPO/cyclohexane) was put on stream. Nipro be-came a 100

%

DSM daughter in 1970.

1.2.3.4. Italy

I n ta y, n1a I I S · V · 1scosa 1 o6-1l3 was pro uctng d · f · b 1 res tn . 1 930 a rea y, an I d d started research on a completely new caprolactam process in the early 1950's. In 1958 the basic nitrosation step was discovered and in 1960 a pilot plant came on stream. In the same year, a decision to build a 12,000 mtpa plant was made. Production started in 1962, but full capacity was

reached only in 1965. Production has since been increased to the current 16,000 mtpa. A new plant based on thi~ process (Chimica Dauna) was started in 1973 with a capacity of 80,000 mtpa. Snia Viscosa (owned mainly by Mon-tedison) has recently developed a process that is completely free of am-monium sulphate, via pentamethylene ketene (PMK) and by extracting

capro-lactam with an alkyl phenol 108•

Besides Snia Viscosa, Rhodiatoce (S0/50 Rhone Poulenc and Montecatini) produced caprolactam until 1963. Edison, who already owned a 20,000 mtpa plant in Porto Marghera (Phenol/Raschig), gradually scaled up to 30,000 mtpa in 1965 and, after the fusion of Montecatini and Edison to Montedison, to 45,000 mtpa in 1967 and 75,000 mtpa in 1969. The present capacity is 80,000 mtpa.

1.2.3.:5.

w.

8,9,114-119

Although continuation of research at I.G. Farben was impeded by the fact that the Ludwigshafen works (BASF) were located in the French sector, the Leverkusen works (Bayer) in the British and the Hoechst works in the Amer-ican sector, caprolactam production was restarted immediately after the war. Production was still batchwise and based on phenol and hydroxylamine. The hydroxylamine at BASF was produced from sodium nitrite, yielding Na

2

so

4 as a by-product. Production was stepped up to 35,000 mtpa in the early fifties and a change-over to a continuous process was made. In 1952

I.G. Farben was divided into BASF, Hoechst and Bayer.

Due to the fact that the demand forecasts for caprolactam were favourable, BASF started extensive research in the early fifties to improve the process.

The items to be improved were:

1. the Beckmann rearrangement and caprolactam purification; 2. oxidation of cyclohexane to cyclohexanol/cyclohexanone;

3.

hydroxylamine production.

The main item was the hydroxylamine process, which had to be changed in view of the various disadvantages of the Raschig route via sodium nitrite. BASF chose NO reduction, which was known from two new Du Pont patents and a pub! ication in 1930 (Butterworth and Partington). From 1951 to 1954 labora-tory research with K. Jochers as research leader was carried out to investi-gate the reaction. From 1954 pilot-plant work was carried out in large batch reactors (approx. 8 m3) and in September 1956 the first commercial plant was started up. The first plant most probably worked batchwise and more research was required to develop a continuous process, which was achiev-ed in the early sixties. In 1958 all hydroxylamine production was changachiev-ed to the NO reduction route and in 1961 the first plant at DOW-Badische was started up, based on NO reduction. BASF increased its production at Lud-wigshafen step-wise to 65,000 mtpa in 1963, 80,000 mtpa in 1964, 100,000 mtpa in 1965, 120,000 mtpa in 1970, and finally the present 140,000 mtpa.

In 1960 the first cyclohexane-based semi-commercial plant was brought on stream and in 1965 all plants were based on cyclohexane oxidation. These developments were achieved witn the help of K. Kahr, who transferred from

lnventa to BASF in 1959.

In 1967 the first plant was constructed in Antwerp (60,000 mtpa), with start-up following in 1968. Capacity was later increased to the present 140,000 mtpa.

After the division of I.G. Farben, Bayer obtained a 1 icense from BASF and also started production of caprolactam in the late fifties on a 15,000 mtpa scale, based on phenol and the Raschig route. Bayer expanded to 40,000 mtpa in 1960 and to 75,000 in 1970. In 1963 Bayer changed from the phenol route to Scientific Design1_{s boric-acid-catalyzed cyclohexane}

oxi-dation process. In 1967 a new plant at Antwerp was started up with a capac-ity of 70,000 mtpa. During start-up of the cyclohexane oxidation and cy-clohexanol-cyclohexanone conversion sections a fire occurred. The damage was repaired and full capacity was subsequently reached in 1970. The plant was later expanded to the present 90,000 mtpa. Bayer investigated NO

re-duction on a laboratory scale, but plants based on this process have not yet been reported. Recently much research has been carried out on gas-phase rearrangement of cyclohexanone oxime to caprolactam, to avoid formation of by-product ammonium sulphate. The process is being tested on pilot-plant scale. The same developments with the gas-phase rearrangement of cyclohex-anone oxime have also been made at BASf.

The Holzverzuckerungs A.G. in Zurich became Ems A.G. after the war, with lnventa as the engineering department. The Emser works functioned as pro-duction plant, and took on the leading Leuna scientists Kahr, Zorn and Giesen, who went on to develop the lnventa process (Raschig route and cyclo-hexane oxidation). At the Emser works caprolactam facilities with an out-put of 10,000 mtpa were started up in 1961; later on, production was in-creased to 16,000 mtpa. lnventa also investigated the NO reduction step on a laboratory scale from 1953, when the first patents were published, and on pilot-plant scale from about 1966.

In 1972 the first semi-commercial hydroxylamine plant producing 5,300 mtpa was started up, and was operated until 1974, when undefined problems were encounterend 117• lnventa also investigated cyclohexane oxidation; in 1962 the first 12,000 mtpa plant based on cyclohexane was started up at the Emser works. These developments mainly stemmed from the Leuna research group (Kahr). lnventa1_{s technology was licensed to Mitsubishi, Ube, Guyarat,} Petkim, Petrokhimiya and Productos Chimicos Esso. All these plants are based on cyclohexane oxidation and the Raschig hydroxylamine route. A plant using the NO reduction route (50,000 mtpa) was I icensed to Polimex (Poland).

In addition to the above-mentioned companies, Zimmer AG have also developed a caprolactam process, which has been 1 icensed to some Eastern countries

(Rumania). Further it has been reported that Chemische Werke Huls has been a transient producer of caprolactam in W. Germany.

1.2.3.6. Other deveZopments

Besides the commercially proven processes, several alternative routes have been proposed. A survey of these routes is given by Rosier, Lunkwitz 90 with details being given in some of the general publications 6

7-BG

1. 3. NYLON 6 IN THE FIBRE INDUSTRY 119-124,. 189

Caprolactam is the monomer of nylon 6, which is used for 85-90 % as a fi-bre. Fibre production can can schematically represented as follows:

Natural cotton wool

ics

silk cellulose acetate rayon

~

glass metal reinforcing Synthetics chemosynthetics

6

6.6 nyl nylons (12,610) polyester (TPA,DMT) acrylics polypropylene

others (e.g. polyvinyl alcohol, polyure-thane, polyvinyl chloride polyvi-nylidene chloride)

The ratio of synthetics to the major natural fibre cotton has changed dramatically during the last 30 years, as can be seen from the following table. 1950 1960 1970 -'· 1980 "'estimated Cotton Production 6 x 10 mtpa 7.0 1

o.o

11.3 12.0 Market % 73 65 54 45 Synthetics share Production 6 x 10 mtpa 0.1 0.7 5.5 9.6 Market share 5 25 36

Of the synthetics, the majority are polyesters (40-45 %), with 30% nylon products, 20% acrylic fibres and 8-9% polypropylene, the special fi-bres taking only 1 %of the market share.

Polypropylene in particular is expected to show considerable growth in the near future. The 30% polyamides are divided between nylon 6 and nylon 6.6.

A negligible share is formed by other polyamides (e.g. laurino-lactam). In 1980 about 60% of the nylon (1.7 x 106 mtpa) will be produced from capro-lactam (nylon

6)

and 40% from hexamethylene diamine.and adipic acid (nylon

6.6).

The major applications of nylon are: textiles (lingerie, socks, nightware, sportsware, upholstery, tights) carpets, technical filament (fish-nets, seat belts, ropes, parachutes, tyre cords, apparel) and engineering plastics

(as parts in toys, machinery, automobiles and electrical goods, further, as pipes, tapes, tubes, sheets or rods), with the emphasis clearly lying on the first two sectors.

After polymerisation to nylon 6, the fibre undergoes a large number of treatments before reaching the consumer, such as spinning, stretching, text-urizing, dyeing and weaving. Depending on the application, the price of the raw material caprolactam therefore seldom constitutes more than 10

%

of the actual retail price paid for the end product, More details are to be

found in the specialist textile literature; a survey of the various treat-ments is given for instance by H. Keller 120

1, 4, DESCRIPTION OF INDUSTRIAL PROCESSES 1.4.1.

The majority

(>

90

%)

of the world's caprolactam is based on the two clas-sical intermediates cyclohexanone and hydroxylamine, and produced by the BASF, DSM, lnventa, Montedison and Allied processes. The remaining 10% of the caprolactam is produced by two other commercial processes (Snia Viscosa, Toray), which will be described further on in the chapter.

Fig. 1.2. Routes to

The processes nor the preparation of hydroxylamine will be dealt with in the next chapter. A survey of the commercial routes to caprolactam via cyclohexanone is given below.

Two petrochemical feedstocks are used in the manufacture of cyclohexanone, toluene and benzene. Production may proceed by either of two routes, via the intermediates cyclohexane or phenol.

Two commercial processes are available for the hydrogenation of benzene to cyclohexane. The I.F.P. process hydrogenates benzene in the 1 iquid phase with a Raney nickel catalyst at 35-50 atm. and 200- 225 °C. (Recently, a new type of catalyst was commercialized). Conversion is quantitative with nearly 100% selectivity. The I.F.P. process is widely applied for the production of cyclohexane. The DSM process proceeds in the gas phase on a Pt catalyst at temperatures above 250 °C and at a somewhat lower pressure than the I.F.P. process. The yield is equal to that of the I.F.P. process. The DSM catalyst is not sensitive to sulphur and has an extremely long

lifetime in comparison. The disadvantage of the DSM process is the need to evaporate benzene.

The cyclohexane obtained is oxidized with air in the liquid phase at 160-180 °C and 10- 16 ats. abs, with a homogeneous Co-salt as catalyst:

@

+

0

2 ___.,...

@-oH

and

@=o

K.A.oil

( Ketone alcohol ) 106

This process, which is based on older Du Pant patents , has been applied on a very large scale since 1960 for the production of cyclohexanone, both for the manufacture of caprolactam and for the production of adipic acid

(precursor for nylon 6.6). The overall yield to K.A. oil is low (70- 75

%),

and conversion is only 4 - 6 % per pass.

The reaction is carried out in a cascade of reactors (CSTR or bubble col-umns). Unconverted cyclohexane is recycled.

The low yield is caused by a further oxidation to acids and esters. A higher yield of 90 % with an increased conversion of 10 % per pass is claimed by the boric acid process developed by Scientific Design

7

8• This process obviates further oxidation of the cyclohexyl hydroperoxide by form-ation of a boric acid ester, which is hydrolysed to cyclohexanol. The

bo-ric acid has to be recovered quantitatively and recycled to the process; this complicates the process considerably, necessitating higher invest-ments. The process is used commercially by Bayer and Mitsubishi for

capro-lactam production and by Rhone Poulenc, lCl, Laporte and Monsanto for adipic acid manufacture 78• Since 1970 no new production facilities have been announced.

OSM has recently developed a process based on non-catalytic selective oxi-dation of cyclohexane to cyclohexyl hydroperoxide, followed by selective catalytic decomposition or hydrogenation of this intermediate to K.A. oil. This process increases the yield substantially, without requiring large changes to be made to the equipment used for the normal oxidation of cyclo-hexane with a Co catalyst. Reactor pollution problems due to Co are, fur-thermore, not encountered in this system 125•

In all the commercial processes, including phenol-based routes, a certain amount of cyclohexanol is produced alongside cyclohexanone. The cyclohex-anol is then converted by dehydrogenation to cyclohexanone. Most processes are comparable and proceed in the gas phase at 400 °C. Copper chromite, nickel, ZnS and ZnO with additions of Fe or Ca are all used as catalysts; the group based on ZnO are said to be best. The Leuna works is still utilizing the Fe/ZnO catalyst originally developed there. DSM use a spe-cially developed catalyst.

In the phenol-based processes the raw material is toluene (DSM) or cumene (Allied/Montedison). Like the cyclohexane oxidation, the toluene oxidation proceeds in the liquid phase with a Co catalyst according to:

Co

150-1600C~ conv. yield = > 90% 30% p.p.

The benzoic acid is purified and oxidized further in the 1 iquid phase by a homogeneous magnesium copper benzoate catalyst, according to:

;;:::0.._

.,.o

~C-OH+ Mg/Cu

250 OC, 2.5 atm

The yield of this step is lower than that of the toluene oxidation.

The phenol route based on cumene is used by Allied to produce phenol as raw material for caprolact~m. lt is one of the major routes to phenol, and finds widespread application on a commercial scale.

The process can be represented by:

cat: H₃P0₄jcarrier

: 250 °C ; 25 ats. abs.

@{H3 _{H + 02} 11S°C/Na2C03

~Ha

_OOH H2S04 _@-oH ;CH3 80°C ;. + c = o

H3 conv.: 25 · 30 % CH 3 \cH₃

yield: 80% _{hydroperoxide phenol}

ol cumene acetone

The phenol obtained is hydrogenated on a Pd-based catalyst, either in the gas phase (DSM) or in the liquid phase (Allied), according to:

2H2 + @-oH Pd _150°C cat.

®=o

(and low amounts anol)

The yield to cyclohexanol/cyclohexanone (anol/anone) is nearly quantitative. After purification of the cyclohexanone obtained, the oximation is carried out by bringing an organic cyclohexanone I iquor into contact with an inor-ganic hydroxylamine salt solution ((NH

30H)2so4). The reaction is carried out at temperatures of 50- 90 °C at pH 4.5 according to:

This is an equilibrium reaction, proceeding completely to the right-hand side at pH 4.5. Due to the fact that the pH decreases with increasing con-version, the solution must be continuously neutralized with NH

3 to keep the pH at 4.5.

The reaction is carried out in rotating disc columns or in a series of con-tinuous stirred reactors. The oxime obtaine~ is purified and contacted with oleum to accomplish the Beckmann rearrangement:

This reaction is highly exothermic. The conditions under which the oleum and oxime are mixed are very critical, and the proportion of oleum to oxime must be kept within narrow limits. The caprolactam/sulphuric acid mixture

is cooled and neutralized with ammonia. Separation of the layers 1 iberates the crude lactam, which is subsequently purified extensively to yield fibre-grade caprolactam. Every caprolactam manufacturer has developed its own caprolactam purification system: solvent extraction, crystallization, distillation, hydrogenation and chemisorption are all commonly applied. Spe-cifications for fibre-grade lactam are extremely stringent. The capro-lactam obtained is polymerized to nylon 6; the following is a typical route:

0

(-

~-{

CH2ls- NH-) 0 nylon-6

Extensive physical data on caprolactam and data on the chemistry and mech-anism of its polymerisation to nylon 6 are given by H.V. Reimschuessel

in J. Polym. Sci., Macromol. Rev. 1977, Series 12, pp. 65-139.

1. 4. 2. Hydr•ox-y lamine processes 1.4.2.1. Introduction

In its free form, hydroxylamine (NH 20H)

126 _{is a white, transparent,}

crys-talline solid (rhombohedral), with a melting point of 33 °C and a boiling point of 57 to 58 °C at 22 mm Hg. Hydroxylamine is very hygroscopic and extremely unstable in its free form; it decomposes explosively upon

heat-ing and is sensitive to metallic compounds and air.

Hydroxylamine shows acidic and basic reactions; it can be easily oxidized and reduced. lt is most commonly found as a salt. These salts are comparable to ammonia salts; with strong alkalis, hydroxylamine can form hydroxylaminates (e.g. NH

20Na).

Hydroxylamine was first synthesized in 1865, by W. Lessen, who found that it was formed in the reaction of nascent hydrogen (tin in hydrochloric acid) with ethyl nitrate. Lobry de Bruyn first isolated free hydroxylamine

in 1890. In 1887, Raschig 127 discovered a new synthesis to hydroxylamine. He reacted nitrite with sulphites, and hydrolysed the product obtained to hydroxylamine. Raschig already realized that this route was one of many possible reactions within this system.

From 1940 up to the present day this synthesis has remained the most important industrial route for the preparation of hydroxylamine.

1.4.2.2. Rasehig route to hydroxylamine The industrial route consists of three steps: 1. Nitrite preparation.

2. Hydroxylamine disulphonic acid (HADSA) salt preparation. 3. HADSA hydrolysis.

ad.l. Preparation of ammonium nitrite. a. Preparation of ammonium carbonate.

The preparation of ammonium nitrite on an industrial scale is a difficult process. lt causes a great deal of air pollution due to the formation of ammonium nitrite mist and to the low efficiency of NOx adsorption facili-ties. The major caprolactam producers operating the Raschig process (Allied,

lnventa, DSM, Montedison) have therefore improved the nitrite preparation step by starting from an ammonium carbonate solution instead of absorption of NOx in an ammonium solution, according to the following reaction: 2NH

3 + C02 + H

2

0~(NH

4

)

2

co

3

+ (NH4Hco

3

).

(C02 is in general abundantly available, from the NH

3

synthesis.) The reaction conditions (e.g. tempera-ture, NH

3;co2 ratio and concentration of NH3) are determining factors for the optimum (NH

4)2co3 solution. b. Nitrous gas absorption.

Nitrous gases are absorbed in the ammonium carbonate solution at a tempera-ture below 10 "C according to: NO+ N0₂ + (NH₄)₂co

3----;-2NH4No2 + C02• Although the reaction equation is simple, it is essential that reaction cconditions (pH, NO/N0

2 ratio, temperature) are very carefully controlled, to minimize the formation of nitrate, N

2 (Piria's reaction) and ammonium nitrite mist, and to maximize NOx absorption efficiency. Notwithstanding a variety of measures, an overall ammonia yield of only about 70% is attained. Ammonia is combusted conventionally on Pt/Rh gauzes at 850-900 °C to yield NOX.

ad.2. so2 adsorption.

Preparation of the hydroxylamine disulphonic acid salt. so

2 is absorbed in the ammonium nitrite solution, under supply of ammonia water. The ammonia salt of hydroxylamine disulphonic acid (HADSA) is ob-tained according to the equation:

NH

4No2 + NH40H + 2S02 -;o-NOH(S03"NH4)2•

The temperature must be maintained at 0 °C, with the NH

4

0H/NH

4

No2 ratio being kept stoichiometric and the pH at about

5.8 - 5.5.

Unconverted ammonium nitrite reacts with hydroxylamine in a subsequent

- +

step according to: N0₂ + NH

30H -;o-N2o + 2H2o.

If conditions are not kept within narrow 1 imits, several by-products are formed, e.g. ammonia salts of nitrilosulphonic acid N(so

3.NH

4

)

3 and imido-sulphonic acid NH(S0

3

NH4)2•

A survey of the chemistry of these various reactions has been published by F S • ee 1 128-132 •

ad,3, HADSA hydrolysis.

The HADSA solution is hydrolysed to ammonium sulphate (A.S.) and hydroxylamine sulphate by boiling the solution for some time. Two consecutive reactions take place:

a) NOH(S0

3NH4)2 + H20 ... NHOH(S03NH4) + NH4Hso4•

ammonium salt of hydroxylamine monosulphonic acid {HAM SA)

b) 2 NHOH(so

3NH

4

)

+ 2H20 ~(NH

3

0H)

2

so

4

+ (NH

4

)

2so4• The by-products are also hydrolysed to A.S.

The hydroxylamine sulphate obtained is partly neutralized and subsequently used for the oximation step.

ad.4, Oximat ion.

The hydroxylamine sulphate solution is reacted with cyclohexanone in a liquid-liquid contactor at a pH of 4.5 and a temperature of 50- 100 °C, according to the equation:

During the oximation, ammonia is supplied to neutralize the acid freed by the converted hydroxylamine. 1.9 tonne A.S./tonne caprolactam is obtained by this route, in addition to 2.6 tonne A.S,/tonne caprolactam from the neutralisation after Beckmann1_{s rearrangement.}

The ammonium sulphate is crystallized, and finds application as a ferti-1 izer.

1. 4. 2. 3. NO reduction to hyd:t!oxyZamine (BASF and Inventa process)

The industrial processes used by lnventa and BASF for the preparation of hydroxylamine from nitric oxide have not been described in detail in the literature. Only K. Jockers gives some features of the BASF process in his article in Nitrogen 144• lt should be remarked that the following descrip-tion of the NO process operated by lnventa and BASF has been largely based on the patent 1 iterature.

The process can be divided into the following steps: 1. NO preparation.

2. H₂

so

₄ purification

3. NO reduction to hydroxylamine

4. Catalyst activation and regeneration. 5. Oximation.

ad.1. NO preparation.

The nitric oxide necessary to prepare the hydroxylamine must be very pure, because N0₂(o

2} deactivates the catalyst even at low concentrations and reacts with hydroxylamine to form N

2

o.

Due to the fact that both hydrogen and nitric oxide are consumed by the reaction and the ratio of nitric oxide to hydrogen is low, it is impossi-ble to use the normal ammonia-air combustion method (inert build-up}. BASF and also lnventa claim the preparation of NO by combustion of ammo-nia with pure oxygen over normal Pt/Rh gauzes at 860 °C; instead of ni-trogen, water vapour is used as inert gas to prevent an explosive mixture forming. After the oxidation, water vapour is condensed and pure NO

obtain-114133 . 134

ed BASF ' cla1ms

o

₂/NH

3 ratios of 1.28- 1.4, lnventa slightly lower ratios of 1.17- 1.26, at gas velocities of 0.06- 0.3 m/sec (NTP). The optimum according to this patent is at o

2/NH3 = 1.25 and 1.2 m/sec. lt is necessary to keep the

o

₂/NH

3 ratios as low as possible, to prevent fur-ther oxidation to N0₂

o

₂), while at the same time the yield of the NH

3 combustion may not decrease (90- 96

%).

The combustion of NH

3

;o

2 with water vapour had previously been described by several authors (e.g. Partington, Wendlandt 135- 137 •140); more

recent-ly Zasorin c.s. 138,139 has described the kinetics of the reaction with water vapour as inert gas instead of nitrogen.

BA5F 141 uses hydrogen to reduce 0 2 and N02 5 ~ 10

%

Ag₂0/25

%

Mn0₂/25

%

Fe₂

o

3 catalyst 200 ~ 300 °C, at a GH5V of about 20,000 h~l further to H₂

o

and NO on a on a

¥

~AJ

₂

o

₃

carrier at ~1 (10,000 ~ 200,000 h is claimed). 5ome·N

20 and N2 are obtained as by-products; the amount of N20 in particular, must be kept as low as possible (NO/N

20/H2 explosive mix-tures). lnventa are not known to possess such a patent; they therefore ob-tain HN0

3 as a by-product. Although other alternatives have been proposed by BA5F (e.g. HN0

3 combustion with NH3

142_{, 5143 or 50}

2), the above-mentioned routes are the most favourable. Schemes are given below for the BASF and lnventa routes.

Ammonia combustion NO:!IO:! converter

200't 1.3. Preparation of NO (BASF) Ammonia combustion 1. 4. Condenser

Tr=%::;:,

_{l, ...}

_,~ Fine purification N0(95%) Caustlcs 33

ad.2. H₂

so

₄

purification

In the NO reduction step a continuous stream of sulphuric acid is supplied

to the precious metal catalyst. As has long peen known, however, precious metal catalysts are very sensitive to all kinds of poisons (e.g. heavy

me-tals, sulphur compounds, selenium and tellurium compounds, CO); high demands are consequently put on the sulphuric acid quality used.

Both BASF and lnventa therefore developed a purification method. BASF

144

described pre-treatment with oxidizing agents (e.g. H

2

o

2, KMn0

4)

to pre-vent an activity decrease by H

2

s

or

so

2• Later on, BASF mentioned that in fact only disti lied

so

3 and condensat_ion water can be used, requiring sep-arate, dedicated apparatus in the normal sulphuric acid production. lt has therefore been proposed to add H

2S to remove the heavy metals by

adsorp-tion as sulphides on activated carbon

145.

In a second activated carbon column excess H₂

s

is oxidized by H₂

o

₂ to sulphate. A later BASF patent 146 uses

so

2 or sulphite instead of H2S to considerably prolong the lifetime of the first activated carbon column. (H

2

s

leads to deposition of elementary

sulphur on the activated carbon.) Regeneration is carried out by caustics (NH

40H, Na2

co

3). Consumption figures of 25-250 mg H202/l H2

so

4 are mentioned.

BASF purifies the dilute sulphuric acid, as it is used for the NO reduction step.

lnventa uses a completely different method

14

7.

1

vol.% hydrochloric acid is added to the concentrated sulphuric acid (75 -

80 %)

to complex the hea-vy metals, especially iron, after which an anionic exchange column (DOWEXI, AMBERLITE lra 93) is used to adsorb these complexes. After passing the ion exchange column, the added hydrochloric acid is removed by vacuum distill-ation/stripping. Important non-metallic impurities such as As and most probably also Se and Te are less effectively removed. lt is also to expected that this corrosive medium will attack the ion exchange material. An active carbon column is probably used to remove organic material from the exchange column before the process 1 iquor enters the catalytic reactor. According

h 1 • 1 4 7, 1 48 d 1 4 9 I I • · 1

to t e 1terature an a patent , nventa use a p ast1cs materia (e.g. PVC) for the lining, which may be to withstand the highly corrosive medium formed by the hydrochloric traces left in the sulphuric acid used. One reason mentioned is that the catalyst is poisoned by silica {ceramic materia 1 s?).

K • oc ers, on t e ot er an , ment1one J k h h h d . d. h' 1n IS pu bl' tcatlon . 114 t at nor-h

that Cr/Ni/Mo/Cu 40/20/2/3 and Cr/Ni/Mo 25/25/2.2 stabilized with Nb, Ta or Ti are favoured.

Actfvated Activated

carbon column carbon column

Ion Exchange vacuum column distillation

Actfvated carbon?

1.5. H₂

so

₄ purifieation BASF above, Inventa beZow

ad.3. NO reduction Reaction medium

The NO reduction proceeds in a liquid phase consisting of a 20% aqueous sulphuric acid solution containing a suspended precious metal catalyst. The overall reaction equation is:

2H+ + 2NO + 3H₂ ~ 2NH;OH

By-products are formed according to: 2H+ + 2NO + 5H

2

~2NHt + 2H₂0

2NO + H

2 ~N

2

0 + H20 2NO + 2H

₂

~N

₂

+ 2H

20

As can be derived from these equations, the sulphuric acid is gradually neutralized by hydroxylamine. When almost all the acid has been neutralized

(pH= 2) a nearly pure hydroxylamine sulphate solution is obtained contain-ing small amounts of ammonium sulphate.

This solution is filtered off and used for the oximation. An obvious dif-ference between the compositions of the Raschig liquor and the NO reduc-tion liquor is the NHt/NH;OH ratio obtained. In the Raschig route this is about 2, in the NO reduction only 0.05- 0.15. This results in 1.8- 2.0 tonne AS/tonne lactam in the Raschig route, compared with about 0.6 tonne