close

Вход

Забыли?

вход по аккаунту

?

Industrial Applications of Photochemical Syntheses.

код для вставкиСкачать
Industrial Applications of Photochemical Syntheses
By Martin Fischer[*]
The principal industrial applications of photochemistry have so far been in the fields of
free-radical chlorination, sulfochlorination, sulfoxidation, and nitrosation. In addition, however,
photochemical reactions are being utilized on an increasing scale for the synthesis of vitamins,
drugs, and fragrances. The present article surveys the various kinds of light-induced reactions
exploited industrially, the equipment developed, and uses of the photochemical products. Furthermore, the problems encountered in designing a photochemical production plant are discussed
for the example of photonitrosation of cyclohexane.
1. Introduction
2.1. Chlorination
Most practical applications of photochemistry nowadays
lie in areas in which the effect of a photochemical primary
process is amplified by a large number of thermal consecutive
reactions[']. This is particularly true of photography, but also
applies to photopolymerizations and radical chain syntheses.
Use is made of advantageous features of photochemical reaction steps, precise energeticand spatial dosage of light, the high
selectivity of individual light reactions, and independence of
temperature, consumption of large amounts of light energy
being avoided.
So far, there are only few examples of industrial photochemical syntheses with stoichiometric light requirements[']. The
ice was broken by the photonitrosation of cyclohexane (cf.
Sections 2.4 and 7).
Two different geometrical arrangements of light source and
reaction space are employed in industrial photochlorination:
1) The lamp, which is generally shaped as a long cylinder,
is fitted with a cooling jacket and immersed in the reaction
solution.
2 ) The reaction solution is contained in a quartz or glass
tube which is irradiated externally with one or more lamps.
The immersion well apparatus has the advantage that no
light is lost to the surroundings. However, as a drawback,
the light intensity falls off very rapidly with increasing distance
from the lamp due to absorption by the reactants and to
divergence of the light rays (Fig. 1).
2. Free-Radical Reactions
The most important applications of photochemical syntheses still lie in the area of radical chain reactions. The
1940's saw the first industrial scale photo~hlorinations~~~.
Since
the light output of the mercury vapor lamps then available
was but a few hundred watts, the use of light for industrial
syntheses apart from vitamin D production was initially restricted to reactions with high quantum yields, i.e. to radical
chain reactions.
Photochemical initiation of radical chains generally permits
relatively low working temperatures, an essential condition
for numerous syntheses involving free radicals. Compared
with radical chain initiation by thermal decomposition of
initiators such as peroxides and azo compounds, light-induced
radical generation is of particular advantage in cases where
the radical chains are very short and very many initiating
radicals are accordingly required.
In the following survey of photochemical chlorination, sulfochlorination, sulfoxidation, and nitrosation the main
emphasis is placed on the description of representative industrial equipmentf4' rather than the extremely facile chemistry
involved. Some details concerning the use of the photochemical
products will also be presented.
p]
16
Dr. M. Fischer
BASF Aktiengesellschaft, Hauptlaboratorium
D-6700Ludwigshafen (Germany)
b
a
Fig. 1. Light intensity distribution in an annular reaction space; L, cylindrical
lamp; R, reaction space. Diminution of intensity from the inside to the
outside because of a) divergence and b) absorption.
The inhomogeneousdistribution of light and the consequent
spatial variations in the generation of initiating radicals raises
the danger of local overheating because most chlorinations
are highly exothermic. Attempts are made to counteract this
by thorough mixing of the reaction mixture.
Ifan emptycylindrical tubeisirradiated by a radially symmetric external arrangement of light sources then the radiation
density increases towards the axis of the tube owing to convergence of the light rays (Fig. 2a). If the tube is filled with
absorbing compounds such as a chlorination mixture, how-
a
b
Fig. 2. Light intensity distribution in a cylindrical reaction space radially
irradiated from an external source. a) Convergence of the light rays and
b) absorption have opposing eflects on the light intensity.
Angew. Chem. Int. Ed. Engl. 17, 16-26 ( 1 978)
ever, the light intensity decreases from the outside to the
inside owing to absorption (Fig. 2 b).
In this case convergence of the light rays and absorption
act in opposite directions and can partly balance each other;
on suitable choice of tube diameter and absorption characteristics of the reaction medium a relatively uniform light distribution can be achieved. (For calculations of photoreactions in
various irradiation apparatus, see ('1.)
Light losses on external irradiation can be minimized by
arranging the lamp and reactor tube along the focal lines
of an elliptical reflector; an internal cooling tube can be fitted
to remove excess heat from the reactor tube (Fig. 3)L6'.
Fig. 4. Industrial photochlorination unit for the production of monochloroalkanes. Two pairs of Vycor tubes are irradiated with a 7.5 kW mercury lamp.
RH = mixture of n-alkanes.
,--coolant
Hg lamp
reactor
.reactants
Hg lamp
coolant
section is only about 20s-a relatively small production unit
of the kind shown in Figure 4 has a production capacity
of some 27 t of chloroparafin per day, corresponding to about
lOOOOt/a.
Compared to the thermal process, however, the photochemical monochlorination of parafins is only of minor significance18*'1.
The chloroalkanes are used for alkylation of benzene to
give alkylbenzenes, which serve as starting materials for alkylbenzenesulfonates, an important class of detergentdgI.
Fig. 3. External irradiation of a tube reactor; the lamp and reactor tube
are located in the focal lines of an elliptic reflector. A cooling tube is located
in the center of the reactor tube.
In practice, radical chain reactions with high quantum yields
are frequently conducted without any great attempt being
made to optimize utilization of the light. An example of industrial chlorination in externally irradiated glass tubes was described several years ago by scientists working at Phillips Petroleum['], viz. the production of monochloroalkanes from a
C1 - 14 n-parafin cut.
RH + Clz b RCI + HCI
Achlorination apparatus in which the lamps are surrounded
by the reaction solution, i.e. in which an annular reaction
space is irradiated from its center, has been developed by
Ethyl Corporation for the photochlorination of benzene to
produce hexachlorocyclohexane[' 1' .
0
f12
c1)5
c1
c1 c1
RH = mixture of n-alkanes
The production apparatus essentially consists of two pairs
of Vycor tubes (7.5 x 150cm) and a water-cooled 7.5 kW mercury lamp. A cylindrical reflecting shield of polished aluminum
reduces dissipation of light to the surroundings.
Removal of the considerable heat of reaction creates problems in photochlorination in externally irradiated glass tubes.
To ensure that the temperature does not rise above 40Yuncontrolled decomposition has been observed at higher temperatures-nly
15 m o l x of chlorine is introduced. Prior
to entering the second pair of reactor tubes the reaction mixture
is first cooled to ca. 5°C.
Further chlorination to form the undesired dichloro compounds is avoided by stopping the reaction at an overall
conversion of 30 %. Owing to the high rate of photochlorination-the residence time of the solution in the irradiated
Angew. Chem. I n t . Ed. Engl. 1 7 , 1 6 2 6 (1978)
The reactor shown schematically in Figure 5 comprises
18 tube segments, of which 17 are represented by lines. Two
40 W fluorescent lamps are located at the center of each segment. The pump at bottom right rapidly recycles the bulk
of the reaction solution in the first 13 segments so that the
heat of reaction is eficiently removed by the cooling jacket.
A side stream passes through segments 14 to 18 (where residual
chlorine reacts) at the same rate as fresh benzene and chlorine
are fed in at bottom left.
Plugging of the tubes by crystalline product is avoided
by maintaining the concentration of hexachlorocyclohexane
in the reaction solution below 15%. This means that only
4 % of the benzene feed can be converted on passage through
the reactor. The chlorine consumption["] of 270 kg/h in the
unit indicates that nearly 3m3 of benzene feed is required
per hour. The quantum yield of this industrial chlorination
amounts to ca. 2500.
17
-
product out
18
17
1
II
16
15
14
2.2. Sulfochlorination
I
I
1,
recycle
drum
I
I
Another well known reaction, photochemical sulfochlorination of paraffins, is of great industrial importan~e['.'~1.In
sulfochlorination the action of light is again limited to the
formation of chlorine atoms from chlorine. The sulfonyl chloride group is distributed almost randomly over all C atoms
of the hydrocarbon
hv
RH SO2 C12 ----j RSOzCl
RH =mixture of n-alkanes
+
chlorine-
L
Fig. 5 . Reactor for photochlorination of benzene to form hexachlorocyclohexane.
The photoaddition of chlorine to benzene serves for production of the y-isomer of hexachlorocyclohexane, a versatile
insecticide marketed as Lindane or Gammexane (y-BHC).
Unlike other chlorinated hydrocarbons, this plant protection
agent still commands considerable interest because it is
degraded biologically with comparative ease. The very modest
yield of ca. 15 % is unsatisfactory, however.
+
Clean photochemical sulfochlorination requires the use of
unbranched alkanes. Side chains favor competing chlorination.
Normal alkanes of high purity are nowadays available in
industrial quantities by means of molecular sieve processes.
Sulfochlorination does not stop at the stage of the monosulfonyl chlorides. Since di- and polysulfonyl chlorides have undesirable technical properties the reaction must be stopped at
conversions of 30 to 50%. Yields of 80 to 90% are then
obtained.
Under industrial conditions the quantum yield is about
2000. This means an energy consumption of ca. 30Wh for
1 kg of sulfonyl chloride on use of fluorescent lamps.
Industrial sulfochlorination has been reviewed by
Lindner" '1.
Most of the alkanesulfonyl chlorides produced industrially
are hydrolyzed by caustic soda to give water-soluble alkanesulfonates which are used mainly as emulsifiers for polymerizations['].
RSOzCl + 2NaOH
The photochlorination of toluene to benzyl chloride, benzylidene dichloride, and benzotrichloride["], known to every
chemist from his practical course work, should not go without
mention. Mixtures of the first two compounds are produced
in a series of four lead-lined photoreactors at 80-110"C['z].
+ HC1
-+
RS03Na + NaCl
+ HzO
Owing to contamination with di- and polysulfonates, which
cannot be removed under industrial conditions, the stringent
demands nowadays placed on detergents can no longer be
fulfilled by this type of alkanesulfonates and they are now
hardly used in detergent manufacture'l6'' '1.
Reaction of alkanesulfonyl chlorides with ammonia leads
to sulfonamides, which are used as textile auxiliaries[l6]and
also reacted further with chloroacetic acid to form a mixture
of sulfonylaminoacetic acids. Compounds of this type serve
as emulsifiers and as anticorrosion agents for mineral
0ils1163 181,
Benzyl chloride, also obtained by purely thermal chlorination of toluene, is converted mainly into benzyl alcohol, a wellknown fragrance. It is also used in the synthesis of drugs,
disinfectants, and emulsifiers["]. O n hydrolysis, or preferably
on reaction with benzoic acid, benzylidene dichloride affords
benzaldehyde; in the latter case a further valuable intermediate,
benzoyl chloride, is also obtained.
Benzotrichloride is a valuable intermediate for dyes of the
triphenylmethane, xanthene, and anthraquinone series and
for plant protection agents.
18
NH3
ClCHzCOzH
RSOzNHCHzCOzH
RS02CI +RS02NH2
NaOH
2.3. Sulfoxidation
All the chlorine used in sulfochlorination is eventually recovered as hydrogen chloride or sodium chloride. A far more
elegant synthesis of alkanesulfonates has been found in sulfoxidation, in which oxygen serves as oxidizing agent instead
of chlorine.
RH
+ SO2 + '/Z
hi'
0 2
+RS03H
RSO3H + NaOH --t RS03Na + HzO
RH = C I 4 - 18-n-alkanes
Angew. Chem. I n t . E d . Engl. 17, 1 6 2 6 (1978)
Photochemical sulfoxidation has been developed as an industrial process at Hoechstrg~
193
The best one of several variants seems to be the so-called light-water process. The primary
sulfoxidation product, peroxysulfonic acid, is trapped by water
before it can undergo radical decomposition.
’’!
of nitrosyl chloride with cyclohexane to form cyclohexanone
oxime.
0+
NOCl --+
h“
#NOH.HCl
r/
1
I
so2
+
I
HiS04
Q-
RS0200H + H20 + SO2 + RSOsH + H2S04
An equimolar amount of sulfuric acid is formed as coproduct
and has to be removed on work-up.
Figure 6 shows a flowsheet of the industrial sulfoxidation
process[”! The photoreaction occurs in an immersed lamp
reactor. A mercury lamp serves as light source. The photoreac-
-
In the 1950s, with the common aim of developing a more
economic synthesis of caprolactam, the monomer of nylon
6, several industrial companies and university departments
in France, Germany, Japan, the USSR, and the USA started
studies on the industrial feasibility of the photonitrosation
of cyclohexane. The Japanese chemical company Toray finally
decided to conduct the photochemical synthesis in a large
production plant[231.Details are given in Section 7.
3. Photochemical Synthesis of Vitamin D and Related
Compounds
extract
3.1. Vitamin D3
02
Paraffin
HzO
Fig. 6. Simplified flowsheet of the light-water process for sulfoxidation.
tor containing paraffin is gassed from below with sulfur dioxide
and oxygen (molar ratio 2 : 1). Thorough mixing is accomplished by circulating the gaseous reactants by means of an
external pump. Water is fed in together with paraffin at the
top of the reactor. Sulfonic acid and sulfuric acid are continuously extracted into the water, thus largely precluding
further reaction of the mono- to di- and polysulfonic acids.
The mixture of aqueous acids is continuously removed in
a separator and the unreacted paraffin returned to the photoreactor (Fig. 6). The rather involved isolation of the sulfonic
will not be considered here.
Compared to conventional photochlorinations or sulfochlorinations the light consumption of sulfoxidation is very high.
About 0.2 kWh of electrical energy is required for lamp operation in the production of 1 kg of alkanesulfonate[201. This
means that only ca. 1600t/a can be produced with one 40 kW
mercury lamp.
Hoechst presently operates production plants having a total
capacity of ca. 50000 t/a in the Federal Republic of Germany
and France”’!
In all the syntheses so far described, radicals, mainly chlorine
atoms, are generated in the primary photochemical process.
One of the oldest non-radical industrial photoreactions is
the synthesis of vitamin D. Whereas vitamin Dz, produced
from ergosterol, formerly took pride of place, interest is nowadays focused mainly on vitamin D3: the latter compound,
unlike the former, is also active in poultry. Most of the vitamin
D3 now produced is not used to prevent or cure rickets
in children but instead as an additive in animal nutrition.
The starting material for the photochemical synthesis of
vitamin D3 is 7-dehydrocholesterol ( I ) , accessible in four
steps from ch~lesterol~’~!
Photochemical ring opening leads
initially to previtamin D 3 (2). O n heating to 50-80°C a
hydrogen shift occurs to give the thermodynamically more
stable vitamin D3 (3).
2.4. Photonitrosation
The light energy costs d o not usually play a decisive role
in industrial radical chain reactions because the light quanta
supplied act with a high multiplication factor. In contrast,
photochemical syntheses with quantum yields of about one
could only be implemented on an industrial scale once powerful and efficient mercury lamps became available. The development of industrially utilizable lamps was prompted by the
photonitrosation ofcyclohexane, i. e. the light induced reaction
Ailgew. Chem. Inr. Ed. Engl. 17.16-26 (1978)
Because of the high absorbance of 7-dehydrocholesterol
(Fig. 7) photolysis must be performed in relatively dilute solution (ca. I %). Ethanol or ether serves as ~ o l v e n t [ ~ ~ ~ ’ ~ !
The absorption spectrum of (2) completely overlaps that
of (I); thus a steadily increasing proportion of the incident
19
light is absorbed by (2) as conversion increases.Unfortunately,
the quantum yield of the consecutive photoreaction of (2)
AcO'
220
240
260
280
LCnmI-
360
-
\II\
AcO
d
c
W
q
2. A
(5)
A
O
320
Fig. 7. UV spectra of 7-dehydrocholesterol ( I ) (-)
and previtamin
D, (2) (------). Emission lines of the medium pressure mercury lamp.
to give tachysterol, the trans isomeric triene, is greater than
that of its formation[261.It is therefore essential to stop irradiation of ( I ) once 30-50 % conversion has been reached and
to isolate previtamin (2).
The emission of a mercury lamp is regrettably very poorly
adapted to the spectrum of 7-dehydrocholesterol (1) (Fig. 7):
only the relatively weak emission lines at 280,289, and 297 nm
can be exploited, which together amount to only 8.5 mol quant a b for a 40kW lamp (total emission over all wavelengths:
I74 mol quanta/h)['l.
Given a quantum yield of 0.31 for transformation of (I)
into (2)LZ6', then 8.5 x 0.31 =2.64mol of ( 2 ) / h should be
obtainable with a 40kW lamp. This means that the synthesis
of 1 kg of vitamin D3 requires an irradiation time of 1 h,
and thus an energy consumption of 40 kWh. The actual energy
required is twice as high: ca. 80 kWh/kg vitamin D3.
3.2. Hydroxy Derivatives of Vitamin D3
that not vitamin D3
Some years ago it was
itself but a metabolite formed in liver and kidney, Iu,25-dihydroxyvitamin D3 ( 4 ) , is the actual active form responsible
for regulating calcium metabolism. Compound (3) exhibits
very high vitamin D3 activity in patients having impaired
kidney function.
ysis ultimately yields la-hydroxy-vitamin D3 (7). A hydroxylated vitamin D3 will probably be marketed before
3.3. Dydrogesterone
The synthesis of the sexual hormone dydrogesterone (1 0)
developed by Philips-Duphar can be regarded as "spin-off'
from intensive research work in the vitamin D
In
analogy to the known photochemical transformation of 7dehydrocholesterol ( I ) into the IOU-methyl isomer lumisterol3, irradiation of the pregnadiene derivative (8) gives the
retrosteroid (9) which is cleaved to form (10) by HCl in
alcohol.
cI H3
c =o
4. Photoisomerization of Vitamin A Acetate
Iu-Hydroxy-vitamin D3 (7) proves just as active as ( 4 ) ,
but is much easier to produce. The intermediate ( 5 ) is accessible from cholesterol in six steps[''* "I.
The same reaction conditions as established for the parent
compound are employed for photochemical ring opening of
( 5 ) to the previtamin D3 derivative and subsequent thermal
hydrogen shift to give the diacetate (6)[28*291.
Alkaline hydrol-
['I
20
The Wittig synthesis of vitamin A acetate, performed on
an industrial scale at BASFL3'1, affords a mixture of two
stereoisomers, the all-trans (12) and the 11-cis form ( I t ) .
Only the all-trans isomer ( 1 2 ) is suitable for use in pharmaceuticals and animal feeds.
A very mild photochemical method has therefore been developed at BASFr33'for converting 1 l-cis- into all-traris-vitamin
A acetate: it consists in irradiating the stereoisomeric mixture
A mol quantum is equal to 6.023 x loz3 photons.
Angew. Chem. Int. Ed. Engl. 1 7 . 1 6 2 6 ( 1 9 7 8 )
OH
with visible light in the presence of a colored sensitizer such
as tetraphenylporphinatozinc or chlorophyll. A photochemical
isomerization process which can also be used for transforming
the 9 4 s isomer into the all-trans form was recently described
in a patent application by Hoffmann-La R ~ c h e ’ ~ ~ ] .
4
(21)
5. Photooxygenation
5.1. Rose Oxide
A further highly promising field for industrial photochemistry is the synthesis of fragrances. Small scale production of
rose oxide ( I 7) is already being conducted at the Firmenich
and Dragoco companies.
--J+
OA c
0
A/&
T T
2. KzC03
OH
(24)
0
Compound (22) can be converted in a series of conventional
steps into 1,2-didehydroaldosteroneacetate ( 2 3 ) which yields
the radioactively labeled aldosterone ( 2 4 ) by catalytic tritiation and subsequent
The American company
New England Nuclear produces ( 2 4 ) on a small scale for
use as a medical diagnostic aid.
Photooxygenation of citronellol ( 1 3 ) with rose bengal as
sensitizer gives a mixture of two isomeric hydro peroxide^['^]
which can be reduced by sulfite to the alcohols ( 1 4 ) and
( 1 5 ) . The main product (15) undergoes allylic rearrangement
to ( 1 6 ) in acid solution; ( 1 6 ) cyclizes with elimination of
water to rose oxide (1 7).
5.2. Ascaridole
Another sensitized photooxidation, i. e. reaction of cr-terpinene (18) with singlet oxygen to give ascaridole (19),
acquired some importance after World War IIr361.It is no
longer performed owing to the toxicity of this anthelmintic.
7. Scale-up of Laboratory Photoreactions to Industrial
Processes: Photonitrosation of Cyclohexane
Further development of the technology of photochemical
processes is of crucial significance for the future of industrial
photochemistry. The present state of the art can best be described with the aid of an example, the photonitrosation of
cyclohexane already mentioned above.
7.1. The Problem
In the interest of a readily comprehensible presentation
of technical aspects, let us consider the decision processes
that an industrial chemist faces and the data he works with
when scaling-upa photochemical reaction from the laboratory
0+
6. The Barton Reaction
The functionalization of saturated carbon atoms by photolysis of nitrites, discovered by D. H . R.
is used
in the production of tritium-labeled aldosterone (24). The
photochemical step of the multi-step synthesis consists in
the transformation of the pregnane derivative ( 2 0 ) into the
C-18-oxime ( 2 1 ) which gives the nitrone ( 2 2 ) on heatingr3*!
Angew. Chem. I n t .
Ed. Engl. 1 7 , 1 6 2 6 (1978)
NOCl
2
O N O H
+
HC1
to a production plant for the case of the photochemical synthesis of cyclohexanoneoxime. Assuming that a photochemical
plant for producing 10000t/a of cyclohexanone oxime were
to be developed: What typically photochemical questions
arise? What parameters play an essential role on scale-up?
And finally: What kind of photoreactor design is particularly
suitable for large-scale manufacture of oxime?
7.2. Laboratory Results as the Basis of Planning
The starting point for the design of an industrial plant
is the results of laboratory experiments on photonitrosation.
21
The various process steps are simulated on a small scale
under conditions corresponding as closely as possible to those
met in practice to permit early detection and mastering of
fundamental technical problems.
Photochemical synthesis of cyclohexanone oxime can be
performed in the usual immersion well apparatus fitted with
a medium pressure mercury vapor lamp (Fig. 8)[411.
coolant
tNOCI*HCI
-
cyclohexone
It is therefore impossible to determine the maximum product
output for a lamp of given wattage in the laboratory.
When beginning to design a production plant on the basis
of the above laboratory results very many questions have
to be answered, for example: Are the starting materials, especially nitrosyl chloride, available in industrial quantities? What
materials of construction apart from glass are sufficiently resistant to the exceptionally corrosive reaction mixture? What
measures can be adopted to ensure optimum temperature
control and mixing in the photoreactor? How is the oxime
best isolated? How can the accumulation of by-products be
avoided?
These questions, which occur in modified form in the design
of very many chemical production plants, are not necessarily
typical of a photochemical process. They will therefore be
put aside in favor of specifically photochemical problems.
7.3. Choice of Lamp
A very important condition for the design of a photoprocess
is a knowledge of commercially available lamps[41.The most
powerful present-day lamps are medium pressure mercury
vapor lamps with a power consumption of 60kW. A schematic
diagram of such a lamp fitted with cooling jacket is shown
in Figure 9. The lamp is ca. 2 m long and 6cm in diameter,
and can be cooled with water or a filter solution.
t
product
Fig. 8. Immersion well apparatus for photonitrosation on a laboratory scale.
The mercury lamp is located at the center of a cold finger, NOCl and
HCI are blown in through a glass frit. The oily bis(hydroch1oride) of the
oxime can be withdrawn from the bottom.
With a view to the design of large-scale industrial plant,
continuous operation should already be aimed at in the laboratory. Cyclohexanone oxime can readily be removed from the
irradiation apparatus as the bis(hydroch1oride) which has a
greater specific gravity and is insoluble in cyclohexane if hydrogen chloride is fed into the reactor together with nitrosyl
chloride[40!
C);NOH
+
2 HC1
--+ H
)...c)
2 HC1
Continuous long-term experiments required for ascertaining
optimum reaction conditions and determining the capacity
of the laboratory apparatus are precluded by formation of
a light-absorbing tarry deposit on the surface of the lamp
coo1er[23340,44!
The short-wavelength emission of the mercury
lamp is mainly responsible for formation of this ~ o a t i n g [ ~ ~ * ~ ’ ] :Fig. 9. 60 kW mercury lamp with cooling jacket.
If the emission between 248 and 366nm is filtered out by
addition of sodium nitrite to the cooling water of the lamp
Figure 10 shows a 60kW lamp shortly after ignition. A
then photonitrosation can be performed for several days with60kW mercury lamp costing about $1500.--[*] without coolout significant
ing jacket can be run for up to a year without significant
Under optimum conditions the yield of oxime is 89 %[411.
loss
of output provided that it is not switched off and on
The most important by-product is chlorocyclohexane, which
too
frequently[23!
has to be removed by distillation to prevent its accumulation.
Owing to the low absorbance of nitrosyl chloride in the
long-wave region (Fig. 1 1 ) and to the limited layer thickness
p] Mercury lamps for industrial processes are available from Quarzlampen
GmbH, Hanau, and Comp. Electrorntcanique Etabl. SCAM, Paris.
in laboratory apparatus, the light is not completely absorbed.
22
Angew. Chem. Int. Ed. Engl. 17, 16-26(1978)
and 366 nm for photonitrosation. Thus ca. 43 % of the overall
light emission of the lamp cannot be used. A shift to the
desired spectral region of 400-600nm can be induced by
addition of small amounts of thallium iodide to the mercury
discharge. This doping causes a very intense emission at 535 nm
at the expense of all mercury lines (Fig. 12).
7.4. Number of Lamps Required
Fig. 10.60kW mercury lamp with cooling jacket. The lamp was photographed
shortly after being switched on before it was burning brightly.
It is important that all the emission lines of the mercury
lamp in the ultraviolet and visible region of the spectrum
be absorbed to a greater or lesser extent by nitrosyl chloride
in order to ensure a good utilization of the light energy (Fig.
11). Overlap is seen to be considerably better than in the
synthesis of vitamin D because of the broad absorption bands
of nitrosyl chloride.
L,
Once the type of lamp required for the production plant-a
T1-doped mercury lamp-has been established[47],it is logical
to ask how many lamps will be required for the projected
capacity of 10000t/a of oxime, or how much product can
be expected per year from a single 60 kW lamp.
The most reliable method of determining this quantity is
undoubtedly to build a pilot reactor with a 60 kW lamp and
to test it. However, it is relatively easy to estimate the amount
of product expected from the light emission of the doped
mercury lamp and the quantum yield of oxime formation.
The output of a TI-doped 60kW mercury lamp at 535nm
is 13.1 kW or 210mol quanta/h. Assuming a quantum yield
of 0.8 under industrial conditions[42s43],168mol or 19 kg of
oxime can be expected per hour.
Formation of up to 0.4 kg oxime/kWh power consumption
of the lamp was determined by experiment; this corresponds
to 24 kg/h for a doped 60 kW lamp. The discrepancy between
this figure and the above estimate comes about because the
weaker lines (Fig. 12b) also contribute to oxime formation
in addition to the emission at 535 nm.
24 kg of oxime per hour corresponds to a capacity of 200 t/a
for one 60kW lamp. Hence it follows that the projected
10000 t/a plant will require 50 lamps.
7.5. Arrangement of Lamps in an Industrial Photoreactor
200
LOO
300
500
600
X.Cnml-Fig. 1 1 . Absorption spectrum of nitrosyl chloride and emission lines of an
undoped mercury lamp. The relative intensity of the emission lines (right-hand
scale) refers to mol quanta/h and not to the wattage.
Unfortunately, however, deposit formation on the immersion well precludes use of the emission lines between 248
200j
a1
180
1
Izl:ooLJ,
160
f
-z 80
0
60
10
20
200
300
,, ,
I
LW
A[nm]
,
500
160
a
I
It would be quite feasible to install 50 photoreactors in
parallel. However, in view of the fact that each reactor would
have to be fitted with numerous supply lines for starting
materials and products, for electrical energy and the filter
solution for cooling the lamps, the construction of 50 individual
reactors appears a very complicated and expensive task. It
is much more reasonable to develop photoreactors containing
say 50 lamps instead of just one.
A photoreactor fitted with 50 industrial mercury lamps
will have to be designed in a completely different way from
the laboratory immersion well apparatus with only one light
source. A question of paramount importance in the construction of a multi-lamp reactor will concern the geometrical
arrangement of the individual 60 kW lamps, which in turn
determines the light distribution.
The mechanical sensitivity of the lamp cooling wells, which
are at least 2.5m long, immediately suggests that the lamps
80
0
20
, ,
600
200
,~i,.
3W
,,,,,,,, .
LW
,,
5W
A[nm] --+
Fig. 12. a) Emission spectrum of an undoped mercury lamp; b) emission
spectrum of a mercury lamp doped with T11.
Angew. Chem. I n ! . Ed. Engl. 17.16-26 (1975)
0
01
0
0 O 0I
Fig. 13. Arrangement of several lamps L (shown in section) in a rectangular
reaction vessel.
23
should be suspended in the reaction solution from above.
What is the optimum distance of the lamps from one another
and from the walls of the reactor (Fig. 13)?
Asked in another way: How far does the light emitted
from a lamp penetrate the reaction solution before it is largely
absorbed?
The thickness of the layer of solution which absorbs 99 %
of the light of wavelength 535 nm can be approximately calculated with the aid of equations (a) and (b) (see Fig. 14). In
then be arranged as shown in Figure 13. Figure 15 shows
a section through a multi-lamp reactor proposed by the Toray
cyclohexane
cooler
coolant-
+
product
Fig. 14. Layer of liquid reached by the light from a lamp. r,=radius of
the lamp cooling jacket; d=thickness of layer of solution absorbing 99 %
of the light; I, Io=intensity of light.
a first approximation the light is assumed to be emitted radially
and light scattering by the heterogeneous reaction mixture
is considered insignificant.
d=r-ro
(b)
If lois the intensity of light at the surface of the lamp cooler,
I = lo-’ I. = intensity of light at a distance d from the lamp,
~ = 0 . 9 ,c=O.O37mol NOCl/l, and ro=lOcm, the value of r
is calculated as 30cm, and the thickness d of the layer in
which the light is almost completely absorbed is 20cm.
The distance between two lamps should accordingly not
exceed 40cm if dark zones are to be avoided. Experiments
confirmed the suitability of this di~tance[*~J.
The region still reached by the light from a given lamp
will of course depend upon the wavelength. Light of shorter
wavelength such as the intense 366-nm mercury line is
absorbed in a much thinner layer of only 0.4 cm, corresponding
to the higher absorbance of NOCl at this wavelength.
Thus the oxime due to short wave light is formed in the
immediate vicinity of the lamp. It is therefore hardly surprising
that the hydrophilic oxime bis(hydroch1oride)which is insoluble in cyclohexane is deposited on the likewise hydrophilic
glass of the cooling well where it resinifies under the action
of light and nitrosyl chloride to form a tarry deposit. As
already mentioned, however, formation of this coating can
be largely suppressed by use of a filter solution for elimination
of UV light.
7.6. Example of an Industrial Multi-lamp Reactor
A rectangular plan appears to be the most suitable for
an industrial multi-lamp reactor. The individual lamps could
24
Fig. 15. Section through a photonitrosation reactor (after [48]).
Apart from the lamps, an opaque wall carrying coolant
is also incorporated in the tapered trough. The reaction
gases NOCl and HCl are introduced into the outer compartment of the reactor. The cyclohexane in this space is carried
along by the rising gases and simultaneously saturated with
NOCl and HCl. The liquid phase flows in direction indicated
by arrows past the lamps, whereupon the oxime bis(hydrochloride) is formed and sinks to the bottom as fine oil droplets.
The droplets coalesce during their downward passage to give
a high density oily liquid which travels down the inclined
floor of the vessel and can be withdrawn from the bottom.
The insolubility of the oxime bis(hydroch1oride) thus permits
facile product removal and hence continuous operation so
important for an industrial synthesis.
Figure 16 shows a view of the upper part of the Toray
industrial photoreactor. The tops of the lamps housing the
electrical supply lines can be recognized.
Fig. 16. Photograph of an industrial photoreactor operated by the Toray
company [23].
The photochemical aspects emphasized in this account
represent but a small part of the enormous task of developing
Angew. Chem. lnt. Ed. Engl. 1 7 , 1&26(1978)
a plant for the photochemical production of cyclohexanone
oxime. The choice of materials of construction was particularly
difficult because the reaction mixture is highly corrosive towards all the usual materials. The only material sufficiently
resistant for use as reactor lining is the very expensive
Problems of cooling, mixing, replacement of burnt-out
lamps, and removal of deposits which cannot be avoided
in long-term operation, can only be mentioned in passing.
A vital question for all commercial scale photochemical
syntheses is that of safety, arising from the close proximity
of the lamp-an 800°C hot source of ignition-and flammable
liquids. In order to control these hazards the lamps are run
in a nitrogen atmosphere and an automatic shut down device
switches them off automatically in the event of a breakdown,
e.g. failure of the cooling water supply or a temperature
rise in the reactor.
Development of a photochemical process for the synthesis
ofcaprolactam is not restricted to the design of a photoreactor.
Economical production processes are also required for nitrosyl
chloride and for the Beckmann rearrangement of the oxime
to the l a ~ t a m [ ~ We
~ , ~shall
~ ! not consider these processes
here.
7.7.
Economics of Photonitrosation
Questions of economics are of supreme importance for the
industrial realization of a large-scale photochemical proc ~ s s [ ~For
~ ] .the photochemical synthesis of cyclohexanone
this means that the production costs should not be higher
than those of conventional processes based on phenol or
toluene. The photochemical synthesis has the advantage of
fewer steps and higher overall yields than competing processes.
However, the fact that only one company has decided to
employ the photonitrosation process suggests that the photochemical lactam synthesis does not have any cost advantages
over conventional syntheses. Some 16OOOOt/a of caprolactam
are now produced by the photonitrosation process at
T~ray‘~~].
In this connection, the cost of energy and lamps incurred
by the photochemical synthesis step is very important: it
has already been mentioned that the production of 1 kg of
cyclohexanone oxime requires 2.5 kWh of electrical energy
for operating the lamps. At a price of 5 cents/kWh, the energy
costs amount to 12.5 cents/kg of oxime. The cost of the lamps
calculated from the capacity, lifetime, and initial purchase
price is 1 cent/kg, which is a relatively low value.
These data clearly show that a photochemical synthesis
step does not have to be expensive, provided that the quantum
yield is not much lower than unity and the photoreactive
starting material also absorbs in the long-wavelength region
of the spectrum. Long-wavelength light is less expensive than
short-wavelength light because the individual quanta contain
less energy, and high-performance sources are available for
visible light.
8. Lauryllactam by Photonitrosation
Apart from the synthesis of caprolactam, photonitrosation
has meanwhile found a second industrial application : the
Anyew. Chem. Int. Ed. Engl. 17.1G-26 ( 1 9 7 8 )
production of lauryllactam, the monomer of nylon I2[’O1.
In place of cyclohexane, cyclododecane (27), accessible from
butadiene ( 2 5 ) in two steps, is converted into the oxime
(28) by nitrosyl chloride and light.
On account of its low density and low absorption of water,
lauryllactam ( 2 9 ) is used for a number of special purposes,
in particular for the production of dimensionally stable plastic
components (automobile construction) and for plastic coating
of metals[’’].
The photochemical lauryllactam process was developed by
A T 0 in France. The photochemical plant having a present
capacity of 8000 t/a comprises several photoreactors, each
containing 27 lamps of 60 kW power con~umption[’~’.
9. The Importance of Industrial Photochemistry
In his comprehensive monograph “Allgemeine Photochemie” written some 40 years ago, Plotnikow predicted that
photochemical syntheses on an industrial sclae would be limited to a few special cases, i.e. to the production of particularly
expensive special tie^['^! Hejustified this statement by pointing
out that light reactions generally require irradiation of large
surface areas which would incur very high costs. In Plotnikow’s
opinion, a tree is an ideal photochemical factory; its leaves
present a maximum absorption area for a minimum volume.
He urgently warned against any, necessarily imperfect, imitation of Nature.
Nowadays, light reactions have acquired a firm place in
the armamentarium of industrial methods. The foundations
of photochemical technology have been laid, for it is now
possible to scale-up photoreactions from the laboratory to
production units employing 40- or 60kW lamps.
The space-time yields of non-radical photochemical syntheses, which are still frequently unsatisfactory, should be
improved by the development of even more powerful lamps;
prototypes of 100kW mercury lamps have already been constructed.
Numerous patent applications indicate that although light
will continue to be used predominantly for initiating radical
chain reactions, its use in the production of synthetically
demanding products for drugs, plant protection, and fragrances will steadily increase. Thus our extensive knowledge
about the manifold possibilities of photochemical
will find more and more practical application.
Received: May, 12, 1977 [A 193 IE]
German version: Angew. Chem. 90, 17 (1978)
25
[l] M . Pape, Pure Appl. Chem. 41,535 (1975).
[2] M . Fischer, L’actualitt Chim. 1974 (5), 7.
[3] W 7: Anderson, Jr., Ind. Eng. Chem. 39, 844 (1947); W H. Shearon,
Jr., H. E. Hall, J. E. Stevens,Jr., ibid. 41, 1812 (1949).
[4] M . Fischer, H. Barzynski in: Ullmanns Encyklopadie der technischen
Chemie. 4th Edit. Verlag Chemie, Weinheim 1973, Vol. 3, p. 305.
[5] R. Thiele, G. Weickert, H . Voss, D. Hiiusler, Chem. Tech. (Leipzig)
27, 5 (1975); R. Thiele, G. Weickert, U . Hertwig, D . Hiiusler, ibid. 27,
595 (1975); G . Weickert, R. Thiele, ibid. 28, 580 (1976); A. E. Cassano,
J . M . Smith, AIChE. J . 12, 1124 (1966); 13, 915 (1967); A. E. Cassano,
P. L. Silueston, J. M . Smith, Ind. Eng. Chem. 59, (1) 19 (1967); S.
M . Jacob, J. S. Dranofi Chem. Eng. Prog., Symp. Ser. 64, (89) 54
(1968); 7: Matsuura, J . M . Smith, AIChE. J . 16, 1064 (1970); E. Bandini,
C. Strarnigioli, F . Santarelli, Chem. Eng. Sci. 32, 89 (1977); B. Boual,
J . M . Smith, ibid. 28, 1661 (1973); M . Ibarra, J. M . Smith, AIChE.
J. 20, 404 (1974); W J. Zollner Ill, J . A. Williams, ibid. 18, 1189
(1972); H. A. Irazoqui, J. Cerda, A. E. Cassano, &id. 19, 460 (1973);
F. P. Ragonese, J . A. Williams, ibid. 17, 1352 (1971); F. B. Hill, R.
M . Felder, ibid. 1 1 , 873 (1965); Chem. Eng. Sci. 25, 1751 (1970); 0.
Wbrz, Chem.-1ng.-Tech.45,1256 (1973); J . E. H U E C.A. Walker, AIChE.
J . 8, 193 (1962).
[6] E Harano, 7: Matsura, Int. Chem. Eng. 12, (1) 131 (1972); C. M .
Doede, C. A. Walker, Chem. Eng. 1955, 159.
[7] 7: Hutson, Jr., R. S . Logan, Chem. Eng. Prog. 68, (5) 76 (1972).
[8] H. Rassaerts, D. Witzel in: Ullmanns Encyklopadie der technischen
Chemie. 4th Edit. Verlag Chemie, Weinheim 1975, Vol. 9, p. 476.
[9] F. Broich, Fette, Seifen, Anstrichm. 72, 17 (1970).
[lo] L. J . Governale, J. 7: Chrke, Chem. Eng. Prog. 52, (7) 281 (1956).
[ l l ] H. G . Haring, H. W Knol, Chem. Process Eng. 45, 560, 619, 690 (1964);
46, 38 (1965).
[12] K . A. Lipper in: Ullmanns Encyklopadie der technischen Chemie. 4th
Edit. Verlag Chemie, Weinheim 1975, Vol. 9, p. 525.
[13] H. I. Joschek, Chem.-Ztg. 93, 655 (1969).
[14] F. Asinger, Ber. Dtsch. Chem. Ges. 77, 191 (1944).
[15] K . Lindner: Tenside, Textilhilfsmittel, Waschrohstoffe. Wissenschaftl.
Verlagsges., Stuttgart 1964, Vol. I, p. 705.
1161 See [15], p. 717.
[173 H. Schuller in: Ullmanns Encyklopadie der technischen Chemie. 3rd
Edit. Urban & Schwarzenberg, Miinchen 1965; Vol. 16, p. 724.
[18] F. Asinger: Die Petrolchemische Industrie. Akademie-Verlag, Berlin
1971, p. 731.
[19] L. Orthner, Angew. Chem. 62, 302 (1950); R. GraL Justus Liebigs Ann.
Chem. 578, 50 (1952).
[20] C. Beermann, Eur. Chem. News, Normal Paraffins Supplement, Dec.
2, 1966, p. 36; H. Hartig, Chem.-Ztg. 99, 179 (1975).
[21] K . Rehn, K . Schimmelschmidt, DOS 1518854 (1965), Hoechst AG; K .
Schimmelschmidt, DBP 907052 (1942), Hoechst AG.
[22] H. Baumann: Waschmittelchemie. Hiithig, Heidelberg 1976, p. 55.
[23] P. Turner, Inf. Chim. 9, (5, 6) 51 (1970); P. Hulme, P. E. 7lirner, Chem.
Process Eng. 1967, (11) 96; E Ito, E Hara, DAS 1468737 (1962), Toray.
[24] J . Fragner: Vitamine. Gustav Fischer Verlag, Jena 1964, Vol. I, p.
61 3.
26
[25] S. B. Greenbaum in Kirk-Othmer: Encyclopedia of Chemical Technology. 2nd Edit. Interscience, New York 1970, Vol. 21, p. 549.
[26] G. M . Sanders, J. Pot, E. Hauinga, Fortschr. Chem. Org. Naturst.
27, 131 (1969).
[27] M . F. Hollick, E. J . Semmler, H. K. Schnoes, H. F. D e Luca, Science
180, 190 (1973).
[28] D. H. R. Barton, R. H. Hesse, M . M . Pechet, E. Rizzardo, J. Am.
Chem. SOC.95, 2748 (1973); N . Miiller, DOS 2400189, BASF AG.
[29] A. Fiirst, L. Labler, W Meier, K.-A. Pfoertner, Helv. Chim. Acta 56,
1708 (1973).
[30] Eur. Chem. News, August 22/29, 1975.
[31] 0. A. de Bruin, H. F. L. Schbler, J . N . Walop, Philips Tech. Rundsch.
28, (3.4) 72 (1967); M . P. Rappoldt, 7: R. Rix, Rec. Trav. Chim. Pays-Bas
90, 27 (1971).
1321 W Re& H. Grassner, Chem.-1ng.-Tech. 45, 646 (1973); H. Pommer,
Angew. Chem. 89, 437 (1977); Angew. Chem. Int. Ed. Engl. 16, 423
(1 977).
[33] M . Fischer, K-W Wiersdorfi A. Niirrenbach, D. Horn, F. Feichtrnayr,
DBP 2210800, BASF AG.
[34] M . Lalonde, H. Stoller, DOS 2548883, Hoffmann-La Roche.
[35] G . Ohlofi E. Klein, G . 0.Schenck, Angew. Chem. 73, 578 (1961).
[36] G . 0.Schenck, Dechema-Monographien 24, 105 (1955).
1371 M . Akthar, Adv. Photochem. 2, 263 (1964).
[38] D . H . R. Barton, N . K . Basu, M . J. Day, R. H. Hesse, M . M . Pechet,
A. N . Starrat, J. Chem. SOC.Perkin Trans. I 1975, 2243.
[39] K. R. Laumas, M . Gut, J . Org. Chem. 27, 314 (1962); D. H. R . Barton,
J . M . Beaton, J . Am. Chem. SOC.83,4083 (1961).
1401 M . Pape, Fortschr. Chem. Forsch. 7, 559 (1967).
[41] E. Miiller, R. Metzger, D. Fries, U . Heuschkel, K . Witte, E. Waidelich,
G . Schmid, Angew. Chem. 71, 229 (1959); E. Miiller, H. G. Padeken,
M . Salamon, G. Fiedler, Chem. Ber. 98, 1893 (1965).
[42] E. Miiller, G. Fiedler, J. Heiss, Chem. Ber. 101, 765 (1968).
[43] H. Miyama, K . Fukuzawa, N . Harumiya, Y. Ito, S. Wakamatsu, J . Phys.
Chem. 73,4345 (1969).
[44] E Ito, S. Matsuda, Ann. N. Y . Acad. Sci. 147, 618 (1969).
[45] E Shintokawa, S. Torimitsu, Y. Ito, J. Chem. SOC.Jpn. (Ind. Chem.
Sect.) 68, 937 (1964).
[46] Y. Ito, S. Torimitsu, E Shimoknwa, F. Nishikawa, E Hara, DOS 1468 730
(1962), Toray.
[47] I! Ito, S. Wakamatsu, E Hara, J . Morino, DAS 1493257 (1964), Toray.
[48] E Ito, K . Ichiro, Y. Hara, DAS 1299600 (1963), Toray.
[49] Eur. Chem. News. Feb. 27, 1976, p. 25; April 30, 1976, p. 24; P.
Hulme, Chem.-Anlagen Verfahren, 1967, (12) 21; M . Tauerna, M . Chiti,
Hydrocarbon Process. 1970, (1 1 ) 137; Eur. Chem. News, Caprolactam
Supplement, April 30, 1969.
[SO] Inf. Chim. 8, (3, 4) 47 (1970).
[Sl] Eur. Chem. News, March 29, 1974, p. 18.
1521 Chem. Ind. (Dusseldorf)27, 735 (1975).
[53] J . Plotnikow: Allgemeine Photochemie. 2nd Edit. de Gruyter, Berlin
1936, p. 78.
[54] Houben-Weyl: Methoden der Organischen Chemie. 4th Edit. Thieme,
Stuttgart 1975, Vol. IVj5.
Angew. Chem. I n t . Ed. Engl. 17, 16-26(1978)
Документ
Категория
Без категории
Просмотров
3
Размер файла
1 011 Кб
Теги
application, synthese, industries, photochemical
1/--страниц
Пожаловаться на содержимое документа