Note: Descriptions are shown in the official language in which they were submitted.
8~
- 1 -
~ESTROYING HAlOGEN-CONTAINING ORGANIC C~MPOUNDS
Polychlorinated biphenyls (PCB's) is a generic term
covering a family of partially or wholly chlorinated
isomers of biphenyl. PCB's are non-conductors of
electricity and have good resistance to high tempera-
tures, so they are widely used as working fluids in
heat exchangers and hydraulic systems and by the
electrical industry in transformers and capacitors.
PCB's are extremely toxic, but are difficult to destroy
on account of their thermal stability and chemical
inertness. The standard destruction method involves
incineration at temperatures around 1500C, but suffers
from several disadvantages. Operation at these very
high temperatures is expensive; and incomplete combustion
can give rise to chlorinated dioxins or furans which
are even more toxic than PCB's. The present invention
provides a method for the destruction of PCB's and
related compounds which involves reaction with metal
rather than combustion, and operates at much lower
temperatures,
C. S. Shultz describes in U.S. Patent 4,469,661 a
method for destroying PCB's by contacting them in
vapour form with molten aluminium metal. This process,
which is not demonstrated by Shultz, is unsatisfactory
for several reasons. The use of a body of molten
aluminium metal is somewhat hazardous, on account of
the risk of explosion, and expensive, on account of the
high temperatures involved (aluminium melts at 660C)
and the difficulty of containing molten aluminium which
aggressively attacks standard materials such as steel
and quartz. It is expected to be more difficult to
ensure intimate physical contact of gaseous PCB's with
molten aluminium than with aluminium in the solid
state, and that the reaction of PCB's with molten
aluminium will produce much more AlCl3 than will
reaction with aluminium in the solid state.
.
~$
~2 ~
-- 2
Shultz discloses as a non preferred embodiment,
but does not claim, a process which involves contacting
PCB's with a solid aluminium surface. As reported in
Examples 1 to 3, his experiments involved heating
transformer oils containing PCB's in the presence of
aluminium foil at various temperatures, but did not
result in complete destruction of the PCB's even over
periods as long as 30 minutes.
Japanese Kokai 51-25471 describes a method of
decomposing PCB's by heating them to ef'fect partial
dissociation to HCl or Cl2, and passing the mixture
over a suitable metal and recovering a chloride salt of
the metal by sublimation from the decomposition system.
There is no indication that air is excluded; and no
evidence that complete decomposition of PCB's is
achieved even after several hours reaction.
This invention is based on the discovery that
PCB's can be rapidly destroyed by heating in the
presence of solid aluminium, but only provided certain
critical parameters are observed. The invention thus
provides a method of destroying halogen-containing
organic compounds by reaction with a metal in the solid
state at elevated temperature, characteriæed by
contacting the compounds in gaseous form in the absence
of oxygen with a metal selected from Al, Mg, Si, Ti and
Be, and alloys thereof having a high specific surface
area at a temperature of at least 450C and a contact
time of from 0.5 to 50 seconds.
Tests have established that the method is capable
3o of destroying a wide range of chlorinated organic
compounds. We know of no reason why any chlorinated
organic compound, which is thermally stable enough to
be heated up to the required reaction temperature,
should not be destroyed by the method. The invention
is also applicable to organic compounds of the other
halogens, fluorine, bromine and iodine. However,
~58
- 3 -
environmental problems do not exist to the same extent
as with chlorine compounds because of the far greater
industrial use of the latter. The invention is, of
course, of particular value in relation to PCB's and
related compounds.
It is necessary that the compounds should be in
the gas phase, for reaction of liquid compounds with
solid metal has not proved effective. The compounds
are preferably entrained in an inert carrier gas, for
example argon or other gas in group 0 of the periodic
table. High purity nitrogen may also be used, and is
regarded for this purpose as an inert gas. However,
its use is not preferred, for nitrogen is known to form
highly toxic compounds with PCB's but the concentrations
so far observed of these are insignificant.
There is no critical upper limit of concentration
of the compounds in carrier gas. The method works
without a carrier gas, provided that sufficient metal
surface area is available for reaction, but would only
be safe with a completely closed reaction vessel. In
industrial practice, safety considerations determine an
upper limit of concentration. There is no critical
lower limit of concentration, but a practical lower
limit is generally determined by economic factors.
The concentration of halogen-containing organic
compounds in carrier gas is preferably from 10 ppm up
to 10%.
It is at all events necessary that oxygen, and
compounds that might generate oxygen in situ, be
3o substantially absent. If the method is performed in
the presence of significant quantities of oxygen, then
destruction of PCB's is incomplete and there is the
risk of formation of chlorinated dioxins or furans.
As metals that can be used for reaction with the
compounds, are specified Al, Mg, Si, Ti and Be, and
alloys of these metals with each other or with minor
~51
- 4
proportions of other metals. The five named metals
have two characteristics in common: they form oxides
having electrically insulating properties; and the
oxides have high thermal and chemical stability.
Berylium presents a toxicity problem in itself, and is
on that account the least preferred. The most
preferred metals are magnesium and, particularly,
aluminium. The metals can be used in the natural
state, i.e. without the need to remove any oxide film
that may be present. Aluminium can be used having an
anodic oxide film which may contain minor proportions
of oxides of other metals such as Co, Ni, Sn, Cu etc.,
in the pores. It is not known with certainty whether
any oxide film remains during operation of the method,
or whether it is removed by reaction with halogen-
containing organic compounds.
The metal can be used in any physical form in
which it has a high specific surface area. Suitable
forms include a packed bed of spheres, chips or
granules, a fluidized bed of powder, honeycomb, wire
mesh or wire wool. Our presently preferred material
is scrap aluminium and alloys thereof in granulated
form, because this is cheap and readily obtainable.
Sufficient metal surface area should be provided to
ensure rapid and complete destruction of the halogen-
containing organic compounds. This is generally 0.1
to 65 m2~ preferably from 1 to 20 m2, of active metal
surface (not necessarily bare metal surface, but
surface not coated with e.g. inactivating carbonaceous
deposits) per gram, of compound to be destroyed.
To achieve sufficiently rapid reaction, the
reaction temperature needs to be at least 450C. An
upper limit on temperature is set by the melting point
of the metal being used. One of the advantages of the
method of this invention is the low temperatures
required, and it is preferred not to use higher
-- 5
temperatures than are necessary in order to a~hieve
reaction at the desired rate. Depending on other
conditions, preferred reaction temperatures are likely
to lie in the range 550C to 650C.
Provided the above reaction conditions: absence
of oxygen; state of the metal; temperature of the
metal, are maintained as described above, destruction
of halogen-containing organic compounds are achieved at
short contact times. We specify a range of up to 50
seconds, preferably 0.5 to 30 seconds contact time.
In a continuous system, this is the average residence
time of gas in the region of the active metal surface.
Clearly reaction time is related to the total surface
area of metal per unit of halogen-containing organic
compounds, and to the reaction temperature. Adjustment
of gas flow to ensure complete destruction of the
compounds is achieved by routine trial and error.
Reaction products resulting from the method appear
to be metal halide (e.g. aluminium chloride), low-
boiling hydrocarbons, halogen, (e.g. chlorine) andcarbon deposited on the metal substrate. As a result
of this deposition, the substrate gradually becomes
inactive. When this happens, the substrate can be
regenerated. With aluminium, this can be achieved by
subjecting the metal to sodium hydroxide solution, or
less preferably, by heating the metal in air to burn
off the carbon deposits. Other treatments for
regenerating aluminium involve exposing the carbonised
surface:-
3o (a) at 580C to oxygen for 30 minutes followed
by chlorine for 1 minute.
(b) hydrogen for 30 minutes at 550C followed by
oxygen for 30 minutes.
(c) oxygen for 30 minutes at 550-580C followed
by a steam/oxygen mixture at 125-175C.
Hydrcgen and chlorine may be diluted with flowing
~s~
- 6 -
argon. Of the above, treatment (a) is preferred. It
seems possible that the regenerated surface is in some
way "re-activated" by the chlorine. Using this
treatment, a 79.7% recovery of usable surface was
obtained. Treatment (b) is more preferred than (c).
Other metal substrates can similarly be regenerated by
removing the carbon deposits under conditions in which
the substrate is not affected.
The following examples illustrate the invention.
Example 1
The metal used was 1100 aluminium alloy chips
(0.5 x 0.5 x 0.1 cm). A bed about 27 cm long of
these chips was positioned in a vertical quartz tube
1.8 cm outside diameter, and maintained at a nominal
temperature of 580C. A vessel containing the
reactant was positioned in the quartz tube below the
bed, and was surrounded by a separate tube furnace
2U whose temperature was raised from ambient to 600C over
a period of 60 minutes. The lower end of the quartz
tube was closed except for an inlet port through which
argon carrier gas was passed at a flow rate of
87.7 ml/min (NTP). As the reactant heated up it
vapourized and became entrained in the carrier gas~
The flow rate was such that the residence time of the
gas in the bed of aluminium chips was about 15 seconds.
The temperature profile of the bed of chips was
measured as 366C at O cm up from the bottom; 473C at
30 5 cm; 563C at 9 cm; 601C at 16 cm and 600C at 27 cm.
The top end of the quartz tube was closed except
for a gas outlet, and the reaction products were
condensed. After the experiment, any remaining
reactant, the material in the bed of aluminium chips
and the reaction products were all analysed for
halogen-containing organic compounds.
G~;~
-- 7 --
In one experiment, the reactant was 0.0078 g of
decachlorobiphenyl. The destruction efficiency was
99.9999%. The section from 14 to 24 cm (measured from
the bottom) of the aluminium bed became black.
In other experiments~ the rate of heating of the
reactant sample was varied so that the concentration of
decachlorobiphenyl in carrier argon gas ranged from
166 ppm to 3048 ppm. In all cases~ the destruction
efficiency was at least 99.999%.
After one experiment, dry air was passed through
the bed of aluminium chips at 580C. The bed, a
section of which had been blackened by carbon deposition,
was partially regenerated by contact with air. Only
~ery little black colour remained on the aluminium
surface.
Comparative experiments were run under the same
conditions but with an empty bed, i.e. without the
aluminium chips. Destruction efficiency was of the
order of 0 to 5%. This demonstrates the excellent
high temperature stability of decachlorobiphenyl under
normal circumstances, and the dramatic effect produced
by the bed of aluminium chips.
Example 2
Other halogen-containing organic compound reactants
were destroyed by the laboratory method described in
Example 1. Destruction efficiency was not measured
with the same accuracy as in Example 1, because of the
3o lack of. analytical techniques for measuring small
amounts of different halogenated organic compounds.
a) The reactant was 0.5 ml of carbontetrachloride,
injected at ambient temperature into the carrier gas.
The bed of aluminium chips was one that had been
regenerated by air as described in Example 1. The
destruction efficiency was not measured accurately, but
~'~5
was high.
b) The reactant was 0.2 ml of ethylene dichloride,
injected at ambient temperature into the carrier gas.
The destruction efficiency was greater than 90%.
c) The reactant was 1.0 ml of Freon-113, injected
into the carrier gas. The destruction efficiency was
greater than 90%.
d) The reactant was 0.0052 g of "Vitar" fluorocarbon.
This was all decomposed in the sample vial and did not
reach the bed of aluminium chips.
e) The reactant was 0.0613 g of iodobenzene. The
destruction efficiency was 99.6%.
f) The reactant was 0.0185 g of pyranol transformer
oil containing 60% by weight of pentachlorobiphenyl.
The concentration of reactant in the argon carrier gas
was 1000 to 2000 ppm. The destruction efficiency was
98%.
Example 3
This example demonstrates the use of different
metal substrates in the laboratory method generally
described in Example 1.
a) A bed of aluminium alloy chips was used (as
described in Example 1). The bed had been previously
used and had been regenerated by treatment with an
aqueous solution containing about 5 g of sodium
hydroxide per litre which was effective to remove all
the carbon residues. The reactant was 0.0261 g of
pentachlorophenol. The destruction efficiency was
greater than 95%.
b) The bed comprised anodized aluminium chips made
from a sheet of 5252 alloy with a 7.5 micron anodic
oxide film which had been coloured electrically with
cobalt. The size of the chips was 0 A 5 x 0.5 x 0.1 cm.
Both sides of the chips were coloured black with cobalt
but the periphery was bare 5252 alloy. The reactant
~5~36
was 0.0049 g of decachlorobiphenyl. The destruction
efficiency was 99.999%.
c) The bed was composed of aluminium fines, that is
to say particles which passed through a 20 mesh sieve
(opening 0.84 mm). The reactant was 0.06 g of pyranol
transformer oil. The destruction efficiency was
99 . 99% .
d) The bed was of 70 to 80 mesh (about 0~2 mm
opening) magnesium metal powder. The reactant was
0.009 g of decachlorobiphenyl. The destruction
efficiency was 99.999%.
Example 4
This example demonstrates how the carrier gas can
be altered or omitted. Where not otherwise stated,
conditions were as described above in Example 1.
a) The bed was of super pure aluminium fines of a
particle size to pass through a 20 mesh sieve (opening
20 0.81l mm). The reactant was about 0.0078 g of
decachlorobiphenyl. No carrier gas was used. The
reactant was heated from ambient temperature to 580C
in 15 minutes and maintained at 580C for a further 15
minutes. The destruction efficiency was 99.999%.
The amount of aluminium used was 1l16 grams per gram of
decachloro-biphenyl, but further experiments
demonstrated that less than 100 grams per gram were
equally ef~ective. While it is not meaningful to talk
about a contact time bewtween reactant and substrate in
3o a laboratory experiment of this kind, in commercial
operation there would always be a flow of gas over the
substrate bed.
b) The bed was of aluminium alloy chips as used in
Example 1. Instead of argon, pre-purified nitrogen
was used as the carrier gas at a flow rate of 87.7 ml/min
(NTP). The reactant was 0.0065 g of decachlorobiphenyl,
~513~8~
- 10 -
and the destruction efficiency was 99.999%.
In a comparative experiment, extra dry air was
used as the carrier gas in place of nitrogen. The
destruction efficiency was less than 80%, and
unidentified and possibly toxic produces were found in
the effluent gas.
