Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
~.~Z~53
PROCESS FOR THE SEPARATION OF AROMATIC COMPOU~DS
FRO~ HYDROCARBO~ MIXTURES
The present invention relates to a process for the separation
in the liquid phase of aromatic compounds from hydrocarbon mixtures.
Hydrocarbon oils which are to be used in the medical and
related fields must meet high purity demands. ~his type of oils,
which are often designated as medicinal oils, are used especially
in the medical field as liquid paraffins and as constituents of
pharmaceutical preparations. In addition, they are used as
lubricating oil for food-processing machinery, where there is a
risk of the lubricating oil coming into contact with the food.
Certain cosmetic products, such as skin ointments, lotions and
sunburn lotions also contain medicinal oils.
Medicinal oils must be substantially free of aromatic com-
pounds, and their purity in this respect must meet certain standard
specifications, such as those according to the DAB VII test (of
the Cerman pharmacopoeia) and the British sulphuric acid test
(of the British pharmacopoeia~.
In order to remove aromatic compounds from hydrocarbon
mixtures several methods have been described. From U.S. patent
specification 2,548,502 it is known to remove unsaturated com-
pounds, including aromatic compounds, from hydrocarbon mixturesin the gaseous phase with the aid of silica gel and activated
carbon as adsorbents.
~2;~S53
In U.S. patent specification No.,2,754,3l~3 a process is
described for the separation of aromatic compounds from non-
aromatic compounds in the liquid phase, which is carried out
by making use of adsorption. The adsorben-t which is preferably
used in said process is silica gel. After the adsorption, a
regeneration solvent, for example xylene or toluene, is used to
separate the adsorbed aromatic compounds from the adsorbent.
This cycle of adso~ption and regeneration is repeated as long
as no deactivation of the silica gel occurs.
In the German "Offenlegungsschrift" (published u~examined
patent application) 2,364,333 there is described a process for
the purification of slack wax by separating in particular the
aromatic compounds therefrom by means of an ion-exchange macro-
porous resin, which may contain a metal of Group Ib.
The methods mentioned are not very efficient.
It has now been found that it is possible to separate very
efficiently, in the liquid phase~ aromatic compounds from a hydro-
carbon mixture by means of adsorption with the aid of a solid
adsorbent comprising an inorganic porous carrier on which a
silver compound has been supported. It has, moreover, been found
that it is possible to regenerate the loaded aasorbent without
reducing its capacity for adsorption of aromatic compounds.
Accordingly, the invention relates -to a process for the
separation of aromatic compounds from a hydrocarbon mixture
which comprises contacting said mixture in the liquid phase
with a solid adsorbent comprising an inorganic porous carrier
on which a silver compound has been supported, and subsequently
separating the adsorbed aromatic compounds from the adsorbent.
As inorganic porous carriers oxidic carriers are very suit-
able, such as aluminas, silica-aluminas (amorphous and
crystalline) and in particular silica.
The adsorbent according to the invention preferably contains
from 10 to 30% by weight of silver. I-t is possible to support the
silver in the form of a silver salt, for example silver nitrate,
~z~553
on the inorganic porous carrier, *or'example by impregnating the
carrier (e.g., silica gel or alumina gel) with an aqueous solution
of silver nitrate. Other methods, such as precipitation of
silver salts or the supporting of silver in the vapour phase on
a carrier, may also be used. Before using the adsorbent, the
silver-containing (e.g., impregnated) carrier is preferably dried
and calcined with an oxygen-containing gas at a temperature o~
from 350 to 500C.
The process according to the invention is very conveniently
carried out by percolating the hydrocarbon mixture to be purified
through a column which contains the inorganic porous carrier on
which a silver compound has been supported. The contacting temper-
ature is preferably from 70 to 150 C. The conditions are prefer~
ably chosen in such a way that the product obtained contains not
15 more than 0.00170wt of aromatic compounds. When the amount of
aromatic compounds in the product obtained after the contact with
the adsorbent exceeds the limits set, the contacting of the
hydrocarbon mixture to be purified with the adsorbent is dis-
continued and the adsorbent is regenerated by removing the
aromatic compounds adsorbed thereon.
The aromatic compounds adsorbed on the adsorbent can readily
be separated therefrom, for example by percolation with an
organic solvent, which preferably comprises polar components.
By this removal of the arom tic compounds the adsorbent is
regenerated.
Suitable solvents may comprise aliphatic alcohols, such as
isopropyl ~lcohol and normal butyl alcohol, ethers, ketones, such
as acetone, methyl ethyl ketone and methyl isobutyl ketone, and
esters, such as ethyl acetate or mixtures of ethyl acetate and
alkanes, for example a mixture of is~tane and ethyl acetate.
The regeneration is preferably carried out at a temperature
of from 50 to 80 C. After the passage of the regeneration solvent,
the adsorbent is dried and subsequently calcined by the passage
of an oxygen-containing gas at an elevated temperature, prefer-
35 ably between 350 and 500 C.
; ' ` : ' ~ ' -
.
~ZZ553
Hydrocarbon feeds which are par~icularly suitable to be
purified by separating aromatics therefrom according to the
process of the present invention, are hydrocarbon mixtures
preferably containing not more than 3% by weight of aromatics.
Hydrocarbon mixtures containing not more than 1% by weigh-t of
aromatics are preferred in particular.
The process according to the invention is in particular
suited for the preparation of medicinal oils from hydrocarbon
mixtures which already have been purified to a great extent,
but in which the content of aromatic compounds, and in particular
polyaromatic compounds, is too high to meet the very stringent
purity requirements set for medicinal oils.
Hydrocarbon mixtures from which medicinal oils can be
manufactured may be prepared in several ways.
Very conveniently such a hydrocarbon mixture, from which
aromatic compounds can very suitably be separated according to
the invention, is prepared by hydrocracking and/or hydro-isomer-
ization of a petroleum fraction. This conversion is very con-
veniently carried out in two stages. In the first stage, a high-
boiling hydrocarbon mixture, for example a distillation residue
or a heavy fraction obtained by pyrolysis of coal, bituminous
shale or tar sand is hydrocracked. Conveniently, use may be
made as feed of petroleum fractions which at least partly have
a boiling point lying above the boiling rangecf lubricating
oils, for example a fraction obtained by vacuum distillation
starting from a petroleum residue obtained during distillation
at atl~ospheric pressure. Such a fraction has a boiling range
lying between 350 and 500 C. Deasphalted residual petroleum
fractions are preferred as feed for the hydrocracking.
The liquid product of the first stage of the hydrocracking
treatment is dewaxed, if desired after distilling off hydro-
carbons having a boiling point below a temperature in the range
from 350 to 400 C. In the second stage, the resultant wax under-
goes a hydroisomerization treatment. The part of the product
- ~ ,
~ . :
~- ~553
of the catalytic hydroisomeriga-tion h~ving a boiling point above
a temperature in the range from 365 to 425C is dewaxed, yielding
a hydrocarbon mixture with a very high viscosity index, of the
order of 130 to 155.
For the hydrocracking and hydroisomerization use may be made
of catalysts known as such, which, e.g., comprise alumina as a
carrier, upon which there have been supported metals of Group
VIII, e.g., cobalt and/or nickel and metals of Group VI B, e.g.,
molybdenum and/or tungsten and/or oxides or sulphides of the said
metals, and if desired non-metals, such as fluorine and/or
phosphorus and/or boron.
If the aromatic compounds are separated from the afore-
mentioned hydrocarbon mixture having a viscosity index from 130
to 155 by means of the process according to the invention,
products are obtained which meet the specifications set for
medicinal oils.
The process according to the invention will further be
elucidated by way of the following ~xamples.
EXAMPLE I
a. Feed:
In order to obtain a suitable aromatic-containing hydro-
carbon feed, a residual petroleum fraction of a North A~rican
crude, deasph~lted with propane, was subjected to a two-stage
hydrocracking treatment. This fraction had the following
properties:
speci~ic weight : o.896
viscosity at 100C: ~5 cS
The fraction was subjected to a catalytic hydrocracking
treatment under the following reaction conditions:
temperature : 390C
pressure ~: 140 bar (absolute pressure)
space velocity : 0.95 kg of oil per hour per litre of
hydrogen/oil ratio: 1,520 n~ of H2 per kg catalyst
of oil.
~t2;~SS3
Of the liquid product of the hy~rocracking treatment, 39% by
weight, representing the fraction of which the boiling point
was below 375 C, was separated by fractional distilla-tion.
The fraction of which the boiling point was above 375 C
(60% by weight) was dewaxed using a mixture of methyl ethyl ketone
and toluene (in equal volumes), at a temperature o~ -27 C. The
solvent/oil ratio was 3:1. During the process, 24 parts by weight
of wax were separated from the fraction. The resultant lubricating
oil had a viscosity index (VIE~ ASTM-D 2270) o~ 130.
The resultant quantity of this lubricating oil amounted to
32% by weight of the original deasphalted residual petroleum
fraction.
The wax released during the dewaxing treatment was then sub-
jected to a catalytic hydroisomerization treatment under the
following reaction conditions:
temperature : 340C
pressure : 140 bar (absolute pre~sure)
space velocity : 0.81 kg of wax per hour per litre
of catalyst
hydrogenloil ratio: 1,660 nl of ~2 per kg of oil.
From the product of the catalytic hydroisomerization con-
version, that fraction having a boiling point b~low 400 C was
separated by fractional distillation.
The resultant quantity of product having a boiling point
above 400 C amounted to 49.4% by weight of the hydrocracking
feed. This oil was dewaxed at -27C with a mixture of methyl
ethyl ketone and toluene (in equal volumes), the sol~ent/oil
ratio used being 8:1.
The resultant quantity of thus dewaxed lubricating oil
amounted to 24.9% by weight ~ the hydrocracking feed.
This lubricating oil had the followillg properties:
viscosity index -
(VIE~ ASTM-D 2270) : 150
kinematic viscosity at 100C: 5.97 cS
- ;
-- -, ~ : :
_ ~ Z2553
As regards this product of cata~ytic hydrocracking con-
version, ultra-violet spectrometric analysis was -then used to
determine what aromatic compounds the reed contained, and in
what quantities.
Table I shows the results of this analysis.
TABLE I
aromatics, m.mol./100 g
benzenes 0.18
naphthalenes 0.01
biphen~ls and
phenanthrenes 0.007
pyrenes 0.008
benzoperylenes
coronenes 0.011
In order to gain an insight into the pharmaceutical quality
of the ~eed, the sulphuric acid test was carried out according
to the test method of the British pharmacopoeia IP 17 and the
U.V. absorption test according to the DAB VII test of the German
pharmacopoeia.
Table II below shows the results of these tests. It also
states the specifications applying to medicinal oils.
TABLE II
DAB VII Test (German pharmacopoeia) Speci~i- Result
cations
_ . ......................... ~ ~ _
absorption o~ 275 nm ultraviolet rays < o.8 1.75
absorption o~ 295 nm ultraYiolet rays < 0.4 11.96
absorption of 300 nm ultraviolet rays < 0.3 13.44
Sulphuric acid test
(British pharmacopoeia)
red < 2.5 11
yellow < 6.5 5o
.~ .,: : . .
- . . : ,
:~ZZ553
b. Adsorbent:
The silver-containing adsorbent was prepared as follows.
Starting from an aqueous solution of silver nitrate, silver
nitrate was impregnated on the porous silica gel having a
specific surface area of 363 m2/g and a pore volume of o.g8 ml/g.
Subsequently, the impregnated silica gel was dried by the
passage of a stream of hot nitrogen at a temperature of 120 C.
Then the adsorbent was calcined by the passage of a stream of
oxygen at a temperature of 425 C and a space velocity of 20 l/h
for two hours. The silver content of the adsorbent prepared ac-
cording to the afore-mentioned process was 20%.
c. Adsorption:
The hydrocarbon feed containing aromatics, described above7
was percolated through a column filled with 100 g of adsorbent
at a temperature of 70C and a space velocity of one part by
volume of oil per part by volume of adsorbent per hour. Then the
production of the column, expressed as the ratio between the
volume of percolated oil and the volume of adsorbent, was
measured. It was found that this ratio could be raised above
25, the volumesof oil leaving the column all corresponding to
the specifications applying to medicinal oils. Table III below
shows the results of the DAB VII test (absorption of ultra-
violet rays) and of the sulphuric acid test for the 10th, 20th
and 25th volume respectively. For greater clarity, it also shows
the results of the analysis of the feed and the specifications
which appl~ to medicinal oils.
: ,
. ~
-` ~lZZ553
TABLE III'
Details vol of DAB VII Test H2S0
adsorbent i 275 nm 295 nm ¦ 300 nm
feed _ l _ 11.96 13.44 red - 6.8
l yello~26
i - __
speci- , <o.8 <.4 ~ <0.3 I red < 2.5
fica- ! ! yellow<6.5
tions I I
. .
per- 10 0.20 1 0.115 0.15 red 1.5
colated , yellow 3.0
oil 20 0.30 0.25 0.20
0.45 ' 0.30 0.25 red 2.5
~i yellow 6.o
EXAMPLE Ia
By means of a comparative test differing only from that in
Example I by the use of non-impregnated silica gel, it was
investigated what quantity by volume (in relation to the volume
of adsorbent) corresponded to the specification of medicinal oils.
Table IV below shows how disappointing the results are if the
adsorbent used is a silica gel not impregnated with silver
nitrate. Even by bringing down the ratio to fol~ i-t was no
longer possible to meet the specifications according to the
DAB VII test.
TABLE IV
Details vol. of feed DAB VII
vol. of adsorbent 275 nm 295 nm 300 nm
percolated oil 0.045 0.99 0.88
specifications <o.8 <0.4 <0.3
~lZZ~53
EXAMPLE Ib
By way of compa~ison, a series of adsorption tests were
then carried out with an adsorbent consisting of an ion-exhange
resin (Amberlyst-15)* charged with silver ions and known from the
German "Offenlegungsschrift" 2,364,333. These tests also involve
the determination of the maximum quantity by volu~e (in relation
to the volume of adsorbent) still corresponding to the specifica-
tion. Table V states the results of these tests.
TABLE_V
¦Details vol. of feed DAB VII
~ol. of adsorbent 275 mm 295 mm 300 mm
_
percolated oil 7 ~0.30 0.48 0.4
specifications ~0.8 ~0.4 ~0.3
... .. _ . ... .
From this Table it follows that the maximum ratio is low-
er than 7.
EXAMPLE II
In order to investigate to what extent the activity of
the adsorbent is influenced by repeated regenerations, 20 re~ener-
ation testswere carried out with isopropyl alcohol as regeneration
solvent. The space velocity of the regeneration solvent was 1
litre per litre per hour. Each regeneration comprised the passage
of three volumes of solvent over one volume of adsorbent. Then the
adsorbent was dried by the passage of a stream of air at 70 C,
after which the adsorbent was calcined for 2 hours at 425 C by the
passage of stream of air for 2 hours. Between the regeneration
tests, adsorption was carried out each time by passing 25 volumes
of feed over one volume of adsorbent
-- 10 --
: . ~ - . :
3L~LZZS53
As Table VI shows, the activity of the adsorbent did
not diminish after repeated regeneration.
lQ~ -
.~;
~: ..
- : :
:: :
~ZZ~53
TABLE VI '
Details DAB VII Test H2SO
275 nm 295 nm 300 nm
Before the 1st
regeneration 0. 45 0.30 0.25
After the 20th red = 2.4
regeneration 0. 45 0.30 0.25 yellow- 6
Specifications <O.o <0. 4 <0. 3 red =< 2.5
L l yell~ 6.5
EXAMPLE III
The influence o~ the silver content on the activity of the
adsorbent was investigated at 70 C for three different adsorbents.
The activity of the adsorbent is expressed here, as in Examples
IV and V, as the maximum ratio between the volume of feed and
the volume of adsorbent, this maximum being established by the
final volume of oil still corresponding to the DAB VII speci-
fication. Apart from the variation in adsorbent, all the con-
ditions of these tests are the same as those in Example I.
TABLE VII
. Adsorbent activity % by weight
vol. of feed o~ silver
vol. of adsorbent
_
~ilica gel, l6.5 5
specific surface area: 28.5 15
363 m2/g
pore volume: o.98 ml/g 33.0 20
9.5 30
~lu=ina 10. 4 5
specific surface area:
211 m2/g 19.8 15
pore volume: o.65 ml/g ! 13.0 20
alumina, 123. 7 10
specific surface area: '
223 m2/g ' 23.1 15
pore volume: o.64 ml/g ¦ 19.4 20
5 3
12
EXAMPLE IV
The influence of the calcination temperature on the same
maximum r~tio was investigated by using the test conditions
described in Example I. Table VIII below shows the results o~
this investigation.
TABLE VIII
_ _
Adsorbent activity calcination
vol. o~ feed temperature,
vol. of adsorbent C
silica gel, 29.5 400
specific surface a~ea:
363 m~/g 24.4 450
pore volume: o.98 ml/g 19.9 5
% by weight of Ag:
12.5 15.8 55
EXA~LE V
The influence of the adsorption temperature on the activity
was investigated, likewise by using the test conditions described
in Example I. Table IX shows the results o~ this investigation.
TABLE IX
Adsorbent activity adsorption
temperature,
vol. of feed o
vol. o~ adsorbent C
alumina 11.2 5o
specific surface area:
223 m2/g 26.7 7o
i pore volume: o.64 ml/g 31.2 90
% by weight of Ag:
12.5 ~6 6 130
: ' ~':
