Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
This invention relates to a process whereby a
c.rystallizable component is separated from a liquid multi-
component system by a process of crystalliza~ion, induced
by progressively cooling said system. This process involves
for example the concentration of fruit juice, vegetable
juices and extracts (such as coffee extract), wine, beer
and other materials which comprise aqueous solutions which
can be concentrated by the ~ormation and removal of ice
crystals. However, also inorganic compounds can be separated
out from aqueous solu~ions. Likewise this process may
involve the resolution of non-aqueous mixtures of organic
compounds.
In continuous, or semi continuous crystallization
processes of this kind a stream of the multicomponent liquid
lS system is progressively passed through zones of progressively
lower temperatures. In cocurrent processes the crystals
formed follow the same direction as the liquid~ In counter-
current processes the crystals formed axe forced in a
direction opposite to the direction of the liquid. In co-
current processes the crystals have to be separated fromthe mother liquox at a point of the last zone where this
mother liquour is percentagewise richest in components not
being the component to be crystallized out and separated. In
contras~, in countercurrent processes the crystals are se-
parated out from the suspension in the first zone, where the
,,1 ~
-- 2 --
liquid is comparatively poor in components which do nothave to be crystallized out. It stands to reason that it is
easier to obtain purity of the component which crystallizes
and which has to be sep~rated outr in the last (counter-
current) case than in the first (cocurrent) case. The con-
tent of impurities adhering to the crystals will be less.
Moreover, in most cases the viscosity in the first zone will
be lower, which facilitates the separatio~ o~ mother liquor
~rom the crystals.
To summarize: generally in cocurrent processes, crystals
have to be separated out from a relatively viscous mother
liquor, with a relatively high concentration of substances
which are to be considered as impurities if they adhere to
the crystals, while in countercurrent processes the reverse
holds true.
Examples of countercurrent processes of this kind are
to be ~ound in U.S.P. 3,283,522 ~Ganiaris), U~S.P. 3,645,699
(Brodie) and U.S.P. 4rl88r797 (Thijs~en~.
In this last patent the process described involves
ireeze concentration, whereby the Gne component to be sepa~
rated out is ice, but an apparatus of the same construction
can also advantageou~ly by used when said one component is
e.g. a crystallizable organic compound.
The present invention provides a process for the
separation of one crystallizable component from a liquid
con~; n; ng a plurality of components comprisin~ passing a
stream of said liquid successively through at least part of
a series of zones of successively lower temperatures in-
cluding the zone of lowest temperature, so as to induce the
formation of crystals of said one component in each of said
zones, passing the crystals formed in each of said zones in
a direction countercurrent to the direction of the stream of
liquid through at least part of said zones, including the
zone of highest temperature, separating the crystals present
in the zone of highest temperature from their mo-ther liq~or
and reco~ering at least part of these crystals, and separating
the crystals in the zone of lowest temperature from their
mother liquor and recovering that mother liquor, character-
ized by recirculating part of said crystals present in one
zone from that zone to a zone of lower temperature~
The advantages of the process according to the present
invention will be discussed hereinafter.
When speaking of "one component" we do not mean that
this one component crys~allizing out should necessarily
have a distinct chemical identity, and certainly not to be
already in a state of high purity. In the case whereby we
have to do with a freeze concentration process and conse-
quently said one component is ice, the crystals can generally
have a high purity. This, however, is e.g. not the case with
many mixtures of organic compounds.
Occlusion of components not desired may occur in the
crystals, and, all dependin~ on the phase diagram, also
"mixed crystals" may crystallize out. In specific cases it
is even possible that the product desired has the composition
gl~
of a mixed crystal. In most cases however a chemically
pure product is desired. During their passage from the zone
of lowes~ to the zone of highest temperature recrystallization
may occur as the crystals formed in the low temperature zones
come in contact, successively in zones of higher tempera~
tures, with liquid with a lower content of t'non-desirables";
e~uilibrium is es-tablished between the crystals and the
liquid and/or occluded and adhering impurities are gradually
removed.
When speaking of "recovering at least part of these
crystals" it does not mean that these have to be fed ollt in
their totality from ~he apparatus as such. One method of
feeding out is in several cases to melt the crystals at
least partly aftPr they have been separated from the mother
liquor. The melt or the melt with crystals still suspended
in it can then be f~d out as a fluid. Part of the melt can
be used as a reflux to "wash" the crystals, as e.g~ described
in U.S.P. 2,854,494 (Thomas) and U.S.P. 3,645,699 (Brodie).
There are several methods by which the crystals can
be induced to follow a path countercurrently to the direction
of the stream li~uid. One such method is described in U.S.P.
3,645,699 ~Brodie). By this method the zones of different
temperatures are not separa-te, but are, so to say, contiguous
to one another. In U.SOP. 3,283,522 (Ganiaris), there are
separate zones and each zone is provided with means to
separate the crystals from the liquid and pass these on in a
different direction. The same occus according to UOS~P.
5~
4,188,797 (Tnijssen).
From the above it follows that the term ~Izone~ has to
be taken in its broadest sense. Though an advantageous
effect can be observed by recirculating part of the crystals
S present in any one of the zones from that zone to a zone of
lower temperature, the best results are generally obtained
by recirculation from the zone of highest to the zone or
lowest temperature. Althou3h the countercurrent process can
in principle be performed while the crystals of any one zone
do not pass all to the subsequent zones o higher temperature
in succession but e.g. are directed immediately to the zone
of highest temperature, preferably the countercurrent process
is operated in such a way that all the crystals pass success-
ively through all of the zones with a temperature higher
than that of said zone.
The liquid can be fed in the zone of highest temper-
ature. This is to be preferred embodiment in those cases
whereby the crystals easily crystallize out in pure con-
dition to wit without occlusion of and not forming mixed
crystals with undesired components e.g. when we have to do
with ice crystals in a process of freeze concentration.
With many mixture of organic compounds the crystals
formed are less pure. In those cases the purity can be in-
creased by feeding the liquid not in the zone of highest
temperature, but in a zone intermediate between the zone of
highest and the zone of lowest temperature. The crystals
of the zone wherein the liquid is fed can in these cases be
-- 6 --
brought in contact ("washed") with a melt of the crystals
of the one component to be recovered. To this purpose the
crystals are partly melted and used as a reflux. One example
- is to be found in U.S.P. 3,645,699 (Brodie). In that specific
case the melt is brought countercurrently in contact with
the impure crystals 50 as to establish equilibrium between
the two phases. Cooling in what is called the "refining"
section takes care that the melted portion again~ crystall-
izes. The refining section described in this patent is com-
bined with an adiabatically operated sectionr designated as"purifying section", which is also used for purposes of se-
parating crystals and liquid by grafity. The refining section
can of course also be combined with e.g. a wash column
instead of a separation by grafity.
While the feed in the processes at issue will be
generally liquid, it may also contain crystals, -thus being
a slurryO
It stands to reason that the effect is the more pro-
nounced the greater the recirculation factor for the crystals
is. When the xecirculated crystals are first separated from
the liquid in which they occur, there is in principle no
other limit to the recirculation factor than the requirement
that in an equilibrated process as much of the "desired"
component has to be remo~ed fxom, as is formed in, the pro-
cess. Of course, for reasons of economy we shall not pushthe recirculation factor beyond what is necessary. Depending
on circumstancesl such as the very nature of the product to
be separated out, the concentration of the liquid~ the cost
of energy, the capital outlay for instrumentation etc., we
shall choose the recirculation such as to obtain an econo-
mical optimization. In most cases recirculation factors of
1-25% are adequate. Generally a recirculation factor of
2-10~ will be preferred. While conveniently and preferen-
tially the recirculation may be accomplished continuously,
intermittent recirculation is not excluded.
According to the present invention there is also pro~
vided an apparatus for applying the process described above,
being an apparatus of the type mentioned above provided with
means to recirculate a predetermined quantity of crystals
per unit of time from any one stage to a stage of lower
temperature. The advantageous effect according to the in-
vention may be understood by the following considerations.
In the zone of highest temperatur2 the crystals willhave the largest size, having had time to grow when passiny
on from the zones of lower temperature to said zone of
highest temperature and will have the highest purity. When
a number of these crystals are,transferred to e.g. the zone
o~ lowest temperature, in which a liquid is present with the
highest content of non~desi-rables, they will have the follow-
ing effect, At the cooling surface of the zone of lowest
temperature nuclei are formed at a high rate, These nuclei
are di5tributed throughout the liquid in that zone by the
scraping and stirring mechanism~ Now, big crystals having a
lower solubility than small crystals, the nuclei are not
~5~
~ 8 --
stable in the presence o~ the recirculated crystals and most
of these nuclei will dissolve again adding to the growth of
the recirculated bigger crystals. This means that the crystals
ultimately obtained in the zone in question will be bigger
than without recirculation as according to the invention.
All this means that the crystals which we will ultimately
get in the zone of highest temperature will be appreciably
bigger than would have been the case without recirculation.
It has to be emphasized that separation processes, whatever
they may be, are the more easy to operate the bigger the
crystals are, and will result in a purer product.
First the bigger the crystals the smaller the surface
per weight unit, and consequently the smaller the quantity
of mother liquor which may tend to adhere to the crystals.
Very important also is the fact that, e.g. in packed bed wash
columns, the pressure to be applied for a certain through~ut
is very dependent on the crystal size, rising sharply as the
crystals get smaller. 'rhis is especially of importance with
crystals of many organic compounds, which, being relatively
soft, are easily de~ormable by prsssure. This results in com~
pression of the crystal layer and thus in lower porosity o~
this layer.
There is one more reason why recirculation according
to the invention will produce a purer product. As the nuclei
at the cooling surfaces of the crystallizer are formed
rapidly there is in many cases an appreciable tendency for
~ 9 --
impurities to be occluded. When the nuclei redissolve due
to the presence of the big recirculated crystals, the mate-
rial present in those nuclei will settle on those recirculated
crystals, but in a more orderly way than by the "shock cooling"
5 at the cold cooling surfaces. Accordingly ~here will be less
tendency for occlusion of impurities.
It may be remarked here to avoid any misunderstanding
that the notion of "desired" and "not desired", as used
above, only means "desired" or "not desired" to crystallize
and does not bear in mind the question whether the component
crystallized out is the most valuable one or, in contrast,
the less valuable one which by crystallization and subse-
quent separation has to be eliminated.
Recirculation of the crystals can ~.g. be accomplished
by withdrawing some of the liquid in the zone of highest
temperature with the crystals suspended in it to the zone of
lowest temperature. As the case may be a thickening may
first take place in order to create a slurry with higher
crystal content. This can e.g. be accomplished in a thicken-
er as descxibed in U.S.P. 4,188,797 (Thijssen).
Many methods for the separation of crystals from their
mother liquor have been proposed in continuous crystalli-
zation processes. Centrifuges are well known, but wash
columns have definite advantages in several instances and
have also been described in several patents, see e.g.U.S.~.
2,854,494 (Thomas), U.S.P~ 3,587,859 (Probstein); U.S.P.
3,872,009 ~Thijssen~ and British Patent Application 79,21808
s~
-- 10 --
(Published British Patent Application No. 2023564A)
(Thijssen)~
Wash columns are appropriate whenever a "washfront"
can be created at about mel~ing temperature with at the one
side the melt of the one component to be separated out and
at the other side the mother liquor trapped in between the
(as the case may be agglomerated) crystals. This has been
extensively discussed in British Patent Application
No. 79,21808 (Published British Application No. 2023564A)
(Thijssen) for aqueous solutions, whereby the one component
to be separated out is ice, but wash columns of the same
construction can also advantageously be used when said one
component is e.g. an organic crystalli7able compound. In
U.S.P. 3,645,69g tBrodie) separa~ion is accomplished by
gravity; the crystals of the component to be separated out
"falling" through a melt of that same component.
The process according to the present invention will
now be further described with reference to the accompanying
drawings, in which~-
Fig. 1 is by way of example a schematic representation
of a flow sheet fox a preerred embodiment o the invention,
in cases wherein an extra refining section is not necessary;
e.g. for freeze concentration; and
Fig. 2 is by way of example a schematic representation
of a flow sheet in the case that-we have to do with amixture
of organic compounds, wherein an extra 1I refining section"
is necessaryO
For details of the apparatus which can be used in
these cases (that is, of course, without recirculation as
according to -the present invention) we refer e.gO to
U.S.P. 4,188,797 (Thijssen) and to the British Patent Appli-
cation No. 79,21808 (Pu~lished British Paten~ Application
No. 2023564A) (Thijssen).
Referring to Fig. 1 the four stages of the crystal-
lizer are shown at 1, 2/3 and 4. Thickeners for crystal
slurry appear at 5, 6 and 7, while the final wash column
is denoted with 8. At 17 we see an apparatus which operates
as thickener or wash column. For this we refer e.~. to
Fig. 8 of U.S.P. 4jl88,797 (Thijss~n). When the apparatus
17 operates as a thickener, some mother liquor of stage 1
will be introduced in stage 4. It has however been found
that generally this does not introduce a major influence
on the efficiency of the opPration.
The liquid to be concentrated enters at 9 and is fed
to stage 1 together with part of the return liquid 10 from
wash column 8. Crystal slurry from sta~e 1 is fed at 11
to wash column 8. At 12 the component sepaxated out is
withdrawn from the wash coll~mn 8, while at 13 the liquid
separated from the crystals is partly returned to stage 1
at 10 and partly, at 14, fed to stage 2. A slurry of
crystals is recirculated at 15 from stage 1. It is thicken-
ed in 17 and led to stage 4, while the mother liquor isrecirculated to stage 1. For the thickening process see
U.S.P. 4,188,797 (Thijssen). The crystals and mother liquor
- 12
are led countercurrently through the various stages, as will
be clear from the figure and as descrihed in detail in the
application m~ntioned above. Mother li~uor with the lowest
content of the crystallizable component is withdrawn at 16.
Further details or the embodiment depicted in Fig. 1 are
given in Example 1 (see especially table 1, in which all the
different streams of material are described).
Referring now to Figure 2, (101) to (106) are crystal-
lizer compa~tments as described in U.S.P. 4,188,797 (Thijssen)
while (107) is a thickener as described in the same patent.
At (108) a wasih column is shown of the constxuction as de-
scribed in British Patent Application No. 79,21808 (Published
British Patent Application No. 2023564A) (Thijssen). As de-
scribed in detail in U.S.P. 4,188,797 (Thijssen) and also
above when discussing Fig. 1, the slurry from each of the
compartments is led to a thickener, the thickened slurry is
passed on to the preceding compartment (of higher temperature),
while the liquid leaving the thic~ener is partly recirculated
to the same compartment and for the rest to the following
compartment (of lower temperature), or finally reco~ered
from the last compaxtment. For reasons of clari~y and sim-
plicity in this case only the general direction of flow of
the liquid and the crystals are shown, the direction of the
liquid being indicated with dotted lines, the direction of
the crystalis with drawn lines.
The feed enters at (109) in crystallizing compartment
(102) and the liquid passes on to (103~ up to (106). Mother
liquid is recovered at (110).
~5~1~
- 13 -
The compartment (101) functions as a refining section,
by a prccess of recrystallization. The crystals arriving in
this compartment from compartment (102) are here brought in
contact with a melt of the one component which has to be
separated out of the multi-component liquid system. The wash
column (108) separates the crystals from their mother liquor.
Part of the crystals after melting are returned to the com-
partment (101), while the rest is recovered at (111). A sus-
pension of crystals (112) is withdrawn from compartment (102)
and fed to the thickener (107). The liquid (113) leaving the
thickener is returned to the compartment (102), while the
thickened slurry of crystals (114) is passed on to the com-
partment (106) of lowest temperature. Of course more than one
refining compartment may be necessary, depending on the
specific composition of the li~uid to be treated and the
purity of the crystallized component desired. Likewise the
number of crystallizer compartmen~s depends on the specific
case at hand.
The present invention will now be fuxther illustrated
by way of the following Examples:-
Example 1
Red wine was treated in an apparatus and according tothe process described above with reference to Fig. 1. Use was
made of a crystallizer with four compartments (1) - (4) with
~5 the ~mpn~ions which were mentioned in U.S.P. 4,188~797
(Thijssen~. A wash colu~n (8) was used as described in Bxitish
Patent Application 7g,21808 (Published British Patent Appli-
95g3~3
~ 14 -
cation No. ~023564A) (Thijssen) while the crystals were re-
circulated as a thickened slurry ob-tained in a thickener (17),
of the construction as described in U.S.P. 4,188,797
(Thijssen). The wine was threefold concentrated. Referring to
Figure 1 the streams of material were as indicated in table 1.
In a reference experiment the process was run in the
same way as described above:, but without recirculation of
crystals. In this case the streams of material were indicated
in table 2.
From these data it is cleax that only about 2.5% of
the ice crystals finally fed out from the apparatus were re-
circulated. The average size of the ice crystals being
treated in the wash column increased by this small percentage
of recirculation two fold, from 70 to 140 micrometer. This
resulted in a decrease of the pressure drop in the wash
column from 7 atm. to 1.5 atm.
TARLE 1 TABLE 2
Streams of m~terial with Streams o~ m~r;~l without
2.5~ recirc~llation of crystals recir~ll~tinn of crystals
Nr.of consistency kg/hour Nr.of consistencykg~hour
5 stream stream
9 liquid 75.00 9 liquid 75O00
12 melted ice 50.00 12 melted ice 50.00
16 liquid 25.00 16 liquid 25.00
11 slurry 200.00 11 slurry 200.00
12 melted ice 50.00 12 melted ice 50,00
13 liquid 150.00 13 liquid 150.00
liquid 56.60 10 liquid 56.60
14 li~uid 93.40 14 liquid 93.40
slurry 157.00 20 sluxry 152.00
31 thick.slurry70.65 31 thick.slurry68.40
21 liquid 86.35 21 liguid 83.60
22 liquid 16.35 22 liguid 13.60
23 liquid 70.00 23 liquid 70.00
24 slurry 105.00 24 slurry 100.00
32 thick.51urry47.25 32 thick.slurry45.00
liquid 57.75 25 liquid 55.00
26 liquid 11.15 26 liquid 8.40
27 li~uid 46.60 27 liquid 46.60
28 slurry 53.00 28 slurry 48.00
33 ~lick.slurry23.85 33 thick.slurry21.60
29 liquid 29.15 29 liquid 26.40
liquid 4.15 30 liquid 1.40
16 liquid 25.00 16 liquid 25.00
slurry 5.00
18 thick.slurry2.25
19 liquid 2.75
~S~
Example 2
The process as generally discussed in Fig. 2 to which
we refer here will now be described as applied to the sepa-
~ ration of p-xylene from a mixture containing it and other
isometric xylenes.
Use is made of a crystallizer with six compartments
(101) to (106) with the dimensions which were mentioned in
U.S.P. 4,188,797 (Thijssen). A wash column (108) is used as
described in British Patent Application No. 79,21808
(Published British Patent Application No. 2023564A) (Thijssen),
while crystals are recirculated as a thickened slurry ob~
tained in a thickener (107) of the construction as described
in U.S.P. 4,188,797. Throughout the apparatus there is a
temperature profile from about -11C in compartment (101) to
-65C in compartment (106). The purified p-xylene is recovered
at (111) as a melt while the rest is returned to compartment
(101). The ratio of recovexed to refluxed melt of p-xylene
is ~:1. The temperature of the melt is about -~15C (the
melting point of pure p-xylene is ~13,26C).
In table 3 the material balance around the total
apparatus i~ given, with a recirculation ratio of crystals
of 20~, which goes to say that 80% of the crystals present
in compartment (102) are passed on to compartment (101) and
20~ is recirculated to compartment (106). In table 3 also
the composition is given of the p-xylene ~ed out at (lll)o
In table 4 the same data are given in the case ~hat no
recirculation of crystals is applied. In both cases the same
- 17 -
pressure drop in the wash column is maintained. Comparing
table 3 and 4 it is seen that the product purity is increased
from 99.5% to 99.9% while at the same time the throughout of
the apparatus by the recirculation is increased with 40%.
T~BLE 3
M;lt~l-; ~1 }~1 ~n~e around tot~l d~ald~U~; with
20~ recir~~ n of crystals
F~ed cryst~l 1; 7.~hl P cn r~nnPnt- m~ther liqu~r
separated out
kg/hour % kg/hour % kg/h~ur %
Ethyl benzene 18.18 19.55 0.001 0.01 18.1822.63
para xylene 19.34 20.80 12.669 99.92 6.67 8.31
methaxylene 43.68 46.96 0.007 0.55 43.6754.37
otho xylene 11.72 12.60 0.002 0002 11.7214.59
others 00.08 00.09 0.000 0.00 0.08 0.10
Total 93.00 100.00 12.679 100.00 80.32100.00
J
~ELE 4
r~t~r;~l h~l~n~e around total d~LCUd~US without
20% recLrclllAti~n of c~ystals
Feed cryst~ Ahl~ cnmr~n~nt ~other liquor
S~ydld~ed out
kg~hour % kg/hour -~ kg/hour %
Et~yl ~enzene 12.89 19.54 0.01 0.12 12.89 22.62
para xylene 13~73 20.80 8.99 99.50 4.74 8.32
meta xylene 31.00 46.97 0.03 0.29 30.97 54.37
oth~ xylene 8.32 12.60 0.01 0.09 8.31 14.59
others 0.0~ 0.. 09 0.00 0.00 0.06 ~.lQ
Total 66.00 100.00 9.04 100.00 56.97 100.00
