Note: Descriptions are shown in the official language in which they were submitted.
108g849
This inv~ntion is dir~ctel to processing rapeseed
or mustard (Brassica species) into oil, flour and hulls with
low loss of oiltand of protein from the flour.
Rapeseed is the most important oilseed crop in
Canada, and contains about 44-45% of an edible ~Jegetable
oil (on a dry basis). The residual meal (seed meats plus
hulls) contains about 40% protein (dry basis). The nutri-
tional value of this protein (protein efficiency ratio) is
good and it is desirable to recover as much of the oil and
protein (in a hull-free form) as possible.
Yellow, oriental and brown mustard are also grown
in Canada, primarily in Saskatchewan, and annual production
is about 100,000 tons. Oil contents of Oriental and brown
types (Brassica juncea) are comparable to rapeseed but the
hull contents are lower and protein levels in the flour are
higher than in rapeseed. The large-seeded yellow mustard
(B. hirta) contains less oil and hull but is exceptionally
high in ~lour yield and protein content in the flour. The
color of protein products from yellow mustard is substan-
tially better than those of other Brassica species. There
is little processing of mustard in Canada; most of the seed
is exported to the United States, Europe and Japan.
Because of their small seed size, rapeseed and
mustard contain a high proportion of hulls that adhere
closely to the cotyledons or meats. Seed sizes among
Brassic3 species ranged from about 2 to 8 mg wt. and hull
contents varied inversely from 23 to 18%. Crude fiber
levels of 12 to 13% in other defatted rapeseed meals or
the values of i4 to 15% commonly reported for commercial
30 ~ meals, are too high for food applications and for most non-
ruminant animal feeds. There is only a limited market for
; high fiber meals in animal nutrition in Canada and profit
:
3~ ~ `'- '`'
~0~9849
margins are small when meal is shippe~ o~er long distances
because the product is bulky relative .o unit value. Gene-
rally, rapeseed meal cannot compete with soybean meal in
monogastric animal feeds because of high fiber levels, the
presence of glucosinolates and low protein levels. The
new double zero varieties will reduce the hazards associa-
ted with the ingestion of glucosinolates but the high pro-
portions of hulls will still lower digestible energy and
protein levels. In the ruminant feed market, rapeseed meal
must compete with inexpensive urea as a source of nitrogen
for rumen bacteria. Therefore, prices for rapeseed meal
are generally depressed relative to soybean meal and other
seed by-products.
Conventional oilseed milling practices for rape-
seed do not permit the separation of hulls sufficiently
to produce a low ~iber;flour for food use or dehulled meal
for non-ruminant nutrition such as in poultry feeds or pet
foods. Processes for dehulling rapeseed and the production
of rapeseed flour have been developed and described in the
literature including patents. The principal proposals are
for dr~ dehulling before oil extraction, wet dehullin~
(aqueous) using rolls and air-aspiration after diffusion ex-
traction process, or air classification of the defatted
meal.
Efficient dehulling rates have been obtained for
"front-end" dehulling by dry and wet procedures. ~he pre-
vious wet dehulling procedure yielded an average of 21%
hulls, and 77~ meats, which was close to the the~retical
~ields of 18% and 80%, respectively, obtained by hand
separation (Table 1~. Unfortunately, the hulls contained
about 24~ oil; due partly to contamination with small par-
ticles of meats. This quantity of oil represented about
--2--
3~
1~G of th~ ori~incll seed oil, a prollibitive loss if the
hulls are not deEatted but rathcr sold directly for live- -
stock feed.
TABLE 1
Yield and composition of products obtained by wet dehulling
~aqueous) using rolls and air aspiration.
Product Yield, % Oil, % Protein content after
defatting, %
. ~
Hulls 18-23 24 28
Meats 76-79 54 55
10 Flour 30-36 1 54-60
Dry dehulling of rapeseed before oil extraction is ~;
slightly less efficient than the above wet dehulling pro-
cess. Dehulling before oil extraction would not be economi-
cal because of oil losses in the hulls and fines which con-
; taminate the hull-rich fraction.
; Pin-milling and air classification of rapeseed
meal, after desolventization, into protein-rich and hull-
rich fractions has been proposed. However, the degree of
hull separation and the decrease of protein loss to the
hull fraction, leave room for improvement. Air classifi-
cations of desolventized and dry milled rapeseed meal after
oil extraction has proven to be effective in increasing
the protein level by 6% and decreasing crude fiber to 7-10~
in the fines or protein-rich fraction. The principal draw-
backs of this dry process are (a) the need for double or
triple milling and air classification to obtain a reasonable
yield of protein-rich fraction, (h) the very fine material
; in both frac~ions is difficult to incorporate into feeds
and (c) the protein-rich fraction is dark in appearance
and its use may be limited to animal feeds (d) the protein
enricllmenl may he too small relative to the cost of the pro-
cess and the competing soybean meal products.
This invention is for the continuous wet processing
of fully or partially defatted rapeseed and mustard marc
(meal-solvent mixture) or meals into a dehulled meal or
edible-grade flour and a hull by-product fraction. The
process is designed for separation of the flour and hulls
from rapeseed or mustard meal or marc immediately after the
complete or partial oil extraction of the seed has been
completed. The dehulled meal or flour is essentially free
of fibrous, pigmented hulls and contains only 5~ of crude
fiber with protein levels of 44 to 54%, depending on culti-
var and process details. The rapeseed or mustard flour
could also be modified during wet processing to improve its
functionality (color, flavor, solubility, texture) in food ~-
uses. The hulls fraction contains over 20~ of protein and
20% of crude fiber, and would have an immediate market in
ruminant feeds. Because flour contamination in the hulls
fraction is quite low, the relatively pure hulls may have
industrial applications such as in adhesives, plastics etc.
In the accompanying flow charts:
Chart ~1 is a flowsheet illustrating a modified
"all-solvent" extraction incorporating wet milling and
liquid cyclone fractionation after oil extraction and marc
filtration.
Chart #2 is a flowsheet illustrating a modified
"all-solvent" extraction using dry fine grinding before oil
extraction and liquid cyclone or decanter centrifuge frac-
tionation in place of filtration.
Chart #3 is a flowsheet of a modified "prepress
plus solvent" extraction incorporating wet milling and liquid
fractionation after oil extraction.
1~3~ 49
Chart #4 is a flo~sheet of a modified "prepress
plus solvent" extraction using dry fine grinding before oil
extraction and liquid cyclone or decanter centrifuge frac-
tionation after solvent extraction.
Chart #~ is a flowsheet illustrating a modifi~d
"prepress plus solvent" extraction using dry fine grinding
and liquid cyclone or decanter centrifuge fractionation in
place ol the normal solvent extraction step.
In the accompanying drawing:
Figure 1 is a schematic flow diagram of one type
of oil extraction and liquid cyclone plus decanter centri-
fuge fractionation illustrating the types of equipment used.
According to this invention, we provide a method
of fractionating rapeseed and mustard seed with high reco~
~eries of oil and protein-rich flour comprising
(a) crushing the seed and removing oil therefrom
to give a meal containing flour and hulls,
(b) providing a suspension of the meal in a non-
aqueous solvent which is a non-solvent for protein, the --
solids content being within about 5 to about 33% by wt.,
(c) providing that the moisture content of this meal
in the suspension is below about 10% by wt., and the parti-
cle size is fine enough to pass a 150 mesh screen (Tyler),
(d) subjecting this meal suspension to a centri-
fugal or gravity liquid separation giving a flour slurry as
overflow and a hull slurry,
and (e) separating solvent from each slurry and reco-
vering oil, flour and hull solids as separate products.
The process is particularly applicable to defatted
meal after all-solvent extraction of oil or prepress plus
solvent extraction of oil from the seed. In these processes,
the extracting solvent is pre~erably used as the medium for
the liquid cyclone fractionation. The process can also be
applied to partially defatted meal followin~ mechanical
(h~draulic, screw or expeller) pressing of the ground seed
in which case the cyclone or centrifuge serves as the second
stage of oil extraction.
The pxocess involves two basic steps (a) Defatted
or partially defatted rapeseed or mustard meal or marc are
disintegrated and fine ground in a colloidal, stone, pin or
other type of grinding mill sufficiently fine to pass a 150-
mesh (Tyler) screen. The grinding may ~e applied to rela-
tively dry meal after the prepress or expeller stage of oil
extraction or in a non-aqueous, inert fluid medium. The
grinding may be applied most conveniently to the solvent-
soaked flakes or marc immediately after oil extraction.
(b) Liquid fractionation of the hull and flour in the fluid
medium using liquid cyclones, liquid or sludge centrifuga-
tion or other gravity separation devices. The principle of
cyclone or e~uivalent dehulling of rapeseed or mustard marc
~meal plus solvent mixture) is based on differential
sedimentation of rapeseed flour (defatted cotyledonary material) ~.
and hulls. Differential sedimentation velocity can occur with
partially or fully defatted meal which is ground to separate
the adhering flour and hull particles and to provide a
differential granulation of flour and hulls. Generally the
desired granulation properties occur when the meal is ground
to less than about 100 microns diameter (< 150 mesh, Tyler).
Partially defatted meal containing no solvent can be fine ground
on a variety of pin, disc and impact mills. Wet milling of
marc requires the use of special grinders such as the explo-
sion-proof colloid mills or stone mills.
;30 With the particle size, suspension solids content,
and moisture c~ntrolled as specified, we find that the hulls
separate so rapidly and the flour so slowly (in a mixing vessel)
-6-
that a Aecantation system would be operative for settling
the hulls and removing flour slurry as overflow.
Solvents or inert liquids are selected from non-
aqueous liquids ~hich do not dissolve the rapeseed or mustard
proteins and which preferably are solvents for the residual
oil (in order to obtain high oil recoveries). Suitable
solvents include hydrocarbon liquids such as pentane, hexane,
octane, decane or highly refined petroleum fractions, alcohols
such as methanol, ethanol, isopropyl alcohol, butanol, ~-
benzene, liquid ethers such as diethyl ether, chlorinated
hydrocarbon liquids, e.g. chloroform, methylene chloride,
trichlorotrifluoroethane, carbon tetrachloride, and mixtures
of these. Hexane and hexane-alcohol azeotropes are particularly
suitable with such azeotropes including hexane-methanol,
hexane-ethanol, hexane-2-propanol.
Repeated experiments have shown that the presence
of about 1% wt. or more water in the solvent seriously inter-
feres with liquid cyclone fractionation. This is believed to
be due to water absorption by meal particles, especially
~0 proteins and carbohydrates, leading to particle swelling and
distorted sedimentation patterns of flour and hull. Wet sugars
also can agglomerate the particles. Therefore, dehydrated
solvent, preferably of less than about 0.5~ wt. water, is
desirably used. In the solvent-meal suspansion, the moisture
content of the meal should be below about 10% ~y wt. Water can
give emulsion-like effects that interfere with the separation.
The meal tends to absor~ moisture from the solvents and if the
meal becomes too moist, the desired separation is not achieved.
Solvent containing oil (miscella) is satisfactory for
cyclon~ fractionation but the ability to defat the marc and
remove residual oil is impaired to the extent of oil contami-
nation in the solvent. Fresh, oil-free sol~ent reduces residual
~ '.t~
oil to as low as 0.1 ~o 0.5% in flour and hulls when the
residual oil is as high as 3 to 4% in the marc and would be
the preferred solvent. For prepress or expeller meals con-
taining 15 to 20% oil, the fresh solvent will effectively reduce
the oil level to 3%, which is approxima~ely the level of
efficiency of current commercial extractors. Thus it is possible
to eliminate the normal solvent extraction step in the prepress
plus solvent method of oil extraction by use of the present
liquid cyclone, centrifuge, or decantation technique.
The solids content of the defatted meal suspension
should be controlled within about 5-33%, preferably within about
16 to about 22% by wt. solids for most effective fractionation
at the liquid cyclone or centrifuge stage.
The granulation characteristics after fine grin-
ding are such that special mixing has been found desirable
to blend additional solvent with the marc or meal, and to
maintain a uniform feed of ground marc to the liquid cyclone.
The fine-ground marc or meal and additional solvent are fed
into the mixer and stirred vigorously for complete suspension
~ of all particles. To avoid settling of the denser particles,
a recirculating pump system operating from the bottom of the
mixer has been found to be effective in maintaining the marc
in suspension. The exit or feed from the mixing vessel ope-
rates most effectively if located SUfficiently a~ove the base
of the vessel to avoid clogging with coarse particles. The
size of mixing vessel must ~e sufficient to accomodate the
ratio of solvent to marc indicated plus allow for enough
residence time to defat the residual oil from the marc.
Residual oil levels of prepress plus solvent or all solvent
extracted marcs are usually less than 3% and levels in the
range of 3-5% are reduced to less than 1% during the minimum
time (less than 20 seconds) that is required for mixing and
_~ _
1~ 3~
passage through the pump and cyclone. Longer residence
time in the mixer, pump, cyclone and centrifuge would be
xequired for partially defatted meal obtained from the pre-
press or expeller which contain 15-20% of residual oil.
The minimum wt. ratio of fine-ground rapeseed
meal or marc solids to solvent has been found to be about 1:5.
That is, for prepress rapeseed meal containing no solvent,
5 parts of solvent were required for effective fluidization
or suspension (and soaking of the meal) and cyclone opera-
tion. For marc which contained 1.7 parts of solvent, 3.3
parts of additional solvent were added to provide the de-
sired 1:5 ratio. After filtration, the marc may contain
as little as 40% solvent, and proportionately more solvent
is required.
Separation of the large volumes of miscella (sol-
vent with low concentrations of extracted residual oil) from
the fractionated flour (overflow) and hulls (underflow) can
be accomplished with either centrifuges or vacuum filters.
The fine flour fraction would be difficult to filter readily
and equipment such as a decanter centrifuge would be more
appropriate. Hulls are more coarse and have a more open
structure when layered on the filter, and filtration may -
prove satisfactory. The decanter centrifuge is designed - -
to remove particles as small as about S microns diameter.
Thus, fines in the miscella should not be a p~oblem. The
decanter centrifuge provides solids fractions with as little
as 10% solvent so that miscella removal is quite complete
and the solids can proceed directly to conventional desol- -
ventizers. The capacity of decanter centrifuges is high,
~30 a large model processes 200 gpm, and a 5:1 ratio of solvent
to meal is within the optimal range to obtain a wet cake
- product and high miscella recovery. The decanter centrifug~
_g_
works particularly well with isopropanol or hexane solvents.
Final desolventization of flour and hulls may be
accomplished by normal commercial desolventizers using heat
(steam) and vacuum to drive off the final traces of solvent.
A dual desolventization system designed to handle the rela-
tive yields of each product would be necessary. The standard
desolventizer-cooker, containing a consecutive series of
kettles, would be satisfactory.
The fine texture of the flour and hulls may be a
disadvantage in handling and feeding, e.g., in blending at
feed formulation plants. The solids removed from the decan-
ter centrifuge would dry into agglomerated chunks that may
pass through the desolventizer without complete disintegra-
tion. The presence of fine dust in the finai product may
be further reduced during steam injection by add-back of
gums or by mechanical pressure (pelleting).
Miscellas drained or centrifuged from the flour
and hulls would contain a low proportion of oil. The large
volume of this miscella would be costly to desolventize
directly. It is proposed that this oil be applied at the
first stage of the percolation extraction where fresh sol-
vent is normally added.
The flow charts of Charts 1-5 illustrate the
sequence of steps in oil extraction using either expeller,
expeller plus solvent or all-solvent methods but incorpora-
ting the liquid fractionation of the meal. Techniques for
inactivation of myrosinase by preliminary dry or wet heat
treatments of the ground seed may be incorporated into the
schemes. Steam volatilization of the hydrolyzed glucosi-
nola-tes in the flour (particularly isothiocyanates) are an
alternative to myrosinase enzyme inactivation for mustard
species and certain rapeseed cultivars. Further extraction
--10--
o~ the flour produst after desolventiza~ion with aqueous
and/or organic solvents to remove glucosinolates, sugars,
adverse flavors, phenols and other pigments may be added
to the Charts 1-5 schemes to produce higher grade rapeseed
and mustard concentrates.
Chart 1 illustrates the introduction of the wet
grinding step as the marc leaves the fil.er in an all-solvent
type of oilseed plant. Solvent is added to the fine mate-
rial in the mixer before liquid cyclone fractionation. Dual
systems for removal of solvent from products are based pre-
ferably on decanter centrifugation of flour and hulls. In
each case, the dilute miscella, containing a low proportion
of seed oil, could be re-introduced into the extraction sys-
tem, e.g., on the miscella filter. Further desolventization
of flour and hulls would pass con~ensed solvent and water to
the solvent recovery system which would provide relatively
pure solvent for use in the mixer.
l'here are certain advantages in dry grinding to
achieve the 150 mesh particle size. In addition to reduction
of explosion hazards, a more uniform grind can be achieved.
In Ch~rt 2, the seed material from the cooker is milled,
e.g., in a wide chamber pin mill before passing into the
extractor. Centrifugation must replace the filter, which
could not handle the pin-milled fines, for separation of -
the miscella. Then fresh solvent is added to make up the
5:1 ratio for liquid cycloning as in Chart 1.
In a prepress-plus-solvent plant, the solvent
extractor contains the filtration unit. The marc from the
extractor would ~e wet mill~d to 150 mesh for the cyclone
~0 fractionation process followed by solvent recovery as shown
in Chart 3.
~v~
An improvement on the above process would be to
mill, preferably pin mill the rapeseed meal after passing
through the expeller because this could be a dry operation.
The solvent extractor could then remove the 15-20% oil in the
expeller meal and, after removal of the miscella, the fine
meal blended with more solvent for cycloning into flour and
hulls (Chart 4).
If the expeller reduces the fat level to about
14-15% of the meal, the liquid cyclone could remove the !-
remaining oil to a residual level of about 1-3% that is within
the range of present commercial meals. Chart 5 illustrates
a prepress-plus-cyclone extraction system which includes the -- -
dry grinding operation and eliminates the solvent extraction
step which is currently used by the industry.
Figure 1 illustrates one general layout plan for
a prepress plus solvent and all-solvent extraction systems
incorporating the liquid cyclone second stage extraction
plus fractionation system with decanter centrifuges used on
both flour overflow and hull underflow streams.
In the example of Figure 1, the seed is ground,
coo~ed and then fed through a press 1, grinder 2 and flaking
rolls 3. The flaked seed then passes to solvent extractor 4
where it is extracted with solvent (usually miscella), with the
extract being fed to distillation column 5 and the oil recovered
at 6. The solvent from column 5 is condensed at 7 and sepa- -
rated from water at ~; the solvent then moves to solvent tank
9. The extracted solid~ (meal) from extractor 4 are conveyed
- to fine grinder 10 (<150 mesh) and then fed to mixing tank 11.Optionally, it may be desirable to bypass the press 1 and to
pass the incoming seed direct to rolls 3, extractor 4 and then
~ -12-
~v~
to grinder 10. The finely ground meal (or marc) and solvent
from 9 are mixed at 11 to form a suspension of the desired
solids content. It has been found desirable to recirculate
settled solids into a baffled zone at lla. The suspension
is then fed to the liquid cyclone stage at 12 from which the
hull underflow and flour overflow move to decanter centrifuge
13 and 14 respectively. The miscella from centrifuges 13 and
14 is recycled to the solvent extractor 4. The hulls and
flour are desolventized and recovered at 15 and 16 respectively.
The following examples further illustrate the
invention.
Example 1
The yield and composition of flour and hulls pro-
; ducts were measured after fine grinding and liquid cyclone
fractionation of the marc obtained from a commercial plant `
as in Chart 1. The marc was from high glucosinolate vari-
eties, and contained 40% hexane. The hexane:meal ratio was
made up to 5:1. The grinder, hexane-pumps, and cyclone were
pilot scale or commercial models. Desolventization was done
in explosion-proof fume hoods. The yields of flour and hulls
can only be expres~ed as a ratio in the continuous run, the
ratio being 67.9% flour to 32.1% hulls (Table 2). In addi-
tion, the marc contain 3.1% of oil, and this was reduced
to 0.7 and 0.3~ in the flour and hulls, showing that another '-
`~ 2.5% of the marc was extracted as oil in the miscella. The
theoretical ratio of flour:hulls was 70:30, indicating only
13-
~:
-,
h ~
TABLE 2
Yield and composition of products from liquid cyclone frac-
tionation of marc from a~ all-solvent plant (high glucosino-
late seed)
Product* Yield Protein Fat Fiber Ash
% % 96 % %
. .
Marc 100 39.6 3.1 11.8 6.8
Flour 67.9 46.6 0.7 6.0 7.2
Hulls 32.1 18.7 0.3 31.4 6.4
Oil (2.5) --- --- --- ---
Reported on a solvent-free basis
2~ loss of flour in the hulls. Protein enrichment was 7,0
and fiber reduction 5.8 percentage units. Because the flour
was essentially free of hull particles, the yield and com-
position of the flour fraction represents the best that
could be obtained from this sample of rapeseed.
Example 2
Commarcial rapeseed exhibits a wide range in che-
mical composition. A second liquid cyclone fractionation
of an all-solvent-extracted marc gave this result (Table 3).
The ability of the fine grinding and liquid cycloning in re-
ducing residual oil in the meal can be very marked.
TAB~E 3
;~ Product Yield Moisture Protein Fat Fiber
% 9~ % % %
Flour 66 4.2 43.9 0.1 6.2
Hulls 34 10.1 24.5 0.1 22.4
Oil (3) --- - -
...
.
-~ 11._
3~
Ex~m~le 3
Miscella containing a significant quantity of oil
from the above experiment (Example 2) was used to liquid
cyclone fractionate a prepress-plus-solvent marc which was
quite high in residual oil. The fractionation was success-
ful in recovery of flours and hulls but deCatting was ineffec-
tive (Table 4). It was concluded that fresh solvent, free
TABLE 4
Liquid cyclone fractionation of prepress-plus-solvent marc -~ `
containing a high level of residual oil
:` ''
Products Yield Protein Fat Fiber Ash
% % 96 % %`
Marc 100 41.0 4.9 12.0 7.6 -
Flour 67 45.1 5.0 5.5 6.6
Hulls 33 26.2 3.9 23.4 6.3
of oil and moisture, was essential if defatting was to con-
stitute one of the functions of the process.
Example 4
Tower rapeseed with low glucosinolate level is
being processed at a commercial plant and liquid cyclone
fractionation of th.e marc from this meal gave these re-
sults (Table 5):
TABLE 5
;~. Products Yield Protein Fat Fiber Ash
,~ 96 % 96 96
Marc 100 41.8 4.7 12.8 7.3
` Flour 66 52.5 0.5 4.8 8.1
Hulls 34 24.1 1.0 28.0 6.0
-- ---
-15- ::
10~ 3
Tower cons,isten-tly yave higher proteir leve s in the flour,
as well as low fiber levels. Combined with low glucosino-
late content, the products would represent a significant
gain in nutritive value for animal feeds.
ExamplP 5
Dry meal obtained from the expeller in a prepress-
plus-solvent plant was pin-milled in the laboratory and frac-
tionated by liquid cyclone means. The oil level in this
partially defatted meal was 20%, which is near the maximum
level for expeller meals. The oil extraction efficiency '
was excellent with flour containing 3% and hulls 1% of resi-
dual oil. There was greater flour contamination in the hull
fraction in this run and an improved grinding system over
that of the pin-mill used is recommended. Also higher solvent:
meal ratios may be desirable for this type of fractionation-
extraction.
Example 6
The first control me,al was obtained after filtra-
tion in an all-solvent plant, and was desolventized at room
temperature in the laboratory. The proximate composition
showad less than 40~ protein, over 3% residual oil, 12% crude
fiber and nearly 7% ash (Table 6).
A second sample of meal, obtained from a prepress- '~
and-solvent plant, was desolventized in the same way. It
contained the same oil and ash composition as the first
sample, but protein content was 2.0% lower and crude fiber
3% higher (Table 6)
The desolventized meals were pln-milled and liquid
~ractionated after reblending with hexane in a single pass
through the cyclone. After filtration and desolventization,
-16-
h4~
~ O ~ ~ ~D~` O O O ~ O
u~ ~ co ~ ~ r co ~ ~ o
O ~ . . . . . . .
r~ D
s~
S~ U~ O O O O ~D Lr)a~ c~
I` O~ ~r1` ~ ~ ~1
dP . . . . . . . . .
~ ,1 ~ u~ r ~r co ~ o
X ~ ,, ~ ~ ,, ~ ,,
a~
O ~D ~D O0~ a~ o ~
~ ~ ~ ~D ~ OCO O ~ er ~ ~r
a~ ~ ~P . . . . . . . . . .
~ ~ ~ o o ~ o ~ ,~ o ,~ o
o ~
aJ ~ n o u~ er
dP . . . . . . . . . .
O a~ D1`1` U~
N ~ e~
a
O
.,1 C) O O O O ~ 0
~. . . . .
~ ~1 ~ O ~ D ~ O ~ ~ ~ ~
Q 0 O O ~ ~ ~ O ~ 1
O '~P~ ~ ~ .,
U) ~
u ~
~ O ~ ~ ~ ~ S~
~ ,1 C)
w ~ ~ ~ o ~ o -~ ~ o ~ o ~l
O ~ ~ ~
a) ~ o ~ ~ ~ ~ ~ ~ ~ .c
~ ~ s~
u ~ ~
~ ~ o o
~,~ ~ ~ ~
~l ~ x~
o -l -1
.
:
a
~l ~ ~ ~
u~ ~ _I ~q a) a
o ~ ~ ~ ul ~ ~
P. a~ ~ a~ ~ ~
~ ~l ~ ~ ~ ~ o o
uo ~ ~ ~ ~ ~
0 O S~ h + + U) ::
.q Iq G) Q~ Q. ~
~> ~ U~ U~ U?
_1 ~ ~ cn o~ 0
:; O O O O O Q~ R
X ~ ~ ~ ~
O ~: ~1 ~ ~: O O S~
~i O -I _l O
C~ C~ ~ ~ ~1
--17--
4~
the products were ~eighed, adjusted to 100% yield, and ana-
lysed for proximate constituents.
The meal:solvent ratios used were
1:20 w/v = 6.6 wt. % solids in slurry
1:5 w/v = 22 wt. % solids in slurry
1:3 w/v = 32 wt. ~ solids in slurry
Meal:hexane ratios were varied from 1:20 to 1:3
in both samples, and similar flour:hull fractionations were
obtained (66-34) with proper adjustment of overflow and
underflow fluid rates (Table 7). These yields are based on
weights of final products and not initial sample weights so
that extraction and process losses are not reflected to spe-
cific levels and that even very low hexane:flour ratios will
function effectively in a liquid cyclone. However, the 1:3
ratio can tend to plug the cyclone sporadically if care is
not taken.
; TABLE 7
Yield of flour and hulls in various runs, %l
~.
Sample Meal to Run 1 Run 2 Run 3
hexane
ratio Flour Hulls Flour Hulls Flour Hulls
Rll-solvent
meal 1:20 66.0 34.0 66.0 34.0 -- --
All-solvent
meal 1:5 66.9 33.1 65.1 34.9 66.0 34.0
Prepress +
solvent meal 1:5 67.0 33.0 66.0 34.0 65.5 33.5
Prepress + ~t.
solvent meal 1:3 65.8 34.2 66.5 33.5 66.1 33.
:~ 30 Losses (or extraction) of oil in hexane and flour and hulls
- during handling, cycloning, etc., are ignored.
::
As mentioned earlier, hand-picking of the hull
fraction indicated that the theoretical maximum yield of
-18-
flour was 70%. However, this would vary with seed size, oilcontent and variety, and there was no attempt to estimate
the potential maximum flour recovery in these expe~iments.
While yields were similar, the appearance of the
fractions was quite different in the ratio experiments. The
flour obtained at the 1:20 ratio was extremely pure and
light, showing no hull contamination; the hull fraction was
very dark, indicating almost no presence of flour particles.
For food product applications of the flour, a high ratio
would be preferred. At 1:5 ratio, the separation of flour
and hulls was very satisfactory for both the all-solvent
and prepress plus solvent material but the flour was darker
and some hull specks occurred in the flour, and more flour
appeared in the hull fraction. In the 1:3 experiment, a
lot of hulls were observed in the flour and the flour losses
into the hull fraction appeared correspondingly higher.
These differences in appearance of the products
were reflected in the proximate analysis (Table 6~. The
protein, fat, fiber and ash in the flour and hulls of the
1:20 samples were essentially those of pure flour and hulls
from the all-solvent meal. The same sample fractionated at
the 1:5 ratio with hexane showed less oil extraction, more
protein in the hulls and less fiber in the hulls. There
appeared to be no change in flour composition, which was
unexpected.
In the case of the prepress meal runs at 1:5 ratio,
higher residual fat, more fiber in the flour and high protein
,, :
in the hulls reflected again, the slight change in fractiona-
tion efficiency due to the use of less liquid medium. How-
ever, both 1:5 flours would ~e entirely satisfactory as
animal feed and pet food products.
~ ~ .
-19-
3~ ~
At th~ 1:3 ratio, the flour fraction showed the
influence of hull contamination in its lower protein ar.d
nigh fiber levels, as well as more residual oil. The hulls
retained sufficient flour to show protein levels of 29.2%
and only 25.0~ of crude fiber. The grind for the prepress
sample was somewhat une~en and this may have contributed
to poor performance at the 1:3 ratio.
~ hese data suggest that a 1:5 ratio of meal to
solvent would be a preferred and economic level for liquid
cyclone fractionation. Since the marc would contain from
1-2 parts of hexane, only about 3.5:1 of hexane would be
added fresh to the marc before the cyclone operation.
Example 7 Yields of flour on cyclone separation
Grinding for first four experiments was done with
the blade grinder.
Sample 1. Rapeseed meal was ground -~o less than 420 microns
(>35 mesh). Sieve analysis sho~ed that 30% of
rapeseed meal was less than 250 microns (>60 mesh)
and 70% larger.
Sample 2. Meal ground to <250 microns (19% larger than 177
microns, 33% was 177-150 microns, and 49~ less
than 150 microns (>100 mesh)).
Sample 3. Meal ground to less than 177 microns (80 mesh),
25% was larger than 150 microns, 75~ less (>100 mesh).
Sample 4. Meal ground to less than 150 microns, 52% less
than 55 microns, 48% larger.
Sample 5. Meal ground on pin-mill to obtain a sample in
which 85% was less than 55 micron diameter and
15% greater than 55 micron.
Each sample was fractionated on the cyclone with
the results shown in Table 8.
-20-
4~
TABLE 8
Meal Ground to less than Yield of Yield of hulls
flour, % mixed with flour, %
420 microns No fractionation
250 " 20.4 79.6
177 " 21.6 78.6
150 " 35.0 65.0
" (85~c) 66.0 34.0
The coarse particles gave no liquid cyclone frac-
tionation, but yields of flour progressively increased from
20.4 to 66.0% as particle size was decreased from-250 to-55
microns. The latter yield represented the practical maximum
yield of pure flour that could be obtained without hull con-
tamination. Results clearly show that greater fineness or
granulation of the meal favors increased flour yield. Grin-
ding to less than 55 microns permitted the recovery of 95%
of the flour in the rapeseed meal and a relatively pure
fraction of hulls as well. The flour particle size range
was found to be less than 37 microns (~40 mesh), while the
hulls fraction from underflow was 79~ greater than 44 microns
and the remainder finer in size. The weight ratio was
checked again and found to be 1.6:1, i.e.,
Density
Flour = 0.34 g/ml
- Wt. Ratio of 1.6:1 of hulls to flour
~ulls = 0.55 g/ml
~;~ For commercial-scale practice, acceptable separa-
tions were achieved at <150 mesh (Tyler), with less than
200 mesh (75 microns) preferred.
Example 8
~;
The degree of hull separation was observed during
30~ static sedimentation by gravity in a hexane medium. Rapeseed
, ~
-21~
meal was ground in a pin mill for 30 seconds after which
84~ wt. OL the sample was reduced to less than 55 microns
(99.4~ less than 150 microns). A standard volumetric
cylinder was used for the sedimentation and the hexane
slurry solids content was 33~. Within 0.5 min. the hulls
had completely settled. After l.0 min., 36~ of the sample
had settled but 92% of the flour fraction remained in sus-
pension and only 8~ of the flour had settled with the hulls.
Fine milling, and sedimentation or settling of the
hulls, is thus able to give good separation and recovery of
rapeseed flour.
C~
c~ ~
s~ - ~
c~
o ~ ~, ~ o
s t o ~ o t
vl --I h S~ n ~ h
.~ Q ~ 3
t ~ ., h
-t
51
0 ~ o ~ ~ o
3 .,1 u) h
o ~ ~
G) X
t l
o ~, ~ ~ o
~ ~ ~ o ~ ~,
.. ~ , .. ..
p ~ 6~ ~ o ~ ~ h h
~,1 ,1 ~) h
Q~
~ ~ .
T
~ ~ .. ;
g ~ ~I h
~J ¢ o
~ ct
_1 ~ h
o ~ ~
.;, ;, _~ X
Q, ~ ~
:: 0 h
: v~ o h
~ S~ o o o
U tJc~
~'
23-
;~
c ~ ~ ~ ~
c> ~
o ,~ o
I C~
24 ~
O ~ ~; C -~
C ,, C~4 C C~ V
\ t
3 b
I ~ C ,o ~ :
h ~ -rl
O S: ~l
T 7
~, ,,
, o ~ ....
U~ ,, ~, ',, s~
E IC .,~
n~ C O ~ ~0
--3'' o ~ 3
4-1
~ ~O~
- O S~ O :~
X ~ Q> ,,~ ,
h C~
~ O : ~ O
C~ O ., .~ e ~ ~ ~
~ ~ ~ Ul ~ O
J ~ S-~
--24--
J
<c - '
h
O V~
h ~ ~S
a~ oq ,~, ~
C ~I I . C
O ~~ S
3 ~ 3 -. ~.
gO ~ ~ ~ e ~ ) ~ a ~ a
~d X
Q C: o
h f o
. ~ ,~,
.
a ~ O ~x
> 9
O f~ 0 ~ ~ O
~ a~ h
a~ :~ a~
o~
~ ~6
, ~
o o
~: '
nl
_.
o ~' O ~ o
: "
--2 5--
lV~
t-.
` _
C~
o V~ . .
'~ ~ . . . ~ . .
;~ o
~,
. C~
c o ~ ~ 3 ~ o o
C
;~ ~ 0
c> o ~ h
C~ C ~ C o ~ ~ S.
Y ~ ~ ~ o ~1 D
/
' ~ ~o
3 - ;~ o 0 0
r~ O O
. t~ o ~d ~ ' o ~ O
0 _ ~ 0 ~ o ~
0
P
~1
E~ '~ ~ :
C~ 0 V
~ t2. ~d
..~
--.
~ .
0 t~ h o
~ o ~
J Y C~
`': -
~ .
-
--26--
, . . .
4~
~1
~ ~ .
-
C~ ~o~
o
j A.
U ~ U
O ~ ~rl ~ O e~
u~ h X ~
~) tlS
T ~
,.C o
a ~ C~ h
y ~ ~
' ~
, ~ C,l &
` -27-
