Volume 32 OCTOBER, 1955 No.

10

Continuous Purification and Acidulation of Vegetable

Oil Soapstock
FREDERICK W. KEITH JR., VINCENT G. BELL, FRED H. SMITH, The Sharpies Corporation,
Philadelphia, Pennsylvania

OAPSTOCK from vegetable oil refining usually pre- phase was clean and brown in color. In a commercia]
s sents either a disposal problem for the refiner or
a processing problem for the user, and frequently
installation of th e process it became refinery practice
to call these treated f a t t y acids Brown Acids to dis-
tinguish them from the Black Acids made by stand-
both conditions prevail. A few years ago an excess of
supply over demand permitted the buyer to choose ard acidulation of soapstock, this terminology will be
his sources on the basis of quality. More recently, used here for convenience.
shortages have favored the refiner with a higher mar-
ket price, but they have also increased the buyer's Process F l o w Sheet
problem of quality control, particularly in the pro- A full scale continuous plant incorporating three
duction of f a t t y acids. stages was installed at the Opelousas Oil Refinery in
Several years ago work was started on methods Of Opelousas, La. This unit was operated on an experi-
p u r i f y i n g soapstock; results of laboratory and pilot mental basis while testing different types of mixing
plant studies have been reported previously (1). It equipment, pumps, and control methods to complete
was found that the nature of the impurities in vege- the scale-up from pilot plant for soapstock purifica-
table oil loots rendered them resistant to treatment tion and from laboratory data for the acidulation.
by acids as in normal acidulation. Reaction with a The final process is outlined in F i g u r e 1. Soapstock
small excess of caustic above that theoretically needed
for saponification could, with prolonged boiling, pro-
duce a fully saponified soap of higher T F A than the
loots but with erratic and often limited reduction in
impurities. Strong caustic however attacked the gums
and made them at least partially soluble in 5 to 10%
aqueous caustic solution. At this concentration a soap
phase of 40-55% T F A could be centrifugally sepa-
rated, and the majority of the impurities or their
degraded products discharged with the lye stream.
Oxidized Acids, as determined by A.O.C.S. Tenta-
tive Method G3-53, and other more or less insoluble
impurities could be reduced to 5% or less in Black
Acids produced from the soap. L a b o r a t o r y distil-
lations showed marked reduction in still losses and
greatly increased color stability for these purified
f a t t y acids without the necessity of Twitchell or
pressure-splitting.
Batch acidulation of the purified soap was not al- FIG. 1. Continuous purification and acidulation of soapstock.
ways readily accomplished, and some of the usual
plant acidulation problems were still present. The
f a t t y acids, for example, were discolored b y contact is withdrawn from the refinery receiving tank and
with strong mineral acid for extended periods, and metered to the system by a variable speed gear pump.
with some stocks a fairly long boil was necessary to Immediately before the mixer 50% caustic and hot
complete acidulation. Also, with most soaps, break- water, regulated by rotameter, are added to the line.
ing of the interracial emulsion or curd was still slow. The mixer is compartmentized and agitated; the bay-
The laboratory study was therefore extended to in- onet-type of steam heaters maintain temperature with
clude methods of continuous acidulation of the puri- a controller on the out-going mix. The hot mix passes
fied soap. The order and rate of addition of the downward through a hold-up tank and is separated
reactants had no consistent effect on the rate of re- into soap and pitch water in the 1st stage centri-
action. It was found however that proper dilution fuge. The soap, usually at 35-50% T F A , is diluted
of the mineral acid plus addition of a small amount in the covers of the centrifuge with hot water and
of one of several commercial dispersing agents caused drops into an open tank where it is grained b y 20%
a fast and complete acidulation reaction with rapid brine added through a rotameter. Agitation is ac-
breaking of the emulsion. The resulting f a t t y acid complished by an over-size gear pump and a high
517
5]8
rate of recycle. A regulator in the recycle line adds
live steam to maintain a t e m p e r a t u r e of 170-185~
A weight controller on the t a n k diverts feed to the
2nd stage centrifuge, where soap and pitch water are
again separated.
The purified soap at 50-62% T F A is diluted with
hot water in the covers of the 2rid stage centrifuge
and drops directly into an agitated open t a n k with
steam coils. A f t e r heating to about 180~ a n d blend-
ing with 2.5% solution of a dispersing agent, which
is added b y rotameter, the mix is fed b y a variable
speed gear p u m p to the bottom of the acidulation
tower. The 66 ~ B& sulfuric acid is diluted to about
half strength b y hot water with rotameter regulation
before being p u m p e d to the bottom of the tower.
Generally no heat is needed in the tower, but it can
be supplied b y live steam if required. A Monel metal
or plastic-lined tower is necessary with alternating
disc and donut Monel baffles spaced every 6 in. The
overflow f r o m the tower is fed directly to the 3rd
stage centrifuge, where it is separated into Brown
Acids and spent mineral acid. The latter partially
neutralizes the 1st stage pitch water in a skim t a n k
before discharge to sewer.
D u r i n g scale-up f r o m the pilot plant two i m p o r t a n t
changes were made in the 1st stage. Originally soap-
stock was blended with water and heated in an open
t a n k before addition to the mixer; changes in soap-
stock viscosity required compensating changes in wa-
ter addition. W i t h the shift to closed feed the s:~stem
became relatively insensitive to viscosity changes.
The hold-up t a n k a f t e r the ]st-stage mixer was a
necessary addition to the process. I t was a vertical
t a n k about 29 in. in diameter a n d 12 ft. high with no
baffles, heating, or agitation. Some soapstocks were
encountered which a p p e a r e d to be grained in the
mixer alone but which had so strong a tendency to-
w a r d emulsification that a clean centrifugal separa-
tion could not be obtained. W h e n these inadequately
treated stocks were passed t h r o u g h a quiescent hold-
u p tank, a stable grain resulted probably because of
the increased reaction time a n d f u r t h e r degradation
of the emulsifying impurities. Agitation of the hold-
up t a n k decreased its effectiveness considerably. E v e n
stocks that could be handled directly f r o m the mixer
benefited by the additional hold-up time. I n most
cases caustic r e q u i r e m e n t s were reduced, a n d the
quality of the soap was improved. A n outstanding
example is shown in Table I for stock f r o m soda
TABLE
E f f e c t of H o l d u p Tank
Holdup
, ~ I Tank ] TFA
~ampm ~ource Time
_ _ _ (rain.) _ _
Soapstoek F e e d ................ [ .... I 15.9
Condition A ..................... [ 0 ]
Condition
Condition
B ..................... I
C .....................
26
29
r 31.9
C o n d i t i o n D ..................... I 0 I
a CA = Oxidized Acids, BA : T F A -}- O x i d i z e d A c i d s .
ash-refined hydraulic cottonseed oil. Salt was used
in this r u n as an auxiliary graining agent b u t was
held constant d u r i n g the run. The results were simi-
lar in other r u n s without salt. I n Condition A of
Table I the hold-up t a n k was not used w h i l e in Con-
dition B the same mix was passed through the hold-up
THE JOURNAL OF THE AMERICAN OIL CHEMISTS' SOCIETY VOL. 32
tank. Condition C represents a reduction in N a O H ,
using the hold-up tank, while D shows that the same
mix as C without the hold-up t a n k was insufficiently
grained to give a split in the centrifuge.
Process Conditions
The effects of the process variables in the 1st stage
of the full scale p l a n t bore out the conclusions reached
in the pilot p l a n t although the sensitivity to changes
in soapstock feed or process conditions was very nmch
reduced by introduction of the hold-up tank. I n gen-
eral, ]st-stage t e m p e r a t u r e s were kept at 200~176
to maintain a high soap T F A . Low ratios of soap
phase to lye phase gave slightly better quality soap,
but high ratios gave better economy. A compromise
ratio of about 1.0-1.2 lbs. of soap produced per pound
of pitch water was set as the nominal operating range.
Operation with the m i n i m u m caustic concentration
necessary for graining and c e n t r i f u g a l separation
tended to decrease T F A and to improve slightly the
removal of impurities. A n y caustic addition higher
t h a n the m i n i m u m produced satisfactory soap with
little effect f r o m changing the lye concentration. Gen-
TABLE II
E f f e c t of 3 r d Stage Spent Acid Concentration
Conc. of H~SO4 i n
Spent Acid (%)
] Brown Acids
Acid_ V a l u e
Distilled Acids
Still Loss 24 h r . C o l o r
(%) (Gardner)
190 24.8 3 D
190 26.0 3 L
:::::::::::::::::::::::::::::::::: 188
188
26.4
28.0
3
5
erally the concentration of caustic in the ]st-stage
pitch water was 7 - ] 2 % with total solids at 18-32%.
P a r t i a l replacement of lst-stage caustic with another
graining agent, such as sodium chloride, reduced the
soap quality. T F A in the pitch w a t e r averaged 0.1%
for expeller or extracted oil soapstocks and 0.3% for
hydraulic stocks with little effect f r o m the t y p e of
refining. Material balances on the NaOH, allowing
for saponification of the n e u t r a l oil content of the
soapstock, indicated that r o u g h l y a quarter of the
caustic might be used in reaction with the impurities.
The 2nd, or washing, stage was relatively insensi-
tive to process variables. I t was necessary to mMntain
a t e m p e r a t u r e of 170~176 as lower t e m p e r a t u r e s
produced f o a m y emulsions instead of clean grains.
The s o a p / p i t c h w a t e r ratio was in the range of 1 to
I 2 lbs./lb, with NaC1 in the pitch water at 6-8%, total
on First Stage Soap solids at 15-20%, and T F A averaging 0.2%.
The variables of the 3rd, or acidulation, stage were
[ Concentrations in
IOA/BAal Pitch Water (%) most interesting as the effect on B r o w n Acid quality
(% ) (%) was quite marked. The result of v a r y i n g the mineral
_ _ NaOH NaC1
acid concentration is shown in Table II.
[ 7.6 I .... ] ....
21.8 I 7.4 I 9.6 I 6.3 I t is seen that still losses decrease and acid values
33.5
I 5.9
15.8]
] 10.1 I 6.2
6.__8__] 6.6
increase as the acid concentration increases, probably
( G r a i n too closed to s p l i t )
due to more complete reaction and better removal of
residual impurities. Color is high at low acid concen-
tration where reaction m a y be incomplete. A f t e r pass-
ing through a m i n i m u m at 3 - 4 % ItfSQ, color tends
to increase for higher acid c o n c e n t r a t i o n s , where
" b u r n i n g " m a y occur. P l a n t operation at about 4 %
H f S O 4 in spent acid is v e r y satisfactory.
The acidulation t e m p e r a t u r e has considerable effect,
as shown in Table I I I , for a set of runs at about 4.5%
spent acid and 10-min. hold-up time in the acidula-
 OCT. 1955 KE1TH ET AL.: CONTINUOUS I:)URIFICATION 519

TABLE III lb. per rain. of soapstock, but these differences were
Effect of A c i d u l a t i o n T e m p e r a t u r e not apparent in the Brown Acids.
D i s t i l l e d Acids Analytical data arc available for t h e 23 tankcars
Acidulation B r o w n Acids
Temperature Acid V a l u e Still Loss 24 hr. Color of B r o w n Acids produced to date, and these are sum-
(~ (%) (Gardner) marized in Table VI. L a b o r a t o r y distillation losses as
170 .............................. 190 25.2 2 determined b y the b u y e r are available for a n u m b e r
180 ............................... 188 23.8 3 of cars and these range f r o m 14-22%.
193 ............................... 188 23.3 3
200 ............................... 23.4 3 D
215 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.0 4
T A B L E VI
S u m n l a r y of Opelousas B r o w n Acid P r o d u c t i o n
tion mixer. At 170 ~ the increased still loss shows in-
complete acidulation although the high acid value a n d Range Average

good color indicate t h a t the cleaner acids in the mix B r o w n Acids :

Moisture, % .................................. 1.3-- 4.6 3.0
m a y react more rapidly. At 180 ~ and 193 ~ the still TFA, .% ......................................... 91.0-- 94.3 92.9
losses are low and color is good, b u t at 200 ~ and above U n s a p o n i f i a b l e s , % ....................... 1 . 7 - 4.5 3.2
Acid V a l u e .................................... 184 - 1 9 1 188
color is adversely affected. At 215 ~ polymerization I o d i n e V a l u e ................................. 99.6--105.0 102.2
L a b o r a t o r y D i s t i l l e d Acids :
increases the still losses. Still Residue, % ............................ 1 8 . 0 - 24.0 21.0
G a r d n e r Color ............... : ............... 1 D-2 2L
Residence time in the acidulation mixer was also 24-hr. G a r d n e r Color .................... 1D-5L 2
tested over the range of 9 to 15 rain. as shown in
Table I V ; still losses and color both decreased some-
The data in Table V I on distilled acid colors were
obtained on a middle cut amounting to 50% of the
TABLE IV Brown Acids still charge. The 24-hr. color reversion
Effect of Residence T i m e i n A c i d u l a t i o n Mixer on this cut at 95oF. was about one-third of a G a r d n e r
D i s t i l l e d Acids
color unit so that the average darkening for produc-
R e s i d e n c e Time B r o w n Acids tion was f r o m 2L to 2. On recent production an accel-
(min.) Acid V a l u e Still Loss 24 hr. Color
(%) (Gardner) erated color stability test was also made b y holding
14.5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 21.8 3
a sample at 250~ for 30 rain. These results are
10.8 ............................. 188 23.8 3 available f o r five tankears as shown in Table V I I .
9.2 .............................. 188 24.6 3 D

T A B L E VII
what with increased hold-up time. These results to-
R e s u l t s of Accelerated Color S t a b i l i t y Test
gether with the t e m p e r a t u r e data indicate t h a t the G a r d n e r Color of 5 0 % Color Cut
best operating condition for the acidulation mixer is
24 hr. @ A f t e r 30 rain,
170~ ~ at 15 rain. hold-up time with satisfactory Fresh 95OF. @ 250oc.
operation up to 195 ~ at 10 rain. 2L ............................................... 2L 4L
I t should be noted t h a t figures for still losses are 2L ............................................... 2 4D
2L ............................................... 2L 5
based on l a b o r a t o r y distillations. Distillation results 1D ............................................... 1D 6
f r o m different laboratories, using the same still a n d 2L ............................................... 2 5L

a p p a r e n t l y the same procedure, have differed appre-

ciably although the data f r o m a n y one laboratory Summary
have usually been consistent. F o r example, the aver- I n s u m m a r y , the results f r o m full-scale continuous
age still loss on the first five tankcars as determined production indicate t h a t all types of cottonseed soap-
by the b u y e r was 17.1% while the average f r o m pro- stocks are sufficiently u p g r a d e d b y t r e a t m e n t with
duction control tests was 20.0%, about 3% higher. strong caustic and graining and washing to permit
F o r consistency as well as availability, production subsequent acidulation. A t 175~176 in the pres-
figures are used throughout. ence of a dispersing agent with spent mineral acid
Reagent consumption is always an i m p o r t a n t eco- at about 4 % H2S04, acidulation of purified soapstock
nomic consideration. D a t a a r e available f r o m spot yields a clean, brown, fast-breaking f a t t y acid phase
checks twice a d a y d u r i n g the first six weeks of regu- with acid value about 188, uusaponifiables about 3%,
lar production following the experimental operation. and an Oxidized Acids content of 3 - 5 % . I n labora-
The a v e r a g e results are given in Table V in pounds of t o r y distillations, still residues are usually in the
range of 18-23% with a 24-hr. G a r d n e r color of
TABLE V about 2 on the color cut.
Reagent Consumption in Production Runs

Usage in
Acknowledgment
Process Stage Reagent lb./lb, of
Reagent 1
Conc. B r o w n Acids The authors wish to express their appreciation to
1 ..................................... 100% 0.31
the m a n a g e m e n t of 0pelousas 0 i l Refinery a n d par-
2 ..................................... NaC1 / 100% 0.06 ticularly to E. D. Gile, production manager, for their
3 ..................................... H2S04 / 66 ~ B6. 0.32
3 ..................................... Dispersing Agent t 100% 0.0024 v e r y fine cooperation and constructive interest dur-
ing the experimental stages of p l a n t operation, a n d
reagent per pound of B r o w n Acids produced. Of the to Wales Newby, chief chemist, for making available
four reagents, caustic was the only one t h a t varied the analytical data on production.
appreciably with t y p e of soapstock. REFERENCE
The unit operated satisfactorily for Brown Acid 1. Keith, F. ~V. Jr., Blachly, F. E., a n d Sadler, F. S., J. Am. Oil
rates r a n g i n g f r o m 6 to 16 lb./rain. The 1st stage Chemists' See., 3 1 , 2 9 8 - 3 0 2 ( 1 9 5 4 ) .
showed a decrease in soap quality at rates over 45-50 [Received February 16, 1 9 5 5 ]