Fat Hydrolysis - Industrial & Engineering Chemistry (ACS Publications)

---

HYDROLYSIS VICTOR MILLS AND H. IC. JICCLAIN The Procter and Gamble Company, Cincinnati 17, Ohio

D a t a are presented on the reaction of tallow and coconut oil with water, in the temperature range used i n continuous fat splitting. The maximum amount of splitting which can occur is determined by the glycerol concentration in the aqueous phase. This value is not affected by changes in temperature. I n the range 90 to 100% completeness, the unsplit fat contains more glycerol than the original fat, a n indication of the presence of a large amount of monoglyceride. The percentage of glycerol in the unsplit fat is constant, which is strong evidence that the reaction proceeds stepwise. The autoclave in which the work was done is suitable for general use. I

Figure 1. Batch Autoclave

HIS paper is a report of the equilibrium conditions which obtain in a system containing fat and water at high temperatures. The results show that the distribution of glycerol in the two-phase system follows a simple uniform rule which explains the difficulties encountered in obtaining complete hydrolysis of fats. The work was done in a batch autoclave with a very simple method of heating which may be of interest in other fields. Fat has been hydrolyzed with water to form fatty acids and glycerol on a commercial scale for many years ( 2 , 4, 6 ) . Originally it was done by prolonged boiling in open vats. Then mineral acids were found to act as catalysts, and much later the Twitchell catalyst was discovered ( 7 ) . These methods produced fatty acids for candles and various chemical uses, but did not produce fatty acids to be used in soap at a cost which could compete with ordinary saponification with caustic. The primary purpose in fat splitting is to cause the following reacbion to go t o completion: Fat

100

98 # #

w 96

z

w

:9 4 -I

n

s 92

0 0

# 90

88

+ water efatty acid + glycerol

The amount of water dissolved in the oil phase varies with temperature. Under normal conditions of hydrolysis at elevated temperatures, the oil phase contains from about 10% water for tallow to about 20% water for coconut oil. At 560" F. all proportions of coconut oil and water form a single phase. All proportions of tallow and water form a single phase at 610.' F. The total amount of water in both the oil phase and the aqueous phase is approximately 60 parts for each 100 parts of oil. Such a, large amount of water present obviously act? as a

0

2

4

6

8

IO

12

14

16

0

2

4

6

8

10

12

14

16

100

98 v)

96

z w

94 -I

0

92

0

$ 90

88 Figure 2. Relation between Free Glycerol in Water Phase and Completeness of

F R E E GLYCERINE

Hydrolysis of Fats

1982

IN W A T E R

PHASE

September 1949

INDUSTRIAL AND ENGINEERING CHEMISTRY

3.0 W

a

2.5

I

a.

-I

-0

2.0

f

1.5

W QI

za

1.0

3

0.5

W 0

*

(3

#

0 W

cn a I a

5

-0

4

-I

5 z a

w

3

W

2

0

3 *

I

8 0

2

I

I

4

6

I

8

10

I

12

Yo FREE GLYCERINE IN WATER Figure 3.

I

14

I

1

EQUIPMENT

PROCEDURE AND ANALYSIS

The desired amounts of water, fat, fatty acid, were charged into the and and heated, with constant stirring, to the desired temperature and pressure. After sufficient time $0 obtain the maximum degree of splitting, the agitator was stopped and the oil and water phases were allowed to separate while constant temperature and pressure were maintained. A portion of each phase was withdrawn into its sampler, still under pressure. After cooling, the entire sample was analyzed. The work reported was on coconut oil and tallow a t 453” and 483” F. It was limited to the range of 90 to 100% split. Standard methods of the American Oil Chemists’ Society were used, with such variations as were required to handle the abnormally large samples obtained ( 1 ). Three analyses deserve special mention because the names do not indicate the meaning precisely. “Completeness” is an expression of the alkali neutralized by free fatty acids compared to the total alkali neutralized by the sample. For practical purposes this is the molecular percentage of the fatty acids which have been split from the glycerides present. “Total fatty acids” i s the amount of material extracted by petroleum ether from a sample which has been saponified and acidulated. I n addition to the fatty acids it also contains the unsaponifiable material which is in all natural fats. ‘fGlycerolJJis the material which is oxidized by potassium dichromate, calculated as glycerol. It may include a very small amount of trimethylene glycol. “Neutral fat” is calculated from total fatty acids, free fatty acids, and combined glycerol,

Relation between Free Glycerol in o i l Phase and in Water Phase

Figure 1 shows the general arrangement of the experimental unit. The autoclave proper consists essentially of a 10-foot length of 4-inch extra-heavy stainless steel pi e built in the form of a square. At one of the lower corners of t i e square an 8-inch extension of the pipe is flanged for mounting a propeller agitator which

serves to circulate the stock through the pipe. The square is mounted in a vertical position and charged with enough fat and water so that the flow is maintained in the top horizontal section but leaves a vapor space to avoid hydrostatic pressures. The autoclave is heated by inducing a current of about 5000 amperes which flows lengthwise through the pipe. This induced current is obtained by building a transformer core around the 4-inch autoclave pipe. Rate of heating is controlled by varying the voltage to the primary windings or by varying the number of windings. After the autoclave reaches the desired temperature, it is controlled by an off-on switch. As Figure 1 shows, the autoclave itself is part of the transformer. The complete square of 4-inch pipe is simply a single-turn short-circuited secondary winding on the transformer. This unit has proved extremely versatile both for studies of the nature described here and for general autoclave work. One of the advantages is that the autoclave can be brought up to operating pressure and temperature ih a very few minutes and with uniformly applied heat. Cocks near the top and at the bottom made it possible to withdraw samples of the fat and glycerol-water layers. The sampler for the latter was a piece of ‘/$-inch pipe with a valve at each end. This was attached in a vertical position to the bottom of the vessel so i t could be filled and cooled, and the sample drained in a normal way. The sampler for the fat layer was made of two pieces of ‘/2-inch pipe joined by a union. One end of the pipe was connected through a valve to the autoclave, the other end was closed by a valve connected to a vacuum system. The I/*inch pipe held a Pyrex tube which was sealed a t one end and had only a I/a-inch opening at the other. The glass tube fitted into the ipe so tightly that practically all of the materia? withdrawn from the vessel was caught in the tube. The entire content of the tube was analyzed so that there was no question of mixing to obtain a representative portion. The vacuum system was necessary to remove air from the sampler before the sample was taken; otherwise the air compressed in the tube forced out part of the sample when the pressure was released.

16

PHASE

diluent for the liberated glycerol as well as a reactant with the fat. Since the fat concentration is small on a mole basis and is constantly decreasing, and the amount of fatty acid is large and constantly increasing, the only concentration over which we have any control is that of glycerol. To remove the glycerol most efficiently, the reaction should be carried out in equipment in which glycerol is washed out countercurrently, On this basis the hydrolysis process was developed. The development work indicated that the equilibrium concentration under the conditions of elevated pressure and temperature was appreciably different from the equilibria existing in the Twitchell process ( 3 , 6). It was apparent that lack of quantitative data on these concentrations would make it difficult to pick out the optimum operating conditions; for this reason the reaction was studied in a small batch unit which could be operated under the same conditions of temperature and pressure as the continuous units.

1983

INDUSTRIAL AND ENGINEERING CHEMISTRY

I984

3.0 2.5

2.o 1.5

w en

a

I

1.0

Q

-

J 0

0.5

f w

0

f

2

4

6

8

IO

12

14

2

4

6

8

10

12

14

4K &I

0

> 3.0 a -I

Q

w

2.5

z

g

Vol. 41, No. 9

of both free and combined glycerol in the oil phase for a given concentration in the water phase than was true of tallow. This result was expected, since coconut oil contains more glycerol than tallow does. In the coconut oil system a smaller proportion of the total glycerol in the oil phase was present as free glycerol. Figure 4 shows the relation between neutral fat and combined glycerol in the oil phase for both fats. From the slope of t,he line the percentage of combined glycerol in the unsplit fat can be calculated. If the reaction occurs in one step, from fat to fatty acid, this percentage of glycerol should be the same as in the original fat. The actual figures are much higher for both fats. The original t,allow contained 10.6 glycerol and t>he unsplit neutral fat conta,ined 19.9%. For coconut oil t'he corresponding values were 13.6 and 22.1%. The higher values are between the calculated values for percentage glycerol in mono- and diglycerides of these fats. i2t values near 100% complet,eness these high values could be explained by the presence of a small amount of low-molecular-weight fatty acids, w-hich would have high percentages of glycerol in their triglycerides. This is impossible over t'he wide range covered by the data because there is not enough lowmolecular-weight fatty acid in either tallow or coconut oil. The unsplit fat must therefore be made up of a

2.0

0 0

$

1.5 1.0

0.5 0

0

Oh NEUTRAL FAT I N O I L PHASE Figure 4.

Relation between Neutral Fat and Combined Glycerol i n Oil Phase

and also from the slope of the combined glycerol-completeness curve. RESULTS WITH TALL0W

The same type of results were obtained on tallow and coconut oil. Since tallow is the more important, discussion will be limited to it and the results shown on coconut oil for comparison. The upper graph of Figure 2 shows the relation between the free glycerol in the water phase and the completeness of hydrolysis of tallow in the range of practical interest. This graph has value as a means of control in batch processes because the free glycerol in the water phase, which can be determined by the specific gravity, limits the degree of hydrolysis. Figure 2 (lower graph) shows the same curve for coconut oil. Figure 3 shows the relation between the free glycerol in the water phase and that in the oil phase for tallow. The latter was the glycerol which was still divided into two portions-first, present in combination Kith fatty acid, and, second, the free glycerol which was evidently in solution in the oil phase. The sum of the two was the total glycerol. It is this total glycerol which must be removed from the oil phase in order to hydrolyze fat to 100% completeness. Figure 3 shows similar results for coconut oil. There was more

200

300 400

500 600

TEMPERATURE, OF.

Figure 5.

Effect of Temperature

on Solubility of Water in Fatty

Acids

considerable proportion of monoglyceride, plus either diglyceride or triglyceride or both. Since the percentage glycerol in the unsplit fat is constant over the entire range, the relative proportions of the glycerides are probably constant. It is believed that the hydrolysis reaction is stepwise, proceeding from triglyceride to diglyceride to monoglyceride to fatty acid. The solubility of water in tallow and coconut fatty acids varies with the temperature, as Figure 5 shows. KO accurate data are available for the solubility of water in tallow and coconut oil since it is not possible to make the determination without causing some hydrolysis. The data already obtained indicate a lower solubility of water in the oils.

September 1949

INDUSTRIAL AND ENGINEERING CHEMISTRY MECHANISM OF HYDROLYSIS

m

4

The foregoing results indicate the following picture of hydrolysis: The reaction occurs in a two-phase system: (1) an oil phase containing glycerides, fatty acid, glycerol, and water, and (2) a water phase containing glycerol and water. In the oil phase there is an equilibrium reaction between the unsplit glycerides (neutral fat) and the glycerol in the same phase. Then there is the equilibrium of the glycerol distribution across the oilwater interface. Water is probably not very important in establishing the equilibrium because, with its low molecular weight, it is present in relatively large molecular proportions. The unsplit fat probably contains mono-, di-, and triglycerides which, above 90% completeness, are present in constant proportions. To determine whether this is a true condition of equilibrium, let us consider how the results obey the phase rule. The experimental points were determined on batches made up of fat-water, fat-fatty acid-water, and water-fatty acid-glycerol. The proportions of the water layer were varied only slightly, but the variation in the make-up of the charge assured that the final conditions were approached from several directions. The system contained three phases-oil, water, and vapor. If all the kinds of

1985

fatty acids present had been considered, there were many components in the system, but the fatty acids could be considered as a single kind if the proportions of the different kinds remained constant. With this concept there were three components. Applying the phase rule, there would be two degrees of freedom. One variable was temperature, which also fixed the pressure. The other variable chosen was glycerol concentration in the water phase. As the graphs have shown, when this concentration was fixed a t any value, it defined the entire system. Therefore it was concluded that it was a true equilibrium. LITERATURE CITED

(1) Am. Oil Chemist’s SOC., “OfficialMethods,” Ea 5-38,Ca 5-40, Cd 3-25,G 2-39. (2) Ittner, U. 5.Patent 2,139,589(1938). (3) Lascaray, Lucio, Fette u. Seifen, 46,628-32 (1939). (4) Mills, V.,U. S. Patent 2,156,863(1939). (5) Tilghman, Ibid., 11,766(1854). (6) Trisler, R. B., J . Oil & Fut I n d s . , 8,141-3 (1931). (7) Twitchell, U.S. Patent 601,603 (1898). RECEIVED July 16, 1948. Presented before the regional meeting of the American Institute of Chemical Engineers, May 9 to 12, 1948, Cleveland. Ohio.

BATCH DISTILLATION Minimum Number of Plates and Minimum Reflux JOHN R. BOWMAN AND MARIO T. CICHELLI Mellon I n s t i t u t e of Industrial Research, P i t t s b u r g h , P a . Binary distillation curves (distillate composition v i . fraction of charge remaining in the pot) are characterized by the (‘pole height,” defined to be equal to the product of the slope of the curve at mid-height and the fraction remaining in the pot. Formulas for minimum number of plates and minimum reflux for given pole height are calculated theoretically:

approximately, where S is the pole height, a is the relative volatility, R‘ is the minimum reflux ratio, and n’ is the minimum number of theoretical plates. The results are applied to comparison of batch distillations under various conditions of operation. 8

c

ONTINUOUS distillation calculations are frequently based on the 5ninimum number of theoretical plates” and the ‘Lminimumreflux’’ for the required separation. These are the limiting values €or each of the parameters as the other approaches infinity, and they can be computed by the well known formulas,

.4

if the relative volatility can be assumed constant. I n this article analogous approximate formulas are derived for batch distillation. All the usual simplifying assumptions, including no holdup, are made. BATCH DISTILLATION CURVES

The data from a batch distillation are most conveniently expressed as curves relating the pot and distillate compositions,

xw and XD, respectively, to the fraction of the original charge remaining in the pot, W . The first of these can be calculated from the Rayleigh equation, (-j-j d

(XWW)

= XD

(3)

which states that the loss of light component from the pot is equal to the gain in the distillate. The integration can be effected if the equation of performance, relating the pot and distillate compositions, is known. Where the reflux ratio is so large that the number of plates alone effectively limits the separation, the equation of performance (2,s) is log

w

,271‘

= a

1-x

. log 1 0 - xw- - a”’ ~

1

-

1 log x w

(4)

On the other hand, if the number of plates is so large that the reflux ratio alone effectively limits the separation, a pinch point must exist, and two cases can arise. If there is a pinch point at the top of the column, the distillate is pure light component; but if it is a t the bottom, the equation of performance follows from the material balance over the entire column and the fact that the liquid and vapor a t the bottom are in equilibrium with each other. The result is (5)

Combined with Equation 3, this yields, as has been shown ( 1 ):

I n the course of a batch distillation with the reflux controlling and with a pinoh point initially a t the top of the column, a stage will be reached a t which a pinch point appears at the bottom, and