Art. LX. - The Distillation of the Fatty Acids for the Manufacture of Candles.
[Read before the Wellington Philosophical Society, 6th July, 1904.]
Many processes have been suggested for the preparation of the fatty acids from tallow. Of these, the only one practised in New Zealand is that in which the tallow is hydrolised by means of a small quantity of sulphuric acid. The acids thus obtained are washed free from the sulphuric acid and glycerine, and then distilled in superheated steam.
The products of the distillation have been investigated by Cahours, who paid attention to the volatile portions of the distillate, consisting essentially of a number of the lower fatty acids. The object of the present investigation was to examine the course of the distillation from a physico-chemical standpoint, and no attention was paid to the volatile by-products.
Only in one factory in New Zealand is the distillation conducted in a copper retort, and in this case, as soon as the charge begins to become yellow it is blown into an iron retort, in which, after four or five charges have been collected, the distillation is concluded. In the factory* where the experiments were conducted the whole charge is distilled from an iron retort; the last portions, being too soft and discoloured for candle-making, are returned to the retort and distilled with the next charge. The fatty acids condense in a number of coils, the first three of which collect practically the whole of the distillate.
The freezing-point is a most important property to the candle-manufacturer, as the magnitude of this constant gives a direct measure of the hardness of the material, and hence its suitability for candle-making. Consequently, these data were studied in great detail. At the outset it was found that it would be much better to ascertain the temperature of freezing rather than the melting point, for in the latter case definite results could not always be obtained: hence in the whole operation the freezing-points were determined. To make the observation the bulb of the thermometer was placed in the melted sample, and then held in a small beaker to avoid the influence of air-currents. The mean of two or three determinations was taken as the final result.
[Footnote] * I beg to take this opportunity of thanking the Directors of the New Zealand Candle Company for the kindness with which I was treated when performing the experiments at their works.
The values thus obtained are given in Table I., and are plotted against the time in figs. 1, 2, and 3.
|Time.||First Coil.||Second Coil.||Third Coil.|
In series 1 are the complete results for the products from three coils during the whole distillation. In series 2 and 3 only the last portions were examined, but in much greater detail.
It can'hardly be expected to obtain similar results in three instances, as the composition of the original charge, the temperature of the distillation, and the rate are all liable to vary. Each distillation, however, was performed with a medium charge–that is, one obtained from about equal amounts of hard and soft tallow. But in series 1 the temperature must have been higher, and the distillation conducted more rapidly towards the end, as the final portions contain a large portion of the products of decomposition. The second distillation
corresponded with the first, although the decomposition in this case is not so great. In the third series the temperature was lower and the operation conducted more slowly. Further, the charge, though medium, was somewhat softer, containing less stearic acid and more palmitic and oleic acid.
From a consideration of the three series of experiments the following conclusions can be drawn: In the case of the runnings from the first coil the freezing-point first falls a little and then rises more and more rapidly until it reaches a maximum just before the end of the operation. After that there is a very rapid fall, followed by an even more rapid rise. This final rise, however, is not noticeable in the third series of experiments, but in this case the charge was softer and the temperature was not allowed to rise so high.
During the early stages of distillation the freezing-points of the product that condenses in the second coil are slightly higher than those of the runnings through the first coil. However, the curves soon cross, and then keep parallel for most of their length; but whereas the first curve falls rapidly, the second suffers only a slight decrease, crossing the first as it is falling and again during its rise (fig. 2). The third curve commences higher than the second, but falls below it just as the second gradually drops below the first.
A complete investigation of this nature does not appear to have been previously made, and the exact nature of the changes have not been studied. By testing with his thumb
the distiller knows that the product is at first soft. It then opens out and becomes harder and harder. Towards the end of the operation it becomes brown in colour, and begins to soften once more.
It is generally stated that the freezing-point gradually rises and then falls rapidly. The initial fall and the final rise appear to have been unnoticed by former observers. Further, the results obtained from the different coils are of distinct interest, and in fact have been found to be of great assistance in understanding the nature of the changes that occur. However, a complete analysis was necessary to ascertain the variation of the composition of the products obtained throughout the investigation. The different samples were analysed by means of the following methods:–
The oleic acid was estimated by finding the amount of iodine absorbed according to the method of Hübl. The iodine-solution was made by dissolving 25 grams of iodine and 30 grams of mercuric chloride in 95 per cent. alcohol. This was standardised against N/10 sodium-thiosulphate solution, the strength of which was known accurately by comparison with deci-normal potassium-bichromate.
About 1 gram of the substance to be analysed was dissolved in 10 c.c. of chloroform and transferred to a large stoppered bottle. 25 c.c. of the iodine-solution was run in and the mixture was left, with occasional shaking, for two hours. At the end of that time 20 c.c. of 10-per-cent. potas-
sium-iodide solution and 400 c.c. of water were added. The excess of iodine was then determined by means of the sodium-thiosulphate. The oleic acid was thus determined by difference.
The experimental results are given in Table II. It is seen that the percentage of oleic acid rapidly rises but finally becomes almost constant. More detailed experiments (series 2) show that after the maximum has been reached there is a slight fall, but at the extreme end of the operation the amount again increases.
|Table II.—Oleic Acid.
Distillation, May, 1903.
|Time.||Coil.||Weight.||c.c. N/10.||Wt. Oleic.||%|
|Distillation, July, 1903.|
The total acid was determined by titrating with N/20 soda in alcoholic solution, using phenolphthalein as the indicator. It was found convenient to standardise the alkali against pure stearic acid under exactly the same conditions. The amount of oleic acid is already known, and from the relations–
1 c.c. N/20 soda = 0.128 gram palmitic acid,
= 0.141 " oleic acid,
= 0.142 " stearic acid,
the amounts of palmitic and stearic acids could be calculated indirectly. When the result is greater than 0.142, then the
excess is caused by the presence of non-acid decomposition products. Thus, in a weight W. if O be the amount of oleic acid and T total acid, then T–O = saturated acid; W–T = non-acid portion.
When W = T, then W = O + P + S, where P and S are the amounts of palmitic and stearic acid respectively.
The experimental numbers are given in Table III. Calculating from these results the amounts of the different constituents, it is possible to determine the ultimate composition of the various fractions. The results thus obtained are presented in Table IV.
|Table III.—Acid Values.
|Time.||Coil.||Weight.||c.c. N/20 Soda.||Equivalent of 1 c.c.|
[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]
|Stearone and hydrocarbon||3||23||33|
|Stearone and hydrocarbon||1.4||10.7||13.9||15.9||17.4|
|Stearone and hydrocarbon||0.7||6.6||9.0||11.8||14.5|
Having thus ascertained the composition of the various fractions, the freezing-point curve can now be considered. At the point X, where the first minimum occurs, the palmitic and stearic acids are in the proportion of 4 to 1. (See fig. 1.) Now, when the pure compounds are mixed in these proportions a eutectic mixture is formed. On adding a fixed quantity of oleic acid to palmitic and stearic acids, it was found that the composition of the eutectic mixture was not altered. Consequently, as the percentage of oleic acid in the neighbourhood of X varies only to a slight extent, the minimum observed in the freezing-point curve must be due to the formation of a eutectic mixture between the palmitic and stearic acids.
The following are the details of the experiments performed to prove this conclusion: Definite mixtures of pure palmitic and stearic acids were taken, and the freezing-points were determined. These results are represented in fig. 4. 25 per cent. of oleic acid was then added, and another series of freezing-point determinations was made. This formed a second curve. By employing 40 per cent, of oleic acid a third curve was obtained, and this was found to be approximately parallel to the other two. Further than this, the eutectic
mixture in each case was found to contain the palmitic and stearic acids in the same proportion.
The next characteristic point in the curve is at Y (see fig. 1), where the freezing-point reaches its maximum value. The sudden fall that follows is due to the formation of new compounds of non-acid nature. This fact is well illustrated by a consideration of the curves in fig. 5, where the freezing-points and the weights of substance required to neutralise 1 c.c. of N/20 caustic soda are plotted on the same diagram. At the point where the freezing-point curve suddenly falls a rapid rise occurs in the other curve.
Experiments were then performed to investigate the nature of the new compound or compounds that had been produced. A sample of the last fraction was extracted with boiling alcohol several times, whereby all the acids were removed. There was left a dark-brown mass which melted under boiling alcohol. This proved to be a mixture which was separated
into two components by fractionally recrystallizing from benzine.
After seven recrystallizations the melting-point rose to 86°. The compound thus obtained was stearone (M.P. 88°), which is produced by the decomposition of two molecules of stearic acid:–
2 C1.8H35O2 = C35H70O + H2O + CO2
The stearone was identified by the preparation of the oxime, which melted constantly at 62° (Kipping gives the melting-point as 63°).
The more soluble product left in the mother-liquor after the stearine had been separated melted constantly at 72.3° after several recrystallizations from benzene. This is evidently the hydrocarbon C35H72 (M.P. 74°) formed from the stearone by reduction.
Soon the amount of stearone in the distillate becomes so great that the eutectic point is reached, and the melting-point begins to rise. This explains the minimum melting-point at Z. That this is the correct explanation was shown by the fact that, on adding a little stearone to the products obtained after this minimum had been reached, in each case the freezing-point was observed to rise. The addition of stearone to the product obtained before the minimum, on the other hand, caused a reduction in freezing-point.
The reason why the melting-point curve of the product from the second coil falls more slowly is because stearone tends to collect in the first coil, as it is less volatile than the stearic and oleic acids.
Towards the end of the operation the amount of oleic acid appears to be constant, but it is probable that some decomposition products, unsaturated in nature, are formed. This decomposition appears to be intimately connected with the unbearable smell which always becomes noticeable when the stearone is formed.