284 ANIMAL OILS AND FATS

CHAPTER 6

CAN SAPONIFICATION BE ACHIEVED WITH POTASSIUM

BICARBONATE AND AMMONIUM CARBONATE1 ?

1056. After having considered saponification with respect to the nature

of the bases that can form salts and the amount of fat that a given weight
of alkali can saponify, it would be interesting to investigate whether
alkalinity that is counterbalanced by the acidity of a weak acid such as
carbonic acid still has the ability2 to liberate the acid from a fat with
which it is brought into contact.

§ 1. THE EFFECT OF POTASSIUM BICARBONATE ON LARD

1057. I took 30 g of lard with a melting point of 27°C and allowed it to

react in boiling water containing 60 g of crystalline bicarbonate; most of
the fat was saponified. The soap was then diluted with a large amount of
warm water. The fat that had not been saponified gathered on the sur-
face of the liquid; it was collected and treated again with 30 g of crystal-
line potassium bicarbonate. It saponified completely, except for one
gram. This gram was isolated and finally saponified with 2 g of bicar-
bonate3.

1058. The soap solutions were combined; they were quite clear. When
they were concentrated, no oily drops were formed on the surface of the
liquid. At a certain degree of concentration, the soap was precipitated
from the water by the action of the carbonated alkali. After cooling, the
material was poured onto a filter. The filtrate caused a violent efferves-
cence with tartaric acid but did not yield any fatty material to speak of.
I therefore concluded that all the soap had remained on the filter. The
liquor was saturated with tartaric acid and evaporated to dryness. When
the residue was extracted with alcohol, it yielded a large amount of
glycerin. The soap was collected from the filter and divided into two
parts.

First part 1059. First of all, as much of the mother liquor as possible was removed
by pressing between filter paper. Then the material was dissolved in
boiling water and the resulting solution was poured into an excess of
barium nitrate. The precipitate was collected and dried and on analysis,
it showed constituent fatty acids4 melting at 36°C and barium oxide5. It
did not contain any noticeable amount of carbonic acid6; the soap produced
by potassium bicarbonate is therefore the same as ordinary soap; consequently,
 BOOK V – CHAPTER 6 285

in the experiment I have just described, the carbonic acid must have been elimi-
nated from the potassium that combined with the constituent fatty acids.

1060. It was dissolved completely in boiling water. When the solution Second part

was left to stand, it yielded: 1. a pearly deposit consisting of potassium

bistearate and potassium bipalmitate; 2. potassium oleate.

1061. When an aqueous solution of potassium bicarbonate was boiled, it

was converted into carbonate, as I have ascertained. The saponification
was therefore caused by this salt and not by the bicarbonate, because it
was carried out in boiling water. It is very likely that it is the excess of
potassium in the carbonate7 that determines the saponification and that,
when this takes place, the carbon dioxide that was combined with the al-
kali that becomes attached to the soap moves back to part of the carbon-
ate and becomes bicarbonate if the temperature allows this combination
to be formed; or, if the temperature makes this impossible, this carbon
dioxide is liberated as a gas.

1062. Having mixed lard with potassium carbonate that had previously
been heated to a dull red in a platinum crucible8 and then dissolved in
some boiling water, I observed that saponification took place without
contact with air in a bell jar resting on mercury and placed between two
heated ovens. I also observed that carbon dioxide was not released until
after a large amount of soap had been formed and when the raw materi-
als were exposed to a temperature that was sufficiently high to decom-
pose the potassium bicarbonate. When fat was treated with an excess of a
strong solution of potassium carbonate, beautiful crystals of potassium
bicarbonate were obtained after some time. It was easy to separate them
mechanically from the non-saponified fat and the soapy mass.

§ 2. THE EFFECT OF AMMONIUM CARBONATE ON LARD

1063. 35 g of lard with a melting point of 27°C mixed with sublimated

ammonium carbonate were placed in a flask with a cork stopper. The
material filled a third of the volume of the vessel. The flask was left in
my laboratory in a place where the sun never shines and where the
temperature varies between zero and 18°C. After five years, the flask was
opened. In some places, the surface of the fat was yellow. The material
was treated with amounts of water totalling some 2 liters. A white
emulsion was obtained in which a large number of pearly flakes could be
observed like those formed by potassium bistearate.

1064. One fifth of the emulsion was poured onto a filter; the filtrate was
very slightly opalescent. When filtered again, it was perfectly clear; it
contained a little acid oil dissolved in it, amongst other substances.
 286 ANIMAL OILS AND FATS

Because the filtration took a very long time, the remaining four fifths of
the emulsion were placed in a long-necked flask that was heated on a
water bath. The emulsion split into three liquid layers. The lower,
aqueous layer was almost clear and the two other layers were oily. The
lightest one had a citrine yellow color and the heavier one was orangey
yellow. After they had been allowed to congeal, the lowest layer was
collected, filtered and combined with the filtrate obtained from the first
fifth of the emulsion. I will examine these liquids, which I will call the
aqueous liquid. Next I will examine the fatty part of the emulsion, that is
to say the two upper layers of the emulsion after it had been heated.

ARTICLE 1

EXAMINATION OF THE AQUEOUS LIQUID

1065. It was yellowish and strongly alkaline; on agitation, it foamed.

Hydrochloric acid caused carbon dioxide to be released and produced a
slight cloudiness and on separating, a little orange colored acid oil.

1066. When heated over a flame, ammonium bicarbonate was given off
and droplets of an orange colored acid oil appeared on the surface of the
liquid. After evaporation of the water, a syrupy residue was left that
weighed almost 2 g. It was orange in color and acid; its taste was at first
sweetish and then became very bitter and pungent. It was highly soluble
in alcohol. Cold water dissolved 1.6 g of a syrupy material and left 0.2 g of
orange colored acid oil.

A. Orange 1067. This oil was still liquid at 3°C but it thickened after having been
colored acid oil
exposed to the air for a few days. It was insoluble in cold water and
released only traces of coloring matter into boiling water. It was very
soluble in alcohol; the alcoholic solution turned litmus tincture red and
the red color turned blue again when mixed with water. It was soluble in
very weak aqueous caustic potash and the resulting solution threw a
deposit of pearly material and retained the oleate and an orange coloring
matter. The orange colored acid oil is therefore a combination of oleic acid,
palmitic acid and an orange coloring principle.

B. Syrupy 1068. It was yellow. Its taste, although bitter, was less so than that of the
material
syrupy residue. It was noticeably sweeter. It consisted of glycerin, oleic acid,
an orange coloring principle, ammonia and water, and, so it seemed to me,
traces of an acid that forms a salt with barium hydroxide that is soluble in
water and insoluble in alcohol. These are the experiments that led me to
this conclusion: when I added a little sulfuric acid to the syrupy material,
a small amount of oleic acid and orange coloring principle separated out.
After filtration, the filtrate was boiled with an excess of baryta water; this
caused the release of ammonia and the precipitation of oleate and an
 BOOK V – CHAPTER 6 287

orange coloring principle combined with baryta as well as a little barium sul-
fate and carbonate. The filtrate was evaporated to dryness and the resi-
due treated with alcohol. White flakes that looked to me like a barium salt
separated out. Then the alcohol was filtered and the filtrate evaporated;
the residue was redissolved in water. By means of sulfuric acid, a little
barium sulfate and orange principle were precipitated from this liquor.
When that liquor was subsequently evaporated, it yielded a syrupy
glycerin that had hardly any color or trace of bitterness in its taste. When
this substance was treated with nitric acid, it produced a large amount of
oxalic acid.

1069. I ascertained that the combination of oleic acid and the orange principle
is the cause of the bitterness of the glycerin by taking the same colored
acid, dissolving it in alcohol and mixing it with glycerin obtained from
olive oil. By evaporating the solution, a syrupy residue was obtained
which first tasted sweet but then became very bitter and sharp, just like
the syrupy residue (1066); remarkably, cold water caused some colored
oleic acid to separate. The combination of the oleic acid and the coloring
principle has little or no taste when pure, probably because it does not
dissolve in saliva.

ARTICLE 2

EXAMINATION OF THE FATTY PART OF THE EMULSION

1070. This material formed an emulsion with water because part of its
mass was combined with ammonia. It seemed to congeal at 39°C but
several degrees below this temperature it was still soft. By removing the
ammonia, hydrochloric acid inhibited the formation of an emulsion with
water and simply whitened the surface of the material. After having been
melted, the material formed a single layer that started to congeal at 43°C
and remained soft at several degrees lower9. When allowed to congeal
slowly and without agitation, it did not present an uneven surface10 as
lard does. It had the orangey color of soft adipocere11. It turned litmus
dark red.

1071. The weight of the material, including the fatty material that had
remained on the filter, came to slightly more than 32 g although some of
it had been lost.

1072. 100 parts of fatty matter were treated with 750 parts of boiling
alcohol with a density of 0.83512 (g/mL). The solution was allowed to
cool and the liquid was decanted together with the fatty material that
had been dissolved and precipitated on cooling. I will now examine:
 288 ANIMAL OILS AND FATS

A. the fatty material that dissolved in the boiling alcohol; B. the fatty material
that did not dissolve in the alcohol.

A. FATTY MATERIAL THAT DISSOLVED IN BOILING ALCOHOL

1073. The fatty material amounted to 62.5 parts after the alcohol had
been removed. It was orange in color and started to congeal at 36°C but
did not become hard until 28°C. It had all the characteristics of consisting
of just free fatty acids because it colored litmus very red, released water
when heated with lead oxide and with barium hydroxide it formed a
soap from which alcohol did not extract a significant amount of neutral
fatty matter, as demonstrated by the following experiments.

1074. The 62.5 parts were allowed to react with barium hydroxide. The
resulting soapy mass was treated with warm alcohol.

A. Material 1075. The material was not acid. When separated from the alcohol, it
dissolved by
warm alcohol
started to congeal at 53°C. It was orangey yellow. Hydrochloric acid split
off the barium and released 37.5 parts of an orangey acidulated fat that
melted at 35°C. When this fat was again neutralized by barium hydro-
xide and treated with slightly warm alcohol, it divided into two parts: a
soluble part amounting to 28.75 parts of free fatty acids with a melting
point of 30°C; an insoluble part amounting to 8.75 parts of free fatty acids
with a melting point of 36.5°C. Both parts of free fatty acids dissolved
completely in weak caustic potash.

B. Material not 1076. After elimination of the barium, it amounted to 25 parts. It melted
dissolved by
warm alcohol
at 37.5°C. On cooling, it crystallized as intertwined needles just like the
free fatty acids from lard. It was almost white.

1077. The proof that all the acid fats were completely hydrolyzed is as
follows. After combining these fats, 100 parts with a melting point of
33.5°C were mixed with 70 parts of caustic potash dissolved in water.
This was heated for eight hours and when the soap was decomposed
with hydrochloric acid, 98 parts of acidulated oil with a melting point of
33°C were obtained, together with an aqueous liquid that left a residue
after evaporation from which alcohol extracted only 3 parts of soluble
material which did not contain a significant amount of glycerin but only
a small amount of potassium tartrate and potassium oleate; the acid in
the oleate amounted to 0.1 part by weight as determined by treating the
residue with very weak acetic acid which did not dissolve the oleic acid.

B. MATERIAL THAT DID NOT DISSOLVE IN BOILING ALCOHOL (1072)

1078. It amounted to 37.5 parts; it was white and melted at 31°C but it
was still soft at a lower temperature. It was slightly acid when tested
 BOOK V – CHAPTER 6 289

with litmus. It was treated with a barium hydroxide solution and then
with boiling alcohol.

1079. This material amounted to 25 parts and melted at about 29°C. It A. Material
dissolved by
contained 5 parts of fatty acids that were removed by treating it with alcohol
hydrochloric acid and then with warm alcohol. The latter retained the
fatty acids and allowed a deposit of a fat to separate out; after this fat had
been dissolved several times in alcohol, it was perfectly free from fatty
acids.

1080. After this material had been treated with hydrochloric acid to B. Material not
dissolved in
remove the barium, 12.5 parts of material were obtained from which boiling alcohol
alcohol removed 1 part of fatty acids. The neutral fat was combined with
the previous sample and given the following treatment.

1081. The material was white and still smelled like lard. It melted at Neutral
material
31°C * . When it solidified slowly and without agitation, it presented an
uneven surface, like lard. It was not more soluble in alcohol than lard
and the solution had no effect on litmus. Treating 2.1 g neutral fat with
potassium hydroxide yielded 1.954 g of oleic and palmitic acid with a
melting point of 44°C and a syrupy glycerin weighing 0.265 g. It was
almost colorless and it tasted very sweet without a hint of bitterness.

1082. I cannot confidently state that oxygen from the atmosphere had
no influence on the saponification by ammonium carbonate.

§ 3. EFFECT OF POTASSIUM AND SODIUM PYROBORATE

ON LARD

1083. An amount of 30 g of potassium pyroborate13 was melted and

then crystallized and exposed to the temperature of boiling water for a
period of about fifty hours, together with some water and 10 g of lard. A
large amount of unchanged fat solidified on the surface of the liquor on
cooling and this was removed. More was recovered by concentrating the
liquor. Finally, an almost clear liquid was obtained. When this was
mixed with tartaric acid it yielded about 0.2 g of fatty acids melting at
about 34°C that were slightly yellow and truly saponified. After this
material had been dissolved in a small amount of warm aqueous
potassium hydroxide, it separated out into a pearly material and
potassium oleate on cooling.

1084. Borax gave the same results

*
Because more olein than stearin was dissolved during the purification treatment with
alcohol.
290 ANIMAL OILS AND FATS

1085. It follows from the above that although the pyroborates of potassi-
um and sodium can cause saponification, it is only very slight, probably
because the boric acid, which cannot be precipitated or volatilized, resists
the energy of the potassium and the sodium.
1
As usual, is not immediately obvious which compounds are meant but Google helps.
According to the Traité de chimie élémentaire, Théorique et pratique de Louis Jacques
Thenard, ‘sous-carbonate d’ammoniaque’ is simply ammonium carbonate since it
results from the combination of 50 volumes of carbon dioxide and 100 volumes of
ammonia. ‘Carbonate de potasse’ is described as the residue resulting from burning
wood. In sub-section (1061), the author mentions that when ‘carbonate de potasse’ is
heated in boiling water, it is converted into a ‘sous-carbonate de potasse’. Similarly,
Ostwald (Grundlinien der anorganischen Chemie, third edition, Verlag von Theodor
Steinkopf, Dresden und Leipzig, 1912, page 503) mentions in his sub-section (467)
(Potassium bicarbonate) that when a solution of potassium bicarbonate is heated,
bubbles are formed before the boiling point has been reached, and that these bubbles
consist of carbon dioxide. Accordingly, the ‘carbonate de potasse’ can be safely
assumed to be potassium bicarbonate, KHCO3.
2
Again, the author uses words like ‘force’ and ‘energy’ that have precisely defined
meanings today, which was not the case at the time the author was writing. Using these
terms here would not give an accurate rendering of what he meant. Accordingly, some
less well-defined terms have been chosen.
3
The unsaponifiable in lard is mainly cholesterol and is present in amounts of 3000 to
4000 mg/kg.
4
See endnote 5, Book IV, Chapter 1.
5
Just as the quantitative analytical data on soaps gave the acid anhydride and the metal
oxide contents, the presence of a metal is also indicated by mentioning its oxide.
6
The mother liquor that was not removed by pressing would have contained carbonate.
Accordingly, a small amount of carbonate ended up in the water used to boil the soaps.
This soapy water was then mixed with a barium nitrate solution, causing precipitates of
barium soaps and some barium carbonate to be formed. Apparently, the amount of
barium carbonate was so small that on acidulation with sulfuric acid, no noticeable
amount of carbon dioxide was liberated.
7
The excess of potassium in the carbonate is in comparison with the potassium content
of the potassium bicarbonate.
8
This treatment provides further support for translating ‘sous-carbonate de potasse’ as
‘potassium carbonate’.
9
Pure substances have a sharp melting point. If a substance remains rather soft on
solidification, this may indicate that it has a melting range and is not a pure substance.
10
Fats that crystallize in the β-polymorph tend to form uneven surfaces. Solidified fully
hydrogenated soybean oil can look like a cauliflower.
11
See Book IV, Chapter 4, § 2.
12
The density of 0.835 (g/mL) corresponds to an ethanol content of 85.3 % by weight or
89.7 % by volume.
 BOOK V – CHAPTER 6 291

13
The text refers to “sous-borate de potasse” and this term is used in the heading as
well. Both Google hits state that “sous-borate de soude” is also known as borax and this
compound (Na2B4O7) is also referred to as sodium pyroborate, presumably because it
can be made by heating the decahydrate. Accordingly, the “sous-borate de potasse” has
been translated as potassium pyroborate.