VEQEiM3i.E OILS CONTAINING GLYCERIDES OF ERUIUGIC AC119.
BY
9. J. Sudborough, H. E. Watson and P. Ramaswami Ayyar.
DR. M. 0. FORSTBK.
F.R.s., CEAIPMAN OF EDITORIAL
BOARD,
Vtl.
S o L x ~ n W l ~14tbl?rr
:~~
L.I.~,V!~
.
h f ~ m l % k+ ~Q
UKTIIVI.
~ % B ~ ~ I ! N .:\ . T
~ i~l f . ! i l ~ ~ t " t ~ ';\:it i ;
CURVXS o r XIH'I.I.KBS
' $ 1 &P~*ANIC ~
~
with N. I<, I kmilf! ...
...
AN])
...
>lRr3<-..8. s'r%:Ai<Arih k i i l . l ~ ? l f$'rt:?;i'
t
i ji ~ lriSXi
i
A ~ l i i i :~ .
'.
,.
.
fit)
Gl)'cei'ir!ua (ti esiicic acid all: !. r .: 1;) large amounts in oils
iron1 thrw ,.:.!., i: ,- c ~ Etiic geliilb IIi.m.~itli(gat. (.)r& Crucifera) and
SO h ih: <);:h i 1 3 1 ihc ~ t ~ <
of.. T,vp<,*,j/bm
h
p,,tzjrts (,gr& ]J/,ay)7E.,
!. L~: .
~ S y g ,273, 371) (Nat. Ord. f ' .
In addition theL. are lnet
with i n i:ert:iirt tibh oils ir:.l~.iI;r~:; wjx& arid cod-liver oils (Bull.,
C/Ifh?. %'it., :S;)t,i, ( ~ ( , ) h j;j ; ' ~ . , t$06, $9, 3570) and accordirIg to
Neville ("1.C%.W. aSiJt., ~ c ) 2,r L01, I xed) approximately 18.6 per cent.
of tile tc.rtai laity 'icitl irrnii ttic dry roots of the common beet (Bet&
i,iO2:iii L,j c'o~isists~.ri
enrcic acid.
.t.
.
8::
' I :
. d );i:.ge :tnlouiil%ot glycerides of erucic acid in an
oil produces a low saponikicatian value am! hence a relatively high
moleculrtr weight ; a iuv: apcciiic gravity and a refractive index-iodine
value c u ~ v efor tire hydrogeitnted oils which lies above that of oils
acids (TiiiicJourrml, 1924, 7, 83).
containiog c:nly C, and I:,,
Other species I~cloii~i~!::
to thc natural order Crucifers: yield oilbcal-it);. sccds, *.ye, garden cress (L@iikiwasativunz, Linn.), radish
(R7//c,/i/i!~
strti'los, I.irin.) and the oils from these have relatively low
saponification values (Crossley and Lc Sueur, J. SOL.Chem. Im'.,
1898, li 7, 9 0 2 ) and hence may contain glycerides of erucic acid.
The three species of rYvassicn yicltling large amounts of erucic
acid glycerkles are rape, lnustard and jamba ; varieties of all three are
common in India. 'The study of varieties of these three plants was
begun in 1 9 2 1 and at that time practically nothing was known as to
the percentage of the various acids present in the mixed acids from the
different oils : the i u v c : ~ l i g i i i o n was undertaken with the object of
ascertaining theac ~~crceritagea
and also of studying the refractive
index-iodine value curves for the hardened oils. During the course
of the work, Tayama (1.
Cham. id. Jafia~~,
1922,25, 1044) published
a paper in which the percentage of erucic acid in the mixed acids
from a sample of Japanese rape oil is given as 65.
T h e three varieties of Bva~sicLzwe have investigated are
I. Brmsi~a:campcstm's, Var. aaptls Linn. The genuine rape or
colza, and usually known in India as sarson.
2 . BvassicR, jwma H. or Sinapis itauea, Linn.
The Indian
black mustard known as rai.
3. Eiuca sntiza, Lam. 'The jamba or rocket seed known
generally in India as tori.
T h e T r n p ~ o b m+us seeds were imported, and obtained fron
the Empress Gardens, Poona. As the oil obtained b y extraction witf
ether was only 7 per cent. of the weight of the seeds, ~t was noi founc
p & l e to undertake a complete examination of the oil.
acids together with ttrncic and oleic acid:;. T h e presence of still lnore
unsaturated i t c i d ~ W d s p v e d , I J U ~ only stnall amounts of insoluble
hcxa-am1 t~trnbromiries could he iiolated, indicating the probable
presetice c r f ,-;-!iric~lcnic.nrd Br-linoiic acids ('rakabasi, J: ToJbKio Chznz.
Scc., ~ r j l y &k,
,
?:;;i).
111 the same year Toyarua (lor. 6if.j examined
the nci(ls h i r ! :i h : : ~ ~ p l i of
:
Japanese oil and gave the erucic acid
c::mlenl ;IS (5.5 1"':. ~.t.iil.., ~ ; t l ~ r r a t eacids
d
less than 2 per ceilt., together
w i t h o!i.ii:, Ii~iibiit.m t J iirldcr~icacids, as proved by oxidation a ~ i d
brcimil~atio~a.
'Thr wiiole of the oi! used in OUT experiments was obtained by
crushing Caii;utkt secch vnr. ictrpiis in an iron ghanni allti was refined
by ;~ik:tll iwcti-ilcnl. Forty-nine po~indsoi seed were used and were
mixed with ro ijcr cent. (.I{ marc?;. 'E'be yield of oil \vas 18-75 lbs. or
38 ,per c w t . "1';~: ~ l k a ! ii l h ~ di n r refinin;;. v a s 20 per cent. aqueous
sorl~un-I
hytlrouiclt: iir slight excess of the ainount required to neutralise
the free fat1.y : ~ c i t l ~ .:Sfi.c:. rt:~liuvalof the soap, the oil was washed
repeatcrlij, wit11 !,:~iiingwater, heated to 1 roo and finallj- clarified by
treatmcrli at 70° with freshly ignitcd luller's earth.
94'1-94'8
l'itxe
... !
Moan m o l a c u l a r ~
1.4635-1.4647
20.0
l?-13
I
10i-107
/ 95'_..6
I
I
...
The refined oil was saponified with alcolrolir- i3 .*-' the dry
potassium soap c?ctr.u lei1 with etlier ti] Lerwovr .
. ,
rlmrtc:
and then decomposed with hydroci~~or~c
acid. A i.,t:Li ril' rgu grams of
mixed acids was used in eight lots of about 2 4 grams caclti. 'rbc
acids were dissolved in 750 cc. ctf gj per cent. (by I crirrmt) alcohcl. the
solution heated to boiling and solution of I Wj gratns i a f !cad acetate in
250 cc. of 95 per cent. alcohol also heated to Ijoil;i,,, added,
'The
mixture was kept at room-ternpcratciri- o r v ~ - ~ i i ~ ~ail$
T - i then at I jo
for four hours, when the prccipitafe was ir.i>~ovcc!and washed wlth
1
lodoinc valar
...
t6'1
I!WO
rwn
.
,
1 hc resuits shcrw &it there wits :s Iuss of less than one per cent.
of fatty acids during ti.1~-I,nx.e--, 'i'he d i d acids have a high iodine
oi erucic acid carried down with the
value due to the largc ;.+ , <,at:.;(,
saturated acids. The iodine v d u c in; pure erucic acid i s 75, so that
86.8 per cent. of the solid acids ccln4its of crucic acid. The high iodine
value of liquid acids I inciirntrs the !,!~..-(.:i(.t~ of highly unsaturated acids
such as linoiic or linolcni:,, naii the tclntiveiy high r~iolecularweight
of liquid acids I1 points t r t thc pirscmce nf erucic acid in this frzction.
The method oE examination was similar to that used by
Jamieson and Bau~hmaii (1-A 1 9 x r " ~ . Cdcm. SOL.,1920, 11& 1372)
in their study of &c liquid acids from mitanaeed oil. About 5
grams of the total liquid acids were dissolved in roo cc. of dry ether
and dry bromine added gradually at oo until the colour persisted;
after further cooling at oo"ior five hours, the precipitate was removed to
a tared filter paper, then washed with chilled ether, dried and weighed.
The melting point was r 78-1 7go corresponding with that for linolenic
Calcnlnrad from tlls raluc. i c r :!.c x l i 2 m d !icu!d ac.ds.
'The nautraliinlo:: ;<lr!rn!r.~::.; w v a ct:erz e;d by sspon:hcarlcn vi'h rre;ss oi olkail
and back titration with sbrrinid x i d , as uiroz: ti!ratior. dvrs nigh msu119.
In order to i ~ > i > : i : ! t ellis residuc Ira nlorrt clctnii, roo grams of
IreshIjr prepared tiquid acids werc irentcd in the s:mc z::annr:r, giving
5.6 grams of l i e s ; 1 ! ~ ! ~ ~ 1 i : i 2 ; : insoluble i n ether a t 11.' nrxi 2.5 grams of
residue irrsaiulle in w r n t light petroletin\. This time the
,l';r.c t was
calourless, nrcitscl at 175-r 7 7 O , coritaiiled 62-3 per cciir. (,I fxcimine
and after one crpstallisatioil frcrnr li~.im*wzneited 31 r X c - - ~ R r ' ~ . (This
gives n total of 2'97 per ecut. rrf linnlenic acid in the Iliriuid acids). It
is thus clear that the hcxabrumide is solrlbk to n r:urtniti c x t c t l i in dry
ether at oo i n the ~ ) w s i ~ iui
i r crucir:
~~
acid clibr<i~triticand tijc c.~mc.lucion
was drawn tlint in thc i : i - l : ; i : i . i ' i i . x t t t 1:51.1.::!
t ! w ~ ~ r o ( l t x! i 1: . ~ ~ ' ~ . i l ! t . in
petrolcum ether was a very impure Ircxni~roniitlc:ctantniriiii~: ~ ~ t i t i : 1 . ; 1 1 ; ~ ,
niatter and pxhaps oxidnticart prodnct~clE high!y rlnsanmtedi acids, a:
the liquid acids had l~cen lcft i n n rlr4ce:1tc.)r for ssmx t i m e !,eiore
investigatiou. The 35-5 per ccrrt. oi ' j t t ? i : : ; r > i . wns iirerdr,~.c~
iiakcti ns
being due to hcxabrurnidc and the :id!:: a i i i ! I t l i flernbroliiide p:z,-enl in
the residue calculated.
6
From the solution in light
a portion (JL the 2itncaiir. acid
tetrabromicle was abtnlned 1.;)
~ t t l i i : ~ ; to
ct" ~i:.~~~-!,j~:l.:.
The
precipitate was collected (317 t i goocti cnrrilk, avzhtlcd, rlried and
weighed. It melted a t 108-roc)" ;rind ;iEtcr u t r ~ ~ ~ r . ; - i . i l : i . . i t i ~ r ~ , at
r ~ z - , rr3", thus corresponding with the tctrabro~nide whic:b melts at
1I
4
After ren'iovaI td the solvet~rfrom the filtrate, the :>].I o:ii!~econtent G£ the residue was deter~ninedboth hg Carius' and Sfcpanoff's
method and was found to be 39-I per cent. The bromi:le i.t,::~p.>:~nds
present are not: only linolic ncid tetrabmtnide and oleic acid dibxomide,
but also erucic acid dibromide ; this cunqAicates the calculation of the
three constituents in the residue.
;i;.i:$~lt,i.;~t
For purposes of calculation the following equations have been
used :(I) x + y + c + a = ioo
(2) 75%
goy 4- r8aa
z74a
= zoo 1:
(3) 338% 2827 -I- 28w C 27862 = roo M
where
+
+
+
I t w a d d :rise 111: pissii.~lic:t t ~d c u l n t e tht: quniitities of crucic,
oleic aild linuiir: ;xi(!.: i'rir~ti rlv: itvight of di.. and tcir;tbromides
i s o l a t e or I
1 r
I of i
i u . T h e results
;vitli Lllosc obtninr:l w;wn tlic iodine ,:slue is
however, do nut :!;re!:
used as :hi: basis t i i the- c~alc111.vioi:. 'l"hi:. i s due to slight experimental crrcxs in the :lot veqt s i n ~ p l cproi:es:; or' isolating :hc bromides
;: .t;nlalI error has s
and as, owing to tlzt: indirect irietl:cxl of c~aic~ilnticm,
large int!uetrcc c r i i the rcl;r.~!t, it ~ : I Jlice11 ctwaidcrrd advisable to use
only the most ewiiy i!etw?iinetl c:onstant:i, viz., the mean ~nolecular
weight and the iodine va!uc.
,,
1he time equation>, r ,
conditions are fultjlied :--
I
.
anti 3, only hold good if the following
A . N o satumted acids are prcsent in the liquid acids. The
saturated acids usually present in liquid acids arc those of low
molecuiar weight as their lead salts are rrlnre soluhle in alcohol than
are lead ste;uratc a i d palmitate, 2nd tbeil presence is indicated by the
low mean molecul:u wcight of ~ h ' liquid acids. As ertlcic acid with
a nlolec~~lar
wcight (!I 538 is present, i t i s n o t possible to detect acids
of low moleculnr weight by tlctermining the neutrdisation value.
T h e results ol thi: distillation ol the iiieth>l estels obtained from
the con~pletely hatdencrl oil (13. 41) indicate the absence of any
appreciable an~ountsof esters derived from lhese acids.
2. T h e absence oi unsaturated acids which would yield
tetrabromo- and hexabron~ot)chenicacids on bromination. All that
can be said is that there is no absolute proof ot their absence, but as
the linolenic acid hexabromide and linolic acid tetrabromide were
readily isolated with correct melting polnts the presence of highly
unsaturated acids derived from behenic acid is unlikely.
T h e results of examining the brorno-derivatives and of calculating
the composition of the liquid acids are given in Table 111,
a
32
TABLE IIJ,
Grams of acids :al:m for brc~nlnctiu:: ..
Grams of hexabrom~decryr;t:lls
...
Grnms insoiuble i s fight pelrole~la~
Percentage o: bromine in tills ros~duc
Grams of hcxabromidc in this I-eslduc,..
Total crams ol hexab~omidc
...
Percentage of linalenic ncid ( u )
...
G r a n s of tetrabromide crystals
.,
Grams of mixed di- and letrabrom~des..
...
Percentage of broiuiue in above
Percentage of erucic acid (r)
Percentage of olcic acld (y) ...
...
Percentage of iinolic acid (2)
...
...
..
.
...
T h e solid acids h a w a high iociine v d ~ i e(ci. 'l'al,!e 11) due
the presence of large amounts of erucic acid, I>ut the value is less tk
that of pure erucic acid owing to the presence uf s~irirated acic
among the solid acids. If it is assumed that the iudinc value
entirely due to erucic acid then the percentage <:ierucic acicl in tl
solidacids is 87, or I O O grams of total mixed ;icicls contain .jo..? grar
of erucic acid present in the solid acids and r ;i.o g r a ~ i ~pi.e~~!lt
s
in :I
liquid acids, i.e., a total of 57.2 per cent.
An attempt to separate the acids present in the soiitl acids 1
conversion into methyl esters anti Iractionni disliiiation tikl not yie
satisfactory results, and it was found marc ; onv!-nit-~-~t
to !ic::xrntc tl
acids present in the complete!)i ilardened eil.
Exaami~za~ioreo f L'he pavfiinliy Pwripifaicd Jbiid Acids.--Twi
che!l (1.
Imi. Ezc. Chew., 1921, 13, 840), in his paper on tl
lead-salt-alcohol method oE separating solid ar?d liquid acids, chin
that, by adding lead acetate insufficient to precipitate the lez
salts of all the saturated acids, a iractional precipitation occu
and the lead salts of the acicjs oi high molecular weight a1
precipitated first. Using this method he succeeded in separatir
from the acids of ground-nut oil 5 per cent. of acids with a mei
molecular weight of 317, that of arachidic acid being 21 a. A s lez
erucate is.more soluble in alcohol than lead stearate, attempts we]
made to separate the saturated acids by partial precipitation. FI
this purpose 1,030 grams of rape oil acids were used in lots of roo
grams. I!r each experinlent the volume of alcohol was 1000 cc, 2nd
the amount of lead acetate jrast sr?ficient to precipitate 5 grams
of acid of n e a n molecular weight 540.
From ehc tot21 lend s d t s so precipitated 50 grams of solid acids
were obtnhccl wit11 nr; iodine value of 49.o I t is thus clear that the
separation in i ~ o conlplele,
t
as this iodine value indicates the presence
of 66 per cent. 01 erucic acid in the solid acids. The acids so obtained
were crystailisetl lrom gg per cent. alco!io!, atid after four crystal11sations s o x i t l with ri. 3~lit'!~,: p i i n ! 7.2'5-75.0" and a mean ~nolecular
weight 0% 355 was obtained. The iaetl:yl ester of this acid was
prcpnretl :lad inelied at $3-5-54'q0 The melting point found
corresponds with that givec uy Ponzio (J Pi,. &hem., 1893, [Ii] 18,
t87) who ii~oughtthat the acltl
arachidic : but tile equivalent of
tne acid and the inolecular weigh: ol the inethyl ester calculated from
its saponiiication value are much higher than those required for
arachidic ncid. On further crystallisation the melting point of the
acid rose to 75.2" and the molecuiar weight to 358, and on still further
crystallisation a small amount of an acid melting at 7 6 - 7 7 0 5 ~was
obtained. These rcsults point to the presence of an acid higher in
the series than arachidic and in all probability lignoceric.
HYDROGENATION
OF RAPE OIL.
T h e oil has been hydrogenated both for the purpose of obtaining
the refractive index-iodine value curve (Sudborough and Watson,
This Joztwzzal, 1 9 2 2 , SI 57 ; 1924, 7, 81) and also with the object of
obtaining large quantities oi completely reduced acid from which the
methyl esters could be obtained and careiully fractionated in order to
determine the percentages of the different acids present.
When the refined oil was heated with a nickel-kieselguhr catalyst
at 180" in a hydrogen atmosphere under conditions exactly similar to
those described in earlier papers (This Jozlmal, 1922,s~62) it was found
that even after four hours there was no reduction in the value of the
refractive index, and the hydrogen escaping possessed a characteristic
odour. After seven hours the iodine value had fallen by 1 2 units only.
The catalyst was then removed by filtration and the oil subjected to
reduction with a fresh catalyst, and it was found that hydrogenation
proceeded smoothly and the oil was completely hardened at the end of
four hours. Apparently the refined oil contains small amounts of
impurities, probably sulphur compounds, which poison the nickelcatalyst. Such compounds react with the nickel or volatilise with
the hydrogen and the oil after one treatment withnickei and separation
of the catalyst can readily be hydrogenated when mixed with a second
batoh of catalyst.
During the reduction small samples of the hardened oil were
removed from time to time by means of a siphon, filtered and the
iodine values determined by \Vink!er's nietkoti and the rclractive
indices with the aid of an Abb6 refractometsr. 'I'hl: d u e s obtained
have already (]bid.,1924, 7, 86) been p:ublishecl and are given in Table
IV, and the corresponding curve in Plate In.
TABLE IV.
SeZective Nydvogmat~orz.-11: has been shown by Moore, Richter
and van Arsdel (1.h d . E?g. Chem., 1917,9,451) that when a mixture
of linolenic, linolic and oleic acids is hydrogenated, the two former
acids are almost completely reduced to oleic acid before appreciable
amounts of stearic acid are formed. I t was thought that similar
selective hydrogenation might occur when erucic acid is present and
to test this hypothesis two partially hardened samples were examined.
acids from sample M (I. V. 60.6) gave
iodine value 49.4. Froln this it may be
calculated that the iodine values of the total fatty acids alld the liquid
21.1 gms. of the fatty
I 3-3 gms. of solid acids with
35
fatty acids arc 63.2 and 86.5 respectively. Assuming tbat erucic and
oleic are the only rmaturated acids present, the percentage of erucic
acid in the solid acids ir; 65.5 and ill the liquid acids 24. The toial
erucic acid present i s consequently 50.5 per cent. of the total fatty acids,
and the oleic acid 2 8 per cent. ' h e amounts of these acids present
in an oil formed by converting t!~c linolic and linolenic acids of the
original oil icto olcic acid ~vouidI>e57.0 and 36.7 so that in the sample
under exanhation 3 2 per ceut. of the oleic acid has been reduced and
only r z pcr cent., of the erucic acid.
Sample N (1. V. 33.9) gave 3.5 grams of liquid acids with iodine
m
of total fatty acids with iodine value 35.4.
value 79.8 from r ~ grams
From this it may be calculated that, assuming the liquid acids to
consist only of erucic and oleic acids, they contam 3 2 per cent. of oleic
acid, a figure which agrees with the value deduced from the equivalent
weight which was found to be 320- Thc iodine value of the solid
acids is, by calculation, 24.8 con-esponding with 3 3 per cent. of erucic
acid. I-lencc the original fatty acids from this sample contained
.to per cent. of erucic acid and 6.5 per cent. of oleic acid so that in
forming this sample from sample M, 77 per cent. of the oleic acid and
only 2 1 per cent. of the erucic acid has been reduced.
The above figures, although necessarily somewhat inaccurate
owing to the indirect method of calculation employed, indicate fairly
clearly that erucic acid is hydrogenated less readily than oleic acid.
It is intended to investigate this point in greater detail at an early
date.
Since the above work was carried out several other papers dealing
with selective hydrogenation have appeared. These are summarised
by Prmstrong and nilditch (Proi. Roy. Snc., 1919, (A), 108, 1 2 1 ) and
ind~catethat selective hydrogenatiou is a very common occurrence.
EXAMINATION
OF ACID FROM
COMPLETELY HARDENED
RAPEOiL.
400 grams of hardened oil with an iodine value of less than mity
were hydrolysed with alcoholic potash ; the alcohol was evaporated,
the potash soaps dried and extracted with dry ether to remove
unsaponifiable matter. The soap was decomposed with hydrochloric
acid and the liberated acids dried and converted into methyl esters
by the Fischer-Speyer method using 4 percent. of dry hydrogen chloride
as catalyst. After remo~ral of methyl alcohol in excess, the esters
were extracted with ether and washed with sodium carbonate solution.
The ethereal solution was dried, the ether removed, the esters
weighed and then subjected to fractional distillation under reduced
pressure, using a Bruhl receiver.
T h e results of the first distillation are given in Table V and the
results of a second fractionation i n Table VI, AS the pressure.: varied
during the redistillation of the various fractions obtained cluri:ig the
first distillation, it was considered inadvisable to ~ n i sarly fractions
and thtls the number of fractions aiter the second tli&lla.tion was 18.
For each fraction the iollowing constants were determined :--(0) titre
of the esters ; (ii) saponification value ; (6) molcculai weight (.!f cster
fraction from saponification value ; ( d ) n~olecuktr weight of :!cids
irom c ; (8) molecular weight of acids from direct neutralisation o i the
isolated acids ; ( f ) titre of thc acids ; (c) melting point of t!:e acids.
All these values are given in Table VII. From the exanlirintlon of
the solid and liquid acids it is clear that the three main acid* pi.crse:lt
in the hardened oil will be belienic (by rcciuction o i erucic), stexic
and palmitic, together with a. little lignoceric and ~ ~ c r h n arachiclic.
ps
Assuming that each fraction contain:; not more t h a n tvjo componenls
it is possible to calculate its composition from :I.
T h e molecular weight of the ester.
2.
T h e molecular weight of the acid.
3. The titre of the esters provided the titre-composition curve
for the two esters is known.
'TABLE V.
First DistilZatioa of Medhyd Esters (296 ,<~n!ms)of H t ~ v d ~ cAcin's.
d
.
III
...
...
i
1
55
I
206-21.5
130.54
/!
20.45
.
IVb
..,
4. T h e titre or ~neltingpoints of the acids provided the curves
are linown,
Methods r and 2 can be regarded as only approximate as the
addition, for example, of 5 per cent. of palmitic acid to stearic acid
only changes the molecular weight from 284.0 to 282.6, a difference
which falls almost within the experimental error of the method used.
The titre curve for mixtures of methyl palmitate and methyl
stearate is known (This JozlrmZ, 1923,6 , 126) and also the melting
point curve and the titre curve for mixtures of palmitic and stearic
acids (Lewkowitsch, Chsmical Technobgy o f Oils, Fat3 and Waxes,
1921, ~ 0 1 .I , p. 1 2 0 ) but no data for mixtures of methyl stearate and
methyl behenate or of the free acids are available. In Appendix I we
give titre curves for the methyl esters derived from stearic acid
melting at 69.8"nd
behenic acid melting at 79.85 and melting
point and titre curves for mixtures ol the acids. These curves have
been used for cietermining the composition oi the fractions i c to Vn.
I n Table V I I the variuus lractions have been arranged in order
of increasing molecular weight as determined by titration of the free
acids after liberation from the esters. These molecular weights are
shown in column 5 and the molecular weights of the esters calculated
from the saponification values are given in colun~nA. Corresponding
values for the same fraction should differ by 1 4 and this is seen to be
the case within the limit of experimental error. I n columns g and 10
are shown the percentages of stearic acid calculated from these
molecular weights assuming that myristic and stearic acids only are
present in the first three fractions and stearic and behenic acids only
in fractions Ic to V6.
I n order to determine which acids were present in addition to
stearic, use was made of the fact that equal weights of myristic,
palmitic and behenic acids when added to stearic acid produce
depressions ol the melting point which are of the same order of
magnitude, or in other words mixtures having the same melting point
contain roughly the same percentage of stearic acid. In the case 01
palmitic and myristic acids in quantities below 40 per cent. the
agreement is close, for behenic acid the divergence is greater and the
relation only holds up to 2 5 per cent. of the acid. As an example,
binary mixtures with myristic, palmitic and behenic acids melting at 65O
contain 80, 78 (Heintz, A m a l e ~ z , 1873, 92, 295) and 85 per cent.
(Appendix I) of stearic acid respectively. If the assumptions are
made that arachidic acid produces a depression similar to that of the
other three acids and that the addition of a mixture of two acids has
the same effect as that of an equal weight of a single acid, then it is
possible to estimate approximately the amount of stearic acid in the
fractions resulting from the distillation of mixed methyl esters in
which stearic acid is the principal constituent.
In column 12 are shown the percentages of stearic acid calculated
in this way. T h e figures for the first three fraEtions are the same
whether myristic or palmitic acid is present. For fractions IIIa to
Vd the figures are derived from the curve for mixtures of methyl
stearate and methyl behenate given in Appendix I, and for fractions
IC to IIc intermediate values have been selected.
Assuming these figures, it is possible to calculate the proportions
of two additional acids since the mean molecular weight is known, If
3
In the case of fractions IITtz ! r i Vli ih:: \-;~li!c:: !iii- Ilic. )xxcelrtage
of stearic acid deduced from (n) tile titrr IJ!' the c:iw (t:oIumn S),
( 6 ) the molecu1:tr weight of the esiers (i:trlu~;ii; (.t). :,! till: nii~iecular
i ~ i ~ <if t l :itids
~
weight of the acid:; (coiumn 10) mrl ( d ) tlxh~ ~ i ( . i l I:(.IXLI
(column 1 2 ) (111 llic ;ts!.,:i!!;!~t:<~~~
tlirlt nil t!li.:ic
I r ; ~ ( , t i < ~ i .ZIT
t f . , liinarY
:i;;rt:v
* r t ~ ,well witllin
mixtures of methyl sieamtc n d incthyl i ~ i * I ? v ; ~ ~
the limit oi cxperiniental error arid it i:: ui11ikt:Ij. tli:~: tliirr! acid is
present. T h e values derived frcm the t i l r c I b
i t.he :ic:ids (c o!unin I 11,
difier very widely and these can only ix i . : . . , ~ ! , : i ~ ! t ~ ' !)y i!lc r.sisience of
some undetected experimental error.
T h e presence of behenic acid in theso fmctiti:is W : i s tlcrnui:.itraieil
by crystal!isation of the acids from 95 per c:elil. ;~l!.tri:oi asiyg 2 5 rc,. of
solvent per gram of acids. After two tri- t!mc ::;.y~r;illi~nfi~rn.s
;ill xicl
with a meltlng point between 78 nnc! Soo :>nil ::I! t.cl:!;~~;.!i ;'.cigl~t of
341-142 could be isolated. The presence r d stearic: ar:iiI w s mirre
difficult to prove. The following ~>:occsswas used tritl; >,itc.ceis in
r
frr>:ii tlic iirst
the case of. several fractions. 'I'he ~ : i o t l ~ eilqiior
crystallisatio~lEron~9s per. cent. alcohul is ci.:i;>:~:;!ct:
!r,
clrylxss and
the residue crystallised from a s m ? l anmm? oZ ctli): :II-ctntc,tht.
crystds are rejected and the acids from the nwther licliior ri+ci~vered
rind
crystallised from zoo cc. of 75 per cent. alcohoi !>ti- g i m i of acid.
T h e crystals are rejected and on concentrating the mirtiiei liclriitl an
acid melting from 67-68O and with an equivaitnt weii~lil2St-290 is
obtained. The presence of both behenic and stearic acid in Iractici:is
IId, 1116, and V n was proved by this process.
11;
T h e case of Lractions Ic to IIc is a difficult one since the figures
in column 5 differ widely from those En r o l u ~ n n7 . T h e only satisfactory explanation is that they consist vE ternary or even quaternary
mixtures. As palmitic acid does not appear to i,c p r r ~ n i ,the acid
of lower molecular weight than stearic must be myristic acid, Tinat
of higher molecular weight rmy be arachidic or behenic or both. f t
is known, k!omever, that only sna.11 amounts of other acids can be
present because there m i s t be at least sufficient behenic acid to
correspond ivitli the er~:cic acid originally present and sufficient
stearic acid t o correspond with tiie other unsaturated acids. This
leaves oniy a small percentage io be accounted for otherwise. The
behaviaur of a mixture of a small quantity of rnetnyl arachidale with
methyl stearatc and behenate on distilhiion is not known and it is
consequently impossible l o say definitely whether arachidic acid is
present or n o t except by further fmctionaiion. In calcu!ating the
co~nposition i? lrav been assumed that arachidic acid is not present
and if this is done, fairly uniiorm results are obtained for the composition o i .the earlier fractions as shown in Table VIII.
TABLE VIII.
-----
Percentage composition
Behenic
Fraction Vc is nearly pure methyl behenate, but even after five
crystallisations from methyl alcohol the melting point did not rise
above 5z0, and the acids derived from this fraction after repeated
crystallisations from alcohol melted at 79.5-80.0' (cf. Appendix I).
Residue I, which contains the acids of highest molecular weight
was hydrolysed, the acids crystallised from alcohol and then fractionally
precipitated with niagnesium acetate, when an acid melting at 77'578.0" and with an equivalent weight of 363 as compared with
lignoceric acid 368 was isolated.
,_-'
.A-T h e compositidn of residues I and 11 have been calculated on
the basis that the only constituents are the methyl esters of behenic
and lignoceric acids.
The compositioxis of the acids from the hardened ctil a ~ x irom
l
the
original oil are given in Tabic IX.
;
I
Acid
I'urw1i:ige p r t L i i i !
1
fro11 i : !
ti![
Myristic
Stcaric
Behenic
Lignoceric
Oleic ...
Erucic
Linolic
Linolenic
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
... 1
...j
...I
... ,I
...
...
...
.../
.
hnrdvnorl oil
1
.
...
I'c?A n t ; i x ~present
in acids Isom
I
-
precipitated with the solid acids. T o obtain tho amount of stearic
acid in the acids from the unhardened oil the stearic acid produced by
the reduction of the oleic, Iinoiic and linolenic acids present in the
original acids is deducted from the total stearic present in the hardened acids. I t will be noted that the percentage of behenic acid
present in the hardened acids corresponds tery closely with the
amount formed from the total crucic acid present in the original acids.
Independent experiments made with the solid acids irom 400
grams of mixed acids resulted in the actual isolation ot 3 grams of
nearly pure behenic acid, so that this acid niust be present to the
extent of at least 0'75 per cent. No trace of stearic or arachidic acids
could be detected and the figure obtained by difference and given in
the table is probably incorrect.
The unsaponifiable matter from the hardened oil gave a sterol
crystallising from alcohol and melting at I 37-1 39O. The acetyl derivative melted at 134-136" and corresponded with the acetyl derivative
of the sterol from the original oil.
From the mother liquors of the first sterol a fraction was
--sbta.h.ed melting at 1 5 0 ~and after recrystallisation at 1700. This
compound forms an insoluble compound with digitonin and may perhaps be the dihydrositosterol melting at 17s0 (Hauth, Lewkowitsch,
vol. i, 282).
*
Indian mustard oil 1s prepared from the seeds of Bynssjra
juwr.a, I).C., a variety o i Simpis ~ ' p vL.. ~
(B~nssicn&pa, Koch).
I t is known i n India a5 15and is manufactured in very largc
quantities in 'dengal in ?;lotor-dl-ivenghannies and also by the solvent
extraction process. 1t IS l ~ r g e l yused for edible purposes and in
general characteristics reseinbles rape oil, the chief constituent being
a glyceride of eruclc acid. Litt!e or nothing appears, however, to be
known about the other acids present or the proportions of the
acids.
Hn Table I are given the analytical data for two samples of oil
examined by us. The one was a sample received from England, it
had a pale yellow colour but the relatively high acid value 27.
The second sample was pressed in these laboratories and was a
sampIe oi Indian mustard seed grown in Mysore State (grown in
Vasantapura village ro miles south of Bangalore, in 1922-23). When
extracted with ether the seeds gave 35.2 per cent. of oil and when
pressed in n power-driven ghanny gave 30.5 per cent. The oil so
obtained was heated to rooo and then filtered. It had a pale brown
colour and a mild taste. Table 1 also gives values found by other
authorities for Indian mustard oil.
T h e Mysore oil has a high saponification value. The iodine
value is also high when compared with values obtained by Csossley
and Le Sueur for oils from Bengal. Its acid value on the other hand
is quite low.
In Table I are also given a few constants for the mixed fatty
acids obtained from the Mysore oil and in Table I I values given by
Huber and van de Wielen (Pev. Ess. Oil Rec., 1915, 6, 341)
showing the variations in the oils from different sources.
...
Origin
...
sp. gr. 18.5'
. . . . . .
Unsaponifiable matter
BeI~narvalue
..
Iodine value
...
173.2
...
101.1
/
IOB.4
Neutralisation value
.
175.7
Mean molecular weight
...
...
1
1
H O I I ~ : ~ ~
1
;:;;;;;-
1'4580
/
i.iii56
7.35
1
/
/
317.8
L4&3.5
/
1
V, i 2 .
0,915.5
...
/
/
1
...
I
173.3
98 8
...
... i
Refractive index calculated
to 2om
.
i
... 1
... 1
i
... 1
...
Iodipe value
1'4739
...1
..
Sa.pouiBcation value
:
...
a.921z
... /
*acaicolaied to 20'
Acid value
/
1
..
106-li:i
..
...
f + f / y Acids.
..
'I
/
1
"'
...
...
"'-w-m--
i
...
...
;
l.s16t;s
1
1.4674
Lp--
TABLE 11.
Vtariatiorts in Secd.i n/td Oi/.
ofi,in
England
No. of Volalile
in
of seed
1 grain
----
1
15
/
.
''K
sapm~iti ~ i n e
catitm wlitc' wlue
'
......
Sicily
...
...
Roumaaia
...
...
..,
...
Bombay
'2;" I
per cent, 1
.
I
...
!
.
'
UU.P-.--..*
'Values from Table I.
"-
TABLE III.
I
I
Solid acids
Liqoid acids
~
Iodine valur:
...
. .I
11.40
11 58
61.6
65.6
la
la 60
1
123'3
6
18.42
122.8
"
.
.
,
,
<
A
"
The iodine value and the mean molecular weight of the solid
acids are high, due to the presence of considerable amounts crf erucic
acd. ':'he values for the liquid acids are for the total liquid acids
obtained by ieoml~iilingliquid acids I and 11.
T h e percentage of erucic acid in the solid acids is 86.9 if the
iodine value ol these acids is due entirely to the presence of erueie
acid, T h e solid acids have much the same iodine value and
molecuiar weight as the solid acids from rape oil.
EXAMINATION
OF LIQUID ACIDS.
T h e method adopted was exactly the same as that used for the
liquid acids from rape oil (p. 6) and the same equations I , 2 and 3
were used for purposes of calculation. The results of the examination are given in Table IV. It will be noticed that the iodine value
of the liquid acids from mustard oil is much higher than that for the
liquid acids of rape oil and that the molecular weight of the same
acids is relatively low, indicating the presence of less erucic acid in
the liquid acids from rape oil,
TABLE iV.
As t h d w w:i% s w ~ iil ~ n l i l:,s t:) !lw p i ? ~ t yoi tlw Ii~x:~l~i.iriniilc
III~!:L~::,YI in t h t e erperi.
ments, t w o fresh delvrmination:. wcro nmic hy m u of ul: (?'.K.A.) u!! ~ l l i k ~ u ijm!p~ril!ioni
lt
of
liquid acids. T h e w Raw the pcrccnt:~;;? r ~ linolenii.
i
x i t i a!. :;'7 nu.1 tL.1
T h e solid acids wcre not examined in detail ic~rthe reasons given
on 11. 32. T h e oil was hardened and the acids from tllc completely
hardened oil were isolated, converted into inethyl esters anti these
esters caxfully Iractionated ant1 each fraction emmineti.
I n artenipting to hydrogenate the saniple of Mysore c d , which had
been alkaii-treated in order to remove lrec fatly acids, it was b u n d
that practically no change had taken piace aiter !n<sing hydrogen
at 180" in presence of niclcel-kieselguhr catalyst ior 1 2 hours. When
however thc oil was steam-disti!lect and dried hefore llyclrogenation the
reduction proceeded smoothly. It appears probable that the ailyl
mustard oil, obtained by steam-distillation, has an inhibiting infiuence
on the activity of the catalyst.
T h e curve connecting the iodine values and the refractive indices
of the hardened samples was found to exhibit certain irregularities and
consequently another set of determinations was made with a rather
larger number of samples. In this case alkali.refined oil only was
used and the difficulty of, hydrogenation overcome by warming the oil
with one batch of catalyst, filtering and reducing with fresh catalyst,
the reduction proceeding quite readily.
I t is interesting to note the considerable change in iod~nevalue
at the beginning of the hydrogenation and the small corresponding
change in refractive index. This has been observed before in the case
of seal oil (TAi.7 Jozr$nn~31924, 7, 81). The very marked break in the
curve at the iodine value 80 indicates the selective nature of the
hydrogenation (d.p. 34) and i s more conspicuous than usual.
4
---.-*
m
-e
--
Fsaclim No,
.
.
;
..--
lhi!iuj: !x<>iut ,,L
c i c ~ r r u< '.
. ..--
I V v ? . i ~ t It!
."
i!iU-200
i
.
Per writ.
tqz:ith.,
2.56
. . ... . ..-
1
2.8
Fractions I II arid I V were refractionated and the results are given
in Table VII.
TABLE VII.
NO. of frnction
1
Boiling point in
degrees C.
Weight in
griims
1
..----"
Per cent. of
total esters
O n cooling to room-temperature (25") it was noticed that fractions
I and I 1 had partia!ly solidified in the fortr, of well-developed plates,
but that a portion ol each remained liquici. The solids and liquids
were separated 'oy suction using a very small paper and the products
examined separately.
T h e n~olecular weights of the liquid portions were determined
from their saponification valucs, and fhe neutralisation equivalents of
the acids were also determined. The acids were then separated as
described on p. 40 and practically pure stearic acid melting a t 68.0" and
an acid melting at 53.5' were isolated. T h e lattcr acid had the same
melting point on further crystallisation and was taken as myristic
acid.
T h c acids flom rcsidue I had a mean molecular weight of 346 and
after two cryitallisations gave an acid melting at 75'4O. The high
molecular weight points to the presecce of lignoceric acid.
'Fable V2II gives the compositmn of the different fractions based
on (a) the moleculav weight of esters calculated from the saponification values, (6) the neutrdlisation equivalents of the liberated acids,
and in the case of the middle fractions (c) the titre temperatures of
the esters and the melting points of the acids.
The molecular weights calculated trom the saponification values
of the esters and the neutralisation values of the corresponding esters
agree quite well, as might be expected with the exception of those for
residue I. I n this case the neutralisation value of the acids was used
for purposes of calculating as the esters evidently contained a little
neutral material, either unsapon;fiable matter from the oil or decomposition products from the esters.
I n calculating the percentage of the diilerent acids small
corrections have been applied for the loss of ester in the condenser and
For the small loss on transferring fractions III and IV to a second
distilling flask.
T h e composition of the mixed hardened acids is shown in Table
IX together with the composition of the original mixed fatty acids
deduced from these figures and those on p. 46. In making this
calculation it was found that the amount of stearic acid which would
be formed on hydrogenating the quantities of oleic, linolic and
h o l e n i c acids given in Table IV would amount to 54.4 per cent. of
the total hardened acids, whereas only 52.1 per cent. was found by
the distillation method. This is evidently an experimental error. It
has been shown on p. 45 that the solid acids form 38.3 per cent. of the:
T A B L E IX.
--..-I-&--~~
acids from
1iai.dcned o i l
-.
Acid
.
-. ...............
iQristic
...
Oleir
..-
Erucic
...
...
"'
1
0.5
I
...
I
... 1
...
I
acids from
oripiud 011
/
0.5
1
41.5
whole and contain 86.9 per cent. of erucic acid. From the distillation
results, the percentages of myristic and lignoceric acids in the solid
acids are 2.9 a i d 1 . 3 respectively. The balance of 9.9 per cent. is in
all probability behenic acid since the mean molecular weight of the
solid acids lies just above 338, the molecular weight of erucic acid.
I t is very unlikely that acids of lower molecular weight are present; 9.9
per cent. of the solid acids corresponds with 3.8 per cent. of the total
acids and there cannot be more behenic acid than this in the mixed
acids. T o obtain 46'3 per cent. of behenic acid in the hardened acids
it may be calculated that the percentage of erucic acid in the liquid
acids would have to be 14.7 instead of I 1.8. As this would only raise
the mean molecular weight from 288.1 to 288.9 the difference is within
the limit of experimental error. In the table, mean values have been
adopted and it has been assumed that there is no stearic acid in the
unhardened fatty acids.
T h e unsaponifiable matter amounting to 1.2 per cent. on the
weight.of the oil, after three crystallisations from go per cent. alcohol,
melted a t 142" and corresponds with the sterol found by Windaus and
Welsch (Bey., 1909, 42, 612) in rape oil.
r.Vith T .
.7.: ? I i w h ~ ~ ~ i m L
jambs is largely grctwn in Nort!m-11 India including the Punjab,
E.\N. Provinces, Sind mil IJnited i'ro~inces. I t is o!ten grown in
associaiioll with cottor? or barley and if cdilva!utl a!iirie requires a
dry soil. i t i s ilsetl as n green iotldcr and i t is stntcd that the oilcake obtained by pressing the seeds is a v;!lrial~le cntge-food although
it contains a g l u c ~ ~ i c lwhich
c
gives rise to ;i lrilngcnt csscntial oil,
(Iials and Cfram, Lrrid. I/rvszrt.hst~rf., t<><ic), 70, 3x0). ' l ' l ~seed is
stated to contain 2 6 per cenl. of oil, but :r sntnple irtrln Sind used by us
gave ;I per cent. ol oil on extraction with cthe: and 2 0 pr: cent. by
crashing. The cake in thc latter case gave :t furtirer ro per cent.
on extraction with ether. T h e oil is iargcly nset:l for burning and
lubricating purposes, but is interior to genuinc tape oil (Sarson oil)
and is sometimes used as an adulterant in the latter on account of its
low price. According to Lewkowitsch (C%cmira/ Y ' ~ d ; i c ~ i ? ? o~ fy Oils,
fiaz!s.aiad Ifixcs, 1922, vol. TI, 2 8 2 ) the oil differs from the other oils
of the rape group as it, docs not lend itself readily to thc manufacture
of 'blown oils ' bccause the specific grnvily docs m ~ rise
t by blowing
as that ol other oils l~elongingto the same ~ Y O U I ) .
Analytical data lor the extracted oil, as weli as for the pressed
oil, are given in Table I together with the values obtained by
Lewkowitsch (vol. 11, p. 283) de Negri ant1 Fabris (Ann. &:.I Lab.
ddlc Ccrbcl/c, 18gr-gz, r37), Thornson and I ) ~ n l : ~ p(r'lmr/jd,
,
1906, 31,
282) and Hals and Gram [Lntzrt! t.'f.vsitdsirr/., ~ ~ u r70,
) , 31I ) .
53
It will be noticed that there are appreciable differences between
the extracted and rxpressed oils particu!ar!y in specific gravities,
refl-aciive indices and iodine values. O n the whole the expressecl oil
ciosely rese~nblesehe indinn rape oil examined in these laboratories
(p. 37). 'The extracted oil has a distinct brown colour and evidently
conlains colouring matter and other products extracted irom the seed.
The expressed oil has a yellow colour, a pcnetrating odour aud a
characteristic bitter taste. A n interesting feature is the comparatively
high acetyl value of both extracted and expressed oils, a phenomenon
already observed in the case ol oils from genuine Indian rape and
mustard seeds. This value is not due to tile presence of hydroxylated
acids, as was proved by the acetyl value (3 : I ) of the methyl esters of
mixed Iatty acids and further confirmed by detailed examimtion of the
acids.
Table I I gives the constants for the mixed fatty acids.
TABLE 11.
..
Lewkowitsch
...
I. NCII and I . U m
Yircliandani
...
.-
1
...
.
---
Hydmzemr/ioa o f OiL-The
oil used for the subsequent work was
the expressed oil and was refined by treatment with a ro per cent.
solution of sodium hydroxide, removal of the soap, thorough washing
with boiling water, drying at ~ o 5 "and final clarifying with fuller's
earth at 70". T h e acid vaiue of the refined oil was 0.2.
Considerable diflicultp was encountered in hydrogenating the
refined oil. Even after 7 hours' treatment at r80° with the usual
nickel-kieselguhr catalyst the refractive index was lowered by 0'0003
only. During the process of reduction the escaping hydrogcn had
a very pungent odour and it was thought probable that a volatile
sulphur compound was present and poisoning the catalyst. T h e
presence of sulphur compounds in the oil was proved by the nitroprusside test. Attempts to remove such compounds by steam-distillatian did not meet with success ; the oil was still not seduced at 180'
in the presence of the catalyst and the presence of sulphur in the
Samples mere taken niter give], intervals and iiie iodine valiles
ant1 refractive iridices determined.
The results are given in
Table 111.
T A B L E IEY.
--Sample
No.
-
--
Original oil
Alcohol refined
(sulphur iree)
26
19
27
28
20
21
72
23
20
24
30
25
...
TABLE V.
Ezan~iizationo f Liquid Acids.
G r a m of acid i a k m for brominatiun
Grams of hexabromide crysliils
Percentage of 1ino:cnic acid ( a )
G r a m of tctr;lbro~nidccrystals
Grams of mixed di- 2nd tetrabromidcs
Pcrceiltage of bromine iu above
Percentage of crucic acid (xi
..
Percentage of olew acid (y)
...
Prrcslltage col l~imlicacid (2)
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
li
4'726
0'617
4.77
0.37
5.91
37'9
I
4'954
0'629
4'64
0'130
7'10
42.5
1'9
65'2
33.2
In the preparation of the bromides great care was taken to dry
all the materials ; the hexabromide obtained was quite white and
melted at 182-1830. After removing ether and excess of bromine, the
residue dissolved completely in warm light petroleum and no residue
was left a s with the acids from rape and mustard oils.
Exnnzinatiolc o f Acids fmaz CompZer'ely Hnrdmcd Oil.-The almost
completely hardened oil (iodine value 1.7) was saponified vith
alcoholic potash, the soap extracted with petroleum ether and the acids
liberated and esterified as described on p. 35.
'The molecula~-weights of the total acids and also of the liquid acids \%.ere determined
by the indxect method or tituxtion or saponification ; cf. LewkowRsch.
'The amount of l i q u ~ dacids obtnined by a second precipitation was smal:, via.. 3'1 par
m ~ r e dancl the total liquid acids examined.
i.*ot., hence the two
5
1
2
i
I
5
6
7
8
Residue 1
TABLE VII.
The saponification value and titre of each fraction was taken, also
the mean molecular weight, the titre and melting point of the mixed
acids obtained lrom each fraction. T h e results are given in Table
VIII.
Fraction No.
Weight
in
grams
A n interesting point is that the molecular weight of no fraction
falls below that required Lor methyl stearate, indicating thi: absence
of any appreciable amounts of esters of acids iowcr i l l tlic series than
stearic. T h e moiecular weight of fraction ! I correspond:i with that
of pure ixethyl stearate although the titre ik. only gz.:! cumpared with
36.7 for pure methyl stearate. When r ~I-:IIX of the fraction was
crystallised from methyl alcohol, o'izt gmm of crystals xelting at
36'5-37-5O and a residue of 0.2 3 grain nielting at 36.0-37'6O were
obtained, indicating that the fraction was nearly 11uremethyl stearate.
By crystallising lractions 7 and 3 and residue I1 from methyl
alcohol, methyl behenate melting at 52' was iwlatcrl, and this on
hydrolysis gave behenic acid melting at j9'g0 to So.ou and having
a molecular weight 339.
Fractions 6 and 6 . when recrpstallised. gave a product melting at
38O and solidifying at 3 0 . 5 O ; on hjdro!ysis an acid melting at 69%O
was obtained.
From fraction r , a small amount of methyl behenate melting at
was isolated and on hydrolysis gave an acid inciting at 79O and
having an equivalent 336. This is sufficient to shorn that the ester
which produces the h ~ g hmolecular weight of the fraction is methyl
behenate and not arachidate.
52O
Fraction n has a molecular weight corresponding with that of
methyl arachidate, but the titre of the ester and the melting point of
the acid prove that the fraction is not pure methyl ararhidate, and on
fractional crystallisation methyl behenate was isolated, indicating that
this fraction is a mixture of methyl stearate and n~etllylbehenate.
The ester of high molecular weight in fraction 8 and residue
has been taken as methyl lignoeerate although so lar no pure lignoceric
acid has been isolated.
In Table V I I I are shown the percentages of stearic acid in the
various fractions calculated by five different methods. In certain
cases the results are in fair agreement, but in others, in spite of the
fact that fractions I to 7 appear to contain only two acids, there are
marked discrepancies. For example, in fraction a, although the
molecular weight of the acids is very close to that of pure stearic acid,
the titre of the ester and the melting point of the acids are both below
the lowest values given by mixtures of stearic and behenic acids, and
33 per cent. of palmitic or myristic acids would be required to
produce a mixture with the same melting point.
The percentages of stearic acid deduced from the titres of the
acids from fractions 5 and 6 are much higher than the values obtained
from the melting points although both these fractions appear to
contain stearic and behenic acids only. A similar difficulty was met
wiih in the case of the acids from hardened rape oil (p. 4oj but the
reason has not yet been ascertailed. A possible explanation might
be that isomeric x i d a are present, as the oil has so far not been
examined for these, a.nd it is hoped that the reason for the discrepantie:, rnaji be iounci on further investigation.
in the absence: of any evidence of the existence of acids other
than stearic, ljehenic and iigi:oct.i.ic in the hardened acids, the composition ol the hardened acids is as shown at the end of Table V I l I
and that of the acids from the original alcoho! washed oil is as follows,
the calculation being made in a similar manner to the one for rape oil.
...
. . .
..,
.....
...
...
. . . . .
...
Stearic
acid
Behenic
. .
Lignocerlc ,,
Oleic
Erucic
,,
Linollc
L~nolenic
.
...
..
.
.
...
...
...
...
...
...
...
...
...
4'2 per cent.
4'5
,,
1.6
28'7
...45. 3
,,
12.4
...
...
...
...
...
2'1
....
..,,
U n s a p o ~ i ~ a dMa2'tev.-The
b
unsaponifiable matter was extracted
by Wilkie's wet process (Agsalyst, 1917, 42, zoo), and amounted to
0.7 per cent. T h e percentage of phytosterol in this was determined
by precipitation with digitonin in go per cent. alcohol (Windaus,
Ber., 1909, 42, ~ 3 8 and
) ~ was found to be 39. T h e phytosterol was
isolated by boiling the additive compound with xylene. After crystallisation from alcohol it melted at 137-138~ and gave an acetyl
derivative melting at 1 2 6 ;~it thus appears to be the ordinary sitosterol
which melts at I 3gUand yields an acetyl derivative melting at I 270.
AZcoliolic EictrocL-As
already stated on page 54, in order to
obtain an oil which could be readily hydrogenated, it was necessary to
extract the oil with hot 88 per cent. alcohol. The results of the
examination of the alcoholic extract are given in Table IX.
TABLE IX.
...
Grams of oil extractea
...
estract ...
...
...
Grams of voIatile oii
Grams of residue after steam-distillat~on
Grams of unsaponifiahle matter
...
...
Grams of fatty acids
Mean molecular we~ghtof fatty acids ...
Iodine value of fatty acids
...
Grains of
...
...
...
-
T h e unsaponifiable matter on examination by the digitonin
method gave the same sterol as obtained by extraction of the potassium
salts with petroleum ether.
Tile amount of glycerides extracied by the alcohol was very
only 0.6 or 0.7 per cent. of the oil. T h e acids derived from
these glycerides have a molecular weight rather lower and an iodine
value rather higher thar the total fatty acids (Table IT) and hence
contain a higher percentage of linoIenic acid than the mixed
fatty acids derived from the original oii. If this is so, then the
percentage of linolenic acid ii; the mixed acids derived from the original
oil would be higher than the valnc given on p. 55, but the difference
is so small as to be inappreciable.
The Volatzb OiL-It
may be seen from Table IX that the
expressed oil yields I per cent. of volatile oil on steam-distillation.
i n order to obtain a larger quantity of this oil, experiments were
conducted with the object of isolating it directly from the seeds.
Joergenseil (Lmn'io. Vers.-Sid,1899, 52, 272) states that jamba
seeds give no volatile oil on steam-distillation, but Hals and Gram
(Zbid..,1909, 70, 310) obtained 1.3 per cent. The discrepancy is
probably due to differences in the experimental conditions, since it
seems likely that the volatile oil is a product of enzyme action upon a
glucoside as it is in the casc of mustard seed.
1883, 28,
Numerous investigations (Dirks, Laadw., Ve~s.-Stat.,
179 ; Koeser, A i i r d y ~ t , 1902, 27 ; 197 ; Joergensen, La~zdw.Vws.Stat., 1910, 52, I ; Bruoux, R ~ LChim.
.
Amdvt., 1912, 17, 6 ; Raquet,
Rc'pzvt. 1-harm., 1912 (3) 24, 145 ; Wehrusen, Wagner, Brannvvorth
~
1,915, 253, (4), 306 ; Vichoever, Jorash
and Neyer, A Y C Phnrm.,
Clevenger and Wing, J. A p z c . Kcs., 1920, 20, 117) have been made
upon the yieid of volatile oil from mustard and rape seeds, but the
results are not in complete agreement. The use of antiseptics such
as thymoi aad sodium fiuoride is stated to be beneficial and the
addition of alcohol appears to improve the yield.
We bave carried out experiments npon two lots of seed, the first
grown in 1924 and consequently only a few months old, the second
grown in 19x9. The results obtained are shown in Table X.
In all these experiments except when otherwise noted, the
powdere2 seeds (roo or 2 0 0 gms.) were shaken with ten times their
wLght of water ax s f 0 and the mixture steam-distilled. I t will be
noticed that the old seeds gave a conslderabiy higher yield of oil than
:he new. that the optimum time of maceration was about two hours,
TABLE X.
i
6
i
1
2.0
I
20
10
1
2.0
6
0.10
1
Seeds g m m d with water,
010
0.25
! Tbis
f o l l o ~ d n gexperiments
rrith 1919 crop.
0.44
, Liquid separated as In 4. 5 u t gave no oil.
0.0
1
r.ld
I
13
::
1
2
'
2
2
2.0
i
2 0
03
(
Solid and liquid s e p a ~ a t e dbefore steamdzstil!ation ; liquid gave no 011.
0'0
3 parts
made
by weight of crater x e d .
,,
that a considerable volume of water was necessary and that the addition
of sodium fluoride improved the yield. As the experiments were
designed to find out suitable conditions for obtaining a good pie!d of
oil and not for a detailed examination of the process of hydiolysis, a
more comprehensive series was not carried out. The method finally
adopted was as follows :200 grams of powdered seeds (1919 crop) were placed in a fivelitre Aask with 2 litres of water containing 2 grams of sodium fluoride.
T h e mixture mas kept at 3j0 for 2 hours, shaken from ti-e to tlme
and then subjected to steam-distillation from the same flask. About
two litres of the distillate mere collected ; the ciistillare was quite
turbid containing oil droplets which settled on allowing the distillate
to stand overnight. it was found necessar] to coilect about 2 litres
of the distillate as the oil does not seem to be easily volatile with
steam. This was also noticed by Hals and Gram (doc. tit.). The
distillate mas saturated with salt and extracted twice with ether. T h e
etilereai soiution of the volatile oil was then dried over anhydrous
sodium sniphare a11d the solvent distilled until about 30 cc. of the
solution were left; it was then poured into a weighed dish and the
solvent allowed to evaporate spontaneously, the last traces of the
solvent being removed by placing the dish in a desiccator and
evacuating the latter from time to time until there was no further
loss in the weight of volati!e oil. Several such experiments were
made and the volatile oil (about 15 grams) collected. T h e mean
yield was 0.80 per cent. whereas Hals and Gram (bc.cit.) obtained
1.30 per cent. and Cartel ( A m . FnZsif., 1909, 2 1 5 ) 0'56 per cent.
T o purify the oil, it was subjected to fractional distillation
under reduced pressure. On re-fractionating the second and third
fractions, an almost constant boiling liquid was obtained weighing
6.0 gms. The results are given in Table XI.
TABLE XI.
Fya&izal Distiltation o f Crude Volatile Oil w d e y 9 mm.
Fraction No.
I
(Tempersrure *C.nncorr.)
/
Weight in grams
1
Residue I.
A
B
C
Residue 11.
Fraction B boiled mainly at 139' and was used for analysis.
T h e oil i s a pale, mobile liquid, has a sharp burning taste, an
irritating effect on the skin and a deep penetrating odour. It contains
nitrogen and sulphur. It decolorises bromine in chloroform and
potassium permanganate in acetone solution. I t is insoluble in water,
but dissolves readily in most organic solvents. I t is optically inactive
(chloroform was used as a solvent). It has been mentioned (page 60)
that Xais and Gram (doc. c2.j isolated the volatile oil and partially
examined it. Their analjitical data are given in Table XII. T h e
authors distilled the seeds and received the distillzte in aln~noniac~l
alcohol which on evaporation left a semi-solid residue, r -48 per cent.
of the untreated seeds. T h e residue was not bomoge~eoussince, 012
treating with alcohol, a p o r t i o ~remained
~
insoluble. This I-esidue was
almost compictely s o l ~ ~ b iine hob water. The yield of oil soluble
in alcohol was 1.3 pcr cent. of the seeds. The authors came to
the conclusion that the nature of the volatile oil could not be
ascertained from their analytical results since the nitrogen and sulphurcontent were lower than :hey ale in the case of the essential oil from
rape seed and the oils have different properties.
Totai Distillate.
1. 48 per cent.
Alcolml insoluble
0.18 per cent.
Nitrogen 15-10
Sulphur 1%32
Nitrogen
Sulphur
15'51
17'24
Alcohol solnble
P30 per cent.
Nitrogen 8'70
Sulphur 20.37
In all determinations, the nitrogen percentage was found by
Kjeldahl's method and the sulphur percentage by oxidation with
alkaline permanganate and precipitation with bari~unchloride.
T h e nitrogen and sulphur in the oil were determined by Kjeldahl's
and Dumas' methods and by Carius' and i'dcssinger's method of
oxidation by alkaline permanganate (Bev., 1888, 21, 2914). The
percentage of nitrogen by Dumas' method was 10.0 and by Kjeldahl's
method distinctly lower. The sulphur determined by permanganate
oxidation was 18.5 per cent. and by Carius' method 37'6 per cent., the
figures in both cases being the meanof several concordant results.
The explanation is not apparent, particularly as it seems probable that
the compound contains 3 atoms of sulphur and hence the oxidation of
half the sulphur is unlikely.
T h e values found are not in any simple molecular proportion and
it is evident that the oil was impure. The quantity was too small to
allow of further purification and consequently no accurate analytical
results are available.
6
The oil does not react with bases as readily as the essential oils
from mustard and rape seeds. No compounds could be obtained
with ammonia or aniline, but with iienzylamine on warnling for two
hours on the water-bath in a well-stoppered tube a solid product was
obtained which, when washed with petroleum ether and crystallised
T h e melting point was
from 70 per cent. alcohol melted a i 83-84'.
unchanged on rccrysta!lisation. The subslance was analgsed and the
weight determined by K. Kast's method (Found : C,58.2 ;
H , 6 7 ; N, 11.0; S, 25.3; M.W. 388. Caic.ior C1,11, N,S,;
C, 58.3 ; H, 6.4 ; N, 10.7 ; S,24.6 ; M. W. 391).
If this is correct, the lormuia for the oil should be C, H, N,S,.
A compound with this formula is known, phenyldimethylethylthio~ransulphide (Braun and Stechele, Ber., 1903, 36, 2 2 8 2 3 , but this
is a sclid melting at 95O, and is obviously not identical with the compouud from oil of jamba.
Sinalbin and sinigrin have been isolated from white and black
mustard seeds by adding dry oil cake to four times its weight of
boiling alcohol (cf., Allen, Cornme~mzb Analysis, 19t3, 7, 105).
This method was adopted with jamba cake but 110 s o l d could be
obtained.
Jowett's method (Wellcome Chem. Research Labs., Pamphlet
No. 8, rooo) of isolating salinigrin from willow bark was also tried,
but with no more satisfactory results. Since the essential oils of rape
and jamba differ considerably in their properties, it is only to be
expected that the glucosides are of a different nature and the isolation
in the latter case may offer considerably more difficulty.
According to Gadamer (Avch. Plzav~tc.,1899,273, 471) the oil
from Trop~o~uzcni
ilzajm, Linn. consists of nearly pare trierucin. He
found that the oil (iodine value 73-81 which had been extracted iron!
the seeds by ether, solidified on cooling a i d ~vher.pressed between
paper gave a product melting at 30.5" and on hydrolysis yielded erzcic
H e concluded that the oil consists mainly
acid of melting point 34'.
of triesucin as the iodine value of this compound is nor very different
from that of the oil.
As no other data-not
even the yield-regarding the oil appear
to be I-ecorded, we have prepared about 50 grams oi the oil and
have determined some of the common constants, and have also
attempted to prepare from the erucic acid pure behenic acid.
The seed used was imported seed obtained through the Empress
Gardens, Poona, and aiter crushing was extracted with ether. Using
375 grams of seed the weight of oil obtajned was (a) 26 9 and (t)
27.3 grams, or a yicld of only 7.2 per cent. The constants of the oil
are given in Table I, and those of the fatty acids free from nnsaponifiable matter in Table 11.
'TABLE I.
Analytical +da fov Trop~olzunOil.
...
...
Sp. gr. at 15.5
.".
...
Saponification value
Iodine value
Refractive i n d a at 40'
...
Unaaponitiable matter (per cent.)
...
"..
...
..
...
...
...
...
...
...
...
...
0'9092
172.6
77.5
1.4568
1'1
T A B L E II.
Fah!y acids //row Tvop~oluvzOil (fvee from mnsnpo&
fiadle nzattev).
Hehner number
...
..
991
...
...
Iodine value
Mean molecular weight
...
Per cent. solid acid
...
...
Per cent. liqnid acids
Iodine value of solid acids
...
Mean molecnlar aelgbt of solid acids
Iodine value of liquid acids
...
Mean molecular weight of liquid acids
...
...
.
...
...
...
...
...
...
...
...
...
...
72%
312.8
45.i
343
729
330'0
72'6
298'0
The method adopted for separating solid and liquid acids was the
one due to Twitchell and as with the acids from rape, mustard and
jamba seeds erucic acid is present in both the solid and liquid acids.
*rhe oil itself has a distinct green colour, but the fatty acids have
Pt is noticeable that the iodine value of the oil is
higher than tilat of the free latiy acids, whereas the reverse is usually
the case.
value oi the oil proves that it is by no means
The
pure trierucin, as this has a saponification value of 160. T h e iodine
value and ~nolecular weight of the solid acids prove that these are
nearly pure erucic acid as this has an iodine value of 75 and a
weight oL 338. 111 addition to erucic acid thew must be
present small quantities of an acid with a lower molecular weight.
a brown tinge.
It is remarkable that the iodim value of the liquid acids is below
that of emcic acid and as the molecular weight is considerably below
that of erucic, it is probable that the liquid acids contain a certain
amount of a saturated acid of low molecular weight.
The quantitj- of oil was not suificient to nlake an exhaustive
study of it, as our main object was to isolate trierucin and reduce it
to tribehcnin.
Tricvmziz.-When piaced in ice-coid water the oii solidified and
was then filtered with the aid of a Buchner funnel, a small amount of
green oil was collected and the solid 1eIt on the funnel melted at 25'.
It was crystallized from light petroleum (b. p. 40-60") ; by cooling
the solution with a freezingmixture, the melting point rose to 28",
but the yield mas only 35 per cent, of the original oil, and the melting
point was not raised by further crystallisation from the same solvent ;
but o c crystallisation from a mixture of benzene and chloroform the
melting point rose to 30.5-31.0" corresponding with the melting
point given by Reimer and Will (Bev., 1887, 20, ~ $ 6 ) . T h i s was
hydrogenated at 180' by means of 5 per cent. of nickel deposited on
kieselguhr and trihehenin melting at 80-81" isolated. After three
crystallisations irom benzene and subsequent crystallisation from
chloroform the melting point remained constant a t 81.0-81.5" and
when saponified with alcoholic potash the glyceride gave a behenic
acid melting at j9.j-80.0' and the melting point was not raised by
t-so crystallisations from alcohol.
Theloiiowing data have been obtained for the two glycerides
together with some values for a sample of tribrassidin
from
kierucir, by the action of nitrous acid :Trie~z~ci~z. T~ib~axsidia.
A s already pointed out (pp. 40, 49, 58 and 66) the behenic acid
obtained from the hardened oils of rape, mustard, jnmba and T~opmZez~m
ma+
seeds melts s t 79.3'-79.8O:
This does not agree with the melting
point given by recent author[tles and attempts have therefore been
made to prepare behenic acid by different methods and see whether
the melting points of different specimens of the acid melted n ~ u c h
above 79*S0.
~,
60, 271) gave the melting
As early as 1846 Walter ( A i ~ m z L e r1846,
point at 52-5 jU,hut in 1S48 Volcker ( I b d . , 1848, 64, 342) gave the
value 76' and this was confirmed by Goldschmidt (K%n Acaid., 1874,
70, 451). In 1Sg4 Talanzeff (J. pr. Chefit., 1894, [Ii] , 50, 72) prepared
the acid by the addition of hydrogen iodide to erucic acid and
subsequent reduction in alcoholic solution with zinc and hydrogen
chloride. The melting point is given as 8q0 in a capillary tube.
Meyer, Brod and Skita (Munalsh., 1913, 34, I 128) prepared behenic
acid by reducing pure erucic acid with hydrogen in the presence of
nickel as catalyst and gave the melting point at 82-84", but state that
with a catalyst which has been used several times a product melting at
7s0 is obtained and that it is impossible to get a product melting at 84"
from this. Flecker and Taylor (1.Chem. Suc., 1922, 121, 1 1 0 2 j
used an acid melting at 91-82' and prepared by reducing erucic
acid and subsequent crystallisation from alcohol, and Toyanla (1. Chem.
Zxd. Jnpnn, 1922, 25, 1053) describes an acid obtained by the same
method as melting at 81-8z0.
It is thus clear that the acid obtained iron: the seeds melts at
a lower temperature than the acid prepared by the reduction of
erucic acid, although the values given for this reduced acid vary
from 81 -84O.
Behenic acid has been prepared by the following methods :The reduction of brassidic acid melting at 59.5-60'0° and
I.
with an equivalent weight 337 (theory) but unknown iodine value by
three per cent. of nickel deposited on kieselguhr. After four hours at
180" the iodine value fell to 4 and the product was repeatedly crpstallised from acetone. After two crystallisations the melting point was
79-3-79-8" and remained constant after further crystallisation.
Kahlbaum's pure erucic acid was twice crystallised from
2.
95 per cent. alcohol and was once crystallised by cooling to rgo to
remove saturated acids. About i o per cent. oi the product was
precipitated as lithium salt, also with the object of removing s a t u ~ a t e d
acids, and the residue was crystallised again from alcohol. T h e acid
melted at 33.5-34.0" and had an equivalent 337. I t was reduced at
180" using a nlckel catalyst and on crystallisatron frorn acetone save a
product nlelting at 79.3-79'80
3. Attempts were made to obtain from Kahlbaum's acid a
product with the correct iodine value (7j) by Holde's (Ar~gew.
Chn,~.,1 9 2 2 , 35, 290) method of fractional precipitation with lithium
acetate. The iodine value obtained for the more soluble portions
was only 71-5, and the behenic acid obtained from this by hydrogenation melted a t 79.3-79'8".
a. -4 fresh specimen of erucic acid obtained from Kahlbaum
had the iodine value 7 5.4 and the solidifying point 33.0:
T h i s was
completely hardened, extracted with petloleunl ether to remove nickel
soap and the acid crystallised from 95 per cent. alcohol and acetone
in succession. I t melted at 80° and the solidifying point was
79'2".
5. A sample of brassidic acid with an iodine value 73.8
(theory 75) and a melting point 59.5-60.0" gave on reduction an acid
melting at 79-3-79.8' after two crystallisations frorn acetone.
6. As already stated, the acid obtained by hydrogenating
trierucin to tribehenin, and subsequent hydrolysis melted at
79'5-80.0".
With
N.R. D a d .
T h e acids used were Kahlbaum's pure stearic acid recrystallised
several times from alcohol and melting at 69.8O, and a bebenic acid
obtained by the reduction of erucic acid and melting at 79.8'. The
methyl esters were prepared from these acids by the Pischer-Speyer
method and melted respectivelv at q2" and ?So. The values obtained
are given in Tables I A d I 1 a& th&e valu& are given as graphs in
Figs. 11 and III.
TABLE 1.
Solidifying points of Mizfiwes ofMethyd Sfenmte and Methyl Behenate.
--.
Mixture No.
/
methyl
behenate.
/
centigade.
'
behenic acid
Begrees C .
Melting pomt In
degrees C.
~
I:
0'0
I:,
I
2,
8.9
1.1
x,
1P1
Y,
203
x,
28.0
Z,
91.4
i
1
69'2
65'4
644
63.6
I
68'8
67'2
85.4
64'6
83'4
1i
62.8
618
1
62 6
77.0
The mixtures designated with the letter Z were examined by
Mr. Nirchandani using different preparations of the acids. T h e
agreement between the two sets of values is satisfactory.
.
Department' o f Gmeral arrd Organis Chemistry,
Indian Zmkdtute o f Science,
[Accepted, 5-r-25.1
Bangabre.
MnteQ *;dl uubllshd by
George Kenneth at the D l w s a n Press, Post Box 4Ss. Madras-1926
C12SZI