View Article Online / Journal Homepage / Table of Contents for this issue

1994 Dicker and Whiting:

407. Synthetical Studies on Terpenoids. Part I . TJze Synthesis

of Squalene.
By D . W. DICKERand M. C. WHITING.
General problems involved in the synthesis of squalene are discussed,
earlier work is reviewed, and a method is described whereby pure squalene
can be obtained synthetically by a Wittig-Scholkopf reaction between 1 : 4-
dibromobutane and pure geranylacetone, followed by isolation of the all-
Published on 01 January 1958. Downloaded by RMIT Uni on 18/08/2015 15:43:48.

trans-isomeride as its thiourea clathrate.

SQUALENE (I) has assumed increased importance with the demonstration that it is an
intermediate in the biosynthesis of the sterols2 While the mode of cyclisation3 now
seems certain, biochemical work could still be facilitated by a satisfactory total synthesis
of the hydrocarbon. Karrer and Helfenstein’s partial synthesis of the squalene hydro-
chlorides (11) from farnesol was of great importance as a confirmation of the structure (I).
Unfortunately it was misinterpreted a t the time as a synthesis of the hydrocarbon (I),
since the dehydrochlorination product of the chlorides (11) was then believed to be squalene.
It is now realised that the regenerated hydrocarbon, whether derived from natural or
PBr,
H*[CH,*CMe:CH.CH,],*OH +{H*[CH,-CMe:CH*CH,],-Br)

CH:CH, + H*[CH,*CMe:CH*CH,],*[CH2CH:CMe*CHz]3.H’
I
H*[CH,*CMe:CH*CH,],*CMe (-30y0?)Squalene (I)
I
( N 70Yo> C H,*[C H ,*C H: Me*CH ,].H

(11) H.[CH ,*CMeC I*CH ,*C H,],fCH,-CH,*C MeCICH, ],.H

synthetic hydrochlorides, can contain, at the most, 3% of squalene (probably much less).
This hydrocarbon, conveniently called psezcdosqualene,” must consist of a mixture of
“

all the possible products of olefin-forming eliminations, which, including stereoisomers,

number 1275.”
There are ten possible stereoisomers of structure (I). The formation of a thiourea
adduct in high yield suggests that most of the natural hydrocarbon, at least, is the all-
trans-form. This has a shape well suited to complex formation; when Catalin models
are used its dimensions are 34-4 x 6.7 x 4.65 A, which, after correction to van der Waals
radii, become about 35 x 7.3 x 5.1 A, the width of the two farnesyl chains (which are not
quite collinear) being only 6.2 A. For adduct formation Schiessler and Flitter 7 found
* In a compound R-R’, in which R and R may each exist in the same n forms, there exist n forms
in which R = R’; for each form, &, of R, there are ( n - 1) unsymmetrical forms (R-R’) ; but, since
R,, - Rtb= Rb - R’, (when optical isomerism is not involved), the number of unsymmetrical forms is
n ( n - 1)/2, and the total number of forms is n ( n +
l ) / Z . For the farnesyl group (2 x 5 x 5) forms are
possible if equilibration of double-bond isomers is permitted, and (1 x 2 x 2) if only stereoisomerism
is allowed. An approximate calculation of the squalene content of the dehydrochlorination product of
the squalene hexahydrochlonde, based on the results of dehydrobrominating 2-bromo-Qmethylbutane
(Ingold, “ Structure and Mechanism in Organic Chemistry,” Bell, London, 1953, p. 442), gave 0.14%.
Langdon and Bloch, (a) J . Amer. Chem. Soc., 1952,74, 1869; ( b ) J . Biol. Chem., 1953,200, 135.
Heilbron, Kamm, and Owens, J., 1926, 1630; Channon, Biochem. J.. 1926, 20, 400.
Woodward and Bloch, J . Amer. Chem. SOC.,1953, 75, 2023.
Karrer and Helfenstein, Helv. Chim. A d a , 1931, 14, 7 8 .
Dauben, Bradlow, Freeman, Kritchevsky, and Kirk, J.Amer. Chem. SOC.,1952,74,4321; Tomkins,
Dauben, Sheppard, and Chaikoff, J . Biol. Chem., 1953, 202, 487.
Nicolaides and Laves, J. Amer. Chem. SOC.,1964, 76, 2596.
Schiessler and Flitter, ibid., 1952, 74, 172.
 View Article Online

[19581 Synthetical Studies on Terpenoids. Part I .

that 5.8 x 6.8 A was the optimal molecular cross-section; models of the 6-cis- and the
10-cis-isomer in their most stable conformations are bent, with minimal y-axes of 12.5 and
13 A, respectively. The all-cis-isomer, with a y-axis of 8.4 A, is the most nearly linear
form after the all-trans-isomer, but even this is larger than that of any of the hydrocarbons
found capable of complex-formation. Nicolaides and Laves,6 however, deduced the
all-trans-configuration not from the mere fact that squalene formed a stable thiourea
clathrate, but from the difference in X-ray spacing constants between the clathrates of
squalene and perhydrosqualene. (This was a wise precaution in view of the fact that in
Published on 01 January 1958. Downloaded by RMIT Uni on 18/08/2015 15:43:48.

urea clathrates, at least, a molecule possessing a long chain of the correct dimensions may
be able to form a complex despite the presence of a " bulge '' over a short distance.') The
numerous postulated * modes of cyclisation of squalene to a variety of natural terpenoids
all require an all-trans-configuration ; indeed the only real question remaining is whether
any other isomer exists in the natural hydrocarbon and, if so, whether it has any biogenetic
importance. Although up to a third of commercial " 90% squalene " fails to form an
adduct-the yield of clathrate from squalene already purified by this method exceeds
SO% under similar conditions-the non-complexed material may well consist largely of
artefacts. Accordingly it seems permissible to d e j n e squalene as the all-trans-form of (I),
and this definition is, in fact, implicit in some recent discussions.
Me,C:CHCH,CH:CMe.CH:CH,, etc.

Me,C:CH*CH,CH,CMe:CH.CH,.OH
trans
+ SOCI, 1- 1- Me,C:CH*CH,*CH,CMe:CH.CH,CI (111)

Cyclic chlorides? 1
,Me,C:CHCH,.CH,*CMeClCH:CH, (IV)

liHeat
Me,C:CH *CH , C H ,.C Me:C H C H ,C I (V)
cis

(111) __t (VI)

Me&: CH*CH,-CH,*CMe:CH*CH2*CH(C0,H)s
trans

(VII I) Me&: CH *CH,*CH,*CMe*CH,*CH ,*CH , Me,C:CH*CH2~CH2*CMe:CH*CH2*CH2*CO2H(VII)

I
0
I
co
trans

11
Me,C:CH~CH,*CH:CMe*CH,CH,~CH,*CO,H, etc.

The mixture of hydrocarbons from which the squalene hydrochlorides were first
prepared synthetically may have been quite rich in squalene. If the (naturally derived)
farnesol used was the pure tram,trarts-isomer-a point on which little definite evidence
has been published-and if configuration, structural and geometrical, at the three double
bonds were quantitatively retained, the C, fraction might have contained as much as 30%
of squalene, according to the results of analogous coupling experiments on the lower
isoprenoid homol~gues.~On the other hand the hydrocarbons prepared by Schmitt l o
and by Farmer and Sutton,ll and examined by Dauben and Bradlow,12 must have been
mixtures of-on the most favourable assumptions-fifteen isomers, and their squalene
Eschenmoser, Ruzicka, Jeger, and Arigoni, Helv. Chim. Acta, 1955, 38, 1890.
Barnard and Bateman, J., 1950, 932.
l o Schmitt, Annalen, 1941, 547, 115.
11 Farmer and Sutton, J., 1942, 116.
l2 Dauben and Bradlow, J . Amer. Chem. SOC.,1952, 74, 5204.
 View Article Online

1996 Dicker and Whiting:

content must have been, and indeed demonstrably was,lbvery low. Trippett's synthesis,18
similar to that described below, was thus the first preparation of a mixture of squalene
stereoisomers almost uncontaminated with prototropic isomers, although as is shown
below the all-tram content is unlikely to have greatly exceeded 25%. On the other
hand, the synthesis by Isler, Ruegg, Chopard-dit-Jean, Wagner, and Bernhard,14 which
involved major modifications of the Karrer synthesis and was submitted for publication
very shortly after our own preliminary note,ls did give isolable, all-tram-squalene.
As a starting-point for the synthesis of squalene the crystalline calcium chloride
Published on 01 January 1958. Downloaded by RMIT Uni on 18/08/2015 15:43:48.

complex of geraniol has obvious advantages. Attempts to transform the pure geraniol
obtained on regeneration into reactive derivatives without loss of configuration were not
successful, toluene-jkmlphonyl chloride in pyridine giving either geranyl chloride or the
pyridinium salt rather than the desired toluene-P-sulphonyl ester. The pyridinium
toluenesulphonate proved to be very unstable and unreactive toward sodiomalonic ester.
The use of a " geranyl halide " therefore seemed necessary. As previous work ~ h o w e d , ~
'' geranyl chloride," obtained by conventional methods, is a mixture of the desired chloride
(111) with '' linalyl chloride " (IV), which on isomerisation or by reaction with nucleophilic

Infrared spectra : 1, Squalene (natural, three times regenerated). 2, Squalene (synthetic, twice
regenerated). 3, Stereoisomers of squalene unaffected by thiourea under usual conditions.
n 1

Wove/ength ( p )
Specha were determined on neat liquids in a celZ of about 0.05 mm. path-length; the " percentage transmitted "

values for nos. 3 and 2 have been increased by 10% and 5%, respectively.

reagents would be converted in comparable proportions into derivatives of geraniol and

nerol. Furthermore, Sorensen et al. have shown l6 that on reductive coupling '' geranyl
chloride " gives a mixture containing cyclic hydrocarbons, which in all probability are
formed from cyclic contaminants in the geranyl chloride. Now, in any synthetic prepar-
ation of squalene, it was necessary to minimise the formation of isomeric, including
stereoisomeric, hydrocarbons, which would interfere with the formation of the squalene-
thiourea clathrate ; and the 6-cis-isomer was a particularly dangerous contaminant because
of the possibility that it might form a (less stable) clathrate itself, having, for much of
its length, the correct molecular cross-section. In any synthesis from geranyl chloride, a
further crystalline intermediate was therefore considered essential. Geranylmalonic acid,
not previously described, and geranylacetone semicarbazone were obvious possibilities.
Geranyl chloride was chosen in preference to the bromide because of the probability
that the latter, but not the former, would equilibrate with its anionotropic isomers at room
l3 Trippett, Chem. and Ind., 1956, 60.
l4 Isler, Ruegg, Chopard-dit-Jean, Wagner, and Bernhard, Helv. Chim. Acta, 1956, 39, 879.
l6 Dicker and Whiting, Ckem. and Ind., 1956, 351.
l6 N. A. Sarensen, Gillebo, Holtermann, and J. S. Sarensen, Acta Chem. S c a d , 1951, 5, 767.
 View Article Online

[1958] Synthetical Stadies on Teq5enoids. Part I . 1997

temperature.17 I n the halogenation much hydrocarbon (presumably myrcene and ocimene,
although cyclic products may also be present) is formed concurrently; to separate the
chloride, distillation has been customary. Indeed Barnard and Bateman say that, on
heating, " a more homogeneous product " results, implying that the equilibration of the
three chlorides gives a mixture containing less of the tertiary isomer than does the product
of the kinetically controlled reaction. This, however, must mean the formation of a large
proportion-probably nearly 50y0-of the cis-isomer (V). We therefore chose to treat
the crude, undistilled chloride-hydrocarbon mixture with sodio-malonic or -acetoacetic
Published on 01 January 1958. Downloaded by RMIT Uni on 18/08/2015 15:43:48.

ester solutions, and distil the stable products (after hydrolysis to geranylacetone in the
latter case).
Geranylmalonic ester has been prepared before, but gave a syrupy acid on
hydrolysis.l&l9 Our specimen was hydrolysed to an acid which solidified on cooling;
about a third could be obtained crystalline, and when pure it melted at 504--51*5". The
low yield from constant-boiling ester reveals the presence of isomeric impurities in the
latter, although no doubt they amounted t o much less than two-thirds of the whole
product ; the non-crystalline acid in the mother-liquors had an infrared spectrum not
greatly different from that of the crystalline product, which at least excluded the presence
of H,C=C< groupings. The crystalline acid absorbed two mols. of hydrogen and gave in
good yield a crystalline perhydro-acid; it is assumed to be the tram-acid (VI).
Decarboxylation gave an 82% yield of a monocarboxylic acid, and a 9% yield of a neutral
isomer which was clearly the 8-lactone (VIII). The acid had m. p. -11" to -lo", raised
inefficiently to a constant -1.5" on recrystallisation. Evidently part of the acidic
product was formed via the lactone, and presumably consisted of up to five isomers. This
complication lowered the overall yield of the pure geranylacetic acid (VII) and, having
obtained poor preliminary results in the reaction with methyl-lithium, we turned our
attention to the known geranylacetone semicarbazone ; however, the acid may prove
useful in other synthetic work.
Despite the precautions taken, geranylacetone, prepared as above and by methods of
Dauben and Bradlow I2 for hydrolysis of the keto-ester, gave a semicarbazone with m. p.
ca. 70°,which was raised to the published value of 96-97' only after about five crystallis-
ations from aqueous ethanol, the best solvent found. The yield, on each recrystallisation,
then reached a constant value, and the melting point also remained constant. Treatment
with 2 : 4-dinitrophenylhydrazine sulphate gave a derivative which crystallised readily
and had m. p. 54-55' (cf. the liquid 2 : 4-dinitrophenylhydraone obtained by
Carroll,2o and by ourselves from the crude ketone *). We therefore assumed that the
semicarbazone was in fact homogeneous. The yield was only 48-62y0 from the crude
ketone, whereas regeneration gave a ketone which was reconverted into a semicarbazone,
m. p. 96-97", after only one crystallisation, in high yield. Again, it is clear that the
crude ketone from geranyl chloride, despite the a priori reasonable methods used, was far
from pure. Since the completion of the present work Stadler et aL21 have reached a
similar conclusion regarding " geranylacetone '' prepared via the bromide, and have
described a method of purification by rigorous fractional distillation.
Preliminary attempts to convert the regenerated tram-ketone into farnesic acid via the
ethoxyacetylenic alcohol were discouraging in that the product apparently underwent
partial cyclisation on acid-catalysed rearrangement ; and neither the crude ethyl farnesate,
* Naves (Helv. CAim. Acta, 1949, 32, 1801) obtained a 2 : 4-dinitrophenylhydra~one~m. p. 72-73',
from a " geranylacetone " obtained by a reverse aldol fission of natural farnesal. This suggests the
presence of a cis-linkage in the latter.
l7 See, inter al., Winstein and Young, J. Amer. Chew. SOL, 1936, 58, 104.
l8 Dupont and Labaune, Wiss. und Indzsstrie Ber., Rour-Bertrand Fils, 1911, (111), 3, 3; Chem.
Zentralblatt, 1911 (11), 15, 138.
lS Forster and Cardwell, J., 1913, 1338.
2 o Carroll, J., 1940, 704.
21 Stadler, Nechvatal, Frey, and Eschenmoser, Helv. Chiin. Acta, 1957, 40,1373.
 View Article Online

1998 Dicker and Whiting:

nor the derived farnesol, formed a thiourea clathrate. Attention was turned to the possible
use of the Wittig reagent from 1 : 4-dibromobutane to give squalene directly. Triphenyl-
phosphine and the dihalide gave a good yield of the 4-bromobutyltriphenylphosphoniuni
bromide in boiling butan-2-one, and this reacted with another molecule of triphenyl-
phosphine in boiling cyclohexanone, giving a salt of which the dihydrate, m. p. 146-149",
was converted with some difficulty into an anhydrous product, m. p. 300-308". Addition
of ethereal butyl-lithium solution to an excess of the salt gave only a transient yellow
coloration; but when the salt was added gradually to the butyl-lithium solution (in
Published on 01 January 1958. Downloaded by RMIT Uni on 18/08/2015 15:43:48.

tetrahydrofuran) it dissolved almost completely, and a deep red solution of the Wittig
reagent resulted. Addition of the ketone decolorised the solution with the formation of
a precipitate, which was decomposed under nitrogen at 60". Extraction with light
petroleum and chromatographic purification gave a hydrocarbon (15% yield) of which the
infrared spectrum was identical with squalene.
Trippett's brief description l3 of the reaction between geranylacetone and 1 : 4-di-
bromobutane appeared when the Wittig reaction was under investigation in these
laboratories, and his use of tetrahydrofuran as solvent prompted us to employ it.
The hydrocarbon obtained gave a crystalline precipitate when treated with thiourea
in methanol-benzene, from which the regenerated all-tram hydrocarbon was obtained in
6-13% yield. Treatment of the mother-liquors with further quantities of thiourea failed
to produce more clathrate; on the other hand the once regenerated hydrocarbon gave
the same yield (75430%) in a further purification cycle as did once regenerated natural
squalene in duplicate small-scale experiments. Infrared spectra of the regenerated
synthetic and the natural substance are illustrated, together with that of the hydrocarbon
which refused to form a clathrate. Such spectra are evidently unsuitable for distinguish-
ing between stereoisomers of the polyisoprene type, although they would reveal the presence
of even a small proportion of any prototropic isomer with a >C=CH, grouping.
Dr. N. Nicolaides and Dr. F. Laves have examined the clathrate of synthetic squalene,
and report that the length of the hydrocarbon chain is the same, within experimental
error of & 0.1 A, as that of natural squalene. He was, however, able to prepare, from a
sample of the hydrocarbon which we had tried to free from the all-trans-form by treatment
with thiourea, a certain amount of thiourea clathrate which showed an inclusion chain-
length measurably shorter than that of the &hydrocarbon. This would correspond to
the presence of one or more cis-linkages,G with some distortion of the molecule from its
most stable conformation which is apparently unsuited to clathrate formation.
Although we believe that, in the combination of infrared scrutiny and the formation
and X-ray examination of clathrates, definitive physical methods for the comparison of
stereoisomeric unconj ugated polyenes are at last available, it remained necessary, for
complete satisfaction, to prove the identity of the natural and synthetic hydrocarbons
toward an enzymic system. Dr. K. Bloch, Dr. T. T. Tchen, and Dr. R. K. Maudgal have
now obtained conclusive evidence on this point, which will be described in detail elsewhere.

EXPERIMENTAL
trans-4 : 8-Dinzethylnona-3 : 7-diene-1 : 1-dicarboxylic (" Geranylmalonic ") Acid.-Geraniol
(25 g.; regenerated from its calcium chloride complex), dry pyridine (13.8 g . ) , and dried ether
(100 c.c.) were stirred a t -20" to - 10"while thionyl chloride (22.3 g. ; redistilled over quinoline,
then linseed oil) was added during 1 hr. Dry ether (100 c.c.) was added, and the solution
was stirred overnight a t room temperature. The supernatant liquid was decanted from the
pyridine hydrochloride, which was washed with ether (2 x 50 c.c.), and the ethereal solution
was washed successively with much water, sodium hydrogen carbonate solution t o neutrality,
dilute sulphuric acid until no more pyridine was removed, sodium hydrogen carbonate solution,
and water, and dried (Na,SO,). Evaporation a t 10 mm. gave a liquid (25.0 g.) which showed
bands a t 835, 895, 925, 990 cm.-l attributable respectively to >C=CH-, unconjugated vinyl,
conjugated vinyl, and unconjugated vinyl groupings. This was added dropwise to a solution of
 View Article Online

[1958] Synthetical Studies on Terjbenoids. Part I .

the sodium derivative of ethyl malonate, prepared from sodium (6.7 g.) and the ester (57.6 g.)
in dry ethanol (200 c.c.). The mixture was heated under reflux during 6 hr. with mechanical
stirring. Water (400 c.c.) was added and the neutral fraction was isolated with ether. After
removal of the solvent, distillation from a Kon flask gave (a) ethyl malonate, (b) a mixture (of
terpene hydrocarbons, geraniol, and C,, ethyl ethers), b. p. 90-97"/0.15 mm., and (c) the
required ester (17.5 g.), b. p. 97-100"/0~15 mm., ng 1-4610 (Barnard and Bateman give b. p.
126-129"/0-2 mm., n g 1.4610).
This was heated with 10% potassium hydroxide solution (150 c.c.) under reflux for
15 hr., the upper layer disappearing. Isolation of the acidic fraction gave a syrup (12.7 g.),
Published on 01 January 1958. Downloaded by RMIT Uni on 18/08/2015 15:43:48.

which crystallised partly a t -78". A solution in methylene dichloride (30 c.c.) was held a t
-78' for 18 hr., and the liquid was removed as completely as possible from the separated solid.
After repetition of this process (necessarily inefficient because of the fibrous crystals obtained)
from 30 c.c., then from 10 c.c., of methylene dichloride the product (5.0 g.; m. p. 41-45") wits
crystallised from light petroleum containing a little ether a t O', giving 4-0 g. of the essentially
pure acid, m. p. 46-48", finally raised to 50.5-51.5" on further crystallisation (Found: C,
64.75; H, 8.25. C13H2004 requires C, 65.0; H, 8.35%). (The overall yield of essentially pure
acid from geraniol varied from 11 to 13yo.)
4 : 8-Dimethylnonane-1 : 1-dicarboxylic A cid.-Crystalline geranylmalonic acid (65 mg.) was
hydrogenated in glacial acetic acid (10 ml.) in the presence of palladium-calcium carbonate
(35 mg.) for 30 min. at 20' [uptake 2.3 mols.]. After filtration and evaporation the product
solidified. Crystallisation from pentane a t -78" gave the acid, m. p. 56-57' (Found: C,
63-85; H, 10.05. Cl3H2,O4requires C, 63.95; H, 9.85%).
trans-5 : 9-Dimethyldeca-4: 8-dienoic (Geranylacetic) A cid.-Geranylmalonic acid (6.6 g. ;
m. p. 46-49") was heated in nitrogen a t 0.05 mm.; decomposition began a t 145" and was
completed by raising the bath-temperature during 1 hr. to 165'. The distillate ( 4 4 g.) was
dissolved in ether and shaken with sodium carbonate solution. The latter was acidified and
the acidic fraction was isolated with ether, giving, after removal of solvent a t 0.05 mm., the
crude acid (4.45 g., 83y0), m. p. -11' to -lo", raised economically to -7" on one crystallis-
ation from pentane, or in ca. 25% yield after five crystallisations and redistillation to -1.5"
p. p. 115' (bath-temp.)/0-05 mm., ng 1.47181. The S-benzylisothiztroniurnsalt formed plates,
m. p. 128-129", from aqueous ethanol (Found: N, 7.40. C,,H3,0,N,S requires N, 7.73%).
Distillation of the neutral fraction gave 5-hydroxy-5 : 9-dimethyldec-8-enoic lactone (450 mg.),
b. p. 100-120" (bath-temp.)/O-01 mm. (Found: C, 73.0; H, 10.1. C,,H,,O, requires C,
73-45; H, 10.2y0).
Similar decarboxylation of the syrup from the geranylmalonic acid mother-liquors (7-16 g.)
gave a monocarboxylic fraction (3-6 g., 61%) which failed to solidify even a t -78' or on treat-
ment with light petroleum at low temperatures and seeding. Its S-benzylisothiuronium salt
had m. p. 119-125" initially, raised to 124-131" after four crystallisations, but still evidently
heterogeneous.
trans-6 : 10-Dimethylundeca-5: 9-dien-2-one (' ' Geranylacetone ").-The crude geranyl chloride
(211 g.) obtained from geraniol (200 g.) was added to a solution prepared from sodium (36.8 g.)
and ethyl acetoacetate (273 g.) in dried ethanol (750 c.c.), and the mixture was heated under
reflux with mechanical stirring for 6 hr. Water (3 1.) was added, and the neutral fraction was
isolated with ether. After removal of the solvent a t 10 mm., the residue (279 g.) was added to a
solution of sodium hydroxide (70 g.) in water (2 1.) and ethanol (3 1.). The mixture was heated
under reflux for 48 hr., water (5 1.) was added, and the neutral fraction was isolated with ether.
Addition of benzene and removal of solvents a t 10 mm., followed by distillation from a Kon
flask, gave a series of fractions (45.5 g.), b. p. 53-61"/0.1 mm., n: 1.4596-1.4640, which were
rejected, followed by others (107.5 g.), b. p. 62-68', n'," 1-4655-1-4686, which were added to
a solution of semicarbazide hydrochloride (67 g.) and sodium acetate trihydrate (100 g.) in
water (650 c.c.) and ethanol (750 c.c.). The solution was warmed a t 50" for 30 min. and cooled
to 0"; the precipitated semicarbazone was collected, then crystallised five times from 1 : 1 v/v
water-ethanol (10 C.C. per g.). The recovery rose gradually from 84% to a constant 94%, the
crystals, originally small and indefinite, became large, transparent plates, and the m. p. rose
from 85-94" to 96.5-97.5" (yield 66 g.) (Forster and Cardwell 1, give m. p. 97").
The sernicarbazone (66 g.) dissolved when shaken with 2~-sulphuncacid (1.5 1.) and light
petroleum (b. p. 40-60"; 1-5 1.) for 24 hr. a t -22'. Evaporation of the solvent and distillation
gave the trans-ketone (48-2 g., 19% from geraniol), b. p. 60-61°/0.05 mm., n z 1.4661. nl,s
 View Article Online

Synthetical Studies on Tm$emids. Part I .

1.4688 (Found: C, 80.35; H, 11.2. Calc. for C,,H,,O: C, 80-3; H, 11.4%). It gave a semi-
carbazone which after one crystallisation had m. p. 96-97'. The 2 : 4-dinitvophenyZhydrazone
formed orange plates, m. p. 64-55', from ethanol (Found: C, 61.2; H, 7-05. C,,H,,O,N,
requires C, 60-9; H, 7.0y0).
4-BromobutyEtvi~henylphos~honium Bromide.-Triphenylphosphine (78 g.), 1 : 4-dibromo-
butane (43 g.), and butan-2-one (300 c.c.) were heated under reflux for 12 hr. After cooling,
the product (93 g., 98%) was collected, washed with ether, and dried. It formed plates, m. p.
207-212', from water (Found: C, 56.3; H, 4-7. C,,H,,PBr, requires C, 55-25; H, 4.8%).
Tetramethylene-1 : 4-bisfri~henyZ~ho~phonium Bvomide.-The above salt (103 g . ) , triphenyl-
Published on 01 January 1958. Downloaded by RMIT Uni on 18/08/2015 15:43:48.

phosphine (173 g.), and cydohexanone (1600 c.c.) were heated under refhx for 48 hr. On
cooling, two layers formed; the lower layer was dissolved in hot water (200 c.c.), the hydrated
salt (84 g.), m. p. 146-149" with resolidification, separating on cooling. Further quantities of
hydrate separated from the cyclohexanone layer, bringing the total yield t o 122 g., i.e., 77% of
the dihydrate. Prolonged drying of the finely ground hydrate a t 100"/0~06 mm. over phosphoric
oxide gave the anhydrous salt, m. p. 300-308".
a&trans-2 : 6 : 10 : 15 : 19 : 23-HexamethyEtetracosa-2 : 6 : 10 : 14 : 18 : 22-hexaena (Squalene).
-A solution of butyl-lithium in ether was prepared by the action of n-butyl bromide on lithium
(6.45 g.) at -lo", and estimated, after filtration through glass wool into a flask filled with
nitrogen, by the double-titration method.Pa A portion (46 c.c., containing 0.062 mole) was
added to purified tetrahydrofuran (100 c.c.), and the mixture was stirred under nitrogen while
the finely powdered dibromide (22.9 g., 0.031 mole) was added during 10 min. The solution
became deep red, but some of the saIt remained undissolved. trans-6 : 10-DimethyIundeca-
5 : 9-dien-2-one (12.0 g.) in anhydrous tetrahydrofuran (50 c.c.) was at once added dropwise;
a white precipitate was formed, the solution becoming almost colourless. The mixture was
heated under reflux for 1hr., the solvent removed a t 10 mm., and the semi-solid residue heated
in nitrogen for 5 hr. at 60'. The resultant material was extracted with light petroleum (b. p.
60-80°; 5 x 100 c.c.), the extract was concentrated, and insoluble matter was removed by
centrifugation. Evaporation gave an oil (11.2 g.) which was chromatographed on alumina
(500 g.; grade '' H ") with light petroleum (b. p. 60-80°) as eluant. The hydrocarbon fraction
so obtained (1.95 g., 15%) had an infrared spectrum indistinguishable from that of squalene;
yields varied from 13 to 22%.
The hydrocarbon mixture in benzene (24 c.c.) was added in one portion to a saturated
solution of thiourea in methanol (160 c.c.) with swirling. After 1 hr. a t 0" the adduct (1.50 g.)
formed fine needles; a further crop (0.97 g.) separated overnight. The first crop was
decomposed by shaking i t with water (50 c.c.) and light petroleum @. p. 40-60'; 60 c.c.), and
the extract was washed with water, dried (MgSO,), and evaporated, giving the hydrocarbon
(122 mg., 6.3%) (in other experiments yields, based on the hydrocarbon mixture, were 104-
13%) (Found: C,87.8; H, 12-2. Calc. for C,,H,,: C, 87-7;H, 12.3%).
Mixture of the AlO-cis- and the A10 :14-Di-cis-isomer.-Recovery of the hydrocarbon by
dilution of the methanol-benzene-thiourea solution with water and extraction with light
petroleum gave a,n oil (1.7 g.) which was again treated with benzene (20 c.c.) and thiourea-
methanol (135 c.c.), giving a slight precipitate which was removed. Re-isolation and re-chrom-
atography gave a colourless liquid (1.40 g.) with the infrared spectrum illustrated (Found: C,
87.6; H, 12.1%).
Infrared spectra were determined with a Perkin-Elmer Model 21 spectrophotometer with
sodium chloride optics.

This work was carried out during the tenure of a Maintenance Grant from the Department
of Scientific and Industrial Research by one of us.
THE DYSONPERRINS OXFORDUNIVERSITY.
LABORATORY, [Received, December 17th, 2957.1
22 Gilman and Hanbein, J . Amer. Chem. Soc., 1944, 66, 1615.