Handbook of Solvents - George Wypych - ChemTech - Ventech!

944 Phillip J. Wakelyn, Peter J. Wan
centage of oil extracted from cottonseed flakes was calculated. From these data, Wan et
al. 85 was also able to estimate the initial rate of extraction and final extraction capacity for
each solvent as fresh and at selected initial miscella concentrations up to 30%.
Bull and Hopper 89 conducted extraction of soybean flakes in a stainless steel batch-extraction
apparatus of the Soxhlet type with petroleum solvents, Skellysolve F (boiling
range, 35 to 58 o C) at 28 o C and Skellysolve B (boiling range, 63 to 70 o C) at 40 o C. The extraction
was carried out to permit the miscella obtained by each flooding of the flakes with solvent
to be recovered separately. Their results showed that iodine number decreased and
refractive index increased slightly with the extraction time which implied that more saturated
fat was extracted during later stage of the extraction. Oils extracted during the later
stages of the extraction were found to contain greater amounts of unsaponifiable matter and
were rich in phosphatides, as high as 18% of the last fraction. Skellysolve B which is a hexane
rich solvent demonstrated a much faster initial rate of extraction than that of Skellysolve
F which is a pentane rich solvent and therefore, it took longer to complete the extraction for
Skellysolve F. The fatty acid profile of each fraction showed a slight increase of saturated
and a slight decrease of unsaturated fatty acid in the later fractions.
Arnold and Choudhury 82 reported results derived from a lab scale extraction of soybean
and cottonseed flakes in a tubular percolation extractor at 135-140 o F with pure, high
purity and commercial hexane, and reagent grade benzene. They claimed that pure hexane
extracted soybean slower than high purity and commercial hexane. During the first 60 minutes
of extraction, benzene extracted more oil than the hexanes. However, at the end of 80
minutes, benzene extracted only slightly more than pure hexane but definitely less than the
commercial hexanes. Similar results were obtained for the four solvents when cottonseed
flakes were extracted.
A laboratory extraction study of cottonseed flakes using various hydrocarbon solvents
was reported by Ayers and Dooley. 84 Soxhlet extractor and Waring blender were used for
these experiments. Among the petroleum hydrocarbon solvents tested were branched, normal
and cyclo-paraffins as well as aromatic hydrocarbons with various degrees of purity.
They were: pure grade (99 mole percent purity) n-pentane, isopentane, cyclohexane, benzene,
and n-heptane; technical grade (95 mole percent purity) neohexane, diisopropyl,
2-methylpentane, 3-methylpentane, n-hexane, and methylcyclopentane; technical grade
(90 mole percent purity) cyclopentane; and commercial grade n-heptane, isohexanes,
n-hexane, isoheptane and n-heptane. To assess the performance of these solvents, they used
the following empirical formula:
Quality-Efficiency Rating = 0.4 (Oil Yield Factor) + 0.4 (Refining Loss Factor)
+ 0.2 (Refined and Bleached Oil Color Factor) [14.10.2]
When comparing the oil yield factor alone, 3-methylpentane was rated the best. When
comparing the solvents based on the empirical Quality-Efficiency Rating formula, they
concluded that methylpentanes (3- and 2-methylpentane) were superior extraction solvents
for cottonseed oil. The normal paraffins, highly-branched isohexanes, cycloparaffins, and
aromatics were progressively rated as less efficient than methylpentanes. Therefore, they
recommended a tailor-made solvent for the extraction of cottonseed should exclude aromatic
hydrocarbons, have low limits on cycloparaffin content, and consist largely of normal
and isoparaffin hydrocarbons.