(191o) Optimization of Supercritical CO2 Extraction to Maintain the Ratio of ?-6 and ?-3 Fatty Acid from Hemp Oil

Optimization
of Supercritical CO₂ Extraction to Maintain the Ratio of ω-6
and ω-3 Fatty Acid from Hemp Oil along with Cross-validation

Vibha Devi¹, Shabina khanam².
Department of Chemical Engineering, Indian Institute of Technology Roorkee, Roorkee
247667 Uttrakhand, India, ¹E-mail:
vibha.info100@gmail.com. ²E-mail: shabifch@iitr.ac.in. 

Abstract:
Hemp
(Cannabis sativa) possesses a rich content of ω-6 linoleic and ω-3
α-linolenic essential fatty acid. It is the exclusive plant possessing the
content of ω-6 and ω-3 in the ratio of 3:1, which is a rare and most
desired ratio that enhances the quality of hemp oil, which is very essential
for vegetarian diets. This ratio is beneficial for the development of cell and
body growth for toddlers, strengthens the immune system, possesses
anti-inflammatory action, lowering the risk of heart problem owing to its
anti-clotting property and a remedy for arthritis and various disorders [1, 2].
Thus, the study proposed its contribution to maintain a human health offering balanced
ratio of ω-6/ω-3. To achieve this ratio, the study
employs supercritical CO₂ extraction (SFE) approach on hemp seed, owing
to its effluent restricted approach. Gas chromatography was carried out to
determine ω-6 the ω-3 fatty acid content. As suggested through
central composite design, the study experimented 32 combinations of SFE
parameters; temperature (from 40 to 80) °C, pressure (from 200 to 350) bar,
flow rate (from 5 to 15) g/min, particle size (from 0.430 to 1.015) mm and
amount of co-solvent (from 0 to 10) % of solvent flow rate. Tolerance limit was
defined as ±0.2. The ratio obtained through SFE found to be in the range of
2-3:1, depending on parameter conditions. Interaction terms were significantly
(P <0.05) influenced on the ratio of ω-6/ω-3. Based on obtained
ratio, a quadratic model was developed to predict ω-6/ω-3 ratio. Close
values of R-square (0.9539), adjusted R-squared (0.8701), predicted R-squared
(0.8109) and insignificant lack of fit indicates excellent fit of developed
model. Best operating conditions for optimized ω-6/ω-3 ratio, i.e.
3.03:1 were achieved at temperature 40 °C, pressure 350 bar,
flow rate 5 g/min, particle size 0.43 mm and without co-solvent.  

Further,
large number of SFE process parameters i.e. temperature, pressure, flow rate,
co-solvent, particle size, time etc. provides a large number of datasets. Therefore,
the present study recommends the application of resampling techniques for
cross-validation to assure the authenticity of obtained results.
Cross-validation refits the model on each data to achieve the information
regarding the error, variability, deviation etc. in the data. Bootstrap and
jackknife are the most popular resampling techniques, which create a large
number of data through resampling from the original dataset and analyze these
data to check the validity of the obtained data [3, 4, 5].
Jackknife is the resampling procedure, which is based on eliminating one
observation from the original sample of size N without replacement [3].
Bootstrap, proposed by Efron [6], is the frequently
used statistical approach for estimating the sampling distribution of an estimator
by resampling with replacement from the original sample [4].

For
bootstrap resampling, the sample size is 32, which is the number of
observations. Number of repetitions was selected as 1000 as per the literature [7].
However, for jackknife resampling, the sample size is 31 (eliminating one
observation), which is repeated by 32 times. Estimands for these resampling
techniques are considered as mean, variance, standard deviation, variation
coefficient and standard error of the mean. Significance level was considered
as 5 %. Mean reflect the central value of the data point while the standard
deviation represents the dispersion of the data from its central value. Using
experimental data of ω-6/ω-3 ratio (true value 2.619), the mean value
for bootstrap and jackknife resampling was observed as 2.621 and 2.619,
respectively, which is the average of the sample mean of all data points.
Variance exhibits the spread out of the data from its mean [8]. The variance of
0.052 (true sample), 0.050 (bootstrap resampling) and 0.052 (jackknife resampling)
indicates that the data points are close to the mean value. Variance
coefficient was found to be low as 0.087 (bootstrap resampling) and 0.087 (jackknife
resampling), which is desirable. Similarly, standard deviation exhibits how
much spread (deviation) of the data points from its mean exists. Standard
deviation of 0.17 and 0.0171 indicates the dispersion of data points over the
range of values [8]. Further, standard error of the mean shows the deviation of
sample mean from its actual mean of all data points. Small errors of 0.041 and 0.042
reflect the accuracy of the sample for prediction. All the estimators’ values
of variance, standard deviation, variance coefficient and standard error of the
mean were found within the lower and upper bound of simple percentile interval,
which is in the 95 % of confidence interval [4]. Thus, small standard error,
variance, variance coefficient and standard deviation confirm the validity of
the data points, which minimizes the likelihoods of error in the prediction.

Keywords

Supercritical extraction, Central composite design, hemp (Cannabis
sativa), fatty acid, ω ratio

Graphical abstract:

References:

[1]  Defence
JL, Pate DW. Hemp seed oil: a source of valuable essential fatty acids. J Int.
Hemp assoc. 1996;3:4–7.

[2]
Borhade SS. Chemical Composition and Characterization
of Hemp (Cannabis sativa)    Seed oil and
essential fatty acids by HPLC Method. Sch. Res. Libr.
Arch. Appl. Sci. Res 2013;5(1):5–8.

[3]
Efron B (1982) The
jackknife, the bootstrap and other resampling plans. Siam. Pennsylvania.

[4]
Efron B, Tibshirani R
(1993) An introduction to the bootstrap. Chapman & Hall/CRC.

[5]
Good P (2006) Resampling methods. A guide to data analysis.
Third Edition. Birkhäuser.

[6]
Efron B (1979) Bootstrap methods: Another look at the
jackknife. Ann Statist 6:1-26.

[7]
Efron B, Gong G (1983) A leisurely look at the
bootstrap, the jackknife, and cross-validation. The American Statistician
37(1):36-48.

[8]
XLSTAT (2015)
http://drjackson.ca/applied_research_methods/xlstat_user_manual.pdf.