The Pareto diagram for the global yield as the response variable
is shown in Fig. 3. The estimated effects and interactions between
the range of variables studied and the analysis of variance are
also shown in Table 2. The degree of significance of each factor
is represented by its p-value; factors with a p-value < 0.05 having
a significant influence on the extraction process for a confidence
level of 0.95. The sign associated with each of the effects indicates
a positive or negative influence on the yield caused by the variable
in question.
Empirical correlations were obtained using the experimental
data and the program Statgraphics Plus 5.1®. These correlations
link the variables with influence on the global extraction yield of
the process using CO2 + 20% ethanol as solvent system.
Analysis of Table 2 and Fig. 3 indicates that the temperature
and the combined interaction of pressure/temperature are
the factors that influence the global extraction yield of the
process using CO2 + 20% ethanol. The combined interaction of pressure/temperature
showed a negative effect on the extraction yield
whereas the effect of temperature was positive. This explains the
high extraction yield obtained at low pressure and high temperature.
The empirical correlation obtained for the global extraction
yield is given below (Eq. (6))
Y = −8.6758 + 0.0304 · P + 0.290167 · T − 0.000726667 · P · T (6)
Y: global extraction yield (g/100 g of dry matter); T: temperature
(◦C); P: pressure (MPa). The resulting determination coefficient(R2)
was 0.9961.