**3. Results and discussion**

As acknowledged, the selection of the most appropriate solvent for extracting the analytes of interest from the plant matrix is a basic step in the development of any method of solvent extraction. Theoretically, solvent would provide not only a background for the extraction process but it would also stabilize the analytes and the transition state species by solvating process. This solvation is due to solvent-analyte interactions during which a solvent acts either as a nucleophile or as an electrophile by donating or accepting electron pairs from the analyte. The research data evidence for hot pepper cultivars indicate that methanol and ethanol are solvents usually used in the extraction of capsaicinoids in various extraction techniques (Barbero et al., 2006; Kirschbamm-Titze et al., 2002; Williams et al. 2007). Studies on the solvent influence on pigments extraction from *Capsicum* fruits ascertained *n*-hexane and acetone as suitable solvent medium for pigments (Boyadzhiev et al., 1999; Feltl et al., 2005; Tepić et al., 2009).

Evidence provided by relevant literature positively confirm recent growing interest in the development of mathematical models that describe the extraction process as a function of various operational variables and, particularly, those that describe their combined effect (Acero-Ortega et al., 2005; Bo et al., 2008; Hismath et al., 2011; Liu et al., 2010).

In order to select the extraction solvent for pungent paprika matrix, experiments were performed with three solvents: ethanol, methanol and *n*-hexane. According to our previous experiences (Rafajlovska et al., 2007), the two variables that could potentially affect the extraction efficiency of the analytes of interest in chosen solvents are extraction temperature and dynamic time. Owing to the significance of interaction between time and temperature, their interactive influence on the extraction efficiency was also considered. Other parameters implicated in the extraction were kept constant, namely the solid:phase ratio and particles size.

### **3.1 Extraction of pungent capsicum oleoresin, capsaicin and capsanthin with ethanol 3.1.1 Model fitting**

Table 1 shows the liner, quadratic and interactive coefficients of the independent variables in the models and their corresponding R2 when ethanol was used as extraction solvent. It can be seen that the R2 values for these response variables are higher than 0.97 where PCO and capsaicin are concerned, indicating that the regression models adequately explained the process. Therefore, the R2 values are 0.9795 and 0.9810, respectively, for PCO yield and capsanthin. The probability (*p*) values of regression models for PCO and capsaicin show no lack-of-fit (p < 0.001). However, since the R2 value of capsanthin is not acceptable (R2=0.7890) this regression model is not suitable to explicate the extraction process for capsanthin, probably owing to the solvent characteristics.


Subscripts: 1 = temperature (° C); 2 = time (min);

114 Mass Transfer in Chemical Engineering Processes

ANOVA was used to evaluate the significances of the coefficients of the models judged by

The influence of the predictors on the responses was also presented using 3-D mesh plots

As acknowledged, the selection of the most appropriate solvent for extracting the analytes of interest from the plant matrix is a basic step in the development of any method of solvent extraction. Theoretically, solvent would provide not only a background for the extraction process but it would also stabilize the analytes and the transition state species by solvating process. This solvation is due to solvent-analyte interactions during which a solvent acts either as a nucleophile or as an electrophile by donating or accepting electron pairs from the analyte. The research data evidence for hot pepper cultivars indicate that methanol and ethanol are solvents usually used in the extraction of capsaicinoids in various extraction techniques (Barbero et al., 2006; Kirschbamm-Titze et al., 2002; Williams et al. 2007). Studies on the solvent influence on pigments extraction from *Capsicum* fruits ascertained *n*-hexane and acetone as suitable solvent medium for pigments (Boyadzhiev et al., 1999; Feltl et al.,

Evidence provided by relevant literature positively confirm recent growing interest in the development of mathematical models that describe the extraction process as a function of various operational variables and, particularly, those that describe their combined effect

In order to select the extraction solvent for pungent paprika matrix, experiments were performed with three solvents: ethanol, methanol and *n*-hexane. According to our previous experiences (Rafajlovska et al., 2007), the two variables that could potentially affect the extraction efficiency of the analytes of interest in chosen solvents are extraction temperature and dynamic time. Owing to the significance of interaction between time and temperature, their interactive influence on the extraction efficiency was also considered. Other parameters implicated in the extraction were kept constant, namely the solid:phase ratio and particles

**3.1 Extraction of pungent capsicum oleoresin, capsaicin and capsanthin with ethanol** 

Table 1 shows the liner, quadratic and interactive coefficients of the independent variables in the models and their corresponding R2 when ethanol was used as extraction solvent. It can be seen that the R2 values for these response variables are higher than 0.97 where PCO and capsaicin are concerned, indicating that the regression models adequately explained the process. Therefore, the R2 values are 0.9795 and 0.9810, respectively, for PCO yield and capsanthin. The probability (*p*) values of regression models for PCO and capsaicin show no lack-of-fit (p < 0.001). However, since the R2 value of capsanthin is not acceptable (R2=0.7890) this regression model is not suitable to explicate the extraction process for

(Acero-Ortega et al., 2005; Bo et al., 2008; Hismath et al., 2011; Liu et al., 2010).

capsanthin, probably owing to the solvent characteristics.

 

*X X XXi j*

  (2)

kk k o i 1 ii 2 ij i1 i1 i1

 

Y

and contour maps.

2005; Tepić et al., 2009).

size.

**3.1.1 Model fitting** 

**3. Results and discussion** 

computing the *F*-value at a probability (*p*) of 0.001, 0.01 and 0.05.

\*Significant at 0.05 level; \*\*Significant at 0.0l level; \*\*\*Significant at 0.001 level.

Table 1. Regression coefficients, *R2*, adjusted R2 and *p* for three dependent variables for pungent capsicum oleoresin obtained by ethanol.

### **3.1.2 Influence of extraction temperature and time**

The influence of extraction conditions on the PCO yield and capsaicin were presented by the coefficients of the second-order polynomials. As shown in Table 2, PCO yield was significantly affected by the positive linear effect (*p* < 0.001) of the temperature and the positive linear effect (*p* < 0.01) of the time. In this case, the temperature and time were relevant variables for the model. However, significant linear interaction between the temperature and time (*p* < 0.01) had a negative sign. Moreover, it was found that the influence in the second-order term for the both variables showed no significant effect (*p* > 0.05). These results suggest that the linear effect of the extraction temperature was the primary determining factor for PCO yield but there is no need for prolonged solid/liquid phase contact. The response surface and contour map were also developed to facilitate the visualization and latter, for predicting the optimum condition for PCO yield and capsaicin in ethanol (Fig. 1).

Fig. 1b shows that the PCO yield increased as the temperature increased. As for the capsaicin content in PCO, the positive interaction among the independent variables (p < 0.001) significantly influenced the capsaicin content. It was also found that quadratic effect of extraction time is negative at p < 0.01. However, the linear term of temperature and time showed no significant effect on capsaicin content in ethanolic PCO. Hence, when analyzing the interactive effect of temperature and time on the extraction efficiency of capsaicin (Fig. 2) in the model developed for ethanol as extraction solvent, it was observed that extended time of extraction is not appropriate under increased temperature condition.

Fig. 3 shows that owing to the capsanthin temperature liability (Ahmeda et al., 2002; Pérez-Gálvez et al., 2005; Schweiggert et al., 2007), capsanthin extraction in ethanolic medium should be performed at decreased temperature of about 40oC at most during extended time.

Extraction of Oleoresin from Pungent Red Paprika Under Different Conditions 117

(a)

(b) Fig. 2. 3-D mesh plot (a) and contour plot (b) of the effects of extraction temperature and

time on capsaicin in ethanolic PCO.

Fig. 1. 3-D mesh plot (a) and contour plot (b) of the effects of extraction temperature and time on PCO yield (%) in ethanol.

(a)

(b) Fig. 1. 3-D mesh plot (a) and contour plot (b) of the effects of extraction temperature and

time on PCO yield (%) in ethanol.

Fig. 2. 3-D mesh plot (a) and contour plot (b) of the effects of extraction temperature and time on capsaicin in ethanolic PCO.

Extraction of Oleoresin from Pungent Red Paprika Under Different Conditions 119

The liner, quadratic and interactive coefficients of the independent variables in the models and their corresponding R2 when methanol was used as extraction solvent are presented in Table 2.

Yield (%) Capsaicin (mg/100g) Capsanthin (mg/100g)

**3.2 Extraction of pungent capsicum oleoresin, capsaicin and capsanthin with** 

bo (intercept) 4.938929 16.501750 - 56.065700 b1 0.282113 6.922010\* 9.785980\*\* b2 0.036924\* 0.324730 0.808700\*\* b12 - 0.002389 - 0.087330\*\* - 0.097400\* b22 - 0.000051 - 0.001700\*\* - 0.001300\* b12 0.000337 0.012620\*\*\* - 0.006600\*\* *R2* 0.9702 0.9391 0.7228 adjusted R2 0.9553 0.9087 0.5843 *p* or probability 0.0000 0.0000 0.0130

Table 2. Regression coefficients, *R2*, adjusted R2 and *p* for three dependent variables for

Table 2 clearly shows that the R2 values for these response variables are higher than 0.93 for both PCO and capsaicin, indicating that the regression models adequately explain the process. Hence, the R2 values are 0.9702 and 0.9391, respectively, for methanolic PCO yield and capsaicin. The *p* values of regression models for PCO yield and capsanthin show no lack-of-fit. However, as expected, the R2 value of capsanthin is low, (R2 = 0.7228) confirming that a high proportion of variability is not explained by the model. We therefore conclude that this regression model cannot offer a satisfactory explanation of the extraction process

The influence of extraction conditions on the PCO, capsaicin and capsathin are presented by the coefficients of the proposed model. As indicated by *p* value, positive linear (*p* < 0.05) effect of time is only confirmed to be significant for PCO yield, while positive linear (*p* < 0.05) effect of temperature is noticed for capsaicin content present in methanolic PCO. Furthermore, it is found that interactive influence of both variables has the prominent positive effect (*p* < 0.001) for capsaicin content. On the other hand, a negative quadratic

Fig. 4 and 5 show the response surface and contour map for PCO yield and capsaicin. It was observed that the capsaicin content rises as the temperature and time increase, but prolonged phase contact at increased temperature will not be acceptable due to the negative quadratic terms at *p* < 0.01. Generally speaking, when a higher extraction temperature was applied to the process, a higher velocity and extraction efficacy were achieved. However, some degradation processes can easily occur at high temperature, resulting in lower analyte

**methanol** 

**3.2.1 Model fitting** 

for capsanthin.

recovery.

Subscripts: 1 = temperature (°C); 2 = time (min);

pungent capsicum oleoresin obtained by methanol.

**3.2.2 Influence of extraction temperature and time** 

effect (*p* < 0.01) has been verified for both variables for capsaicin.

\*Significant at 0.05 level; \*\*Significant at 0.0l level; \*\*\*Significant at 0.001 level.

Fig. 3. 3-D mesh plot (a) and contour plot (b) of the effects of extraction temperature and time on capsanthin in ethanolic PCO.
