*3.2. Bioactive Compounds*

The quantification of total phenols, vitamin C, β-carotene, and antioxidant activity evaluated by two methods (FRAP and DPPH), was carried out in the FOP and after being freeze-dried for each of the 12 conditions evaluated.

The PLS-R indicated that TP is not projected on the bi-plot (Figure 1), which means that it is relatively less impacted than the other bioactive compounds (TP vector very close to the origin 0.0). In fact, according to the ANOVA, the TP was not affected by working pressure and freezing rate (*p* > 0.05) and was better preserved when the freeze-drying was carried out at 30 or 50 ◦C as compared to 40 ◦C

(*p* < 0.05), despite the temperature factor also having a low F-Value (8.07). In fact, the most marked difference was observed with the factor temperature, which was 5% of preservation between 30 and 40 ◦C (Figure 4). No significant interactions between the factors were observed (*p* > 0.05).

**Figure 4.** Percentage (%) of preserved total phenols (mean and Tukey's HSD) of samples according to the shelf temperature factor (30, 40 or 50 ◦C). Different letters indicate different homogeneous groups for the shelf temperature factor (*p* < 0.05). Data of both freezing rates and both pressures are considered in the mean values.

Figure 4 shows TP preservation by shelf temperature, considering both freezing rates and pressures in the mean values. It seems that TP may be affected by the ratio of time and temperature of processing, as reported by other authors [40]. In this case, mild 40 ◦C heating for more than 7 h seems to compromise TP preservation.

The presence of VC in the final product is used as a reference of high nutritional quality for the different industrial processes, due to its relative instability to heat, oxygen and light [41,42]. The impact of temperature on VC can be clearly observed on the PLS-R according to axis 2 (Figure 1). Vector T50 and T40 are projected in the same direction (negative side of axis 2), while T30 is anti-correlated to them (projected on the positive side of axis 2). This confirms that a higher shelf temperature along the process preserved the vitamin C of the samples. According to the ANOVA, it can be confirmed that VC was affected by the shelf temperature and the pressure during freeze-drying (Figure 5, *p* < 0.05). A significant interaction between both factors indicated that heating the freeze-drier shelves to 40 or 50 ◦C promoted samples with higher vitamin C content than those freeze-dried at 30 ◦C. Despite VC being reported to have thermal stability [41], the length of time required when the freeze-drying is carried out at 30 ◦C (25 h) may cause VC loss. However, the lower content of VC of samples freeze-dried at 30 ◦C was even lower when higher pressure was applied (Figure 5). This means that for a long expected process time, oxygen presence should be maximally avoided.

**Figure 5.** Percentage (%) of preserved vitamin C (mean and Tukey's HSD) according to the interaction between shelf temperature (30, 40 or 50 ◦C) and pressure (P5: 5 Pa and P100: 100 Pa) factors. Different letters indicate different homogeneous groups for the temperature\*pressure interaction (*p* < 0.05). Data of both freezing rates are considered in the mean values.

Certain carotenoids are highly coloured compounds that also exhibit provitamin A activity. BC has the highest vitamin A activity [43]. The PLS-R (Figure 1) underlined, in particular, the BC preservation by low pressure, as it is projected on the positive side of axis 1 and highly correlated with vector P5. In this case, the ANOVA indicated a significant effect of the three factors considered (*p* < 0.05, Figure 6), the higher the pressure, the higher the temperature, and the slower the freezing rate, the greater the loss of BC. Nevertheless, the F-Values were 88, 6 and 6 for pressure, shelf temperature and freezing rate, respectively. Again, a low F-Value in the ANOVA is correlated with no significant effect detected by the PLS-R analysis. However, the ANOVA also revealed a significant interaction between the pressure and both the shelf temperature and the freezing rate (*p* < 0.05). Figure 6 shows the interaction of pressure and shelf temperature for each FR.

According to these interactions, the pressure effect is no longer significant at 50 ◦C, when a significant part of BC has already been degraded by the effect of the shelf temperature. In addition, the effect of the shelf temperature was not significant at higher pressure. Furthermore, when almost no oxygen is present (P5), BC is conserved quite well, regardless of freezing rate. It is in the greater presence of oxygen (P100) when the effect of the freezing rate is significant in relation to the better preservation of BC at the FR-F. From the ANOVA results, the PLS-R analysis can be nuanced in the sense that the most recommendable way to keep the maximum carotenoids present in the orange puree during freeze-drying is when the drying stage is carried out at the lowest pressure studied and heating the freeze-dryer shelves to 30 or 40 ◦C, without the freezing-rate being relevant in this case.

**Figure 6.** Percentage (%) of preserved β-carotene (mean and Tukey's HSD) according to the interaction between shelf temperature (30, 40 or 50 ◦C) and pressure (P5: 5 Pa and P100: 100 Pa) factors for each freezing-rate (FR-S and FR-F: slow and fast freezing rates, respectively). Different letters indicate different homogeneous groups for the temperature\*pressure interaction for both freezing-rate independently (*p* < 0.05).

As regards antioxidant activity, values of DPPH between 86.5 ± 1.8% and 94.3 ± 1.5% were observed. The PLS-R revealed that the vector DPPH is projected on the lower right corner of the graph, which means a moderate positive correlation with both the lowest pressure and the highest temperature. According to the ANOVA analysis, no significant effect of pressure was detected (*p* > 0.05). Although freeze-drying carried out at 30 ◦C leads to samples with lower DPPH than those processed at 40 ◦C or 50 ◦C (p < 0.05), once again, with a low F-Value (9.51) for shelf temperature factor. On the other hand, FRAP values between 91 ± 2% and 103 ± 6% of preservation for all the conditions studied were analysed. Neither the PLS-R nor the ANOVA analysis showed a significant effect of any of the freeze-drying process variables on FRAP (*p* > 0.05).

From the Pearson correlation, only a significant and positive correlation (0.5774, *p* < 0.05) was obtained between values of DPPH and vitamin C. Despite AOA being correlated in a positive way with the total phenolic, vitamin C content and carotenoids, it has been suggested that VC contributes to antioxidant capacity more than others antioxidant constituents, such as phenols or carotenoid in fruits with high VC content [41,44]. This can also be observed on the PLS-R projection as VC, DPPH and FRAP are projected on the same direction.
