3.1.2. Structure

Freeze-dried purees showed high porosity, typical of freeze-dried products, with values ranging from 86.4–87.4%. The PLS–R indicated a correlation between FR-S and porosity (mainly projected in the positive side of axis 2, Figure 1), which means that a slow rate of freezing leads to bigger ice crystals and so a bigger expansion of gas cells in the structure (higher porosity) during the freeze-drying process. Nevertheless, they are in the inner circle, which means the effect is not significant. In fact, according to ANOVA, porosity of the samples was not significantly affected by any factor evaluated (*p* > 0.05).

As a result of the mechanical analysis, the force versus distance curves were obtained for each sample freeze-dried under each of the 12 conditions (Figure 3). The linear regression of the first part of the curve was taken to calculate the slope before Ff, related to the rigidity of the sample. As can be observed in the PLS-R (Figure 1), factors T50 and rigidity are both partially correlated and projected on the negative side of axis 2, which means that a higher temperature leads to an increase in the rigidity of the freeze-dried samples. However, because the rigidity is projected in the inner circle, it indicates a moderate positive correlation. In regards to the ANOVA results, the shelf temperature had a significant effect on the rigidity (slope) of the freeze-dried samples (*p* < 0.05); however, its low F-Value (8.65) makes it possible to confirm its low significance.

**Figure 3.** Examples of force–distance curves obtained from the freeze-dried purees frozen at a slow rate and freeze-dried at 5 Pa, and with different shelf temperatures (30, 40 and 50 ◦C). Ff: fracture force.

Values of the slope 18–26 N/mm were obtained when heating the freeze-drier shelves to 50 ◦C, as compared to 11–18 N/mm values obtained at 30 and 40 ◦C. As a greater slope indicates less deformation of the sample by exerting an effort on it, it can be concluded that freeze-drying heating the shelves to 50 ◦C promotes the mechanical rigidity of the freeze-dried samples before they fracture. This can be considered desirable as it would be related to a higher mechanical resistance of the sample during its handling before consumption.

Fracture force of all the samples varied between 12.1 ± 1.6 N and 19 ± 5 N. The PLS-R revealed a moderate positive correlation between fracture force and both lower pressure (P5) (both projected in the positive side of axis 1) and T50 (both projected in the negative side of axis 2). This means that working with lower pressure and higher temperature seems to promote a freeze-dried puree that is more resistant to fracture. According to the ANOVA, neither the temperature nor the freezing rate showed a significant effect on the fracture force (*p* > 0.05). The pressure did show *p* < 0.05, but again, with a very low *F*-Value (7.17).

The water content of the samples ranged between 2.2 ± 0.3 and 4.2 ± 0.2 g water/100 g sample. The PLS-R indicated a high correlation between water content (%), high pressure (P100) and low temperature (T30), circled in blue colour in Figure 1. In fact, this statement was confirmed by the ANOVA that indicated a significant effect on the final water content of the freeze-dried samples by the shelf temperature and the pressure (*p* < 0.05). In this way, drying at the highest pressure (100 Pa) and the lowest shelf temperature (30 ◦C) promoted samples with the highest water content. On the other hand, Figure 1 shows a clear negative correlation between water content and fracture force, which means the smaller the water content of the freeze-dried samples, the greater the mechanical resistance to fracture. In this case, the final water content of the samples was lower than 4.2%. However, it seems that small changes outside this range, related to working conditions, will have an important impact on the mechanical properties of the sample.
