In the last few years, there has been a noticeable consumer trend towards low fat or fat free products that has forced the snack industry to produce foods with these characteristics, while maintaining their traditional flavour and texture properties. For this reason, research aimed at reducing the fat content in fried foods or snacks has gained force, either by modifying process conditions, by changing the frying medium or by the use of pre-treatments such as blanching, immersion in sugar solutions or edible coatings.  At the same time, the food industry is also searching for new products which may  increase its competitiveness.
Vacuum frying stands out among the techniques considered to meet these quality improvement requirements. In this case, the food is processed under reduced pressure in a closed system, which lowers the boiling point of the water in the food and allows frying at a lower temperature than in conventional deep fat frying. In this way, food moisture can be removed rapidly once the oil reaches the boiling point of water. Another advantage of vacuum frying is that food colour and flavour are better preserved, because the food is heated at lower temperature and oxygen content than when using atmospheric frying. This also reduces the deterioration and oxidation of the frying oil, thus extending oil shelf life. Additionally, acrylamide formation is prevented in products prone to it.
The aim of this work was to study the influence of the temperature and vacuum level on mass transfer during the frying process. Specifically, this work mainly focused on the moisture loss of Granny smith apples during vacuum frying and on oil absorption as a function of the operating conditions.  
In order to carry out this study, first of all a vacuum frying system was designed and built. The system functionality allowed the experiments necessary to study the moisture loss kinetics to be carried out under various operating conditions, and the effective moisture coefficient and the activation energy to be determined, by analyzing the influence of operating conditions, such as temperature and vacuum level, on these parameters.
In order to study the pressure effect on the kinetics of water loss and for comparison purposes, the work was carried out in two parts: in the first one, the samples were processed at atmospheric pressure and in the second one, at sub-atmospheric pressure or vacuum.
In the experiments at atmospheric pressure (101.3 kPa), the oil was heated at 140, 150, 160 and 170 ºC and the frying times were 30, 60, 90, 120 and 150 s. In the experiments at subatmospheric pressure, the apple slabs were processed at three pressure levels (18.6, 13.3 and 8 kPa). For each one of these pressures, experiments were performed at oil temperatures of 80, 90, 100, 110 and 120 ºC and for the following times: 30, 60, 90, 120, 150, 180 and 300 s. Each experiment was duplicated by using four 2mm thick slabs in each replication.
The curves of moisture loss obtained using the vacuum frying process showed the typical falling rate period shape of drying curves, similar to the ones obtained at atmospheric pressure and to the ones reported in the literature for vacuum frying.  
To model the water loss kinetics in the two kinds of processes, at atmospheric and sub-atmospheric pressure, the diffusion model for a slab was used. The experimental moisture values and the ones computed with the model showed a close correlation (%var > 99%); for this reason, it can be stated that the second Fick’s law for an infinite slab geometry satisfactorily describes the water lost in this process under the operating conditions considered. 
The effective moisture diffusivity values for apple slabs are in the general range for food dehydration, between 10-9 and 10-11 m2·s-1. Regardless of the pressure, when the temperature increased, the drying rate increased and thus the frying time decreased. Furthermore, at a given temperature, when the pressure decreased, the evaporation rate increased and, for this reason, the drying rate, too. 
The influence of temperature on the effective diffusion coefficient was explained through the Arrhenius equation, where the logarithm of the diffusivity showed a linear tendency versus the inverse of the absolute temperature (r2 = 0.99). From these results, the activation energy for each pressure was computed. 
The activation energy values computed for the apple slabs fried at atmospheric pressure were lower than those obtained by other authors, not only for dehydrated apple but also for other fried products. This would mean that it is easier to remove water in apple frying compared to other dehydration processes. 
As regards the activation energy values obtained in vacuum frying, when the vacuum level decreased, Ea decreased too. Furthermore, it was observed that Ea values obtained at sub-atmospheric pressure were higher than the value at atmospheric pressure, although no significant differences (p > 0.05) between the Ea values obtained were observed. Thus it cannot be stated from these results that there is an influence of frying pressure on the activation energy. 
In order to check the proposed model, experiments were carried out on 5 mm thick slabs. When plotting the dimensionless moisture content computed using the Deff obtained for the 2 mm thick slabs versus the ones obtained experimentally for 5mm thick slabs, an explained variance of 98% was obtained. This indicates the validity of the moisture loss model regardless of the sample thickness under the assayed conditions. 
At sub-atmospheric pressure, when changing the temperature, no differences in the final oil content of the samples were detected. As regards the pressure effect, at sub-atmospheric pressures the oil content decreased compared with the slabs processed at atmospheric pressure, in which case the oil content increased considerably, mainly at high temperature 
With the aim In order to analyze these findings, which are supposed to be related with the product microstructure, microphotographs were taken by Cryo SEM. When comparing the images obtained, differences in the final microstructure were appreciated. At atmospheric pressure, the sample did not collapse as much as it did after vacuum frying. Those slabs submitted to vacuum frying presented a more shrunken surface, the cells lost their original polyhedrical shape and the cell walls were much more distorted. Moreover, when processing the samples at sub-atmospheric pressure, the intercellular spaces were smaller and the cross section was also finer than when processing at atmospheric pressure. These findings explain the smaller amount of oil absorbed when vacuum frying as opposed to  atmospheric frying, since the cells are more collapsed and there is less space for oil absorption, which is hindered. 
The vacuum frying process responds to similar phenomena to those observed at atmospheric pressure and high temperature. This can be deduced since, in all cases, the values of Deff and Ea were of the same order of magnitude. Diffusion coefficients were significantly different, although the activation energy was not. 
In order to elaborate fried foods In order to elaborate fried foods, vacuum frying appears as an alternative of enormous significance when compared to conventional frying. Therefore, it is important to continue with the appropriate studies in order to increase our knowledge of this process and, consequently, to contribute with scientific data that will allow the technological situation in this field to be understood and improved.