Influence of brine concentration on sugar and sodium chloride diffusion during the processing of the green olive variety Arauco

The green olive variety Arauco were debittered using lye concentrations of, 2.50% NaOH.They were then subjected to two rinsing processes with tap water. Next, the olives were cured with brines at 7%, 10% and 13% sodium chloride concentration. During this curing process, the loss of reducing sugars from the olives, and the diffusion of sodium chloride into the olives were quantified. Effective diffusion coefficients of both solutes in the skin and the flesh were calculated for this period using a diffusion model for a composite hollow sphere. The skin effective diffusion coefficients for both solutes ranged from entre 8,27x10 -14 - m 2 /s to 4,04x10 -13 m 2/ s. The flesh coefficients varied from 2,92x10-9 m 2 /s to 3,99x10 -9 m 2 /s for sodium chloride and from 9,76x10 -11 m 2 /s to 2,22x10- 10 m 2 /s for reducing sugars.


INTRODUCTION
The diffusion of solutes in food products plays an important role in many food processes, such as extraction of lipids and sugars, drying, chemical peeling, etc., as well as the processing of fermented green olives.
Several researchers have studied the diffusion of NaCl by soaking vegetable pieces in salt solutions at different temperatures and determining the salt intake by the tissues.Liu (1992) determined the apparent diffusion coefficient of salt in potato tissue immersed in different brines with sodium chloride concentrations between 1 and 5% (w/v), at temperatures between 50 °C and 120 °C.
During salting, sensory, chemical, and physical changes occur in vegetables due to the high ionic strength of the medium.(Liu, 1992).Drusas, et al. (1988) quantified the diffusion of sodium chloride into green olives placed in brines of various concentrations.They studied untreated olives and olives pretreated with lye at 1.8 % for 6 hours and calculated salt effective diffusion coefficients assuming a hollow sphere geometry and negligible external resistance to mass transfer.They measured the absorption of salt from changes in brine concentration.
The final in the fermentation of green olives is the lactic fermentation process that takes place during curing in brine.Due to the diffusion of sugars and other nutrients out of the olives, the brine becomes an appropriate growth medium for the microorganisms responsible for lactic fermentation (Bobillo and Marshall, 1991;Marsilio and Lanza, 1998) which provides the acidity necessary for the stability and preservation of the olives (Marsilio, 1990).
The factors that influence this diffusion process are olive variety, lye treatment, rinsing, olive/brine ratio and brine concentration, among others (Fernández Diez, 1985).
Brine concentration is of upmost importance in the fermentation process.A high concentration will inhibit the growth of microorganisms responsible for causing defects, but also hinder the growth of lactic bacteria.Whereas a lower brine concentration facilitates an earlier and more vigorous growth of these bacteria, impeding the development of other microorganisms.(Fernández Diez, 1985).
According to Durán Quintana et al. (1997) practically inhibit the growth of lactic bacteria entirely.They indicate that a concentration of 6% of NaCl inhibits about 33% of the bacteria, at a 7% concentration of salt the inhibition reaches 50%, while at 8% only 10% of bacteria survive and a 9% NaCl concentration causes total inhibition.Recently, Maldonado et al. (2008) studied the influence of different concentrations of lye on the diffusion of reducing sugars during curing.
Due to the importance of the availability of reducing sugars in the brine for the fermentation process and considering that, to date, the diffusion phenomena of reducing sugars has not been quantified from a mass transfer standpoint in olives in terms of brine concentration, the aim of the present work is to evaluate the diffusion of sugar and sodium chloride during the curing process of green olives of the local variety "Arauco", as affected by different brine concentrations.

Sampling
Green olives of the variety "Arauco" were used in this study.They were harvested with a Maturity Index (M.I.) 2 (Fernández et al. 1991), corresponding to a green-yellow skin color, with few red spots.The M.I. scale ranges from 0 for olives with a skin color of intense green to 7 for olives with black skin and totally purple flesh.A random sample of 50 kg of olives previously graded was selected for the experiments, from these, 150 olives were measured giving the following average dimensions: weight = 7.433x10 -3 kg, = 0.915x10 -3 kg; equatorial diameter = 20.50x10 - m, = 1.59x10 -3 m; length = 31.38x10-3 m, = 2.20x10 -3 m.Twenty five olives gave the following skin thickness = 4.0x10 -5 m.The skin was removed with the help of a sharp thin blade, taking care to eliminate any mesocarpic cells attached and to avoid cuts and breakage of the skin.The skin thickness was measured with a digital caliper Palmer Helios (1:100 mm).
Batches of 4.5 kg olives each were taken from the previous sample and placed in 10-liter plastic containers with lids, into which 3 liters of treatment solutions (lye, water and brine) were added at each stage.The temperature was kept constant a 20 °C ± 1 °C, throughout the processes.Each treatment was applied in triplicate.
The concentration of NaOH studied was 2.50% (w/v).As a customary industry practice, the debittering process was terminated when the lye had penetrated 3/4 of the flesh thickness, which was visually determined by a colorimetric reaction with phenolphthalein.The debittering process took 10 hours.
After the debittering process, the lye was removed from the containers and the olives were rinsed with tap water for two consecutive periods of six hours each.At the end of each rinsing period, the water was completely removed from the containers.Finally, 3 liters of brine with 7%, 10% y 13 % sodium chloride concentration, acidified with 0.2% (v/v) of each acetic and hydrochloric acid (Merck, PA), were added to the containers.In order to inhibit microbial activity 1000UI/ml of nisine (Chrisin™) and 50 ppm of chloranphenicol (Avicloran™) were also added to the brines.The microbial inhibition was checked by periodical microbial analysis (ICMSF, 1986).
The containers with the solutions and the olives were maintained in continuous agitation in order to achieve a large surface mass transfer coefficient (large Biot number).
Brine samples were taken periodically.The samples were placed in plastic test tubes, sealed and kept in a freezer at -18 °C until they were analyzed.

Analytical Methods
The concentration of NaCl in the brines was measured by Mohr's method (Skoog and West, 1976;Maldonado et al., 2008).
Miller's technique was adapted to determine reducing sugars in olive flesh (Maldonado et al., 2008).
The salt absorbed by the olives was calculated from the changes in brine concentration.
The concentration of reducing sugars in the olives was measured before the debittering process and at the end of the second rinsing.During the curing process, it was calculated from the increase in sugars in the brine.

Theoretical Considerations
In order to calculate the effective diffusion coefficients of both solutes in the skin and the flesh during the curing process, a diffusion model for a composite hollow sphere was adjusted to the experimental data, using the following assumptions (Maldonado and Zuritz, 2004).See Figure 1. 1.The flesh and skin of the olives are homogeneous and isotropic with a uniform initial concentration (c i ). 2. The olive flesh is a hollow sphere with internal radius r i and external radius r o , surrounded by the skin of internal radius r o and external radius r s .3. The thickness of the skin is much less than the internal and external radii, therefore, the difference in surface areas can be neglected.4. Provided that (r s -r o )<< (r o -r i ), the accumulation of sodium in the skin can be neglected allowing the assumption of a quasi-steady state diffusion process through the skin.
INFLUENCE OF BRINE CONCENTRATION ON SUGAR AND SODIUM...

5.
The pit is a concentric solid sphere impervious to the diffusion of solutes.6.As a simplifying assumption, the model considers constant effective diffusion coefficients in both the skin (D S ) and the flesh (D F ). 7. The system is well-agitated (large mass Biot number) therefore, the outer skin surface (r= r s ) reaches the solution average concentration (c s ). (3) where: (4) and in which: (6) Where ␣ 1 are the eigenvalues that satisfy the following equation: (10) where: The flesh and skin effective diffusion coefficients (D F and D S ) and the eigenvalues (α 1 ) for each treatment were estimated using the least squares method, which minimizes the function: Since the sum of squares of the residuals given as equation ( 11) is a non-linear function of the flesh and skin effective diffusion coefficients (D F and D S ) and the eigenvalues a 1 , the interactive non-linear regression method implemented in the program Microsoft Excel Solver " was used minimizing equation ( 11) simultaneously subjected to the restriction imposed by equation (10).Under the above conditions, the onedimensional diffusion process with constant effective diffusion coefficients through the flesh (D F ) and the skin (D S ) is expressed in dimensionless form as: Equation ( 1) is subjected to the following initial and boundary conditions: Where: The olive dimensions were adjusted to the radial dimensions of an equivalent sphere of equal volume giving the following values: r i = 4.29x10 -3 m; r o = 11.77x10 - m; r s = 11.81x10 - m and skin thickness (r s -r o ) = 4x10 -5 m.
The average volumetric concentration <C> of the solution to equation (1) subjected to equations (2a-c) for large processing times was presented by Maldonado and Zuritz (2004) as:  (1979) studying brine concentrations between 4% and 14%.
The diffusion of NaCl into the olive flesh is shown in Figure 4.The average equilibrium concentration of NaCl in the olive flesh was about 19, 27 and 32 g NaCl/kg of flesh in brines at 7%, 10% and 13% respectively.Nevertheless, the time required to reach the equilibrium concentration was inversely proportional to brine concentration, i.e.: 42 hours with brine at 13% NaCl, 56 hours with brine at 10% NaCl and 64 hours with brine at 7 % NaCl.

RESULTS AND DISCUSSION
The initial content of reducing sugars in the olives was 74.94 g/kg.After 10 hours of debittering and two rinsing processes of 6 hours each, the remaining concentration of reducing sugars was 22.03 g/kg, which corresponds to a loss of 70.6%.
Figures 2 and 3 show respectively the experimental data of the average concentrations of sodium chloride and reducing sugars in the brines for the three treatments.
After the second day, the concentration of NaCl remains almost constant for the rest of the curing process.The salt concentration decreases around 59, 58 and 60% for brines at 7%, 10 and 13% respectively.On the other hand, the three treatments extracted practically the same amount of reducing sugars (11.76, 11.86 and 11.68 g/l for at 7%, 10 % and 13% of NaCl), but this amount was reached at 120, 144 y 288 hours, respectively.However, at a given time, the brine with lower salt concentration had extracted more sugar than the other brines, e.g. at 88 hours the concentration of sugar in the brines was respectively 8.48, 5.75 and 3.24 g/l for brines with   Experimental and calculated concentration of sodium chloride in olive flesh (g/kg) until equilibrium, for curing treatments with brines at 7, 10 and 13 % NaCl.
Figure 5 shows the decrease of reducing sugars in the flesh during the curing process, up to the time of maximum concentration in the brine.The remaining average concentration of sugars were 14.19, 14.12, 14.24 g/kg of flesh for brines at 7%, 10% and 13% respectively.These values represent around 64.5% of the concentration of sugar at the beginning of the curing process.Both figures 4 and 5 show the values calculated with the model (equation 3).This difference in solute diffusivity could be due to the difference in molecular size between the glucose and the NaCl.Since the molecule of glucose is much larger than the molecule of NaCl, it would be expected that the movement out of the flesh of the former would be slower than the entry of the later into the olive.
Other reasons postulated are that during osmotic processing, the rate of diffusion of water from any vegetable tissue depends on the concentration and temperature of the osmotic solution, size and geometry of the material, solution to material mass ratio, and level of agitation of the solution (Raoult-Wack et al., 1989;Raoult-Wack, 1994;Rastogi and Niranjan, 1998;Rastogi et al., 1999).
The effective diffusion coefficients of reducing sugars and NaCl calculated for the flesh (D F ) and the skin (D S ) are presented in Table 1.
In the flesh, the diffusion coefficients (D F ) for reducing sugars were between one and two order of magnitude smaller than the coefficients for sodium chloride.For instance, in treatment for brine concentration at 7.0% NaCl, the D F for reducing sugars was 2,216x10 -10 m 2 /s, while for NaCl this value was 2,920 x10 -9 m 2 /s.As indicated above, the difference in magnitude between the flesh diffusion coefficients could be attributed to the difference in molecular size between the glucose and the NaCl.
In addition, the data seem to indicate that for reducing sugars, an increase in brine concentration gives rise to smaller values of D F .For instance, D F decreased from a value of 2.216x10 -10 m 2 /s for a brine concentration at 7.0% NaCl, to 1.57x10 -10 m 2 /s and 9.759x10 -11 m 2 /s for brine concentrations at 10% and 13%, respectively.However, an opposite behavior was observed for the D F values of NaCl.This difference in behavior could be explained by the effect that the brine's osmotic pressure probably exerts on the diffusion of solutes in and out of the olive flesh.An increase in osmotic pressure on the outside would provoke a faster diffusion of solutes (NaCl) into the olive flesh, but would have an inverse effect with regard to solutes diffusing out of the flesh (reducing sugars).
With regard to the skin effective diffusion coefficient (D S ), while their values were of the same order of magnitude for both solutes, they were three orders of magnitude lower than the D F values for reducing sugars and four orders of magnitude lower than the D F for NaCl.These values are consistent with those reported by Maldonado et al. (2008), for the diffusion of reducing sugars and NaCl in olives treated with different lye concentrations.
Although in the skin the effect is not as noticeable as in the flesh, the skin presented a greater barrier effect for the diffusion of reducing sugars than for NaCl, and the diffusion coefficients for NaCl were slightly larger than for reducing sugars.For instance, for olives treated with a brine concentration at 13.0% NaCl, the D S for reducing sugars was 8.278x10 -14 m 2 /s, while for NaCl it was 4.035x10 -13 m 2 /s.
As with the flesh diffusion coefficients, the same effect of brine osmotic pressure was observed with the skin diffusion coefficients.
This difference in diffusivity values is probably due to the different physical structures between the flesh and the skin, which, being the natural barrier of the fruit, has a more impervious structure to the diffusion of different chemical compounds, such as reducing sugars and NaCl.Drusas et al. (1988) reported values of effective diffusion coefficients for sodium chloride in green olives of the variety Konservolea at 20 °C, in the order of 1.95x10 -10 m 2 /s for a brine concentration of 10.7 % in olives pretreated with lye at 1.8 % for 6 hours.The difference in the reported values and the ones presented in Table 1 is probably due to the different variety studied and lower lye concentration and debittering time used.The structural difference of the two varieties could also be reflected in the time required to attain a constant concentration of NaCl in the brine.They reported 50 days, while in the present case it took only two days.Furthermore, they calculated values of "overall" effective diffusion coefficients (which include the effect of the skin), while the present values differentiate the resistance of the flesh from the resistance of the skin to the diffusion process.
In closing, it can be said that since the nutrients, particularly reducing sugars, needed for the growth Figure 1 Schematic of olive cross section showing radial dimensions.

Figure 2
Figure 2Concentration of sodium chloride in brine (% w/v) until equilibrium, for curing treatments with brines at 7, 10 and 13% NaCl.

Figure 3
Figure 3Concentration of reducing sugars in brine (g/l) until maximum concentration, for curing treatments with brines at 7, 10 and 13% NaCl.
Figure 5Experimental and calculated concentration of reducing sugars in olive flesh (g/kg), for curing treatments with brines at 7, 10 and 13 % NaCl.

Influence of brine concentration on sugar and sodium chloride diffusion during the processing of the green olive variety Arauco
M. B. MALDONADO, C. A. ZURITZ AND N. MIRAS MALDONADO, C. A. ZURITZ AND N. MIRAS 7%, 10% and 13% NaCl.Similar results were found by Rodriguez de la Borbolla y Alcalá y Rejano