Effect of phosphate addition and homogenization pressure on particle size distribution in pasteurized milk

The aim of the current study is to investigate the effect of analysis times and stabilizing salt blends (LD88, LD89 and KM5), at different doses and homogenization pressures, on particle size distribution (PSD) in pasteurized whole milk. PSD has evidenced that samples not subjected to stabilizing salt addition, heating and homogenization did not show significant variations in storage time. Sixty-seven (67) of the 81 evaluated experiments subjected to stabilizer addition recorded Dv90 values (volume, where 90% of particles were found) higher than those of their corresponding treatments without salt addition. The LD89 blend was the stabilizer recording the lowest Dv90 values at hydration times 0 and 24 hours, regardless of dose level, at homogenization pressure equal to 20 MPa. This very same same blend - at the dose of 1.0 g.L -1 , 0-hour hydration and homogenization pressure equal to 80 MPa - was the stabilizer recording the lowest overall Dv90 value (0.926 µm) among all experimental conditions. Results enabled concluding that analysis time did not influence PSD, although stabilizer type had influence on PSD, under the homogenization conditions adopted for whole milk in the current study.


INTRODUCTION
Using phosphate salts in milk-producing industries helps saving time, electric power, manpower and chemical products used to clean equipment. In addition, phosphate salts enable increased thermal stability and help minimizing issues such as geling and sediment formation (BRASIL, 1997;WALSTRA et al., 2005;FOX, 2015). The addition of stabilizing salts, mainly of phosphate salts such as sodium monophosphate, diphosphate and triphosphate, to Ultra High Temperature (UHT) products is authorized and regulated in Brazil by Ordinance n. 370 -from September 4 th , 1997. However, these salts cannot be added to UHT products in amounts exceeding 0.1 g.100 mL -1 . Stabilizing salts are used to increase milk thermal stability, as well as to delay gelling in UHT products. These salts are chelating agents capable of sequestering and complexing calcium ion; consequently, they change the balance between caseins and minerals (AUGUSTIN; CLARKE, 1990;SINGH et al., 2021).
Heating milk to high temperatures in ultra-pasteurization processes can affect milk constituents and likely lead to protein denaturation reactions and to complex formations between whey proteins and caseins. Thus, using these phosphate salts can help maintaining the colloidal stability of these proteins by changing their structure (BUENO et al., 2020;NIEUWENHUIJSE;HUPPERTZ, 2018).
Accordingly, homogenization is a mandatory step in UHT milk production processes, since it provides better product stability under storage conditions, as well as longer shelflife, by applying high pressures capable of reducing the size of milk constituents such as fat globules (D'INCECCO et al., 2018). Homogenization is a processing stage carried out to change the functional or sensory properties of milk with little or no effect on this product's nutritional value. A new membrane is immediately redone during homogenization due to the action of surfactants found in milk; consequently, it reduces the surface tension between globules and the soluble phase. The newly formed membrane presents higher protein concentrations and lower relative phospholipids' participation (MCCRAE et al., 1994;CANO-RUIZ;RICHTER, 1997;YE;ANEMA;SINGH, 2008).
The technology called milkmusion was developed by focusing on the milky matrix in order to get protein-lipid nanoparticles with hydrodynamic radii centered at up to 200 nm. Reducing the size of these fat globules leads to unique emulsification abilities, improves products' digestibility by increasing nutrient absorption and by enabling new mechanisms for active compounds' release in the body (QUEIROZ et al., 2021;DE PAULA et al., 2021;INOVALEITE, 2021).
Fat globules' natural size is approximately 5 micrometers; however, homogenization can reduce their size to approximately 1 micrometer, which is a viable size to avoid lipid separation. Nevertheless, these fat globules are too big to enable the proper stabilization of emulsions as complex as milk (BAARS et al., 2016;GALLIER, 2017;LERMA et al., 2018).
The process adopted to determine particle size distribution (PSD) through phosphate addition to casein and skim milk is described in the scientific literature The current study is based on the hypothesis that homogenization pressure and stabilizing salt addition to whole milk can influence its particle size distribution. The literature has few studies focused on investigating the way different phosphate salts can influence the particle size of whole milk homogenized at different pressures. Thus, the aim of the present study was to investigate the influence of type, concentration and addition time of different phosphate salts on the PSD of milk pasteurized and homogenized under different pressures.
Particle displacement velocity is often calculated by taking into account milk fat density. However, samples in the current study were subjected to homogenization process at different pressures; this process has changed fat globules' structure, a fact that does not guarantee the same milk fat density in all treatments. Fat density increases after the homogenization process because the membrane formed around the fat globule is formed by proteins. Figure 3 -Flowchart outlining the methodology adopted for all three treatments carried out in the current study; in total, 180 experiments were subjected to particle distribution analysis in order to find parameter Dv90 (hydrodynamic diameter, expressed in μm).
Source: Elaborated by the authors. Table 2 shows results recorded for Dv90 in the PSD analysis applied to samples that were not added with salts, at the three analyzed times, in all three treatments. The analyzed times (0, 24 and 48 h) did not influence the hydrodynamic size corresponding to 90% of particles accounting for values lower than the recorded ones (Table 1). In addition, heating (80°C ± 2°C) had low influence on Dv90 values. On the other hand, homogenization led to particles' reduced hydrodynamic diameter.

RESULTS AND DISCUSSION
Homogenization at pressure of 20 MPa was enough to reduce particle size by approximately 3 times in comparison to that recorded for non-homogenized samples.
However, the smallest hydrodynamic diameters were observed when homogenization pressure increased to 50 MPa and 80 MPa. Among the samples subjected to pressure of 50 MPa, only the one subjected to hydration time 0h presented Dv90 in the submicrometric region (<1µm). Finally, increasing the pressure from 50 MPa to 80 MPa enabled 100% of analyzed samples to have particle size allocated in the sub-micro region (<1µm). Table 3 shows the Dv90 parameter of samples added with phosphate salts at different times and different doses, based on PSD of Treatment 3. Only one sample recorded hydrodynamic diameter value higher than 2 μm; it was the one added with 5 g.L -1 of KM5 phosphate salt, at time 0 h and pressure of 80 MPa. On the other hand, the lowest hydrodynamic diameter values observed in the sub-micro region (<1 μm) were recorded for samples subjected to homogenization pressure of 80 MPa, and added with KM5 (24h/1 gL -1 and 48 h/1 gL -1 ), LD89 (0h/1 gL -1 ) and LD88 (48h/1 gL -1 ). Table 3 shows Dv90 values and their respective standard deviations, and it enabled seeing the confidence interval between the minimum and maximum values of such hydrodynamic diameter, at each time and homogenization pressure. Based on results presented in this table, it is possible comparing Dv90 values recorded for samples added with different salt types and concentrations to those recorded for samples without salt addition, at the very same homogenization times and pressures. It is worth mentioning that the current study comprises a significantly large number of variables (combination among 3 salt types, 4 salt doses, 3 homogenization pressures and 3 different hydration times), a fact that highlights its exploratory nature. All results recorded for hydrodynamic diameter (Table 3) were either within or above the confidence interval shown in Table 2. Therefore, it is noteworthy that no sample recored hydrodynamic diameter value lower than the confidence interval determined for samples without phosphate salt addition. In total, 83% (67 experiments) of the 81 results (  (2000) conducted a study to investigate the effects of adding calcium salts and chelating agents in skim milk to help better understanding the reversibility of changes induced in casein micelle at fixed pH. Adding a mix of phosphates (Na2HPO4 and NaH2PO4), at concentrations of 10, 20 and 30 mmol.kg -1 , to skim milk did affect the effective diameter of micelles -which was approximately 197 nm -in the current study. However, the addition of 30 mmol.kg -1 of these phosphates, which was followed by the addition of 10 mmol.kg -1 of CaCl2, has increased the effective diameter of micelles in milk suspensions to approximately 209 nm. Pitkowski, Nicolai and Durand (2008) investigated casein dissociation resulting from chelating salts' addition by using static and dynamic light scattering and small angle X-ray scattering (SAXS). A mix of sodium polyphosphates was used as chelating agent; it comprised concentrations of each polyphosphate ranging from 10% to 15%, namely: ortho-, pyro-, tri-and tetraphosphate. Different polyphosphate mix concentrations (1.0, 2.0, 3.0, 5.0 and 7.0 g.L -1 ) were added to 11 g.L -1 of casein solution; casein complexes were fully dissociated in all cases. Casein that got fully dissociated after polyphosphate addition has formed small micellar particles that presented hydrodynamic radius of approximately 10 nm (PANOUILLÉ; NICOLAI; DURAND; 2004). Furthermore, a mix of intact and dissociated micelles were observed in casein solutions after polyphosphate addition at concentration of 0.5 g.L -1 .
Sodium hexametaphosphate (SHMP) has the potential to bind to up to three calcium atoms, since its six homogeneously distributed phosphate molecules enable this salt to directly interact with casein amino acid residues through electrostatic interactions.
SHMP triggers electrostatic repulsions in micelles and this process leads to κ-casein dissociations (ANEMA, 2015;DE KORT et al., 2011). De Kort (2012) has analyzed variations in casein micelles' particle size at SHMP concentrations ranging from 0 to 100 mmol.L -1 . The threshold SHMP concentration, at which smaller particles were observed, was 25 mmol.L -1 . The distribution peak started to change towards particles with smaller diameters (from 30 mmol.L -1 on) and reached values lower than 10 nm at SHMP concentrations higher than 50 mmol.L -1 . The same analyses were performed, after Na2HPO4 addition, also at concentrations ranging from 0 to 100 mmol.L -1 . Results have shown that Na2HPO4 has poor ability to dissociate casein micelles into smaller particles, since Na2HPO4 addition has increased the diameter of intact micelles from 190 to 220 nm. This outcome suggested that the volume of casein micelles has increased after Na2HPO4 addition, which is likely associated with decreased free calcium levels that, in their turn, induced increased repulsion between casein molecules (DE KORT, 2012). Renhe, Indris and Corredig (2018) have evaluated the stability of micellar casein concentrates subjected to heat with, and without, the addition of calcium chelators (trisodium citrate and Na2HPO4). In order to do so, particle size measurements were taken, by taking into consideration pH and chelating salts, after heating at 120°C, for 10 minutes; pH 6.5 did not show significant differences in casein micelle diameter, which ranged from 165 to 190 nm. On the other hand, particle size at pH 6.7 has increased as 13 calcium chelator concentration also increased, mainly when the citrate-phosphate combination was used. In this case, micelle diameter at concentration of 60 mmol.L -1 of citrate-phosphate was 240 nm. Particle size increase at pH 6.9 was more significant at phosphate concentrations higher than 30 mmol.L -1 , as well as at citrate-phosphate combinations ranging from 264 to 524 nm. According to Sauer & Moraru (2012), such increase in micellar casein particle size at pH 6.9 is correlated to heating temperature. Table 4 shows the time (in days) required for particles to cover 200 mm (average size of 1 liter UHT milk carton packs in Brazil).