Journal of Dairy Science
Volume 92, Issue 5 , Pages 1864-1875, May 2009

Effect of high-pressure homogenization on droplet size distribution and rheological properties of ice cream mixes

Department of Food Science, Faculty of Agriculture, University of Udine, Udine, Italy

Received 9 October 2008; accepted 16 December 2008.

Article Outline

Abstract 

The effect of different homogenization pressures (15/3MPa and 97/3MPa) on fat globule size and distribution as well as on structure-property relationships of ice cream mixes was investigated. Dynamic light scattering, steady shear, and dynamic rheological analyses were performed on mixes with different fat contents (5 and 8%) and different aging times (4 and 20h). The homogenization of ice cream mixes determined a change from bimodal to monomodal particle size distributions and a reduction in the mean particle diameter. Mean fat globule diameters were reduced at higher pressure, but the homogenization effect on size reduction was less marked with the highest fat content. The rheological behavior of mixes was influenced by both the dispersed and the continuous phases. Higher fat contents caused greater viscosity and dynamic moduli. The lower homogenization pressure (15/3MPa) mainly affected the dispersed phase and resulted in a more pronounced viscosity reduction in the higher fat content mixes. High-pressure homogenization (97/3MPa) greatly enhanced the viscoelastic properties and the apparent viscosity. Rheological results indicated that unhomogenized and 15/3MPa homogenized mixes behaved as weak gels. The 97/3MPa treatment led to stronger gels, perhaps as the overall result of a network rearrangement or interpenetrating network formation, and the fat globules were found to behave as interactive fillers. High-pressure homogenization determined the apparent viscosity of 5% fat to be comparable to that of 8% fat unhomogenized mix.

Key words: ice cream mix, high-pressure homogenization, particle size distribution, rheology

 

Back to Article Outline

Introduction 

Homogenization is widely employed in the food industry for emulsion stabilization and to improve the texture, taste, and flavor of many products including milk, milk cream, and ice cream mixes. Homogenization leads to a reduction in size and an increase in number of solid or liquid particles of the dispersed phase. During this process, the fluid is subjected to several simultaneous force-induced phenomena such as cavitation, turbulence, shear, friction, heat, compression, acceleration, rapid pressure drop, and impact (Phipps, 1975;Fellows, 1988;Paquin, 1999;Floury et al., 2000;Roach and Harte, 2008). All these forces are significantly increased as pressure increases. For this reason, homogenization technology has evolved from systems operating at less than 50MPa (conventional pressure, CP) to devices working at up to 200MPa [high pressure (HP) and ultra-high-pressure homogenization (UHP); Floury et al., 2000, 2002, 2003;Desrumaux and Marcand, 2002;Sandra and Dalgleish, 2005;Bouaouina et al., 2006;Roach and Harte, 2008].

The forces involved in HP homogenization may produce different effects on food macromolecules (fat, proteins, and polysaccharides). With regard to fat, the homogenization process results in smaller fat globules and a more uniform size distribution, thus limiting the rate of phase separation. The mean fat drop size (d) is reduced, with increasing pressures (P), following the exponential relationship d ∝ P−m, with m related to the applied pressure and the fat content of the system (Phipps, 1975;Kessler, 1981;Floury et al., 2000, 2003;Desrumaux and Marcand, 2002). The value of m has been reported as equal to 0.6 in emulsions with a low dispersed phase fraction, whereas it decreases at higher fat contents (Floury et al., 2000).

With reference to milk proteins, there are few works in the literature on dynamic HP homogenization effects. In oil emulsions, no change has been observed in the primary and secondary structure of β-LG and α-LA after HP treatment >200MPa (Subirade et al., 1998;Paquin, 1999;Desrumaux and Marcand, 2002). Nevertheless, it has been suggested that the protein architecture is stabilized by slightly different interactions following the treatment (Subirade et al., 1998). In contrast, Bouaouina et al. (2006) did not observe significant changes in the native structure of α-LA and β-LG when they were in solution. In regard to caseins, their micelle sizes have been found to be modified through HP homogenization treatment depending on the pressure applied (Hayes and Kelly, 2003;Sandra and Dalgleish, 2005;Roach and Harte, 2008).

With regard to polysaccharides, Paquin (1999) found a decrease in the average molecular weight of xanthan after homogenization because of an irreversible disruption of the biopolymer. A subsequent change in rheological properties was observed. Similarly, HP homogenization treatment significantly reduced the average molecular weight of methylcellulose. The emulsions stabilized by the disrupted molecules were found to lose their shear thinning behavior and to undergo a large decrease in viscosity (Floury et al., 2003).

Ice cream is a multicomponent system in which air bubbles, ice crystals, and fat globules are dispersed in a freeze-concentrated continuous watery phase consisting of sugars and polysaccharides, milk proteins, and salts (Goff, 1997;Muse and Hartel, 2004). Creaminess, texture, and meltdown behavior of ice cream are strongly affected by partial coalescence of the destabilized fat globules occurring during the freezing process (Marshall and Arbuckle, 1996;Goff, 1997;Goff et al., 1999). Fat globule destabilization is promoted by the modification of the membrane, which is broken down by homogenization. A subsequent rearrangement of the fat globule membrane occurs first with the adsorption of milk proteins, which are then partially displaced by the emulsifiers. The new recombined membrane becomes weaker and thus, much less stable to freezing shear forces. This favors a partial coalescence of the destabilized fat globules leading to a network that stabilizes the air bubbles and the foam structure (Marshall and Arbuckle, 1996;Goff, 1997).

Homogenization pressures employed in ice cream mixes generally range between 6 and 20MPa depending on fat content (Marshall and Arbuckle, 1996); extensive literature is available with regard to these CP (Schmidt and Smith, 1988,1989;Koxholt et al., 2001;Goff, 2002 among others). High pressure and UHP have been only recently tested on ice cream mixes and very little has been published (Hayes et al., 2003).

Ice cream mixes are hydrocolloidal dispersions in which proteins and polysaccharides are present in concentrations above the phase separation threshold, determining a 2-phase aqueous system, according to the concept of thermodynamic incompatibility (Tolstoguzov, 2003). Gelation of phase-separated systems can lead to gels filled with liquid or gel-like dispersed particles. Gels are multicomponent systems having a substantial liquid phase but exhibiting solid-like behavior. Ice cream mixes may be considered emulsion-filled gels; that is, macromolecular gels containing dispersed fat particles (filler). The latter may exhibit different filler-gel matrix interactions, strongly influencing structure and rheological properties (van Vliet, 1988). No literature is available with regard to the effects of different homogenization pressures on this complex structure.

In regard to structure, some information on the type of interactions between the conformation and functional properties of molecules and biopolymers can be obtained by rheological analysis such as steady shear flow and dynamic tests.

The aim of this study was to investigate the effect of different homogenization pressures on fat globule size and distribution as well as on structure-property relationships of ice cream mixes. For these purposes, dynamic light scattering, steady shear, and dynamic rheological analyses were performed on mixes with different fat contents (5 and 8%) homogenized at 15/3MPa and 97/3MPa and then aged for 4 and 20h at 4°C. Unhomogenized mixes were used as controls.

Back to Article Outline

Materials and Methods 

Ingredients and Ice Cream Mix Formulation 

The following ingredients were used to prepare the ice cream mixes: commercial homogenized and pasteurized whole milk and milk cream (35% fat) and sucrose (all purchased in a local market), maltodextrins (Natural World s.r.l., Ravenna, Italy), 30 dextrose equivalent glucose (Cerestar France, Haubourdin Cedex, France), skim milk powder (Bayerische Milchindustrie, Landshut, Germany), whey protein concentrate (Borculo Domo Ingredients, Zwolle, the Netherlands), carboxymethylcellulose (Comiel s.r.l., Milan, Italy), guar gum (Indian Gum Industries Ltd., Mumbai, India), locust bean gum (LBG Sicilia s.r.l., Ragusa, Italy), and mono- and diglycerides of fatty acids (Natural World s.r.l.). Two formulations were prepared, containing 5 and 8% fat, respectively. Table 1 reports their overall composition.

Table 1. Composition of 5 and 8% fat ice cream mixes
Item15% fat8% fat
Total sugars (g/100g of mix) from:19.4219.42
Sucrose (% of total sugars)76.3476.34
30 DE glucose (% of total sugars)21.3721.37
Maltodextrins (% of total sugars)2.292.29
Total NMS (g/100g of mix) from:10.8310.58
Milk (% of total NMS)55.4049.62
SMP (% of total NMS)28.9929.68
WPC (% of total NMS)10.9911.25
Milk cream (% of total NMS)4.629.45
Total fat (g/100g of mix) from:5.347.98
Milk cream (% of total fat)54.6873.31
Milk (% of total fat)43.6325.56
WPC (% of total fat)1.501.00
SMP (% of total fat)1.090.13
Emulsifiers (g/100g of mix)0.50.5
Total stabilizers (g/100g of mix) from:0.250.25
CMC (% of total stabilizers)6060
Guar gum (% of total stabilizers)2020
Locus bean gum (% of total stabilizers)2020
Total solids (g/100g of mix)36.3538.73

1DE = dextrose equivalents; NMS = nonfat milk solids; SMP = skim milk powder; WPC = whey protein concentrate; CMC = carboxymethylcellulose.

Ice Cream Mix Processing 

The dry ingredients were blended together and added to the liquid ingredients (milk and milk cream) at 65°C. The mix was then pasteurized at 82°C for 8min and divided into 3 equal portions. One portion was not homogenized (CON) and the 2 remaining portions were immediately subjected to homogenization in a 2-stage mode homogenizer (Panda 2K, Niro Soavi s.p.a., Parma, Italy). Conventional homogenization (CP) was performed at a primary pressure of 15MPa and a secondary pressure of 3MPa with an inlet temperature of 65 to 70°C. High-pressure homogenization (HP) was conducted at 97MPa and 3MPa with an inlet temperature of 45 to 48°C (Hayes et al., 2003). The outlet temperatures of all samples did not exceed 80°C. The CON, CP, and HP mix samples were cooled to 4°C and aged for 4 and 20h without stirring.

Particle Size Distribution Analysis 

Particle size distribution of the mix samples was measured by dynamic light scattering, using a Nicomp 380 ZLS analyzer (Particle Sizing System Nicomp, Santa Barbara, CA). All mix samples were diluted 1:1,000 with a dissociation medium (aqueous solution of 1% wt/vol SDS; Sigma-Aldrich, Steinheim, Germany). The measurement parameters were set at 25°C constant temperature, 1.333 refractive index, scattering angle of 90° for CON and CP mix samples and 170° for HP samples. Data were integrated over 3min. For size analysis, the Nicomp volume-weighted distribution was used.

Microstructure 

Magnifications (1,000×) of the ice cream mixes were obtained by using an optical microscope (Axiophot, Carl Zeiss, Oberkochen, Germany) at ambient temperature, using differential interference contrast mode. Samples were placed on a glass microscope slide, covered with a glass coverslip, and immediately observed.

Rheological Analysis 

Rheological tests were carried out by means of a controlled stress rheometer (StressTech rheometer, Reologica Instruments AB, Lund, Sweden) at 4.0±0.2°C, using a cone-plate sensor geometry (cone angle 4°, 40mm diameter). Before analysis, the sample placed in the rheometer cell rested for 5min to allow the stress induced during loading to relax.

Steady Shear Flow 

Steady shear flow curves were determined at shear rates ranging from 0.3160s−1 to 152.5s−1. The power law equation was applied to samples exhibiting shear thinning and the Herschel Bulkley to samples with yield stress (Stress Tech 2.24 software for Windows, Reologica Instruments AB). Apparent viscosities (ηapp) were taken at a shear rate of 20s−1.

Dynamic Oscillatory Measurements 

Linear viscoelastic regions were determined by stress sweeps at a fixed frequency of 1Hz. Mechanical spectra were performed in the frequency range 0.1 to 10Hz, at shear stresses within the linear viscoelastic region.

Statistical Analysis 

For each mix sample, at least 3 experiments were independently performed and all analyses were carried out at least 3 times on the samples. Thus, the data shown are the averages of at least 9 values. Student's t-test was used to compare 2 means, and one-way ANOVA (F-test) and Tukey's Honestly Significant Difference test were used for multiple comparisons. In both cases, the differences between the means were considered statistically significant for P-values <0.05. All statistical analyses were conducted using the software Statistical Discovery JMP 3.0 for Windows (SAS Institute Inc., Cary, NC).

Back to Article Outline

Results and Discussion 

Particle Size Distribution 

As an example, Figure 1 shows the particle size distributions of the CON, CP, and HP homogenized ice cream mixes. The CON sample showed a bimodal distribution with a main group of particles above 1,000nm and a second group at about 100 to 150nm (Figure 1A). Both the CP and HP samples were characterized by monomodal distributions with smaller mean diameters with respect to the main group of CON (Figure 1B, C). In the CON mix, the main group is attributed to the fat globules derived from milk and milk cream, whereas, presumably, the second group refers to casein micelles. Indeed, commercial milk homogenization at 20 and 10MPa followed by UHT treatment leads to a casein micelle diameter of 110 to 270nm (García-Risco et al., 2002). Moreover, it has been reported that casein micelles are disaggregated only by homogenization pressures of 100MPa (Roach and Harte, 2008). The CP treatment of ice cream mixes could allow casein micelles to disaggregate because of the higher viscosity of the mixes with respect to commercial milk and milk cream. In fact, with the increase in medium viscosity, an increase in the resulting pressure has been reported (Phipps, 1975). Moreover, adsorption of caseins to the fat by the new surface area created following CP and HP may also occur.

  • View full-size image.
  • Figure 1. 

    Representative volume-weighted distributions of ice cream mix samples: A) unhomogenized (CON); B) homogenized at 15 and 3 MPa pressure (CP); C) homogenized at 97 and 3 MPa pressure (HP).

As shown in Figure 2, fat globule mean diameter was reduced by homogenization even though the milk and milk cream used as ingredients had already been subjected to industrial homogenization. As noted above for casein micelle disappearance, the conventional homogenization of mixes would cause a further decrease in fat globule size with respect to the CON mixes, because of the higher viscosity of mixes. Moreover, although mean diameter was consistently decreased by CP, HP caused only a small additional decrease. Fat globule size is known to vary as a function of the applied pressure with an exponential relationship (Kessler, 1981;Floury et al., 2000). In most cases, the homogenized mixes with 5% fat showed smaller globules than the mixes with 8% fat. Indeed, fat content was found to influence the relationship between mean diameter and homogenization pressure, with a decrease in the absolute value of the exponent for greater fat amounts (Kessler, 1981;Floury et al., 2000). Different aging times did not affect globule size distribution as also observed by Gelin et al. (1994) and Alvarez et al. (2005). Gelin et al. (1994) attributed this lack of difference to the fact that no noticeable destabilization of the fat globule membranes would occur during aging until freezing takes place with the consequent physical changes.

  • View full-size image.
  • Figure 2. 

    Mean fat globule volume-weighted diameters of ice cream mixes with 5% and 8% fat content and aged for 4 h (panel 1) or 20 h (panel 2). CON = unhomogenized; CP = homogenized at 15 and 3 MPa pressure; HP = homogenized at 97 and 3 MPa pressure. Statistical analysis (Tukey's HSD test, P<0.05) was carried out within groups at different fat content; lowercase letters refer to 5% fat content mixes, and uppercase letters refer to 8% fat content mixes. Asterisks mark statistically different pairs (Student's t-test, P<0.05).

Microstructure 

Figure 3 shows the microstructures of CON, CP, and HP mixes at 5 and 8% fat content. Spheroidal elements are fat globules that were smaller in the CP and HP samples compared with the CON mix. A further reduction in size was not as evident as it was when observed with dynamic light scattering measurements. The 8% fat HP sample showed a more pronounced 3-dimensional structure.

  • View full-size image.
  • Figure 3. 

    Optical light microscope images of mixes at 5% (left column) and 8% (right column) fat content. CON = unhomogenized; CP = homogenized at 15 and 3 MPa pressure; HP = homogenized at 97 and 3 MPa pressure. Magnification 1,000×; scale bar = 25μm.

Steady Shear Flow Behavior 

Steady shear flow determinations allow the shear stress curves as a function of shear rate to be determined. Figure 4 shows representative steady shear flow curves of 5 and 8% fat ice cream mixes aged for 4 and 20h. All samples were non-Newtonian. The CON and CP samples were shear-thinning and hence, were described by the power law equation (R2 = 1). The HP sample showed a yield stress, which is the characteristic behavior fitted by the Herschel-Bulkley equation (R2 = 0.999). In shear-thinning systems (n<1), dispersed particles may disaggregate or orient in the direction of flow, opposing less resistance (Lapasin and Pricl, 1995;Aguilera and Stanley, 1999). Both fat globules and stabilizers (locust bean gum and guar gum) may be responsible for the shear-thinning behavior of the mixes (Cottrell et al., 1980;Goff et al., 1994). With regard to HP mixes, the logarithmic plot of shear stress versus shear rate reported in Figure 5 enables the plastic behavior of these samples to be further highlighted (Lapasin and Pricl, 1995). In systems with a yield point, a network structure has to be broken prior to flow. Higher yield stress values marked the 8% fat mixes (Figure 4), indicating more firmly structured systems and, according to Rao (2007), more stable emulsions. In particular, mean yield stress values increased from about 20Pa for 5% fat HP mixes to about 80Pa for 8% fat HP mixes. In general, the magnitude of the yield stress increases with increasing particle volume fraction, decreasing particle size, and increasing magnitude of interparticle forces (Poslinski et al., 1988;Genovese et al., 2007). Flow curves of the 5% fat samples (Figure 4) changed following aging, and a shift to higher stress values was observed. Analogously, apparent viscosity at 20s−1 of 5% fat CON and CP mixes significantly increased after 20h of aging (Table 2). This was expected because the hydration of proteins and stabilizers during aging causes a viscosity increase (Marshall and Arbuckle, 1996;Goff, 1997) possibly because of protein displacement from the surface back to the serum phase. No aging effect was observed in 5% fat HP mixes or in any of the 8% fat mixes. The latter had more surface area and hence, more adsorbed proteins, resulting in a faster stabilizing effect.

  • View full-size image.
  • Figure 4. 

    Representative steady shear flow curves of 5% fat and 8% fat ice cream mixes aged for 4 and 20h. CON = unhomogenized; CP = homogenized at 15 and 3MPa pressure; HP = homogenized at 97 and 3MPa pressure.

  • View full-size image.
  • Figure 5. 

    Logarithmic plot of shear stress versus shear rate of ice cream mixes. CON = unhomogenized; CP = homogenized at 15 and 3MPa pressure; HP = homogenized at 97 and 3MPa pressure. Inner rectangle: Theoretical shear thinning (a) and plastic (b) behaviors reported in Lapasin and Pricl (1995).

Table 2. Mean values and standard deviations of apparent viscosity (Pa·s; shear rate: 20s−1) of 5% fat and 8% fat ice cream mixes subjected or not to homogenization and aged for 4 and 20h
5% fat8% fat
Ice cream mix14h20h4h20h
CON0.85ab±0.67*1.44b±0.05*2.07b±0.062.06b±0.32
CP0.31b±0.07*0.40c±0.03*0.53c±0.050.52c±0.01
HP1.85a±0.792.28a±0.554.70a±0.724.57a±0.18

a-cDifferent letters within the same column refer to statistical differences (Tukey's HSD test, P<0.05).

1CON (control) = unhomogenized mix; CP (conventional pressure) = mix homogenized at 15 and 3MPa; HP (high pressure) = mix homogenized at 97 and 3MPa.

*Within the same row, asterisks indicate statistical differences between 4 and 20h aging time (Student's t-test, P<0.05).

Apparent viscosities were lower for CP and higher for HP compared with CON for both 5% fat and 8% fat samples (Table 2). Schmidt and Smith (1989) observed a viscosity decrease in mixes homogenized at CP close to the values used in the present study. Those authors attributed the lowering of viscosity to changes in the size of the fat globules. With smaller fat globules, internal resistance is less than with larger ones or linear chains of globules. High-pressure homogenization caused a marked increase in the ηapp of mixes (Table 2), as was reported by Hayes et al. (2003), who attributed the viscosity increase to greater amounts of absorbed proteins or more tightly packed proteins at the oil-water interface. Viscosity changes after processing were more pronounced in high fat content samples, which also had a higher initial ηapp. Similarly, Hayes et al. (2003) reported higher viscosities for the 8% fat HP and CP mixes compared with the 5% fat mixes. They also reported comparable viscosities for 5% fat HP and 8% fat CP ice cream mixes. In the present case, the viscosity of the 5% fat HP mix was comparable to that of the 8% fat CON mix. This difference can be attributed to the mix formulations used in the present study, which differed from those of Hayes et al. (2003) mainly because of a greater amount of stabilizers and emulsifiers.

Dynamic Properties 

Mechanical spectra were performed in the frequency range 0.1 to 10Hz, at shear stresses within the linear viscoelastic region (previously determined by frequency stress sweeps). Linear viscoelastic ranges were seen to be wider and at higher stress values for HP-treated samples (data not shown). Critical deformations; that is, the maximum strain characterizing the limit of the linear viscoelastic regimen, were always less than approximately 6%. In the literature, linear viscoelastic strain is reported to extend to approximately 20% for strong gels, whereas for weak gels it is usually <5%, but can be very much smaller (up to 1,000 times smaller; Ross-Murphy, 1995). An investigation of gel properties was performed by means of oscillatory measurements at a constant stress in the frequency range 0.1 to 10Hz. Figure 6 shows a representative mechanical spectrum of an HP mix in which the elastic modulus (G′) is always higher than the viscous modulus (G″) and both are slightly frequency dependent; the phase angle is within 20 to 30° and also slightly frequency dependent; complex viscosity decreases with increasing frequency. Systems with these characteristics are referred to as weak gels (Lapasin and Pricl, 1995;Ross-Murphy, 1995). Similar trends, which can be mainly attributed to the colloidal phase, were obtained for all samples, including the CON mixes. Viscoelastic properties may arise from pasteurization, which causes some protein denaturation and contributes to gel formation.

Although all samples exhibited gel-like properties, differences in the absolute value (compared at 1Hz) of dynamic moduli were observed. Figure 7 shows the mean values and the standard deviations of complex moduli (G*) of CON, CP, and HP ice cream mixes at 5 and 8% fat content aged for 4 and 20h. The complex modulus G* was chosen as an indication of the system strength; G* of 5% fat CP mixes was the lowest and G* of HP 8% fat the highest. All 5% fat mixes showed a G* increase at 20h compared with 4h. Conversely, in the 8% fat samples, G* was almost constant suggesting that higher fat contents had a stabilizing effect on the viscoelastic properties of mixes, as was also observed for apparent viscosity (Table 2). The HP samples, both 5 and 8% fat, showed higher G*. As described above, ice cream mixes may be considered as emulsion-filled gels. The elastic modulus G′ in HP mixes markedly increased from 258Pa in 5% fat mixes to 1,804Pa in 8% fat mixes; therefore it can be inferred that, in these systems, fat behaves as an interactive filler. This may arise from HP treatment, because CP and CON did not show an increase in G′.

  • View full-size image.
  • Figure 7. 

    Mean values and standard deviations of complex modulus of ice cream mixes at 5% and 8% fat content aged for 4 and 20h. CON = unhomogenized; CP = homogenized at 15 and 3MPa pressure; HP = homogenized at 97 and 3MPa pressure.

Another useful parameter to characterize a viscoelastic system is the ratio G″:G′, named loss tangent (tan δ), which is a measure of the viscous/elastic ratio for a material at frequency ω (Ross-Murphy, 1995). The loss tangent curve did not show variations over the frequency range 0.1 to 10Hz (data not shown). Values of tan δ at 1Hz frequency are compared in Table 3. Both 5 and 8% fat HP mixes showed tan δ lower than those of CP and CON, indicating a more structured gel-like system. The HP homogenization would give structure to the system and strengthen the elastic component, possibly as a consequence of pressure-induced colloidal interactions and network formation. The Cox-Merz superposition rule constitutes a further criterion for the distinction between gel systems and biopolymer entangled solutions, by combining oscillatory and steady shear measurements. In the Cox-Merz superposition the curves of η versus shear rate and η* versus ω are plotted on the same diagram. The results, reported in Figure 8, show that in all cases the curves are not superposed, which is typical of colloidal and particulate gel networks (Ross-Murphy, 1995). In CON samples, the curves tend to cross with increasing shear rate and frequency, whereas in CP mixes, this tendency was less clear. In HP mixes the 2 curves were in parallel. The convergence of the 2 curves refers to weak gels, whereas the parallelism is typical of strong gels (Ross-Murphy, 1995). According to these data, CON and CP mixes behaved as weak gels and HP samples as stronger gels. High-pressure homogenization on a model oil-in-water emulsion was used by Floury et al. (2000) and they related the effect of HP homogenization not only to the change in emulsion droplet size but also to the properties of the stabilizer molecules. The authors hypothesized an unfolding or partial denaturation of the globular proteins caused by treatments at high pressure and simultaneously high temperature. A similar mechanism may be suggested to explain the HP structuring effect on the ice cream mixes; that is, the HP treatment could cause the unfolding or partial denaturation of both whey protein and caseins leading to a network formation or rearrangement or to interpenetrating networks within the colloidal phase. In contrast, CP pressures would not enable changes at such extents. Similarly, Floury et al. (2002) reported a strong gel-type behavior for soy protein-stabilized emulsions when homogenized at UHP (250 and 350MPa). The authors suggested that disulfide bonding and noncovalent bonds such as hydrophobic interaction and ionic and hydrogen bonding may be involved in the gel-like formation when produced at homogenizing pressures >200MPa.

Table 3. Mean values and standard deviations of calculated tan δ of ice cream mixes at 5% and 8% fat content aged for 4 and 20h
5% fat8% fat
Ice cream mix14h20h4h20h
CON0.70±0.080.57±0.020.66±0.090.54±0.04
CP0.72±0.020.53±0.030.55±0.020.53±0.02
HP0.38±0.030.36±0.040.33±0.010.35±0.02

1CON (control) = unhomogenized mix; CP (conventional pressure) = mix homogenized at 15 and 3MPa; HP (high pressure) = mix homogenized at 97 and 3MPa.

  • View full-size image.
  • Figure 8. 

    Cox-Merz superposition of steady shear flow viscosity (η) versus shear rate (γ, ■) and complex viscosity (η*) versus frequency (ω, □) for samples at 5% fat (left column) and 8% fat (right column) aged for 20h.

Back to Article Outline

Conclusions 

Homogenization of ice cream mixes resulted in a change from bimodal to monomodal particle size distributions with a progressive reduction in the mean particle diameter as pressure increased from conventional to high values. Because the milk and milk cream used as ingredients in mix formulation were already conventionally homogenized, the further modification in the particle sizes and distributions following the CP process of the mixes could be attributed to the effects of the increased viscosity of ice cream mixes. Fat globule mean diameters were reduced at higher pressure, but the homogenization effect on size reduction was limited with the highest fat content.

The rheological behavior of ice cream mixes was determined by both the dispersed and the continuous phase. In most cases, the dispersed phase accounted mainly for viscosity, whereas the colloidal phase was responsible for the viscoelastic properties and gel behavior. Higher fat contents resulted in both higher viscosity and dynamic moduli. Homogenization at CP mainly affected the dispersed phase and determined a more pronounced viscosity reduction in the higher fat content mixes. This effect was related to fat globule size reduction, whereas the viscoelastic properties were only slightly influenced by the process. Homogenization at HP caused a further slight decrease in particle size with respect to CP, but a strong enhancement of the viscoelastic properties and apparent viscosity. Hence, the HP treatment acted mainly on the colloidal phase.

Rheological results indicated that unhomogenized and CP mixes behaved as weak gels, whereas HP treatment led to stronger gels possibly as the overall result of a network rearrangement or interpenetrating network formation and the fat globules were found to behave as active filler. It should be noted that the HP process produced an apparent viscosity in the 5% fat mix comparable to that of the 8% fat CON mix. Further research needs to be undertaken to determine whether HP homogenization of ice cream mixes could enable the production of lower fat ice creams with similar characteristics to those of higher fat ice creams.

Back to Article Outline

Acknowledgments 

The authors are grateful to Enrico Maltini(Department of Food Science, University of Udine) for his helpful advice.

Back to Article Outline

References 

  1. Aguilera JM, Stanley DW. Microstructural principles of food processing and engineering. 2nd ed.. Gaithersburg, MD: Aspen Publishers Inc.; 1999;
  2. Alvarez VB, Wolters CL, Vodovotz Y, Ji T. Physical properties of ice cream containing milk protein concentrates. J. Dairy Sci. 2005;88:862–871
  3. Bouaouina H, Desrumaux A, Loisel C, Legrand J. Functional properties of whey proteins as affected by dynamic high-pressure treatment. Int. Dairy J. 2006;16:275–284
  4. Cottrell JIL, Pass G, Phillips GO. The effect of stabilizers on the viscosity of an ice cream mix. J. Sci. Food Agric. 1980;31:1066–1070
  5. Desrumaux A, Marcand J. Formation of sunflower oil emulsions stabilized by whey proteins with high-pressure homogenization (up to 350MPa): Effect of pressure on emulsion characteristics. Int. J. Food Sci. Technol. 2002;37:263–269
  6. Fellows P. Food Processing Technology. Principles and Practice. Chichester, UK: Ellis Horwood Ltd; 1988;
  7. Floury J, Desrumaux A, Axelos MAV, Legrand J. Effect of high pressure homogenisation on methylcellulose as food emulsifier. J. Food Eng. 2003;58:227–238
  8. Floury J, Desrumaux A, Lardières J. Effect of high-pressure homogenization on droplet size distributions and rheological properties of model oil-in-water emulsions. Innov. Food Sci. Emerg. 2000;1:127–134
  9. Floury J, Desrumaux A, Legrand J. Effect of ultra-high-pressure homogenization on structure and on rheological properties of soy protein-stabilized emulsions. J. Food Sci. 2002;67:3388–3395
  10. García-Risco MR, Ramos M, López-Fandiño R. Modifications in milk proteins induced by heat treatment and homogenization and their influence on susceptibility to proteolysis. Int. Dairy J. 2002;12:679–688
  11. Gelin JL, Poyen L, Courthaudon JL, Le Meste M, Lorient D. Structural changes in oil-in-water emulsions during the manufacture of ice cream. Food Hydrocolloids. 1994;3/4:299–308
  12. Genovese DB, Lozano JE, Rao MA. The rheology of colloidal and noncolloidal food dispersions. J. Food Sci. 2007;72:11–20
  13. Goff HD. Colloidal aspects of ice cream-A review. Int. Dairy J. 1997;7:363–373
  14. Goff HD. Formation and stabilisation of structure in ice-cream and related products. Curr. Opin. Colloid. Interface Sci. 2002;7:432–437
  15. Goff HD, Davidson VJ, Cappi E. Viscosity of ice cream mix at pasteurization temperatures. J. Dairy Sci. 1994;77:2207–2213
  16. Goff HD, Verespej E, Smith AK. A study of fat and ice structures in ice cream. Int. Dairy J. 1999;9:817–829
  17. Hayes MG, Kelly AL. High pressure homogenisation of raw whole bovine milk (a) effects on fat globule size and other properties. J. Dairy Res. 2003;70:297–305
  18. Hayes MG, Lefrancois AC, Waldron DS, Goff HD, Kelly AL. Influence of high pressure homogenization on some characteristics of ice cream. Milchwissenschaft. 2003;58:519–523
  19. Kessler HG. Food Engineering and Dairy Technology. Freising, Germany: Verlag A. Kessler; 1981;
  20. Koxholt MMR, Eisenmann B, Hinrichs J. Effect of fat globule sizes on the meltdown of ice cream. J. Dairy Sci. 2001;84:31–37
  21. Lapasin R, Pricl S. Rheology of industrial polysaccharides. Theory and application. Glasgow, UK: Blackie Academic & Professional; 1995;
  22. Marshall RT, Arbuckle WS. Ice Cream. 5th ed.. New York, NY: Chapman and Hall; 1996;
  23. Muse MR, Hartel RW. Ice cream structural elements that affect melting rate and hardness. J. Dairy Sci. 2004;87:1–10
  24. Paquin P. Technological properties of high pressure homogenizers: the effect of fat globules, milk proteins and polysaccharides. Int. Dairy J. 1999;9:329–335
  25. Phipps LW. The fragmentation of oil drops in emulsions by high-pressure homogenizer. J. Phys. D Appl. Phys. 1975;8:448–462
  26. Poslinski AJ, Ryan ME, Gupta RK, Seshadri SG, Frechette FJ. Rheological behaviour of filled polymeric systems. I. Yield stress and shear thinning effects. J. Rheol. (N.Y.N.Y.). 1988;32:703–735
  27. Rao MA. Rheology of fluid and semisolid foods. Principles and applications. 2nd ed.. New York, NY: Springer; 2007;
  28. Roach A, Harte F. Disruption and sedimentation of casein micelles and casein micelle isolates under high-pressure homogenization. Innov. Food Sci. Emerg. 2008;9:1–8
  29. Ross-Murphy SB. Rheology of biopolymer solutions and gels. In:  Dickinson E editors. New Physico-Chemical Techniques for the Characterization of Complex Food Systems. Glasgow, UK: Blackie Academic & Professional; 1995;p. 139–156
  30. Sandra S, Dalgleish DG. Effects of ultra-high-pressure homogenization and heating on structural properties of casein micelles in reconstituted skim milk powder. Int. Dairy J. 2005;15:1095–1104
  31. Schmidt KA, Smith DE. Effects of homogenization on sensory characteristics of vanilla ice cream. J. Dairy Sci. 1988;71:46–51
  32. Schmidt KA, Smith DE. Effects of varying homogenization pressure on the physical properties of vanilla ice cream. J. Dairy Sci. 1989;72:378–384
  33. Subirade M, Loupil F, Allain A-F, Paquin P. Effect of dynamic high pressure on the secondary structure of β-lactoglobulin and on its conformational properties as determined by Fourier transform infrared spectroscopy. Int. Dairy J. 1998;8:135–140
  34. Tolstoguzov V. Some thermodynamic considerations in food formulation. Food Hydrocolloids. 2003;17:1–23
  35. van Vliet T. Rheological properties of filled gels. Influence of filler matrix interaction. Colloid Polym. Sci. 1988;266:518–524

PII: S0022-0302(09)70501-6

doi:10.3168/jds.2008-1797

Journal of Dairy Science
Volume 92, Issue 5 , Pages 1864-1875, May 2009

Article Tools

* Image Usage & Resolution