- Email: [email protected]
Influence of milk pre-heating conditions on casein–whey protein interactions and skim milk concentrate viscosity
International Dairy Journal 69 (2017) 19e22
Contents lists available at ScienceDirect
International Dairy Journal journal homepage: www.elsevier.com/locate/idairyj
Influence of milk pre-heating conditions on caseinewhey protein interactions and skim milk concentrate viscosity Suresh G. Sutariya a, Thom Huppertz a, b, Hasmukh A. Patel a, c, * a
Dairy Science Department, South Dakota State University, Brookings, SD, USA NIZO Food Research, P.O. Box 20, 6710 BA, Ede, The Netherlands c Dairy Foods Research and Development, Land O'Lakes, Inc., Arden Hills, MN, USA b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 18 December 2016 Received in revised form 16 January 2017 Accepted 16 January 2017 Available online 4 February 2017
The viscosity of concentrates (50e55% total solids) prepared from skim milk heated (5 min at 80 or 90 C) at pH 6.5 and 6.7 was examined. The extent of heat-induced whey protein denaturation increased with increasing temperature and pH. More denatured whey protein and k-casein were found in the serum phase of milk heated at higher pH. The viscosity of milk concentrates increased considerably with increasing pH at concentration and increasing heating temperature, whereas the distribution of denatured whey proteins and k-casein between the serum and micellar phase only marginally influenced concentrate viscosity. Skim milk concentrate viscosity thus appears to be governed primarily by volume fraction and interactions of particles, which are governed primarily by concentration factor, the extent of whey protein denaturation and pH. Control and optimization of these factors can facilitate control over skim milk concentrate viscosity and energy efficiency in spray-drying. © 2017 Elsevier Ltd. All rights reserved.
1. Introduction Evaporation and spray drying are two major steps involved in the manufacture of skim milk powder. Since removal of water by evaporation requires ~10e20 times less energy than by spraydrying (Fox, Akkerman, Straatsma, & De Jong, 2010), maximizing the solids content of concentrate prior to spray-drying is desired economically and environmentally. Commonly, skim milk powder manufacturers concentrate skim milk to 45e50% total solids prior to drying. At higher solids content, viscosity increases dramatically with further increases in total solids. The viscosity of concentrated milk depends primarily on the solids content and composition of the milk, heating conditions, evaporation temperature and holding time of the concentrate (Bloore & Boag, 1981; Fernandez-Martin, 1972; Karlsson, Ipsen, € , 2005; Snoeren, Damman, & Klok, 1982; Ve lezSchrader, & Ardo novas, 1998). Snoeren et al. (1982) showed that Ruiz & Barbosa-Ca increases in viscosity as a result of heating of milk are related to an increase in milk protein voluminosity, because denaturation increases the voluminosity of the whey proteins; as a result,
* Corresponding author. Tel.: þ1 651 375 1497. E-mail address: [email protected] (H.A. Patel). http://dx.doi.org/10.1016/j.idairyj.2017.01.007 0958-6946/© 2017 Elsevier Ltd. All rights reserved.
concentrate viscosity was found to correlate strongly with the extent of whey protein denaturation (Snoeren et al., 1982). More recently, Anema, Lowe, Lee, and Klostermeyer (2014) suggested that not only the extent of denaturation, but also the distribution of denatured whey proteins, and k-casein, between the serum and micellar phase affects the viscosity of milk concentrates for solids content <40%, but that at higher total solids content, viscosity was less influenced by the distribution of denatured whey proteins and k-casein. Hence, for reducing skim milk concentrate viscosity in industrial operations, tailoring caseinewhey protein interactions does not appear to be a key route based on the results of Anema et al. (2014). However, the studies performed by Anema et al. (2014) only studied solids contents up to 45%, which are at the lower end of industrial processing. For investigating options to improve spray-drying efficiency by reducing milk concentrate viscosity, effects at solids contents >50% should be considered. Another factor that is crucial to consider in these studies is interparticle interactions. In unconcentrated milk, where the interparticle distance is several times larger than the particle diameter, such effects may appear to be limited. However, in concentrated milk the inter-particle distance becomes considerably smaller than particle diameter and particle interactions become much more important in governing rheological behavior (Karlsson et al., 2005), particularly in milk concentrates of >50% total solids.
20
S.G. Sutariya et al. / International Dairy Journal 69 (2017) 19e22
Hence, the objectives of the current work were to evaluate the role of whey protein distribution, as well as pH, on the rheological behavior of skim milk concentrates (>50% dry matter) and evaluate whether protein distribution or other factors govern the rheological behavior of the concentrates. 2. Materials and methods 2.1. Sample preparation Low-heat skim milk powder (Associated Milk Producers Inc., New Ulm, MN, USA; whey protein nitrogen index > 6; 36%, w/w, protein on a dry matter basis) was reconstituted in deionized water at 10% (w/v) for 4 h at room temperature. Sodium azide (0.02%) was added to prevent bacterial growth and the pH was subsequently adjusted to 6.5 and 6.7 by addition of 1 M HCl or 1 M NaOH with continuous stirring. The pH-adjusted skim milk samples were then stored overnight at 4 C. Samples were brought to room temperature the next day and pH was checked and readjusted if required. Samples were the heated in a hot water bath to 80 or 90 C, held at this temperature for 5 min and subsequently cooled to room temperature in an ice-water bath. For samples at pH 6.5 and 6.7 that were heated at 80 C, subsamples were subsequently adjusted to pH 6.7 and 6.5, respectively. For each sample variant described above, 3 replicate samples were prepared. 2.2. Determination of whey protein denaturation and whey protein distribution
180e190 mbar. Concentration was continued until a solids content of 50e55% (m m1) was reached. Solids content was determined by oven drying. Immediately after concentration, apparent viscosity was measured at 55 C, over a shear rate profile of 50e500 s1, using a STRESSTECH rheometer (ATS Rheosystems, Bordentown, NJ, USA), fitted with an CC 25 CCE SS cup and bob attachment. Shear rate was increased by 50 s1 at 10 s intervals. Since milk concentrates showed non-Newtonian viscosity behavior, viscosity profiles were evaluated using a Power Law model:
s ¼ K*gn
(1)
where, s is the shear stress (Pa), g is the shear rate (s1), K is the consistency coefficient (Pa sn) and n is the dimensionless flow behavior index. K and n were determined from a plot of log s versus log g, which has log K value as the intercept and the n as slope. For shear thinning liquids, 0 n 1, with lower values indicating a greater degree of shear thinning. 2.4. Statistical analysis The results were analyzed by SAS (version 9.3; SAS Institute Inc., Cary, NC, USA). Analysis of variance (ANOVA) was performed to examine the statistical difference between the samples, and the differences were considered significant when p-values were less than 0.05. All experiments were replicated on three individual milk samples. 3. Results
To determine the level of whey protein denaturation in heated milk samples, the pH 4.6-soluble fraction of milk samples was prepared as described by Gazi and Huppertz (2015). The pH 4.6soluble fraction and whole sample were subsequently analyzed by sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDSPAGE) as described by Meletharayil, Patel, and Huppertz (2015) and the stained gels were scanned using a Bio-5000 Microtek scanner (Microtek, Hsinchu, Taiwan) and the images were analyzed to quantify the intensity of whey protein bands using the Bio-image Intelligent Quantifier v 3.3.7 system (Bio Image Systems, Jackson, MI, USA). The level of whey protein denaturation was calculated from the difference in band intensities of the pH 4.6-soluble fraction unheated milk sample (residual native whey protein in the unheated samples) and the pH 4.6-soluble fraction of the heated sample (residual native whey protein in the heated sample). To determine the distribution of whey proteins in heated milk, samples were centrifuged at 25,000 g for 1 h at 20 C, followed by SDS-PAGE analysis, scanning and image quantification of the supernatants as described above. 2.3. Viscosity of concentrated milk samples Samples of 1.5 L of skim milk were concentrated at 60 C under vacuum using a rotary evaporator (Rotovapor Hei-VAP Value/G3, Heidolph, Schwabach, Germany) with an absolute pressure of
3.1. Whey protein denaturation and distribution in heated milk samples Heat treatment of samples at 80 C denatured ~43% and 50% of native whey protein present in samples at pH 6.5 and 6.7, respectively, whereas heat treatment of the same samples at 90 C denatured ~84% and 93% of total native whey protein present, respectively (Table 1). In addition, the distribution of whey proteins and k-casein between the serum phase and colloidal phase was also affected by the pH at which heat treatment was carried out. Heat treatment at pH 6.7 resulted in higher levels of whey proteins and k-casein in the serum phase of milk than heat treatment at pH 6.5 (Table 1); this, again, is in agreement with previous studies (Anema, 1998; Anema & Klostermeyer, 1997; Anema et al., 2014; Cassandra & Dalgleish, 2006; Renan et al., 2006; Vasbinder & de Kruif, 2003). 3.2. Viscosity of concentrated milk samples Samples were concentrated to a solids content of 50e55% dry matter by evaporation (Table 2) which was accompanied by a reduction in pH of the samples, by ~0.4e0.5 pH units, irrespective of starting pH. These decreases in pH are in agreement with previous studies and can be attributed to a concentration of acids by
Table 1 Extent of whey protein (WP) denaturation and individual proteins in the serum phase of milk heated at pH 6.5 or 6.7 at 80 or 90 C for 5 min.a pH at heat treatment
Temperature ( C)
6.5 6.7 6.5 6.7
80 80 90 90
Denatured WP (% total) 43 50 84 93
± ± ± ±
5a 2a 7b 2b
b-Lactoglobulin
a-Lactalbumin
(band intensity)
(band intensity)
0.80 1.06 0.32 0.79
± ± ± ±
0.01a 0.02b 0.00c 0.05a
0.47 0.74 0.24 0.33
± ± ± ±
0.01a 0.01b 0.01c 0.04d
k-Casein (band intensity) 0.37 0.62 0.28 0.55
± ± ± ±
0.05a 0.05b 0.04c 0.09d
a Values are means ± standard deviation of measurements on three individual milk samples; means within the same column not sharing a common superscript letter are significantly different (P < 0.05). Band intensity values are in arbitrary units.
S.G. Sutariya et al. / International Dairy Journal 69 (2017) 19e22
21
Table 2 pH and dry matter content of milk concentrate prepared from milk pre-heated at pH 6.5 or 6.7 at 80 or 90 C for 5 min followed by concentration to 50e55% dry matter at pH 6.5 or 6.7.a pH at heat treatment
pH at concentration
Heating temperature ( C)
pH after concentration
6.5 6.7 6.5 6.7 6.5 6.7
6.5 6.5 6.7 6.7 6.5 6.7
80 80 80 80 90 90
6.06 6.03 6.26 6.30
± ± ± ±
Dry matter after concentration (%, m m1)
0.03a 0.05a 0.02b 0.03b
54.2 54.9 52.6 52.5 52.5 50.7
± ± ± ± ± ±
0.0a 0.7a 0.1b 0.2b 0.1b 0.1c
a Values are means ± standard deviation of measurements on three individual milk samples; means within same column not sharing common superscript are significantly different (P < 0.05).
evaporation, as well as the reduced solubility of calcium phosphate, leading to insolubilization of calcium phosphate and a concomitant release of Hþ (Nieuwenhuijse, Timmermans, & Walstra, 1988). Viscosity of samples was determined as a function of shear rate profiles are shown in Fig. 1, showing notable differences in viscosity between different milk samples. Analysis of the viscosity profiles with the Power Law also highlighted large differences in the flow behavior index, n, and the consistency coefficient, K (Table 3). The differences in solids content of the concentrates (Table 2) were not the cause of differences in viscosity as actually a negative rather than positive correlation was found between viscosity and solids content (data not shown). Hence, effects of variables studied, i.e., pH at pre-heating, and concentration and temperature at preheating, had sufficiently strong effects on viscosity to ensure that variations in viscosity as a result of solids content were readily overruled (Fig. 1, Table 3). Skim milk concentrate viscosity was increased when the temperature of pre-heating was higher and when the pH of concentration was higher (Fig. 1). In addition, for samples pre-heated at the same temperature and concentrated at the same pH, viscosity was slightly higher when pre-heating was carried out at a higher pH, but such effects were considerably smaller than those for pre-heating temperature and pH at concentration. Along with increases in viscosity, increases in K and decreases in n were observed (Table 3) suggesting that, as expected, samples became more shear-thinning as viscosity increased.
4. Discussion When skim milk is concentrated, viscosity of the concentrate increases. In the very dilute region, where relative viscosity increases in linear fashion with increasing concentration of milk protein, these changes can be related to increases in volume fraction of suspended particles alone; however, in more concentrated systems relative viscosity increases more strongly with increasing protein concentration or volume fractions, due to hydrodynamic interactions between casein micelles. Data from Jeurnink and De Kruif (1993) suggest this is already the case in milk system and this effect increases further with increasing concentration factor. With increasing concentration factor of milk, particularly at solids contents >40%, milk concentrates start displaying shear-thinning lez-Ruiz & Barbosa-Ca novas, 1998) indicative of behavior (Ve (non-permanent) clustering of particles (Jeurnink & De Kruif, 1993). Increases in viscosity with concentration have been described by several empirical models, of which those by Eilers (1941), Mendoza and Santamaría-Holek (2009), Snoeren et al. (1982) and derivatives thereof are most commonly used. Although these models often show good fits to datasets, the physicochemical validity of the application of these models to milk may be questioned, because of their assumption of a constant voluminosity of the particles in the milk, i.e., the casein micelles and (denatured) whey proteins, which is unlikely to be the case. With increasing concentration factor of milk, ionic strength increases and pH decreases (Nieuwenhuijse et al., 1988), both of which are known to affect the hydration or voluminosity of casein micelles (Snoeren, Brinkhuis, Damman, & Klok, 1984). In addition, a decrease in pH will reduce the netnegative charge on the proteins, whereas an increase in ionic strength reduces the range over which the charges persist; hence, the decreases in pH and increase in ionic strength on concentration of milk will lead to compression of the electrical double layer and hydrodynamic radius (Karlsson et al., 2005), and thus volume fraction. Therefore, considering particle hydration and voluminosity to remain constant during concentration appears to be unjustified, Table 3 Power Law-derived flow behavior index (n) and consistency coefficient (K) of concentrates prepared by evaporation of milk pre-heated at pH 6.5 or 6.7 at 80 or 90 C for 5 min followed by concentration to 50e55% dry matter at pH 6.5 or 6.7.a
Fig. 1. Effect of shear rate on the viscosity of concentrates prepared by evaporation of milk pre-heated at pH 6.5 ( , , ) or pH 6.7 ( , , ) at 80 C ( , , , ) or 90 C ( , ) for 5 min followed by concentration to 50e55% dry matter at pH 6.5 ( , , ) or pH 6.7 ( , , ). Values are means of measurements on three individual milk samples, with the standard deviation indicated by vertical error bars.
pH at heat treatment
pH at concentration
Heating temperature
n ()
6.5 6.7 6.5 6.7 6.5 6.7
6.5 6.5 6.7 6.7 6.5 6.7
80 80 80 80 90 90
0.75 0.73 0.60 0.61 0.54 0.30
± ± ± ± ± ±
K (Pa sn) 0.13a 0.09a 0.07b 0.06b 0.06c 0.11d
1.8 ± 0.6a 2.4 ± 0.5a 12.8 ± 2.2b 15.9 ± 2.2b,c 18.4 ± 2.8c 155.7 ± 35.6d
a Values are means ± standard deviation of measurements on three individual milk samples, means within same column not sharing common superscript are significantly different (P < 0.05).
22
S.G. Sutariya et al. / International Dairy Journal 69 (2017) 19e22
as also shown experimentally by Liu, Dunstan, and Martin (2012) and also explains why apparently very high maximum volume fractions (fmax, i.e., 0.79) are described when using aforementioned models (Snoeren et al., 1982), whereas these appear unrealistically high when containing skim milk as a slightly polydisperse suspension of near-spherical particles (Farr & Groot, 2009). From viscosity data (Fig. 1, Table 3) it is clear that pre-heating intensity and concentration pH had the largest effect of skim milk concentrate viscosity, whereas the effect of the pH at which preheating was carried out was considerably smaller. The increased viscosity with increasing pre-heating intensity is in line with previous results (Snoeren et al., 1982) and has been attributed primarily to the higher voluminosity of denatured whey proteins, and aggregates thereof, than of native whey proteins (Snoeren et al., 1982). In addition, the association of denatured whey protein with casein micelles also affects the interactions between micelles (Jeurnink & De Kruif, 1993), leading to further increases in viscosity with increasing degree of denaturation. The lower viscosity observed at lower concentration pH (Fig. 1, Table 3) cannot be correlated to the effects of the aforementioned slightly higher degree of whey protein denaturation on total volume fraction. The slightly higher increase in the extent of whey protein denaturation after heating at pH 6.7 compared with that at pH 6.5 would be expected to increase protein voluminosity by ~0.5e1.0%, considering typical voluminosity of 1.5, 3.0 and 4.0 mL g1 for native whey protein, denatured whey protein and casein, respectively, and whey protein and casein representing 20% and 80% of total protein in skim milk, respectively. However, for samples heated at pH 6.7, solids content attained after evaporation was lower (Table 2) and differences in solids content would outweigh small increases in voluminosity. Hence, reduced viscosity for samples concentrated at pH 6.5 compared with pH 6.7 cannot be attributed to the extent of whey protein denaturation. However, the aforementioned effect of pH on protein hydration should certainly be taken into account. A reduction in pH by 0.2 units, which was retained largely after concentration (Table 2) would imply a reduction in net negative charge and a compression of the electric double layer, as seen by a considerable reduction in the zeta-potential with decreasing pH in this pH range (Anema & Klostermeyer, 1996) and reduced micellar hydration (Van Hooydonk, Boerrigter, & Hagedoorn, 1986). A compression of the electric double layer would reduce the hydrodynamic volume fraction and hence viscosity. The higher level of denatured whey protein associated with casein micelles in milk heated at pH 6.5 compared with milk heated at pH 6.7 is unlikely to have a major effect here, as the effect of pH at which heating was carried out was far smaller than that of pH at which concentration was carried out (Fig. 1, Table 3). Similar trends were discernable from the work of Anema et al. (2014) at lower concentrations. Hence, it appears that tailoring caseinewhey protein interactions has only limited effect on skim milk concentrate viscosity in the range of solids content relevant for industrial practice, i.e., >45% (Fig. 1, Table 3) (Anema et al., 2014). Control of protein voluminosity/hydration and interparticle interactions appears to yield far larger effects. 5. Conclusions Overall, it thus appears that, at solids content >50%, the two major factors determining skim milk concentrate viscosity are the pH and the extent of whey protein denaturation, whereas the distribution of whey proteins between the casein micelles and the serum phase had a far smaller effect. At the high solids contents studied here, particle volume fraction is suggested to be the dominant contributing parameter, with increases in particle volume fraction as a result of heat-induced whey protein denaturation
and reductions on particle volume fraction as a result of compression of the electrical double layer on reducing pH.
Acknowledgements The authors would like to thank Midwest Dairy Food Research Center (MDFRC), Dairy Research Institute (DRI), and Research and Agricultural Experiment Station (AES) for providing financial support.
References Anema, S. G. (1998). Effect of milk concentration on heat-induced, pH-dependent dissociation of casein from micelles in reconstituted skim milk at temperatures between 20 and 120 C. Journal of Agricultural and Food Chemistry, 46, 2299e2305. Anema, S. G., & Klostermeyer, H. (1996). z-Potentials of casein micelles from reconstituted skim milk heated at 120 C. International Dairy Journal, 6, 673e687. Anema, S. G., & Klostermeyer, H. (1997). Heat-induced, pH-dependent dissociation of casein micelles on heating reconstituted skim milk at temperatures below 100 C. Journal of Agricultural and Food Chemistry, 45, 1108e1115. Anema, S. G., Lowe, E. K., Lee, S. K., & Klostermeyer, H. (2014). Effect of the pH of skim milk at heating on milk concentrate viscosity. International Dairy Journal, 39, 336e343. Bloore, C. G., & Boag, I. F. (1981). Some factors affecting the viscosity of concentrated skim milk. New Zealand Journal of Dairy Science and Technology, 16, 143e154. Cassandra, R. A., & Dalgleish, D. G. (2006). Structures and some properties of soluble protein complexes formed by the heating of reconstituted skim milk powder. Food Research International, 39, 472e479. Eilers, H. (1941). Die Viskosit€ at von Emulsionen hochviskoser Stoffe als Funktion der Konzentration. Kolloid-Zeitschrift, 97, 313e321. Farr, R. S., & Groot, R. D. (2009). Close packing density of polydisperse hard spheres. Journal of Chemical Physics, 131. Article 244104. Fernandez-Martin, F. (1972). Influence of temperature and composition on some physical properties of milk and milk concentrates. II. Viscosity. Journal of Dairy Research, 39, 75e82. Fox, M., Akkerman, C., Straatsma, H., & De Jong, P. (2010). Energy reduction by high solids concentration and drying. New Food, 2, 60e63. Gazi, I., & Huppertz, T. (2015). Influence of protein content and storage conditions on the solubility of caseins and whey proteins in milk protein concentrates. International Dairy Journal, 46, 22e30. Jeurnink, T. J. M., & De Kruif, K. G. (1993). Changes in milk on heating: Viscosity measurements. Journal of Dairy Research, 60, 139e150. € , Y. (2005). Relationship between Karlsson, A. O., Ipsen, R., Schrader, K., & Ardo physical properties of casein micelles and rheology of skim milk concentrate. Journal of Dairy Science, 88, 3784e3797. Liu, D., Dunstan, D. E., & Martin, G. J. O. (2012). Evaporative concentration of skimmed milk: Effect on casein micelle hydration, composition, and size. Food Chemistry, 134, 1446e1452. Meletharayil, G. H., Patel, H. A., & Huppertz, T. (2015). Rheological properties and microstructure of high protein acid gels prepared from reconstituted milk protein concentrate powders of different protein contents. International Dairy Journal, 47, 64e71. Mendoza, C. I., & Santamaría-Holek, I. (2009). The rheology of hard sphere suspensions at arbitrary volume fractions: An improved differential viscosity model. Journal of Chemical Physics, 130. Article 044904. Nieuwenhuijse, J. A., Timmermans, W., & Walstra, P. (1988). Calcium and phosphate partitions during the manufacture of sterilized concentrated milk and their relation to the heat stability. Netherlands Milk Dairy Journal, 42, 387e421. Renan, M., Mekmene, O., Famelart, M.-H., Guyomarc'h, F., Arnoult-Delest, V., ^quet, D., et al. (2006). pH-dependent behaviour of soluble protein aggregates Pa formed during heat-treatment of milk at pH 6.5 or 7.2. Journal of Dairy Research, 73, 79e86. Snoeren, T. H. M., Brinkhuis, J. A., Damman, A. J., & Klok, H. J. (1984). Viscosity and age-thickening of skim milk concentrate. Netherlands Milk and Dairy Journal, 38, 43e53. Snoeren, T. H. M., Damman, A. J., & Klok, H. J. (1982). The viscosity of skim milk concentrates. Netherlands Milk and Dairy Journal, 36, 305e316. Van Hooydonk, A. C. M., Boerrigter, I. J., & Hagedoorn, H. G. (1986). pH induced physico-chemical changes of casein micelles in milk and their effect on renneting. 2. Effect of pH on renneting of milk. Netherlands Milk and Dairy Journal, 40, 297e313. Vasbinder, A. J., & de Kruif, C. G. (2003). Caseinewhey protein interactions in heated milk: The influence of pH. International Dairy Journal, 13, 669e677. lez-Ruiz, J. F., & Barbosa-C Ve anovas, G. V. (1998). Rheological properties of concentrated milk as a function of concentration, temperature and storage time. Journal of Food Engineering, 35, 177e190.
Contents lists available at ScienceDirect
International Dairy Journal journal homepage: www.elsevier.com/locate/idairyj
Influence of milk pre-heating conditions on caseinewhey protein interactions and skim milk concentrate viscosity Suresh G. Sutariya a, Thom Huppertz a, b, Hasmukh A. Patel a, c, * a
Dairy Science Department, South Dakota State University, Brookings, SD, USA NIZO Food Research, P.O. Box 20, 6710 BA, Ede, The Netherlands c Dairy Foods Research and Development, Land O'Lakes, Inc., Arden Hills, MN, USA b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 18 December 2016 Received in revised form 16 January 2017 Accepted 16 January 2017 Available online 4 February 2017
The viscosity of concentrates (50e55% total solids) prepared from skim milk heated (5 min at 80 or 90 C) at pH 6.5 and 6.7 was examined. The extent of heat-induced whey protein denaturation increased with increasing temperature and pH. More denatured whey protein and k-casein were found in the serum phase of milk heated at higher pH. The viscosity of milk concentrates increased considerably with increasing pH at concentration and increasing heating temperature, whereas the distribution of denatured whey proteins and k-casein between the serum and micellar phase only marginally influenced concentrate viscosity. Skim milk concentrate viscosity thus appears to be governed primarily by volume fraction and interactions of particles, which are governed primarily by concentration factor, the extent of whey protein denaturation and pH. Control and optimization of these factors can facilitate control over skim milk concentrate viscosity and energy efficiency in spray-drying. © 2017 Elsevier Ltd. All rights reserved.
1. Introduction Evaporation and spray drying are two major steps involved in the manufacture of skim milk powder. Since removal of water by evaporation requires ~10e20 times less energy than by spraydrying (Fox, Akkerman, Straatsma, & De Jong, 2010), maximizing the solids content of concentrate prior to spray-drying is desired economically and environmentally. Commonly, skim milk powder manufacturers concentrate skim milk to 45e50% total solids prior to drying. At higher solids content, viscosity increases dramatically with further increases in total solids. The viscosity of concentrated milk depends primarily on the solids content and composition of the milk, heating conditions, evaporation temperature and holding time of the concentrate (Bloore & Boag, 1981; Fernandez-Martin, 1972; Karlsson, Ipsen, € , 2005; Snoeren, Damman, & Klok, 1982; Ve lezSchrader, & Ardo novas, 1998). Snoeren et al. (1982) showed that Ruiz & Barbosa-Ca increases in viscosity as a result of heating of milk are related to an increase in milk protein voluminosity, because denaturation increases the voluminosity of the whey proteins; as a result,
* Corresponding author. Tel.: þ1 651 375 1497. E-mail address: [email protected] (H.A. Patel). http://dx.doi.org/10.1016/j.idairyj.2017.01.007 0958-6946/© 2017 Elsevier Ltd. All rights reserved.
concentrate viscosity was found to correlate strongly with the extent of whey protein denaturation (Snoeren et al., 1982). More recently, Anema, Lowe, Lee, and Klostermeyer (2014) suggested that not only the extent of denaturation, but also the distribution of denatured whey proteins, and k-casein, between the serum and micellar phase affects the viscosity of milk concentrates for solids content <40%, but that at higher total solids content, viscosity was less influenced by the distribution of denatured whey proteins and k-casein. Hence, for reducing skim milk concentrate viscosity in industrial operations, tailoring caseinewhey protein interactions does not appear to be a key route based on the results of Anema et al. (2014). However, the studies performed by Anema et al. (2014) only studied solids contents up to 45%, which are at the lower end of industrial processing. For investigating options to improve spray-drying efficiency by reducing milk concentrate viscosity, effects at solids contents >50% should be considered. Another factor that is crucial to consider in these studies is interparticle interactions. In unconcentrated milk, where the interparticle distance is several times larger than the particle diameter, such effects may appear to be limited. However, in concentrated milk the inter-particle distance becomes considerably smaller than particle diameter and particle interactions become much more important in governing rheological behavior (Karlsson et al., 2005), particularly in milk concentrates of >50% total solids.
20
S.G. Sutariya et al. / International Dairy Journal 69 (2017) 19e22
Hence, the objectives of the current work were to evaluate the role of whey protein distribution, as well as pH, on the rheological behavior of skim milk concentrates (>50% dry matter) and evaluate whether protein distribution or other factors govern the rheological behavior of the concentrates. 2. Materials and methods 2.1. Sample preparation Low-heat skim milk powder (Associated Milk Producers Inc., New Ulm, MN, USA; whey protein nitrogen index > 6; 36%, w/w, protein on a dry matter basis) was reconstituted in deionized water at 10% (w/v) for 4 h at room temperature. Sodium azide (0.02%) was added to prevent bacterial growth and the pH was subsequently adjusted to 6.5 and 6.7 by addition of 1 M HCl or 1 M NaOH with continuous stirring. The pH-adjusted skim milk samples were then stored overnight at 4 C. Samples were brought to room temperature the next day and pH was checked and readjusted if required. Samples were the heated in a hot water bath to 80 or 90 C, held at this temperature for 5 min and subsequently cooled to room temperature in an ice-water bath. For samples at pH 6.5 and 6.7 that were heated at 80 C, subsamples were subsequently adjusted to pH 6.7 and 6.5, respectively. For each sample variant described above, 3 replicate samples were prepared. 2.2. Determination of whey protein denaturation and whey protein distribution
180e190 mbar. Concentration was continued until a solids content of 50e55% (m m1) was reached. Solids content was determined by oven drying. Immediately after concentration, apparent viscosity was measured at 55 C, over a shear rate profile of 50e500 s1, using a STRESSTECH rheometer (ATS Rheosystems, Bordentown, NJ, USA), fitted with an CC 25 CCE SS cup and bob attachment. Shear rate was increased by 50 s1 at 10 s intervals. Since milk concentrates showed non-Newtonian viscosity behavior, viscosity profiles were evaluated using a Power Law model:
s ¼ K*gn
(1)
where, s is the shear stress (Pa), g is the shear rate (s1), K is the consistency coefficient (Pa sn) and n is the dimensionless flow behavior index. K and n were determined from a plot of log s versus log g, which has log K value as the intercept and the n as slope. For shear thinning liquids, 0 n 1, with lower values indicating a greater degree of shear thinning. 2.4. Statistical analysis The results were analyzed by SAS (version 9.3; SAS Institute Inc., Cary, NC, USA). Analysis of variance (ANOVA) was performed to examine the statistical difference between the samples, and the differences were considered significant when p-values were less than 0.05. All experiments were replicated on three individual milk samples. 3. Results
To determine the level of whey protein denaturation in heated milk samples, the pH 4.6-soluble fraction of milk samples was prepared as described by Gazi and Huppertz (2015). The pH 4.6soluble fraction and whole sample were subsequently analyzed by sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDSPAGE) as described by Meletharayil, Patel, and Huppertz (2015) and the stained gels were scanned using a Bio-5000 Microtek scanner (Microtek, Hsinchu, Taiwan) and the images were analyzed to quantify the intensity of whey protein bands using the Bio-image Intelligent Quantifier v 3.3.7 system (Bio Image Systems, Jackson, MI, USA). The level of whey protein denaturation was calculated from the difference in band intensities of the pH 4.6-soluble fraction unheated milk sample (residual native whey protein in the unheated samples) and the pH 4.6-soluble fraction of the heated sample (residual native whey protein in the heated sample). To determine the distribution of whey proteins in heated milk, samples were centrifuged at 25,000 g for 1 h at 20 C, followed by SDS-PAGE analysis, scanning and image quantification of the supernatants as described above. 2.3. Viscosity of concentrated milk samples Samples of 1.5 L of skim milk were concentrated at 60 C under vacuum using a rotary evaporator (Rotovapor Hei-VAP Value/G3, Heidolph, Schwabach, Germany) with an absolute pressure of
3.1. Whey protein denaturation and distribution in heated milk samples Heat treatment of samples at 80 C denatured ~43% and 50% of native whey protein present in samples at pH 6.5 and 6.7, respectively, whereas heat treatment of the same samples at 90 C denatured ~84% and 93% of total native whey protein present, respectively (Table 1). In addition, the distribution of whey proteins and k-casein between the serum phase and colloidal phase was also affected by the pH at which heat treatment was carried out. Heat treatment at pH 6.7 resulted in higher levels of whey proteins and k-casein in the serum phase of milk than heat treatment at pH 6.5 (Table 1); this, again, is in agreement with previous studies (Anema, 1998; Anema & Klostermeyer, 1997; Anema et al., 2014; Cassandra & Dalgleish, 2006; Renan et al., 2006; Vasbinder & de Kruif, 2003). 3.2. Viscosity of concentrated milk samples Samples were concentrated to a solids content of 50e55% dry matter by evaporation (Table 2) which was accompanied by a reduction in pH of the samples, by ~0.4e0.5 pH units, irrespective of starting pH. These decreases in pH are in agreement with previous studies and can be attributed to a concentration of acids by
Table 1 Extent of whey protein (WP) denaturation and individual proteins in the serum phase of milk heated at pH 6.5 or 6.7 at 80 or 90 C for 5 min.a pH at heat treatment
Temperature ( C)
6.5 6.7 6.5 6.7
80 80 90 90
Denatured WP (% total) 43 50 84 93
± ± ± ±
5a 2a 7b 2b
b-Lactoglobulin
a-Lactalbumin
(band intensity)
(band intensity)
0.80 1.06 0.32 0.79
± ± ± ±
0.01a 0.02b 0.00c 0.05a
0.47 0.74 0.24 0.33
± ± ± ±
0.01a 0.01b 0.01c 0.04d
k-Casein (band intensity) 0.37 0.62 0.28 0.55
± ± ± ±
0.05a 0.05b 0.04c 0.09d
a Values are means ± standard deviation of measurements on three individual milk samples; means within the same column not sharing a common superscript letter are significantly different (P < 0.05). Band intensity values are in arbitrary units.
S.G. Sutariya et al. / International Dairy Journal 69 (2017) 19e22
21
Table 2 pH and dry matter content of milk concentrate prepared from milk pre-heated at pH 6.5 or 6.7 at 80 or 90 C for 5 min followed by concentration to 50e55% dry matter at pH 6.5 or 6.7.a pH at heat treatment
pH at concentration
Heating temperature ( C)
pH after concentration
6.5 6.7 6.5 6.7 6.5 6.7
6.5 6.5 6.7 6.7 6.5 6.7
80 80 80 80 90 90
6.06 6.03 6.26 6.30
± ± ± ±
Dry matter after concentration (%, m m1)
0.03a 0.05a 0.02b 0.03b
54.2 54.9 52.6 52.5 52.5 50.7
± ± ± ± ± ±
0.0a 0.7a 0.1b 0.2b 0.1b 0.1c
a Values are means ± standard deviation of measurements on three individual milk samples; means within same column not sharing common superscript are significantly different (P < 0.05).
evaporation, as well as the reduced solubility of calcium phosphate, leading to insolubilization of calcium phosphate and a concomitant release of Hþ (Nieuwenhuijse, Timmermans, & Walstra, 1988). Viscosity of samples was determined as a function of shear rate profiles are shown in Fig. 1, showing notable differences in viscosity between different milk samples. Analysis of the viscosity profiles with the Power Law also highlighted large differences in the flow behavior index, n, and the consistency coefficient, K (Table 3). The differences in solids content of the concentrates (Table 2) were not the cause of differences in viscosity as actually a negative rather than positive correlation was found between viscosity and solids content (data not shown). Hence, effects of variables studied, i.e., pH at pre-heating, and concentration and temperature at preheating, had sufficiently strong effects on viscosity to ensure that variations in viscosity as a result of solids content were readily overruled (Fig. 1, Table 3). Skim milk concentrate viscosity was increased when the temperature of pre-heating was higher and when the pH of concentration was higher (Fig. 1). In addition, for samples pre-heated at the same temperature and concentrated at the same pH, viscosity was slightly higher when pre-heating was carried out at a higher pH, but such effects were considerably smaller than those for pre-heating temperature and pH at concentration. Along with increases in viscosity, increases in K and decreases in n were observed (Table 3) suggesting that, as expected, samples became more shear-thinning as viscosity increased.
4. Discussion When skim milk is concentrated, viscosity of the concentrate increases. In the very dilute region, where relative viscosity increases in linear fashion with increasing concentration of milk protein, these changes can be related to increases in volume fraction of suspended particles alone; however, in more concentrated systems relative viscosity increases more strongly with increasing protein concentration or volume fractions, due to hydrodynamic interactions between casein micelles. Data from Jeurnink and De Kruif (1993) suggest this is already the case in milk system and this effect increases further with increasing concentration factor. With increasing concentration factor of milk, particularly at solids contents >40%, milk concentrates start displaying shear-thinning lez-Ruiz & Barbosa-Ca novas, 1998) indicative of behavior (Ve (non-permanent) clustering of particles (Jeurnink & De Kruif, 1993). Increases in viscosity with concentration have been described by several empirical models, of which those by Eilers (1941), Mendoza and Santamaría-Holek (2009), Snoeren et al. (1982) and derivatives thereof are most commonly used. Although these models often show good fits to datasets, the physicochemical validity of the application of these models to milk may be questioned, because of their assumption of a constant voluminosity of the particles in the milk, i.e., the casein micelles and (denatured) whey proteins, which is unlikely to be the case. With increasing concentration factor of milk, ionic strength increases and pH decreases (Nieuwenhuijse et al., 1988), both of which are known to affect the hydration or voluminosity of casein micelles (Snoeren, Brinkhuis, Damman, & Klok, 1984). In addition, a decrease in pH will reduce the netnegative charge on the proteins, whereas an increase in ionic strength reduces the range over which the charges persist; hence, the decreases in pH and increase in ionic strength on concentration of milk will lead to compression of the electrical double layer and hydrodynamic radius (Karlsson et al., 2005), and thus volume fraction. Therefore, considering particle hydration and voluminosity to remain constant during concentration appears to be unjustified, Table 3 Power Law-derived flow behavior index (n) and consistency coefficient (K) of concentrates prepared by evaporation of milk pre-heated at pH 6.5 or 6.7 at 80 or 90 C for 5 min followed by concentration to 50e55% dry matter at pH 6.5 or 6.7.a
Fig. 1. Effect of shear rate on the viscosity of concentrates prepared by evaporation of milk pre-heated at pH 6.5 ( , , ) or pH 6.7 ( , , ) at 80 C ( , , , ) or 90 C ( , ) for 5 min followed by concentration to 50e55% dry matter at pH 6.5 ( , , ) or pH 6.7 ( , , ). Values are means of measurements on three individual milk samples, with the standard deviation indicated by vertical error bars.
pH at heat treatment
pH at concentration
Heating temperature
n ()
6.5 6.7 6.5 6.7 6.5 6.7
6.5 6.5 6.7 6.7 6.5 6.7
80 80 80 80 90 90
0.75 0.73 0.60 0.61 0.54 0.30
± ± ± ± ± ±
K (Pa sn) 0.13a 0.09a 0.07b 0.06b 0.06c 0.11d
1.8 ± 0.6a 2.4 ± 0.5a 12.8 ± 2.2b 15.9 ± 2.2b,c 18.4 ± 2.8c 155.7 ± 35.6d
a Values are means ± standard deviation of measurements on three individual milk samples, means within same column not sharing common superscript are significantly different (P < 0.05).
22
S.G. Sutariya et al. / International Dairy Journal 69 (2017) 19e22
as also shown experimentally by Liu, Dunstan, and Martin (2012) and also explains why apparently very high maximum volume fractions (fmax, i.e., 0.79) are described when using aforementioned models (Snoeren et al., 1982), whereas these appear unrealistically high when containing skim milk as a slightly polydisperse suspension of near-spherical particles (Farr & Groot, 2009). From viscosity data (Fig. 1, Table 3) it is clear that pre-heating intensity and concentration pH had the largest effect of skim milk concentrate viscosity, whereas the effect of the pH at which preheating was carried out was considerably smaller. The increased viscosity with increasing pre-heating intensity is in line with previous results (Snoeren et al., 1982) and has been attributed primarily to the higher voluminosity of denatured whey proteins, and aggregates thereof, than of native whey proteins (Snoeren et al., 1982). In addition, the association of denatured whey protein with casein micelles also affects the interactions between micelles (Jeurnink & De Kruif, 1993), leading to further increases in viscosity with increasing degree of denaturation. The lower viscosity observed at lower concentration pH (Fig. 1, Table 3) cannot be correlated to the effects of the aforementioned slightly higher degree of whey protein denaturation on total volume fraction. The slightly higher increase in the extent of whey protein denaturation after heating at pH 6.7 compared with that at pH 6.5 would be expected to increase protein voluminosity by ~0.5e1.0%, considering typical voluminosity of 1.5, 3.0 and 4.0 mL g1 for native whey protein, denatured whey protein and casein, respectively, and whey protein and casein representing 20% and 80% of total protein in skim milk, respectively. However, for samples heated at pH 6.7, solids content attained after evaporation was lower (Table 2) and differences in solids content would outweigh small increases in voluminosity. Hence, reduced viscosity for samples concentrated at pH 6.5 compared with pH 6.7 cannot be attributed to the extent of whey protein denaturation. However, the aforementioned effect of pH on protein hydration should certainly be taken into account. A reduction in pH by 0.2 units, which was retained largely after concentration (Table 2) would imply a reduction in net negative charge and a compression of the electric double layer, as seen by a considerable reduction in the zeta-potential with decreasing pH in this pH range (Anema & Klostermeyer, 1996) and reduced micellar hydration (Van Hooydonk, Boerrigter, & Hagedoorn, 1986). A compression of the electric double layer would reduce the hydrodynamic volume fraction and hence viscosity. The higher level of denatured whey protein associated with casein micelles in milk heated at pH 6.5 compared with milk heated at pH 6.7 is unlikely to have a major effect here, as the effect of pH at which heating was carried out was far smaller than that of pH at which concentration was carried out (Fig. 1, Table 3). Similar trends were discernable from the work of Anema et al. (2014) at lower concentrations. Hence, it appears that tailoring caseinewhey protein interactions has only limited effect on skim milk concentrate viscosity in the range of solids content relevant for industrial practice, i.e., >45% (Fig. 1, Table 3) (Anema et al., 2014). Control of protein voluminosity/hydration and interparticle interactions appears to yield far larger effects. 5. Conclusions Overall, it thus appears that, at solids content >50%, the two major factors determining skim milk concentrate viscosity are the pH and the extent of whey protein denaturation, whereas the distribution of whey proteins between the casein micelles and the serum phase had a far smaller effect. At the high solids contents studied here, particle volume fraction is suggested to be the dominant contributing parameter, with increases in particle volume fraction as a result of heat-induced whey protein denaturation
and reductions on particle volume fraction as a result of compression of the electrical double layer on reducing pH.
Acknowledgements The authors would like to thank Midwest Dairy Food Research Center (MDFRC), Dairy Research Institute (DRI), and Research and Agricultural Experiment Station (AES) for providing financial support.
References Anema, S. G. (1998). Effect of milk concentration on heat-induced, pH-dependent dissociation of casein from micelles in reconstituted skim milk at temperatures between 20 and 120 C. Journal of Agricultural and Food Chemistry, 46, 2299e2305. Anema, S. G., & Klostermeyer, H. (1996). z-Potentials of casein micelles from reconstituted skim milk heated at 120 C. International Dairy Journal, 6, 673e687. Anema, S. G., & Klostermeyer, H. (1997). Heat-induced, pH-dependent dissociation of casein micelles on heating reconstituted skim milk at temperatures below 100 C. Journal of Agricultural and Food Chemistry, 45, 1108e1115. Anema, S. G., Lowe, E. K., Lee, S. K., & Klostermeyer, H. (2014). Effect of the pH of skim milk at heating on milk concentrate viscosity. International Dairy Journal, 39, 336e343. Bloore, C. G., & Boag, I. F. (1981). Some factors affecting the viscosity of concentrated skim milk. New Zealand Journal of Dairy Science and Technology, 16, 143e154. Cassandra, R. A., & Dalgleish, D. G. (2006). Structures and some properties of soluble protein complexes formed by the heating of reconstituted skim milk powder. Food Research International, 39, 472e479. Eilers, H. (1941). Die Viskosit€ at von Emulsionen hochviskoser Stoffe als Funktion der Konzentration. Kolloid-Zeitschrift, 97, 313e321. Farr, R. S., & Groot, R. D. (2009). Close packing density of polydisperse hard spheres. Journal of Chemical Physics, 131. Article 244104. Fernandez-Martin, F. (1972). Influence of temperature and composition on some physical properties of milk and milk concentrates. II. Viscosity. Journal of Dairy Research, 39, 75e82. Fox, M., Akkerman, C., Straatsma, H., & De Jong, P. (2010). Energy reduction by high solids concentration and drying. New Food, 2, 60e63. Gazi, I., & Huppertz, T. (2015). Influence of protein content and storage conditions on the solubility of caseins and whey proteins in milk protein concentrates. International Dairy Journal, 46, 22e30. Jeurnink, T. J. M., & De Kruif, K. G. (1993). Changes in milk on heating: Viscosity measurements. Journal of Dairy Research, 60, 139e150. € , Y. (2005). Relationship between Karlsson, A. O., Ipsen, R., Schrader, K., & Ardo physical properties of casein micelles and rheology of skim milk concentrate. Journal of Dairy Science, 88, 3784e3797. Liu, D., Dunstan, D. E., & Martin, G. J. O. (2012). Evaporative concentration of skimmed milk: Effect on casein micelle hydration, composition, and size. Food Chemistry, 134, 1446e1452. Meletharayil, G. H., Patel, H. A., & Huppertz, T. (2015). Rheological properties and microstructure of high protein acid gels prepared from reconstituted milk protein concentrate powders of different protein contents. International Dairy Journal, 47, 64e71. Mendoza, C. I., & Santamaría-Holek, I. (2009). The rheology of hard sphere suspensions at arbitrary volume fractions: An improved differential viscosity model. Journal of Chemical Physics, 130. Article 044904. Nieuwenhuijse, J. A., Timmermans, W., & Walstra, P. (1988). Calcium and phosphate partitions during the manufacture of sterilized concentrated milk and their relation to the heat stability. Netherlands Milk Dairy Journal, 42, 387e421. Renan, M., Mekmene, O., Famelart, M.-H., Guyomarc'h, F., Arnoult-Delest, V., ^quet, D., et al. (2006). pH-dependent behaviour of soluble protein aggregates Pa formed during heat-treatment of milk at pH 6.5 or 7.2. Journal of Dairy Research, 73, 79e86. Snoeren, T. H. M., Brinkhuis, J. A., Damman, A. J., & Klok, H. J. (1984). Viscosity and age-thickening of skim milk concentrate. Netherlands Milk and Dairy Journal, 38, 43e53. Snoeren, T. H. M., Damman, A. J., & Klok, H. J. (1982). The viscosity of skim milk concentrates. Netherlands Milk and Dairy Journal, 36, 305e316. Van Hooydonk, A. C. M., Boerrigter, I. J., & Hagedoorn, H. G. (1986). pH induced physico-chemical changes of casein micelles in milk and their effect on renneting. 2. Effect of pH on renneting of milk. Netherlands Milk and Dairy Journal, 40, 297e313. Vasbinder, A. J., & de Kruif, C. G. (2003). Caseinewhey protein interactions in heated milk: The influence of pH. International Dairy Journal, 13, 669e677. lez-Ruiz, J. F., & Barbosa-C Ve anovas, G. V. (1998). Rheological properties of concentrated milk as a function of concentration, temperature and storage time. Journal of Food Engineering, 35, 177e190.