- Email: [email protected]
Effects of composition and relative humidity on the functional and storage properties of spray dried model milk emulsions
Journal of Food Engineering 169 (2016) 196e204
Contents lists available at ScienceDirect
Journal of Food Engineering journal homepage: www.elsevier.com/locate/jfoodeng
Effects of composition and relative humidity on the functional and storage properties of spray dried model milk emulsions Krystel Li, Meng Wai Woo, Cordelia Selomulya* Department of Chemical Engineering, Monash University, Clayton Campus, Victoria 3800, Australia
a r t i c l e i n f o
a b s t r a c t
Article history: Received 11 June 2015 Received in revised form 28 August 2015 Accepted 3 September 2015 Available online 5 September 2015
This study investigates the effect of relative humidity (RH) and milk composition on the changes in solubility, lactose crystallinity and protein structural form on a commercial infant formula, by identifying the interactions between water, lactose and protein at different storage conditions. Two spray dried model milk emulsions were studied for comparison, by storing these powders under varying relative humidity (RH) levels between 11 and 94% at room temperature. The properties of powders were evaluated from the degree of insolubility, the rate of browning, and the extent of protein denaturation. The model samples were found to be more stable than the commercial powder, even at high relative humidity (>50% RH). The difference was attributed to the presence of casein, which did not denature to the same extent as whey protein. The results also suggested that minerals present in the commercial powder could be responsible in enhancing protein denaturation, thus accelerating the rate of browning and decreasing solubility. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Infant formula Relative humidity Storage Protein denaturation Casein
1. Introduction Although human milk is very complex in nature, the dairy industry has attempted to replicate it using a complex combination of proteins, carbohydrates, fats, and vitamins to produce infant formula for bottle-feeding (Jacobsen, 2013). Previously, milk has been extensively dried using thin films on heated rollers, however this method was largely replaced by spray-drying in the 1960s (Bhandari et al., 2013), as a more efficient and economical method for dairy powder production. Commercialised dairy powders such as skimmed milk powder (SMP) or whole milk powder (WMP) are manufactured through processes such as evaporation, atomisation, spray-drying, and fluidised bed drying/cooling. However, the production of powdered infant formula differs from that of SMP or WMP due to the relatively high level of protein content in the products, preventing direct spray drying of the formulation. Generally, they are manufactured via two types of processes, i.e. dry mixing process or wet mixing e spray drying process. The former consists of mixing together dehydrated powdered ingredients to achieve a uniform blend of macro and micronutrients necessary for a complete infant formula product. Wet blending on the other hand
* Corresponding author. E-mail address: [email protected] (C. Selomulya). http://dx.doi.org/10.1016/j.jfoodeng.2015.09.002 0260-8774/© 2015 Elsevier Ltd. All rights reserved.
constitutes of blending different ingredients together with the addition of water, homogenizing, pasteurizing, and spray drying to produce the final powdered product. Much work has been done over the years to improve the properties of dairy powder. There are standard methods to quantify different properties such as foaming properties, flowability, instant properties, and heat stability. However, the amount of studies done on infant formula pales in comparison to those on skimmed milk or whole milk powder. Little is known on the parameters shortening the shelf-life of this complex dairy powder, besides the occurrence of lipid oxidation (Almansa et al., 2013; Angulo et al., 1998). The aim of this study is to quantify the influence of different relative humidity (11e91% RH) on the physical and functional properties of spray-dried milk model emulsions and commercial formula for comparison. These model emulsions have general compositions (fat, protein, lactose) similar to that of breast milk during the first six months of lactation. 2. Materials and methods 2.1. Preparation of model emulsions Two model emulsions; model emulsion 1 and model emulsion 2 (ME1 and ME2) were prepared using a-Lactose monohydrate (SigmaeAldrich, Australia), whey protein isolate (WPI) (Mullins
K. Li et al. / Journal of Food Engineering 169 (2016) 196e204
197
Whey, USA), sunflower oil (Crisco, Australia) and milk protein concentrate (MPC) (MG Nutritionals, Australia) based on the dry basis milk composition. Their basic compositions (fat, protein, lactose) were based on those of breast milk in the first fortnight and from three to six months of lactation, respectively. On dry basis, WPI (Mullins Whey, USA) contains 0.7 wt% lactose, 94.5 wt% protein, 1.8 wt% fat and 3.0 wt% contents while MPC (MG Nutritionals, Australia) contains 4.5 wt% lactose, 86.0 wt% protein, 1.6 wt% fat and 7.9 wt% ash. All ingredients used in this study were used as purchased. To make up an emulsion of 500 g and of 25 wt. % concentration, specific amounts of lactose, sunflower oil, MPC and WPI were used with 350 g of deionised water added to the dry mixture. The suspension was placed in a water bath on a magnetic stirrer/hot plate to allow it to be constantly stirred for 1 h at 50 C. The mixture was then passed through a high pressure homogenizer (Emulsiflex C5, Avestin, Canada) at two passes - 350 bars and 10 bars to yield a homogenous emulsion. 2.2. Microfluidic jet spray drying Monodisperse droplets were formed by a micro-fluidic aerosol nozzle system with an orifice diameter of 100 mm. The emulsion was fed into a standard steel reservoir with dehumidified instrument air to force the liquid to jet through the nozzle and the jet was broken up by disturbance from vibrating piezoceramics. The droplet formation was monitored using a digital SLR (Nikon D90) with a speed light (Nikon SB-400) and micro-lens (AF MicroNikkon 60 mm f/2.8D), while droplet spacing was optimized via adjustment of frequency and period of stimulation of piezoelectric nozzle under observation using high-speed photography, until monodisperse droplets were formed. The monodisperse droplets were well dispersed and dried in a micro-fluidic-jet spray dryer (MFJSD) at an inlet temperature of 180 C and an outlet temperature of 80 C. The average inlet air flow pressure was at 10 psi, while the inlet and outlet air relative humidity were on average between 52 and 58%. 2.3. Storage study of spray-dried milk powder The equilibrium moisture content was determined using the static gravimetric method, employing the use of saturated salt solutions to maintain a fixed relative humidity at room temperature. Desiccators were used to ensure the required airtight environment. Excess salt were added to obtain saturated salt solutions, hence creating the required equilibrium relative humidity. Ten different environment with relative humidity values in the range of 11.3e93.6% were created using salts (sodium hydroxide, potassium acetate, magnesium chloride, potassium carbonate, sodium bromide, potassium iodide, sodium chloride, ammonium sulphate, potassium chloride and potassium nitrate with RH of 11.3, 22.5, 32.8, 43.2, 57.6, 68.9, 75.3, 81.0, 84.3 and 93.6% RH, respectively). The spray dried powders and the commercialised infant formula were placed in a vacuum oven at 30 C for 3e4 days to remove as much moisture as possible. 2.0 g of the dried ME1, ME2 and commercialised powder (NAN H.A. 1 Gold Infant Formula, Nestle, Australia) were weighed in a plastic petri dish and then placed inside ten desiccators. The experiments were done in triplicate and were run until the constant weights (equilibrium) of the samples were reached (within 0.001 g) after 7 weeks of storage.
Fig. 1. Moisture sorption isotherm of model emulsion 1 (ME1) and 2 (ME2) and commercial infant formula at room temperature.
Emission Electron SEM. The microscope was operated between a HV of 2e5 kV and with a spot size of 2.0. The X-ray diffraction patterns of the dairy powders were measured using a Rigaku MiniFlex 600 XRD, in the range of 5e40 2q at a step size of 2q ¼ 0.02 and reading speed of 2q ¼ 2/min. Additionally, the insolubility index measurement was done according to GEA Niro method No.A3a (IDF, Standard 129), whereby 3 ± 0.01 g of spray dried powder was reconstituted in 150 mL of deionised water at 50 C. The solution was stirred for 30 min before being centrifuged at 4000 rpm for 10 min. Instead of only measuring the volume of supernatant left after centrifugation, the supernatant was removed and the insoluble matter was dried to constant weight at 50 C for a period of 24 h and insolubility index was recorded in milligram (mg) Attenuated total reflectance (ATR) mid-infrared spectra were acquired using a Fourier transform infra-red spectroscopy (FTIR, PerkinElmer, Australia), over a wavenumber range of 600e4000 cm1, with all measurement performed in duplicate. As for the measurement of denaturation of protein, a differential scanning calorimeter (DSC) was employed, whereby approximately 10 mg of the sample solution of concentration 10 g protein/100 ml
2.4. Characterisation of samples Scanning Electron Microscope (SEM) images of the spray-dried samples were examined using a FEI Nova NanoSEM 450 Field
Fig. 2. XRD Pattern of fresh commercialised infant formula (with arrows indicating characteristic peaks of a-lactose monohydrate), ME1 and ME2.
198
K. Li et al. / Journal of Food Engineering 169 (2016) 196e204
Fig. 3. XRD patterns of ME1 and ME2 at different relative humidity when stored for (a) (c) one month and (b) (d) two months.
of solution was hermetically sealed in a DSC stainless steel pan and analysed using a PerkinElmer DSC. An empty pan was used as reference. 3. Results and discussion Moisture sorption isotherm Fig. 1 depicts the moisture sorption isotherm of model emulsions 1 (ME1) and 2 (ME2), with the powders being stored at room temperature. The trend of the graph for ME1 and ME2 is typical for an amorphous powder, with sharp inflection points in the isotherm indicating a phase transition
which is the crystallisation of lactose. The amount of moisture adsorbed by both the model emulsions at low water activities (<22%) were similar, increasing from 0 to 0.72 g/100 g of dry powder, due to the highly hygroscopic amorphous protein present in both powders. The critical aw of ME1 appears to be 0.32, whereas lactose crystallisation occurs at 42% RH for ME2, indicated by the inflection points on the graph. However, sorption isotherms alone are not sufficient to assert the amorphous or crystalline state of lactose. X-ray diffraction was also performed as discussed in the subsequent section. These results were different from those reported by Thomas et al. (2004) who recorded a lower critical aw for
K. Li et al. / Journal of Food Engineering 169 (2016) 196e204
199
Fig. 4. XRD patterns of (a) ME1 and (b) ME2 at different RH after six months of storage.
Pearce and Kinsella, 1978) have demonstrated the tendency of whey protein to denature upon heat treatment. Heat causes the protein's structure to be altered, exposing its hydrophobic groups and causing sulfhydryl/disulphide exchange chain reactions to take place between the exposed cysteine residues. Casein on the contrary tend to form relatively thick open interfacial layers that are resistant to heating (Guzey and Mcclements, 2006). Hence, the lower casein content in ME1, although it has a higher overall protein count, could be responsible for the earlier onset of lactose crystallisation. Fig. 1 also illustrates the isotherm of a commercial infant formula. The graph depicts a gradual increase in moisture content, which suggests moisture absorption solely due to protein as according to XRD analysis (discussed in the Section 3.1.1), all lactose were in their crystalline form in fresh CIF and hence no lactose crystallisation occurred during the storage study. 3.1. Characterisation of samples Fig. 5. SEM image of ME1 stored at 57.6% RH for two months.
powders having higher lactose to protein ratios. For lactose to protein (b-lactoglobulin) ratio of 60:40, the critical water activity was of 0.55 aw as opposed to 0.39 aw for a ratio of 90:10. A possible cause for the higher critical water activity for ME2 could be the higher casein micelles content. It is believed that proteins (casein micelles and whey) delayed lactose crystallisation due to molecular interactions between lactose and proteins. Carpenter and Crowe (1989) have demonstrated that sugars interact with polar groups of globular proteins via hydrogen bonding. Therefore the competition between proteins/lactose interaction (hydrogen bonds) and lactose/lactose interaction (nucleation) might reduce the tendency of lactose to crystallise. Although ME1 contained 17.5 wt % more protein than ME2, the ratio of whey to casein differs from 90:10 to 70:30. Several studies (Lee et al., 1992; Patel and Kilara, 1990;
3.1.1. X-ray diffraction analysis (XRD) X-ray diffraction (XRD) analysis was performed on the freshly spray-dried model emulsions, on the fresh commercialised powder, as well as powders stored at different relative humidity for a period of 1 and two months. Fig. 2 depicts an amorphous lactose state in the fresh ME1 and ME2, judging from the broad peak with a single maximum. In the commercial infant formula, crystalline lactose was present in the fresh sample, as illustrated by the peak intensity of 2 theta values at 12.39e12.53 , 16.2e16.38 and 19 . The presence of minerals is also confirmed from XRD analysis, with peaks in the vicinity of 30e38 indicating calcium or iron (Sun et al., 2006; naite-Felsen _ Trinku et al., 2012). From Fig. 3, it can be observed that the amorphous nature of lactose in the model emulsion was maintained after a month of storage at different RH levels, even those as high as 84.3%. However, after two months, the powders stored at RH greater than 43.2% indicate some degrees of crystallinity, as characterised by the peaks
200
K. Li et al. / Journal of Food Engineering 169 (2016) 196e204
Fig. 6. SEM images of (a) fresh CIF and (b) CIF after two months storage demonstrating changes in surface appearance.
in the XRD patterns (Fig. 3). The same phenomenon is observed for ME1 and ME2 when stored for six months (Fig. 4). Low RH conditions such as 22.5% do not trigger any crystallisation and can be confirmed by the pattern obtained from the MSI, whereby for aw<0.33, the adsorption of water is similar to standard milk powder moisture adsorption rate, displaying a high degree of amorphous state. The XRD patterns show the presence of crystallised lactose in the range of 32.8e43.2% RH for both powders (Fig. 4). These results confirm those obtained from the moisture sorption isotherms on the occurrence of lactose crystallisation in ME1 and ME2 (Fig. 5) 3.1.2. Scanning electron microscopy (SEM) Scanning electron microscopy (SEM) was used to analyse the morphology of dairy powders. Indeed, lactose crystals were only visible after two months of storage when kept at high humidity conditions. No sign of crystallised lactose was detected for powders stored at low humidity (22.5% RH). SEM images also reveal a change in the appearance of the particles, from a relatively smooth surface in the fresh samples to a rough exterior when stored at high RH (Fig. 6). High agglomeration is also noted for high aw (aw>0.3). The presence of lactose crystals in ME1 at RH 57% after two months of storage, and in the fresh sample of the commercial infant formula was also confirmed from the SEM images, as indicated by the arrows.
Fig. 7. Insolubility Index of ME1 and ME2 at RH of 22.50%, 57.57% and 84.30% stored over 8 weeks.
3.1.3. Insolubility index The insolubility was measured on a weekly basis for powders stored at 22.5, 57.6 and 84.3% RH. Fig. 7 depicts the variation in the insolubility index of the three powders (ME1, ME2 and Commercial infant formula (CIF)) stored at the aforementioned RH, over a period of 8 weeks. From Fig. 7, it appears that the insolubility index of the two spray dried powders (ME1 and ME2) did not vary much over the course of the storage study, irrespective of relative humidity. The initial insolubility index of these powders was
Table 1 Compositions of model emulsions 1 and 2 and of the commerical infant formula. Powder
Average amount per 100 ml of formula Casein
Model emulsion 1 Model emulsion 2 Commercial Infant formula a
Whey
Lactose
Fat
Iron
Copper
(g)
%a
(g)
%a
(g)
%a
(g)
%a
(mg)
%a
(mg)
%a
0.15 0.38 e
1.32 2.85 e
1.39 0.89 1.30
12.2 6.68 10.4
6.34 7.62 7.80
55.7 57.2 62.4
3.50 4.44 3.40
30.8 33.3 27.2
e e 0.70
e e 0.01
e e 58
e e 4.6 104
% composition is calculated based on weight dry basis.
K. Li et al. / Journal of Food Engineering 169 (2016) 196e204
approximately 0.11 ± 0.002 g and varied marginally over the 8 weeks of storage. The data indicates a negligible influence of humidity and time on the solubility of these powders. In contrast, the insolubility index of the commercial powder reveals that the solubility is significantly reduced over the storage period and at different humidity conditions. The initial insolubility index of the fresh commercialised powder is 0.06 ± 0.007 g, which qualifies it as an ‘instant powder’ (Pisecky, 1997). Over 8 weeks of storage, the insolubility index almost doubles (0.11 ± 0.005 g) on average when stored at 22.5% RH. A similar trend could be observed
201
for powders stored at 57.6% and 84.3%, with the average insolubility index increasing to 0.14 g and 0.15 g, respectively (with a relative error of 0.005 g). This was attributed to the rapid denaturation of whey protein in CIF, which consequently lead to the increase in insolubility index. Denaturation of whey protein was demonstrated from the FTIR results and DSC results discussed in Section 3.1.5 and 3.1.6. 3.1.4. Non-enzymatic browning Table 2 demonstrates the change in colour of the three different
Table 2 Appearance of CIF, ME1 and ME2 at different relative humidity, stored for a period of six months. Relative humidity, %
Powder Commercial infant formula
11.3
22.5
32.8
43.2
57.6
68.9
75.3
81.0
84.3
93.6
Model emulsion 1
Model emulsion 2
202
K. Li et al. / Journal of Food Engineering 169 (2016) 196e204
milk powders, stored at relative humidity (Hogan and O'Callaghan, 2013) from 11.3 to 93.6%, over a period of six months. As seen from the table, browning occurs at different water activities for different powders. For the commercial infant formula, a change in colour is noticed at aw higher than 0.43. ME1 changes to a dark brown colour at a much elevated RH of 81.0% whereas ME2 maintained its original white colour at RH as high as 84.3%. Though not quantitative, the visual inspection revealed that CIF was more prone to nonenzymatic browning at a lower moisture content. The change from a white/pale yellow to a brown colour is due to the Maillard reaction, and occurs during storage. During Maillard reaction, lactose and lysine-rich protein in milk react with each other through the condensation of amino acid (lysine) with a reducing sugar. The pictures in Table 2 show that CIF, which contained crystallised lactose, has undergone non-enzymatic browning faster and at lower water activities as compared to the in-house spraydried powders. The model emulsion powders did not suffer from extensive browning, even at high RH and for a long storage time (six months). The results obtained in this study were in contrast to those of Mcsweeney and Fox (2009) who reported that the rate of browning was significantly lowered in whey powders containing pre-crystallised lactose. A recent study by Hedegaard and Skibsted (2013) demonstrated that lactose crystallisation in dairy powders resulted in increasing rates of non-enzymatic browning, due to release of water which is initiated by crystallisation of amorphous lactose in the milk powder, as the reactants become more mobile for Maillard reaction. Although the commercial powder was precrystallised (as shown by the XRD results) and the fact that ME1 and ME2 were observed to undergo lactose crystallisation between
the first and second months of storage as confirmed by the moisture sorption isotherm and the XRD results (Section 3.1.1), this did not demonstrated the same outcome as literature. As a matter of fact, browning was more pronounced in CIF, occurring at lower RH, than in the model emulsion milk powders as illustrated in Table 2. In order to identify such browning discrepancies between the three powders, the compositions of the formulations were analysed. While the model emulsions contain a mixture of whey and casein as protein source, CIF contains solely whey as protein source in addition to added minerals as shown in Table 1. The enhanced browning could therefore be attributed to presence of uniquely whey protein in CIF. Whey protein has a high tendency to denature upon intensive heat treatment as compared to casein, and the DSC results in Section 3.1.6 reveals a complete denaturation of whey protein upon storage for CIF. By unfolding, these whey proteins increase the accessibility of lysine (amino acid), which in turn is more prone to react with lactose in the Maillard reaction. This also explains why ME2 could better withstand the storage conditions, with browning occurring at higher RH than ME1. The former contained 0.38 g of casein per 100 ml of milk as opposed to 0.15 g in ME1 per 100 ml of formula. Moreover, the presence of minerals (salts) in CIF should also not be excluded as being a catalyst in promoting non-enzymatic browning in the dairy powder. Indeed, Potman and Van Wijk (1989) reported an increase of 15 fold in the reaction rate of browning in the presence of phosphate, while Kato et al. (1981) associated iron and copper ions to increased browning rate as these cations cause an increase in pH that enhances the extent of browning in food (Baxter, 1995). These finding are in agreement with the results obtained from
Fig. 8. (A) Mid-infrared spectra (1720e1600 cm1; protein bands) of CIF (a), ME1 (b) and ME2 (c) stored at 22.5, 57.6 and 84.3% RH for a period of six months. (B) Mid-infrared spectra (1200e800 cm1; lactose bands) of CIF (a), ME1 (b) and ME2 (c) stored at 22.5, 57.6 and 84.3% RH for a period of six months.
K. Li et al. / Journal of Food Engineering 169 (2016) 196e204
the insolubility test. The Maillard reaction destroys amino acids (mainly lysine) and/or forms cross-links between protein chains, which eventually leads to an overall reduction in protein solubility and digestibility (Hedegaard and Skibsted, 2013). This concurs with the significant loss of solubility (insolubility index doubles from 0.06 g to 0.10 g irrespective of RH) after only 1 week of storage for CIF. From these results, a distinct direct relationship can be drawn between browning and insolubility of dairy powders. 3.1.5. Fourier transform infrared spectroscopy (FTIR) Fig. 8(A) (a) shows the FTIR spectra of the fresh commercial powder, as well as powders that were stored at RH of 22, 57 and 84%, for a total storage time of six months. The spectra region between 1600 and 1700 cm1 (amide I region) provides useful information in the study of protein secondary structure, and can therefore allow the identification of denatured protein. The FTIR demonstrates that the protein in the fresh commercial powder was in its native state, with the band being slightly asymmetric and having a peak maximum around 1650 cm1, representative of alpha-helical structure. In contrast, powders that were stored at the different RH for the prolonged period of time show sign of protein denaturation, with peaks in the vicinity of 1630, 1640 and 1658 cm1, indicative of the predominance of beta-sheet and unordered structures. Fig. 8(A) (b) and (c) demonstrate the FTIR spectra of fresh ME1 and ME2, together with those being stored under the same conditions as CIF as mentioned above. A different trend was observed with these powders. The proteins for the spraydried powders were still in their native forms without any denaturation detected. This is illustrated by the broad singled peak spectra, with a maximum in the vicinity of 1650 cm1. Multiple studies had proved that presence of sugar, in this case lactose, can curtail protein denaturation. While Carpenter et al. (1992), Carpenter and Crowe (1989) and Wolkers et al. (1998) suggested that sugar molecules substitute water through a direct interaction with the protein molecules during the removal of water during drying, Franks et al. (1991) asserted that the glass-forming properties of the sugar could be responsible for keeping denaturation to a minimum. However, those findings do not explain why the MEs powders were able to withstand protein denaturation over storage time and at high RH, since both powders (MEs and CIF) had approximately the same lactose content. A hypothesis to explain such occurrence of denatured protein in CIF would be due to the tendency for ions (especially those with high atomic weights) to break proteineprotein bonds, thus unfolding the protein molecules. The moisture sorption isotherm and the XRD pattern show that the lactose in fresh CIF was in the crystalline form, and thus the water that was absorbed at the different aw was mainly due to the proteins. Excess sorption of moisture could induce different minerals (which exist in salt form) present in CIF to convert into their aqueous forms. This results in mobile ions that disrupt disulphide bridges in proteins, resulting in the unfolding of protein molecules and subsequently denaturation. On the other hand, the absence of minerals in MEs could explain why the proteins in those powders were still structurally intact and in their functionally native forms. The spectral region between 800 and 1200 cm1 provides information on overlapping bands, mostly resulting from CeO and CeC stretching vibrations, which corresponds to carbohydrate (lactose) (Ottenhof et al., 2003). From Fig. 8(B) (a), it is seen that the FTIR spectrum changes drastically with RH and storage time when compared with that of the fresh commercial powder. This suggests the occurrence of degradation of the lactose present and further modification of its chemical structure. As for ME1 and ME2 Fig. 8(B) (b) and (c), the crystallinity of lactose is still not detected at RH of 22.5%, even after six months of storage, but powders stored at medium (57.6% RH) and high (84.3% RH) relative humidity
203
underwent lactose crystallisation, as shown by the pronounced increase in spectral resolution due to the three-dimensional repeated structured of the crystalline system (Ottenhof et al., 2003). This result suggests that lactose crystallisation is highly dependent on relative humidity, but to a lesser extent on storage time. 3.1.6. Differential scanning calorimetry (DSC) The Differential Scanning Calorimeter (DSC) is used to confirm protein denaturation in the commercial infant formula. As the temperature of the sample is gradually increased in the DSC, structural rearrangements caused by the redistribution of noncovalent bonds will arise due to heat adsorption. The enthalpy of protein unfolding is the area under the concentrated-normalised DSC peak (Malvern, 2015). Fig. 9 depicts the DSC results of fresh commercial infant formula, in addition to commercial powders being stored at 57 and 87% RH for a period of six months. A small and narrow peak can be seen between the temperature range of 55 and 65 C, denoting the unfolding of protein to its denatured form. This demonstrates that the fresh commercial infant formula contained protein in its native (folded) conformations. However, for powders stored at high relative humidity for a period of six months, no peak could be detected from the DSC results. This means that all the protein found in those powders were completely unfolded, i.e. denatured. This then explained rapid decrease in the solubility of the commercial infant formula after just a week of storage as discussed in Section 3.1.3.
Fig. 9. DSC results for fresh CIF and CIF (a) stored at 57 and 84% RH for a period of six months (b and c).
204
K. Li et al. / Journal of Food Engineering 169 (2016) 196e204
4. Conclusion This study investigated the effects of relative humidity and milk composition on moisture sorption, lactose crystallisation, morphology and solubility of model milk emulsions and of a commercial infant formula. The findings suggested that higher casein content retarded the occurrence of lactose crystallisation and lowered the rate of browning in milk powder. While the solubility of spray dried model emulsions did not decrease significantly over two months of storage at different relative humidity, the commercial powder showed a decrease by two-fold after only one week of storage. This was attributed to the rapid denaturation of whey protein in this powder, leading to the rise in insolubility. The absence of casein together with the presence of certain minerals such as iron or copper, also explains the fast and premature browning of the commercial infant formula. This study also demonstrated that the model milk powders were more stable (from the insolubility index and browning analysis) than the commercial powder, even when subjected to high storage humidity. The outcome provides a better understanding on how relative humidity and composition affects the functional properties of milk powders, and could help improve the future processing of infant formula. Acknowledgement This project is part of the dairy research activities at Monash University, supported by the Australian Research Council (ARC) through the Linkage program (LP140100922). References ~ o, E., Silvestre, D., 2013. Lipid peroxidation in infant Almansa, I., Miranda, M., Jaren formulas: longitudinal study at different storage temperatures. Int. Dairy J. 33, 83e87. Angulo, A.J., Romera, J.M., Ramirez, M., Gil, A., 1998. Effects of storage conditions on lipid oxidation in infant formulas based on several protein sources. JAOCS. J. Am. Oil Chemists' Soc. 75. Baxter, J.H., 1995. Free amino acid stability in reducing sugar systems. J. Food Sci. 60, 405e408. Bhandari, B., Bansal, N., Zhang, M., 2013. Handbook of Food Powders Processes and Properties. Woodhead Publishing Ltd, Cambridge. Carpenter, J.F., Arakawa, T., Crowe, J.H., 1992. Interactions of stabilizing additives with proteins during freeze-thawing and freeze-drying. Dev. Biol. Stand. 74, 225e238 discussion 238. Carpenter, J.F., Crowe, J.H., 1989. An infrared spectroscopic study of the interactions of carbohydrates with dried proteins. Biochemistry 28, 3916e3922. Franks, F., Hatley, R.H.M., Mathias, S.F., 1991. Material science and the production of
shelf-stable biologicals. BioPharmaceuticals 38e42. Guzey, D., McClements, D.J., 2006. Formation, stability and properties of multilayer emulsions for application in the food industry. Adv. Colloid Interface Sci. 128e130, 227e248. Hedegaard, R.V., Skibsted, L.H., 2013. 16-Shelf-life of food powders. In: Bhandari, B., Bansal, N., Zhang, M., Schuck, P. (Eds.), Handbook of Food Powders. Woodhead Publishing. Hogan, S.A., O'Callaghan, D.J., 2013. Moisture sorption and stickiness behaviour of hydrolysed whey protein/lactose powders. Dairy Sci. Technol. 93, 505e521. Jacobsen, L.N., 2013. In: Ingredients, A.F. (Ed.), Learning from Nature for a Healthier Start (Denmark). Kato, Y., Watanabe, K., Sato, Y., 1981. Effect of some metals on the Maillard reaction of ovalbumin. J. Agric. Food Chem. 29, 540e543. Lee, S.-Y., Morr, C.V., Ha, E.Y.W., 1992. Structural and functional properties of caseinate and whey protein isolate as affected by temperature and pH. J. Food Sci. 57, 1210e1229. Malvern, 2015. Differential Scanning Calorimetry. Confident Characterization of Protein Stability [Online]. Available. http://www.malvern.com/en/products/ technology/differential-scanning-calorimetry/. McSweeney, P.L.H., Fox, P.F., 2009. Advanced Dairy Chemistry. Ottenhof, M.-A., Macnaughtan, W., Farhat, I.A., 2003. FTIR study of state and phase transitions of low moisture sucrose and lactose. Carbohydr. Res. 338, 2195e2202. Patel, M.T., Kilara, A., 1990. Studies on whey protein concentrates. 2. Foaming and emulsifying properties and their relationships with physicochemical Properties1. J. Dairy Sci. 73, 2731e2740. Pearce, K.N., Kinsella, J.E., 1978. Emulsifying properties of proteins: evaluation of a turbidimetric technique. J. Agric. Food Chem. 26, 716e723. Pisecky, J., 1997. Handbook of Milk Powder Manufacture. Niro A/S, Denmark. Potman, T.H., van Wijk, T.A., 1989. Thermal Generation of Aromas. American Chemical Society, pp. 182e195. Sun, Y.-P., Li, X.-Q., Cao, J., Zhang, W.-X., Wang, H.P., 2006. Characterization of zerovalent iron nanoparticles. Adv. Colloid Interface Sci. 120, 47e56. Thomas, M.E.C., Scher, J., Desobry, S., 2004. Lactose/b-lactoglobulin interaction during storage of model whey powders. J. Dairy Sci. 87, 1158e1166. naite-Felsen, _ Trinku J., Zalga, A., Kareiva, A., 2012. Characterization of Naturally Derived Calcium Compounds Used in Food Industry. CHEMJA. Wolkers, W.F., Van Kilsdonk, M.G., Hoekstra, F.A., 1998. Dehydration-induced conformational changes of poly-l-lysine as influenced by drying rate and carbohydrates. Biochimica Biophysica Acta (BBA) e General Subj. 1425, 127e136.
Glossary ATR: Attenuated total reflectance CIF: Commercial infant formula DSC: Differential scanning calorimeter FTIR: Fourier transform infra-red spectroscopy ME1: Model emulsion 1 ME2: Model emulsion 2 MFJSD: Micro-fluidic-jet spray dryer MPC: Milk protein concentrate RH: Relative humidity SEM: Scanning Electron Microscope WPI: Whey protein isolate XRD: X-ray diffraction
Contents lists available at ScienceDirect
Journal of Food Engineering journal homepage: www.elsevier.com/locate/jfoodeng
Effects of composition and relative humidity on the functional and storage properties of spray dried model milk emulsions Krystel Li, Meng Wai Woo, Cordelia Selomulya* Department of Chemical Engineering, Monash University, Clayton Campus, Victoria 3800, Australia
a r t i c l e i n f o
a b s t r a c t
Article history: Received 11 June 2015 Received in revised form 28 August 2015 Accepted 3 September 2015 Available online 5 September 2015
This study investigates the effect of relative humidity (RH) and milk composition on the changes in solubility, lactose crystallinity and protein structural form on a commercial infant formula, by identifying the interactions between water, lactose and protein at different storage conditions. Two spray dried model milk emulsions were studied for comparison, by storing these powders under varying relative humidity (RH) levels between 11 and 94% at room temperature. The properties of powders were evaluated from the degree of insolubility, the rate of browning, and the extent of protein denaturation. The model samples were found to be more stable than the commercial powder, even at high relative humidity (>50% RH). The difference was attributed to the presence of casein, which did not denature to the same extent as whey protein. The results also suggested that minerals present in the commercial powder could be responsible in enhancing protein denaturation, thus accelerating the rate of browning and decreasing solubility. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Infant formula Relative humidity Storage Protein denaturation Casein
1. Introduction Although human milk is very complex in nature, the dairy industry has attempted to replicate it using a complex combination of proteins, carbohydrates, fats, and vitamins to produce infant formula for bottle-feeding (Jacobsen, 2013). Previously, milk has been extensively dried using thin films on heated rollers, however this method was largely replaced by spray-drying in the 1960s (Bhandari et al., 2013), as a more efficient and economical method for dairy powder production. Commercialised dairy powders such as skimmed milk powder (SMP) or whole milk powder (WMP) are manufactured through processes such as evaporation, atomisation, spray-drying, and fluidised bed drying/cooling. However, the production of powdered infant formula differs from that of SMP or WMP due to the relatively high level of protein content in the products, preventing direct spray drying of the formulation. Generally, they are manufactured via two types of processes, i.e. dry mixing process or wet mixing e spray drying process. The former consists of mixing together dehydrated powdered ingredients to achieve a uniform blend of macro and micronutrients necessary for a complete infant formula product. Wet blending on the other hand
* Corresponding author. E-mail address: [email protected] (C. Selomulya). http://dx.doi.org/10.1016/j.jfoodeng.2015.09.002 0260-8774/© 2015 Elsevier Ltd. All rights reserved.
constitutes of blending different ingredients together with the addition of water, homogenizing, pasteurizing, and spray drying to produce the final powdered product. Much work has been done over the years to improve the properties of dairy powder. There are standard methods to quantify different properties such as foaming properties, flowability, instant properties, and heat stability. However, the amount of studies done on infant formula pales in comparison to those on skimmed milk or whole milk powder. Little is known on the parameters shortening the shelf-life of this complex dairy powder, besides the occurrence of lipid oxidation (Almansa et al., 2013; Angulo et al., 1998). The aim of this study is to quantify the influence of different relative humidity (11e91% RH) on the physical and functional properties of spray-dried milk model emulsions and commercial formula for comparison. These model emulsions have general compositions (fat, protein, lactose) similar to that of breast milk during the first six months of lactation. 2. Materials and methods 2.1. Preparation of model emulsions Two model emulsions; model emulsion 1 and model emulsion 2 (ME1 and ME2) were prepared using a-Lactose monohydrate (SigmaeAldrich, Australia), whey protein isolate (WPI) (Mullins
K. Li et al. / Journal of Food Engineering 169 (2016) 196e204
197
Whey, USA), sunflower oil (Crisco, Australia) and milk protein concentrate (MPC) (MG Nutritionals, Australia) based on the dry basis milk composition. Their basic compositions (fat, protein, lactose) were based on those of breast milk in the first fortnight and from three to six months of lactation, respectively. On dry basis, WPI (Mullins Whey, USA) contains 0.7 wt% lactose, 94.5 wt% protein, 1.8 wt% fat and 3.0 wt% contents while MPC (MG Nutritionals, Australia) contains 4.5 wt% lactose, 86.0 wt% protein, 1.6 wt% fat and 7.9 wt% ash. All ingredients used in this study were used as purchased. To make up an emulsion of 500 g and of 25 wt. % concentration, specific amounts of lactose, sunflower oil, MPC and WPI were used with 350 g of deionised water added to the dry mixture. The suspension was placed in a water bath on a magnetic stirrer/hot plate to allow it to be constantly stirred for 1 h at 50 C. The mixture was then passed through a high pressure homogenizer (Emulsiflex C5, Avestin, Canada) at two passes - 350 bars and 10 bars to yield a homogenous emulsion. 2.2. Microfluidic jet spray drying Monodisperse droplets were formed by a micro-fluidic aerosol nozzle system with an orifice diameter of 100 mm. The emulsion was fed into a standard steel reservoir with dehumidified instrument air to force the liquid to jet through the nozzle and the jet was broken up by disturbance from vibrating piezoceramics. The droplet formation was monitored using a digital SLR (Nikon D90) with a speed light (Nikon SB-400) and micro-lens (AF MicroNikkon 60 mm f/2.8D), while droplet spacing was optimized via adjustment of frequency and period of stimulation of piezoelectric nozzle under observation using high-speed photography, until monodisperse droplets were formed. The monodisperse droplets were well dispersed and dried in a micro-fluidic-jet spray dryer (MFJSD) at an inlet temperature of 180 C and an outlet temperature of 80 C. The average inlet air flow pressure was at 10 psi, while the inlet and outlet air relative humidity were on average between 52 and 58%. 2.3. Storage study of spray-dried milk powder The equilibrium moisture content was determined using the static gravimetric method, employing the use of saturated salt solutions to maintain a fixed relative humidity at room temperature. Desiccators were used to ensure the required airtight environment. Excess salt were added to obtain saturated salt solutions, hence creating the required equilibrium relative humidity. Ten different environment with relative humidity values in the range of 11.3e93.6% were created using salts (sodium hydroxide, potassium acetate, magnesium chloride, potassium carbonate, sodium bromide, potassium iodide, sodium chloride, ammonium sulphate, potassium chloride and potassium nitrate with RH of 11.3, 22.5, 32.8, 43.2, 57.6, 68.9, 75.3, 81.0, 84.3 and 93.6% RH, respectively). The spray dried powders and the commercialised infant formula were placed in a vacuum oven at 30 C for 3e4 days to remove as much moisture as possible. 2.0 g of the dried ME1, ME2 and commercialised powder (NAN H.A. 1 Gold Infant Formula, Nestle, Australia) were weighed in a plastic petri dish and then placed inside ten desiccators. The experiments were done in triplicate and were run until the constant weights (equilibrium) of the samples were reached (within 0.001 g) after 7 weeks of storage.
Fig. 1. Moisture sorption isotherm of model emulsion 1 (ME1) and 2 (ME2) and commercial infant formula at room temperature.
Emission Electron SEM. The microscope was operated between a HV of 2e5 kV and with a spot size of 2.0. The X-ray diffraction patterns of the dairy powders were measured using a Rigaku MiniFlex 600 XRD, in the range of 5e40 2q at a step size of 2q ¼ 0.02 and reading speed of 2q ¼ 2/min. Additionally, the insolubility index measurement was done according to GEA Niro method No.A3a (IDF, Standard 129), whereby 3 ± 0.01 g of spray dried powder was reconstituted in 150 mL of deionised water at 50 C. The solution was stirred for 30 min before being centrifuged at 4000 rpm for 10 min. Instead of only measuring the volume of supernatant left after centrifugation, the supernatant was removed and the insoluble matter was dried to constant weight at 50 C for a period of 24 h and insolubility index was recorded in milligram (mg) Attenuated total reflectance (ATR) mid-infrared spectra were acquired using a Fourier transform infra-red spectroscopy (FTIR, PerkinElmer, Australia), over a wavenumber range of 600e4000 cm1, with all measurement performed in duplicate. As for the measurement of denaturation of protein, a differential scanning calorimeter (DSC) was employed, whereby approximately 10 mg of the sample solution of concentration 10 g protein/100 ml
2.4. Characterisation of samples Scanning Electron Microscope (SEM) images of the spray-dried samples were examined using a FEI Nova NanoSEM 450 Field
Fig. 2. XRD Pattern of fresh commercialised infant formula (with arrows indicating characteristic peaks of a-lactose monohydrate), ME1 and ME2.
198
K. Li et al. / Journal of Food Engineering 169 (2016) 196e204
Fig. 3. XRD patterns of ME1 and ME2 at different relative humidity when stored for (a) (c) one month and (b) (d) two months.
of solution was hermetically sealed in a DSC stainless steel pan and analysed using a PerkinElmer DSC. An empty pan was used as reference. 3. Results and discussion Moisture sorption isotherm Fig. 1 depicts the moisture sorption isotherm of model emulsions 1 (ME1) and 2 (ME2), with the powders being stored at room temperature. The trend of the graph for ME1 and ME2 is typical for an amorphous powder, with sharp inflection points in the isotherm indicating a phase transition
which is the crystallisation of lactose. The amount of moisture adsorbed by both the model emulsions at low water activities (<22%) were similar, increasing from 0 to 0.72 g/100 g of dry powder, due to the highly hygroscopic amorphous protein present in both powders. The critical aw of ME1 appears to be 0.32, whereas lactose crystallisation occurs at 42% RH for ME2, indicated by the inflection points on the graph. However, sorption isotherms alone are not sufficient to assert the amorphous or crystalline state of lactose. X-ray diffraction was also performed as discussed in the subsequent section. These results were different from those reported by Thomas et al. (2004) who recorded a lower critical aw for
K. Li et al. / Journal of Food Engineering 169 (2016) 196e204
199
Fig. 4. XRD patterns of (a) ME1 and (b) ME2 at different RH after six months of storage.
Pearce and Kinsella, 1978) have demonstrated the tendency of whey protein to denature upon heat treatment. Heat causes the protein's structure to be altered, exposing its hydrophobic groups and causing sulfhydryl/disulphide exchange chain reactions to take place between the exposed cysteine residues. Casein on the contrary tend to form relatively thick open interfacial layers that are resistant to heating (Guzey and Mcclements, 2006). Hence, the lower casein content in ME1, although it has a higher overall protein count, could be responsible for the earlier onset of lactose crystallisation. Fig. 1 also illustrates the isotherm of a commercial infant formula. The graph depicts a gradual increase in moisture content, which suggests moisture absorption solely due to protein as according to XRD analysis (discussed in the Section 3.1.1), all lactose were in their crystalline form in fresh CIF and hence no lactose crystallisation occurred during the storage study. 3.1. Characterisation of samples Fig. 5. SEM image of ME1 stored at 57.6% RH for two months.
powders having higher lactose to protein ratios. For lactose to protein (b-lactoglobulin) ratio of 60:40, the critical water activity was of 0.55 aw as opposed to 0.39 aw for a ratio of 90:10. A possible cause for the higher critical water activity for ME2 could be the higher casein micelles content. It is believed that proteins (casein micelles and whey) delayed lactose crystallisation due to molecular interactions between lactose and proteins. Carpenter and Crowe (1989) have demonstrated that sugars interact with polar groups of globular proteins via hydrogen bonding. Therefore the competition between proteins/lactose interaction (hydrogen bonds) and lactose/lactose interaction (nucleation) might reduce the tendency of lactose to crystallise. Although ME1 contained 17.5 wt % more protein than ME2, the ratio of whey to casein differs from 90:10 to 70:30. Several studies (Lee et al., 1992; Patel and Kilara, 1990;
3.1.1. X-ray diffraction analysis (XRD) X-ray diffraction (XRD) analysis was performed on the freshly spray-dried model emulsions, on the fresh commercialised powder, as well as powders stored at different relative humidity for a period of 1 and two months. Fig. 2 depicts an amorphous lactose state in the fresh ME1 and ME2, judging from the broad peak with a single maximum. In the commercial infant formula, crystalline lactose was present in the fresh sample, as illustrated by the peak intensity of 2 theta values at 12.39e12.53 , 16.2e16.38 and 19 . The presence of minerals is also confirmed from XRD analysis, with peaks in the vicinity of 30e38 indicating calcium or iron (Sun et al., 2006; naite-Felsen _ Trinku et al., 2012). From Fig. 3, it can be observed that the amorphous nature of lactose in the model emulsion was maintained after a month of storage at different RH levels, even those as high as 84.3%. However, after two months, the powders stored at RH greater than 43.2% indicate some degrees of crystallinity, as characterised by the peaks
200
K. Li et al. / Journal of Food Engineering 169 (2016) 196e204
Fig. 6. SEM images of (a) fresh CIF and (b) CIF after two months storage demonstrating changes in surface appearance.
in the XRD patterns (Fig. 3). The same phenomenon is observed for ME1 and ME2 when stored for six months (Fig. 4). Low RH conditions such as 22.5% do not trigger any crystallisation and can be confirmed by the pattern obtained from the MSI, whereby for aw<0.33, the adsorption of water is similar to standard milk powder moisture adsorption rate, displaying a high degree of amorphous state. The XRD patterns show the presence of crystallised lactose in the range of 32.8e43.2% RH for both powders (Fig. 4). These results confirm those obtained from the moisture sorption isotherms on the occurrence of lactose crystallisation in ME1 and ME2 (Fig. 5) 3.1.2. Scanning electron microscopy (SEM) Scanning electron microscopy (SEM) was used to analyse the morphology of dairy powders. Indeed, lactose crystals were only visible after two months of storage when kept at high humidity conditions. No sign of crystallised lactose was detected for powders stored at low humidity (22.5% RH). SEM images also reveal a change in the appearance of the particles, from a relatively smooth surface in the fresh samples to a rough exterior when stored at high RH (Fig. 6). High agglomeration is also noted for high aw (aw>0.3). The presence of lactose crystals in ME1 at RH 57% after two months of storage, and in the fresh sample of the commercial infant formula was also confirmed from the SEM images, as indicated by the arrows.
Fig. 7. Insolubility Index of ME1 and ME2 at RH of 22.50%, 57.57% and 84.30% stored over 8 weeks.
3.1.3. Insolubility index The insolubility was measured on a weekly basis for powders stored at 22.5, 57.6 and 84.3% RH. Fig. 7 depicts the variation in the insolubility index of the three powders (ME1, ME2 and Commercial infant formula (CIF)) stored at the aforementioned RH, over a period of 8 weeks. From Fig. 7, it appears that the insolubility index of the two spray dried powders (ME1 and ME2) did not vary much over the course of the storage study, irrespective of relative humidity. The initial insolubility index of these powders was
Table 1 Compositions of model emulsions 1 and 2 and of the commerical infant formula. Powder
Average amount per 100 ml of formula Casein
Model emulsion 1 Model emulsion 2 Commercial Infant formula a
Whey
Lactose
Fat
Iron
Copper
(g)
%a
(g)
%a
(g)
%a
(g)
%a
(mg)
%a
(mg)
%a
0.15 0.38 e
1.32 2.85 e
1.39 0.89 1.30
12.2 6.68 10.4
6.34 7.62 7.80
55.7 57.2 62.4
3.50 4.44 3.40
30.8 33.3 27.2
e e 0.70
e e 0.01
e e 58
e e 4.6 104
% composition is calculated based on weight dry basis.
K. Li et al. / Journal of Food Engineering 169 (2016) 196e204
approximately 0.11 ± 0.002 g and varied marginally over the 8 weeks of storage. The data indicates a negligible influence of humidity and time on the solubility of these powders. In contrast, the insolubility index of the commercial powder reveals that the solubility is significantly reduced over the storage period and at different humidity conditions. The initial insolubility index of the fresh commercialised powder is 0.06 ± 0.007 g, which qualifies it as an ‘instant powder’ (Pisecky, 1997). Over 8 weeks of storage, the insolubility index almost doubles (0.11 ± 0.005 g) on average when stored at 22.5% RH. A similar trend could be observed
201
for powders stored at 57.6% and 84.3%, with the average insolubility index increasing to 0.14 g and 0.15 g, respectively (with a relative error of 0.005 g). This was attributed to the rapid denaturation of whey protein in CIF, which consequently lead to the increase in insolubility index. Denaturation of whey protein was demonstrated from the FTIR results and DSC results discussed in Section 3.1.5 and 3.1.6. 3.1.4. Non-enzymatic browning Table 2 demonstrates the change in colour of the three different
Table 2 Appearance of CIF, ME1 and ME2 at different relative humidity, stored for a period of six months. Relative humidity, %
Powder Commercial infant formula
11.3
22.5
32.8
43.2
57.6
68.9
75.3
81.0
84.3
93.6
Model emulsion 1
Model emulsion 2
202
K. Li et al. / Journal of Food Engineering 169 (2016) 196e204
milk powders, stored at relative humidity (Hogan and O'Callaghan, 2013) from 11.3 to 93.6%, over a period of six months. As seen from the table, browning occurs at different water activities for different powders. For the commercial infant formula, a change in colour is noticed at aw higher than 0.43. ME1 changes to a dark brown colour at a much elevated RH of 81.0% whereas ME2 maintained its original white colour at RH as high as 84.3%. Though not quantitative, the visual inspection revealed that CIF was more prone to nonenzymatic browning at a lower moisture content. The change from a white/pale yellow to a brown colour is due to the Maillard reaction, and occurs during storage. During Maillard reaction, lactose and lysine-rich protein in milk react with each other through the condensation of amino acid (lysine) with a reducing sugar. The pictures in Table 2 show that CIF, which contained crystallised lactose, has undergone non-enzymatic browning faster and at lower water activities as compared to the in-house spraydried powders. The model emulsion powders did not suffer from extensive browning, even at high RH and for a long storage time (six months). The results obtained in this study were in contrast to those of Mcsweeney and Fox (2009) who reported that the rate of browning was significantly lowered in whey powders containing pre-crystallised lactose. A recent study by Hedegaard and Skibsted (2013) demonstrated that lactose crystallisation in dairy powders resulted in increasing rates of non-enzymatic browning, due to release of water which is initiated by crystallisation of amorphous lactose in the milk powder, as the reactants become more mobile for Maillard reaction. Although the commercial powder was precrystallised (as shown by the XRD results) and the fact that ME1 and ME2 were observed to undergo lactose crystallisation between
the first and second months of storage as confirmed by the moisture sorption isotherm and the XRD results (Section 3.1.1), this did not demonstrated the same outcome as literature. As a matter of fact, browning was more pronounced in CIF, occurring at lower RH, than in the model emulsion milk powders as illustrated in Table 2. In order to identify such browning discrepancies between the three powders, the compositions of the formulations were analysed. While the model emulsions contain a mixture of whey and casein as protein source, CIF contains solely whey as protein source in addition to added minerals as shown in Table 1. The enhanced browning could therefore be attributed to presence of uniquely whey protein in CIF. Whey protein has a high tendency to denature upon intensive heat treatment as compared to casein, and the DSC results in Section 3.1.6 reveals a complete denaturation of whey protein upon storage for CIF. By unfolding, these whey proteins increase the accessibility of lysine (amino acid), which in turn is more prone to react with lactose in the Maillard reaction. This also explains why ME2 could better withstand the storage conditions, with browning occurring at higher RH than ME1. The former contained 0.38 g of casein per 100 ml of milk as opposed to 0.15 g in ME1 per 100 ml of formula. Moreover, the presence of minerals (salts) in CIF should also not be excluded as being a catalyst in promoting non-enzymatic browning in the dairy powder. Indeed, Potman and Van Wijk (1989) reported an increase of 15 fold in the reaction rate of browning in the presence of phosphate, while Kato et al. (1981) associated iron and copper ions to increased browning rate as these cations cause an increase in pH that enhances the extent of browning in food (Baxter, 1995). These finding are in agreement with the results obtained from
Fig. 8. (A) Mid-infrared spectra (1720e1600 cm1; protein bands) of CIF (a), ME1 (b) and ME2 (c) stored at 22.5, 57.6 and 84.3% RH for a period of six months. (B) Mid-infrared spectra (1200e800 cm1; lactose bands) of CIF (a), ME1 (b) and ME2 (c) stored at 22.5, 57.6 and 84.3% RH for a period of six months.
K. Li et al. / Journal of Food Engineering 169 (2016) 196e204
the insolubility test. The Maillard reaction destroys amino acids (mainly lysine) and/or forms cross-links between protein chains, which eventually leads to an overall reduction in protein solubility and digestibility (Hedegaard and Skibsted, 2013). This concurs with the significant loss of solubility (insolubility index doubles from 0.06 g to 0.10 g irrespective of RH) after only 1 week of storage for CIF. From these results, a distinct direct relationship can be drawn between browning and insolubility of dairy powders. 3.1.5. Fourier transform infrared spectroscopy (FTIR) Fig. 8(A) (a) shows the FTIR spectra of the fresh commercial powder, as well as powders that were stored at RH of 22, 57 and 84%, for a total storage time of six months. The spectra region between 1600 and 1700 cm1 (amide I region) provides useful information in the study of protein secondary structure, and can therefore allow the identification of denatured protein. The FTIR demonstrates that the protein in the fresh commercial powder was in its native state, with the band being slightly asymmetric and having a peak maximum around 1650 cm1, representative of alpha-helical structure. In contrast, powders that were stored at the different RH for the prolonged period of time show sign of protein denaturation, with peaks in the vicinity of 1630, 1640 and 1658 cm1, indicative of the predominance of beta-sheet and unordered structures. Fig. 8(A) (b) and (c) demonstrate the FTIR spectra of fresh ME1 and ME2, together with those being stored under the same conditions as CIF as mentioned above. A different trend was observed with these powders. The proteins for the spraydried powders were still in their native forms without any denaturation detected. This is illustrated by the broad singled peak spectra, with a maximum in the vicinity of 1650 cm1. Multiple studies had proved that presence of sugar, in this case lactose, can curtail protein denaturation. While Carpenter et al. (1992), Carpenter and Crowe (1989) and Wolkers et al. (1998) suggested that sugar molecules substitute water through a direct interaction with the protein molecules during the removal of water during drying, Franks et al. (1991) asserted that the glass-forming properties of the sugar could be responsible for keeping denaturation to a minimum. However, those findings do not explain why the MEs powders were able to withstand protein denaturation over storage time and at high RH, since both powders (MEs and CIF) had approximately the same lactose content. A hypothesis to explain such occurrence of denatured protein in CIF would be due to the tendency for ions (especially those with high atomic weights) to break proteineprotein bonds, thus unfolding the protein molecules. The moisture sorption isotherm and the XRD pattern show that the lactose in fresh CIF was in the crystalline form, and thus the water that was absorbed at the different aw was mainly due to the proteins. Excess sorption of moisture could induce different minerals (which exist in salt form) present in CIF to convert into their aqueous forms. This results in mobile ions that disrupt disulphide bridges in proteins, resulting in the unfolding of protein molecules and subsequently denaturation. On the other hand, the absence of minerals in MEs could explain why the proteins in those powders were still structurally intact and in their functionally native forms. The spectral region between 800 and 1200 cm1 provides information on overlapping bands, mostly resulting from CeO and CeC stretching vibrations, which corresponds to carbohydrate (lactose) (Ottenhof et al., 2003). From Fig. 8(B) (a), it is seen that the FTIR spectrum changes drastically with RH and storage time when compared with that of the fresh commercial powder. This suggests the occurrence of degradation of the lactose present and further modification of its chemical structure. As for ME1 and ME2 Fig. 8(B) (b) and (c), the crystallinity of lactose is still not detected at RH of 22.5%, even after six months of storage, but powders stored at medium (57.6% RH) and high (84.3% RH) relative humidity
203
underwent lactose crystallisation, as shown by the pronounced increase in spectral resolution due to the three-dimensional repeated structured of the crystalline system (Ottenhof et al., 2003). This result suggests that lactose crystallisation is highly dependent on relative humidity, but to a lesser extent on storage time. 3.1.6. Differential scanning calorimetry (DSC) The Differential Scanning Calorimeter (DSC) is used to confirm protein denaturation in the commercial infant formula. As the temperature of the sample is gradually increased in the DSC, structural rearrangements caused by the redistribution of noncovalent bonds will arise due to heat adsorption. The enthalpy of protein unfolding is the area under the concentrated-normalised DSC peak (Malvern, 2015). Fig. 9 depicts the DSC results of fresh commercial infant formula, in addition to commercial powders being stored at 57 and 87% RH for a period of six months. A small and narrow peak can be seen between the temperature range of 55 and 65 C, denoting the unfolding of protein to its denatured form. This demonstrates that the fresh commercial infant formula contained protein in its native (folded) conformations. However, for powders stored at high relative humidity for a period of six months, no peak could be detected from the DSC results. This means that all the protein found in those powders were completely unfolded, i.e. denatured. This then explained rapid decrease in the solubility of the commercial infant formula after just a week of storage as discussed in Section 3.1.3.
Fig. 9. DSC results for fresh CIF and CIF (a) stored at 57 and 84% RH for a period of six months (b and c).
204
K. Li et al. / Journal of Food Engineering 169 (2016) 196e204
4. Conclusion This study investigated the effects of relative humidity and milk composition on moisture sorption, lactose crystallisation, morphology and solubility of model milk emulsions and of a commercial infant formula. The findings suggested that higher casein content retarded the occurrence of lactose crystallisation and lowered the rate of browning in milk powder. While the solubility of spray dried model emulsions did not decrease significantly over two months of storage at different relative humidity, the commercial powder showed a decrease by two-fold after only one week of storage. This was attributed to the rapid denaturation of whey protein in this powder, leading to the rise in insolubility. The absence of casein together with the presence of certain minerals such as iron or copper, also explains the fast and premature browning of the commercial infant formula. This study also demonstrated that the model milk powders were more stable (from the insolubility index and browning analysis) than the commercial powder, even when subjected to high storage humidity. The outcome provides a better understanding on how relative humidity and composition affects the functional properties of milk powders, and could help improve the future processing of infant formula. Acknowledgement This project is part of the dairy research activities at Monash University, supported by the Australian Research Council (ARC) through the Linkage program (LP140100922). References ~ o, E., Silvestre, D., 2013. Lipid peroxidation in infant Almansa, I., Miranda, M., Jaren formulas: longitudinal study at different storage temperatures. Int. Dairy J. 33, 83e87. Angulo, A.J., Romera, J.M., Ramirez, M., Gil, A., 1998. Effects of storage conditions on lipid oxidation in infant formulas based on several protein sources. JAOCS. J. Am. Oil Chemists' Soc. 75. Baxter, J.H., 1995. Free amino acid stability in reducing sugar systems. J. Food Sci. 60, 405e408. Bhandari, B., Bansal, N., Zhang, M., 2013. Handbook of Food Powders Processes and Properties. Woodhead Publishing Ltd, Cambridge. Carpenter, J.F., Arakawa, T., Crowe, J.H., 1992. Interactions of stabilizing additives with proteins during freeze-thawing and freeze-drying. Dev. Biol. Stand. 74, 225e238 discussion 238. Carpenter, J.F., Crowe, J.H., 1989. An infrared spectroscopic study of the interactions of carbohydrates with dried proteins. Biochemistry 28, 3916e3922. Franks, F., Hatley, R.H.M., Mathias, S.F., 1991. Material science and the production of
shelf-stable biologicals. BioPharmaceuticals 38e42. Guzey, D., McClements, D.J., 2006. Formation, stability and properties of multilayer emulsions for application in the food industry. Adv. Colloid Interface Sci. 128e130, 227e248. Hedegaard, R.V., Skibsted, L.H., 2013. 16-Shelf-life of food powders. In: Bhandari, B., Bansal, N., Zhang, M., Schuck, P. (Eds.), Handbook of Food Powders. Woodhead Publishing. Hogan, S.A., O'Callaghan, D.J., 2013. Moisture sorption and stickiness behaviour of hydrolysed whey protein/lactose powders. Dairy Sci. Technol. 93, 505e521. Jacobsen, L.N., 2013. In: Ingredients, A.F. (Ed.), Learning from Nature for a Healthier Start (Denmark). Kato, Y., Watanabe, K., Sato, Y., 1981. Effect of some metals on the Maillard reaction of ovalbumin. J. Agric. Food Chem. 29, 540e543. Lee, S.-Y., Morr, C.V., Ha, E.Y.W., 1992. Structural and functional properties of caseinate and whey protein isolate as affected by temperature and pH. J. Food Sci. 57, 1210e1229. Malvern, 2015. Differential Scanning Calorimetry. Confident Characterization of Protein Stability [Online]. Available. http://www.malvern.com/en/products/ technology/differential-scanning-calorimetry/. McSweeney, P.L.H., Fox, P.F., 2009. Advanced Dairy Chemistry. Ottenhof, M.-A., Macnaughtan, W., Farhat, I.A., 2003. FTIR study of state and phase transitions of low moisture sucrose and lactose. Carbohydr. Res. 338, 2195e2202. Patel, M.T., Kilara, A., 1990. Studies on whey protein concentrates. 2. Foaming and emulsifying properties and their relationships with physicochemical Properties1. J. Dairy Sci. 73, 2731e2740. Pearce, K.N., Kinsella, J.E., 1978. Emulsifying properties of proteins: evaluation of a turbidimetric technique. J. Agric. Food Chem. 26, 716e723. Pisecky, J., 1997. Handbook of Milk Powder Manufacture. Niro A/S, Denmark. Potman, T.H., van Wijk, T.A., 1989. Thermal Generation of Aromas. American Chemical Society, pp. 182e195. Sun, Y.-P., Li, X.-Q., Cao, J., Zhang, W.-X., Wang, H.P., 2006. Characterization of zerovalent iron nanoparticles. Adv. Colloid Interface Sci. 120, 47e56. Thomas, M.E.C., Scher, J., Desobry, S., 2004. Lactose/b-lactoglobulin interaction during storage of model whey powders. J. Dairy Sci. 87, 1158e1166. naite-Felsen, _ Trinku J., Zalga, A., Kareiva, A., 2012. Characterization of Naturally Derived Calcium Compounds Used in Food Industry. CHEMJA. Wolkers, W.F., Van Kilsdonk, M.G., Hoekstra, F.A., 1998. Dehydration-induced conformational changes of poly-l-lysine as influenced by drying rate and carbohydrates. Biochimica Biophysica Acta (BBA) e General Subj. 1425, 127e136.
Glossary ATR: Attenuated total reflectance CIF: Commercial infant formula DSC: Differential scanning calorimeter FTIR: Fourier transform infra-red spectroscopy ME1: Model emulsion 1 ME2: Model emulsion 2 MFJSD: Micro-fluidic-jet spray dryer MPC: Milk protein concentrate RH: Relative humidity SEM: Scanning Electron Microscope WPI: Whey protein isolate XRD: X-ray diffraction