Composition and crystallization of milk fat fractions

Milk fat was fractionated by solvent (acetone) fractionation and dry fractionation. Based on their fatty acid and acyl-carbon profiles, the fractions could be divided into three main groups: high-melting triglycerides (HMT), middle-melting triglycerides (MMT), and low-melting triglycerides (LMT). HMT fractions were enriched in long-chain fatty acids, and reduced in short-chain fatty acids and unsaturated fatty acids. The MMT fractions were enriched in long-chain fatty acids, and reduced in unsaturated fatty acids. The LMT fractions were reduced in long-chain fatty acids, and enriched in short-chain fatty acids and unsaturated fatty acids. Crystallization of these fractions was studied by differential scanning calorimetry and X-ray diffraction techniques. In this study, the stable crystal form appeared to be the β′-form for all fractions. At sufficiently low temperature (different for each fraction), the β′-form is preceded by crystallization in the metastable α-form. An important difference between the fractions is the rate of crystallization in the β′-form, which proceeds at a much lower rate for the lower-melting fat fractions than for the higher-melting fat fractions. This may be due to the much lower affinity for crystallization of the lower-melting fractions, due to the less favorable molecular geometry for packing in the β′-crystal lattice.     

Temperature and concentration dependent effect of partial glycerides on milk fat crystallization

Two diglycerides (distearin and diolein) and two monoglycerides (monostearin and monoolein) were added to milk fat in a concentration of 0.5% and 1%. The isothermal crystallization behavior was evaluated at 22 °C, 23.5 °C, 25 °C and 26.5 °C by DSC and pNMR. The crystallization kinetic was quantified by means of two models. It was noticed that the effect of the minor components on the crystallization behavior depends on temperature and concentration. The type of esterified fatty acids and the polar head of the amphiphilic molecule determine to what extent partial glycerides influence the nucleation and crystal growth of triglycerides. Moreover the degree of insolubility of partial glycerides in the melt determines which effect (on growth or on nucleation) predominates. Stearic acid based partial glycerides enhance nucleation at low temperatures, while at higher temperatures an interaction with the crystal growth predominates. Oleic acid based partial glycerides have an effect on the nucleation process while no interaction with the crystal growth was observed.     

Effect of processing conditions on crystallization kinetics of a milk fat model system

Crystallization behavior of three blends of 30, 40, and 50% of high-melting fraction (MDP=47.5°C) in low-melting fraction (MDP=16.5°C) of milk fat was studied under dynamic conditions in laboratory scale. The effect of cooling and agitation rates, crystallization temperature, and chemical composition of the blends on the morphology, crystal size distribution, crystal thermal behavior, polymorphism, and crystalline chemical composition was investigated by light microscopy, differential scanning calorimetry (DSC), X-ray diffraction (XRD) and gas chromatography (GC). Different nucleation and growth behavior were found for different cooling rates. At slow cooling rate, larger crystals were formed, whereas at fast cooling rate, smaller crystals appeared together. Slowly crystallized samples had a broader distribution of crystal size. Crystallization temperatures had a similar effect as cooling rate. At higher crystallization temperatures, larger crystals and a broader crystal size distribution were found. Agitation rate had a marked effect on crystal size. Higher agitation rates lead to smaller crystal size. Cooling rate was the most influential parameter in crystal thermal behavior and composition. Slowly crystallized samples showed a broader melting diagram and an enrichment of long-chain triacylglycerols. Crystallization behavior was more related to processing conditions than to chemical composition of blends.     

TAG composition of ewe's milk fat. Detection of foreign fats

The TAG composition of 45 samples of ewe's milk, collected throughout the year from five Spanish breeds, was analyzed according to their carbon number by using short capillary column GC. The TAG content had a bimodal distribution with maxima at C₃₈ (12.8%) and C₅₂ (8.4%). The TAG composition did not vary significantly with respect to the time of year of sampling but was affected by the breed. Multiple regression equations based on TAG content are proposed to detect foreign fats in ewe's milk fat. Analysis of known mixtures of lard, palm oil, and cow's milk fat with ewe's milk fat have experimentally confirmed the accuracy of the equations.     

Polymorphism of milk fat studied by differential scanning calorimetry and real-time X-ray powder diffraction

The crystallization behavior of milk fat was investigated by varying the cooling rate and by isothermal solidification at various temperatures while monitoring the formation of crystals by differential scanning calorimetry (DSC) and X-ray powder diffraction (XRD). Three different polymorphic crystal forms were observed in milk fat: γ, α, and β′. The β-form, occasionally observed in previous studies, was not found. The kind of polymorph formed during crystallization of milk fat from its melted state was dependent on the cooling rate and the final temperature. Moreover, transitions between the different polymorphic forms were shown to occur upon storing or heating the milk fat. The characteristic DSC heating curve of milk fat is interpreted on the basis of the XRD measurements, and appears to be a combined effect of selective crystallization of triglycerides and polymorphism.     

The effect of minor components on milk fat crystallization

Milk fat is composed of 97–98% triacylglycerols and 2–3% minor polar lipids. In this study triacylglycerols were chromatographically separated from minor components. Isolated diacylglycerols from the polar fraction were also added back to the milk fat triacylglycerols. The crystallization behaviors of native anhydrous milk fat (AMF), milk fat triacylglycerols (MF-TAG), and milk fat triacylglycerols with diacylglycerols added back (MF-DAG) were studied. Removal of minor components and addition of diacylglycerols had no effect on dropping points or equilibrium solid fat contents. Presence of the minor components, however, did delay the onset of crystallization at low degrees of supercooling. Crystallization kinetics were quantified using the Avrami model. Sharp changes in the values of the Avrami constant k and exponent n were observed for all three fats around 20.0°C. Increases in n around 20.0°C indicated a change from one-dimensional to multidimensional growth. Differences in k and n of MF-DAG from AMF and MF-TAG suggested that the presence of milk fat diacylglycerols changes the crystal growth mechanism. Apparent free energies of nucleation (ΔG_c,apparent) were determined using the Fisher-Turnbull model. (ΔG_c,apparent) for AMF was significantly greater than ΔG_c,apparent for MF-TAG, and ΔG_c,apparent for MF-DAG was significantly less than those for both AMF and MF-TAG. The microstructural networks of AMF, MF-TAG, and MF-DAG, however, were similar at both 5.0 and 25.0°C, and all three fats crystallized into the typical β′-2 polymorph. Differential scanning calorimetry in both the crystallization and melting modes revealed no differences between the heat flow properties of AMF, MF-TAG, and MF-DAG.     

Effect of milk fat fractions on fat bloom in dark chocolate

Anhydrous milk fat was dissolved in acetone (1∶4 wt/vol) and progressively fractionated at 5°C increments from 25 to 0°C. Six solid fractions and one 0°C liquid fraction were obtained. Melting point, melting profile, solid fat content (SFC), fatty acid and triglyceride profiles were measured for each milk fat fraction (MFF). In general, there was a trend of decreased melting point, melting profile, SFC, long-chain saturated fatty acids and large acyl carbonnumbered triglycerides with decreasing fractionation temperature. The MFFs were then added to dark chocolate at 2% (w/w) addition level. In addition, two control chocolates were made, one with 2% (w/w) full milk fat and the other with 2% (w/w) additional cocoa butter. The chocolate samples were evaluated for degree of temper, hardness and fat bloom. Fat bloom was induced with continuous temperature cycling between 26.7 and 15.7°C at 6-h intervals and monitored with a colorimeter. Chocolate hardness results showed softer chocolates with the 10°C solid fraction and low-melting fractions, and harder chocolates with high-melting fractions. Accelerated bloom tests indicated that the 10°C solid MFF and higher-melting fractions (25 to 15°C solid fractions) inhibited bloom, while the lowermelting MFFs (5 and 0°C solid fractions and 0°C liquid fraction) induced bloom compared to the control chocolates.     

Crystallization and pressure filtration of anhydrous milk fat: Mixing effects

Melt crystallization of anhydrous milk fat and subsequent filtration of the slurry is a common process for obtaining milk fat fractions with different physical and chemical properties. The crystallization mechanism is very complex and little is known about how the crystallizer conditions and the crystal size distribution (CSD) affect the filtration process. The objective of this study was to characterize the fractionation process and determine which geometric parameters of the crystallizer affect the filtration step. Two scales of fractionation were studied, 0.6 L and 3.6 L, with crystallization at 28°C. The slurry was pressure-filtered after 24 h at 500 kPa in a 1-L chamber. Impeller diameters and speeds were varied for both scales. Photomicroscopy and spectrophotometry were used to characterize the crystallization process, and filtration rates were measured by weighing the amount of filtrate passing through the filter. Filtration resistance values, calculated using the constant pressure filtration equation, as well as photomicroscopy results indicated that the agglomerates and crystals that formed had different morphological characteristics for the different mixing and flow regimes in the crystallizer. Crystallization conditions that provide an optimal filtration time, a solid fraction with minimal liquid entrainment, and a CSD with an intermediate range of sizes (80–500 μm) having good packing properties for filtration were found.     

Thermal analysis of isothermal crystallization kinetics in blends of cocoa butter with milk fat or milk fat fractions

The kinetics of isothermal crystallization of binary mixtures of cocoa butter with milk fat and milk fat fractions were evaluated by applying the Avrami equation. Application of the Avrami equation to isothermal crystallization of the fats and the binary fat blends revealed different nucleation and growth mechanisms for the fats, based on the Avrami exponent. The suggested mechanism for cocoa butter crystallization was heterogeneous nucleation and spherulitic growth from sporadic nuclei. For milk fat, the mechanism was instantaneous heterogeneous nucleation followed by spherulitic growth. For milk fat fractions, the mechanism was high nucleation rate at the beginning of crystallization, which decreased with time, and plate-like growth. Addition of milk fat fractions did not cause a significant change in the suggested nucleation and growth mechanism of cocoa butter.     

Effect of cooling rate on crystallization behavior of milk fat fraction/sunflower oil blends

The effect of cooling rate (slow: 0.1°C/min; fast: 5.5°C/min) on the crystallization kinetics of blends of a highmelting milk fat fraction and sunflower oil (SFO) was investigated by pulsed NMR and DSC. For slow cooling rate, the majority of crystallization had already occurred by the time the set crystallization temperature had been reached. For fast cooling rate, crystallization started after the samples reached the selected crystallization temperature, and the solid fat content curves were hyperbolic. DSC scans showed that at slow cooling rates, molecular organization took place as the sample was being cooled to crystallization temperature and there was fractionation of solid solutions. For fast cooling rates, more compound crystal formation occurred and no fractionation was observed in many cases. The Avrami kinetic model was used to obtain the parameters k _n and n for the samples that were rapidly cooled. The parameter k _n decreased as supercooling decreased (higher crystallization temperature) and decreased with increasing SFO content. The Avrami exponent n was less than 1 for high supercoolings and close to 2 for low supercoolings, but was not affected by SFO content.     

Solid immiscibility in milk fat fractions and methods to determine their occurrence

In this work, the occurrence of solid phase immiscibility during the fractionation process of milk fat is documented. It is shown that solid phase immiscibility occurs in normal fractionation procedures, upon crystallisation at low temperatures and in particular on further fractionation of narrow melting fractions. This phenomenon results in distinct phases of crystallisation. In some cases, the initial phase yields intermediate solid levels independent of increasing supersaturation. The subsequent increase of solids content is solely a contribution of a second, independent crystallisation event. Furthermore, it could be shown that the induction time for the second crystallisation event is not influenced by the presence or absence of the first triacylglyceride group. These results clearly indicate the independence of the crystallisation of the different immiscible fractions in terms of kinetics and thermodynamics. The different processes during the crystallisation process are monitored simply by viscosimetry.     

Effect of processing conditions on physical properties of a milk fat model system: Rheology

The effect of processing conditions on rheological behavior of three blends of 30, 40, and 50% of high-melting fraction [melting point measured as Mettler dropping point (MDP)=47.5°C] in low-melting fraction (MDP=16.5°C) of milk fat was studied. The effects of cooling and agitation rates, crystallization temperature, chemical composition of the blends, and time of storage on complex, storage and loss moduli were investigated by dynamic mechanical analysis (DMA). Compression tests were performed on samples using frequency values within the linear viscoelastic range (1 to 10 Hz). Loss modulus was, on average, 10 times lower than elastic modulus and was generally not affected by processing conditions. Samples showed a more solid-like behavior that was better described by storage modulus. Storage modulus varied with all processing conditions used in this study, and even for the same solid fat content, different rheological properties were found. Storage and complex modulus increased with temperature of crystallization (25 to 30°C), even though solid fat contents of samples measured after 24 h at 10°C were the same. Moduli were higher for samples crystallized at slow cooling rate, decreased with agitation rate, and were lower for the 30–70% blend at all processing conditions used. Storage moduli also increased with storage time. Shear storage modulus was calculated from the DMA experimental data, and the results were in agreement with the values reported in literature for butter systems. Fractal dimensions calculated for these systems showed a significant decrease as agitation rate increased in agreement with the softening effect reported for working of butter.     

Effect of processing conditions on physical properties of a milk fat model system: Microstructure

The effect of processing conditions on the microstructure of three blends of 30, 40, and 50% high-melting fraction [Mettler dropping point (MDP)=47.5°C] in the lowmelting fraction (MDP=16.5°C) of milk fat was studied. The effect of cooling and agitation rates, crystallization temperature, chemical composition of the blends, and storage time on crystalline microstructure (number, size, distribution, etc.) was investigated by confocal laser scanning microscopy (CLSM). To improve resolution, a mix of Nile blue and Nile red dyes was dissolved in the melted samples in proportions that did not modify the nucleation kinetics. Samples were then crystallized by cooling (0.2 or 5.5°C/min) to crystallization temperature (25, 27.5, and 30°C). After 2 h at crystallization temperature, a slurry was placed on a microscope slide and samples were stored 24 h at 10°C. During this period, more material crystallized. Slowly crystallized samples (0.2°C/min) formed different structures from rapidly crystallized samples (5.3°C/min). Crystals were sometimes diffuse and hard to distinguish from the liquid. Samples were darker as a result of this solid-mass distribution. However, rapidly crystallized samples had well-defined crystals and seemed to be separated by a distinct liquid phase. These crystals were not in touch with each other as was the case for slowly crystallized samples. Higher agitation rates led to smaller crystal size due to enhanced nucleation. Larger crystals were formed when crystallization occurred at higher temperatures. Storage time resulted in an increase of crystal size. Larger crystal size and structures with more evident links had a more elastic behavior with higher elastic modulus E’.     

Effect of phospholipids on isothermal crystallisation and fractionation of milk fat

Milk fats with different concentrations of water and phospholipids (PL) were crystallised isothermally under static conditions and their crystallisation behaviour was monitored by Differential Scanning calorimetry (DSC) and pulsed nuclear magnetic resonance (pNMR). The Avrami and the Gompertz models, which were fitted by non‐linear regression, described the crystallisation process. A significant effect of phospholipid concentration was observed using both techniques (DSC and pNMR). Especially the induction time and the Avrami growth rate constant were altered: higher amounts of PL delayed the onset of static crystallisation. A similar effect of PL on the crystallisation kinetics was observed in a small‐scale fractionation. Moreover, the filtration time of the crystal suspension and melting properties of the stearin were strongly affected by the presence of higher concentrations of PL. These observations emphasise the importance of the adequate removal of PL during anhydrous milk fat production.     

Monitoring milk fat fractionation: Filtration properties and crystallization kinetics

The effect of fractionation temperature, residence time, and agitation rate on the chemical composition of the stearin and olein milk fat fractions was studied. During fractionation, filtration properties of the crystal suspension were monitored; crystallization kinetics was determined by ¹H NMR. Higher fractionation temperatures result in a lower stearin yield, more oil entrapment, and a lower final solid fat content of the crystal suspension. On the other hand, the chemical composition of the resulting fractions is not influenced. Longer residence times lead to longer filtration times and lower oil entrapment, whereas the yield is not affected. Longer residence times induced lower growth rates, but chemical composition is not influenced. Agitation rates varying from 10 to 15 rpm have no influence on the chemical composition of stearin and olein milk fat fractions. Higher agitation rates decrease the filtration quality and increase stearin yield, causing a softer stearin. In designing and monitoring milk fat fractionation, filtration experiments and the assessment of crystallization kinetics are valuable techniques, but compositional chemical analysis is not favorable.     

Milk fat fractionation by solid-layer melt crystallization

The layer crystallization process has the potential to produce the same milk fat fractions as can be obtained by the suspension crystallization process. That is, milk fat fractions with solid fat content melting profiles similar to those obtained by suspension fractionation can be produced with this technique. The fatty acid profiles as well as the melting enthalpies of the different fractions confirm the separation of milk fat by the layer technique. Furthermore, there is potential to improve the results of separation presented in the first part of this paper. The two sources of improvement, temperature control of the process and controlled nucleation, lead to (i) a smooth crystalline layer with a low amount of entrapped mother liquor, contrary to the layers composed of agglomerated needles, and (ii) a good quality of attachment of the crystalline layer to the cooled surface. Moreover, the product quality can be increased using sweating as a postcrystallization step. “Sweating by warm gas” seems to have a better outlook concerning handling and controlling the process than “sweating by warm tube” because sloughing of the crystal layers can be avoided. Further investigations of the mass ratio of sweating fraction and amount of product as well as the aspect of energy consumption will determine the technical feasibility of solid-layer crystallization for fractionation of milk fat.     

Effects of minor lipids on crystallization of milk fat-cocoa butter blends and bloom formation in chocolate

Minor lipids, such as diacylglycerols, monoacylglycerols, cholesterol, and phospholipids play a key role in crystallization of fats. In this study, the effects of minor lipid components on crystallization of blends of cocoa butter (CB) with 10% milk fat or milk-fat fractions, and on bloom formation of chocolate were investigated. Both removing the minor lipids from milk fat and doubling the level of minor lipids from milk fat resulted in longer nucleation onset time, slower crystallization rate, and rapid bloom development in chocolate. Removal of minor lipids resulted in the formation of irregular primary and secondary crystals with inclusions of liquid fat, whereas the crystals were spherical and uniform in shape in the presence of minor lipids. Minor lipids from milk fat, even at the low concentrations typically found in nature, affected the crystallization of milk fat-CB blends, impacted the chocolate microstructure, and affected bloom development in chocolate.     

Relationship between crystallization behavior, microstructure, and mechanical properties in a palm oil-based shortening

In this study, the effects of cooling rate, degree of supercooling, and storage time on the microstructure and rheological properties of a vegetable shortening composed of soybean and palm oils were examined. The solid fat content vs. temperature profile displayed two distinct regions: from 5 to 25°C, and from 25°C to the end of melt at 45–50°C. A peak melting temperature of 42.7°C was determined by DSC. Discontinuity in the crystallization induction time (determined by pulsed NMR) vs. temperature plot at 27°C also suggested the existence of two separate groups of crystallizing material. Isothermal crystallization kinetics were characterized using the Avrami and Fisher-Turnbull models. In using DSC and powder X-ray diffraction, the α polymorph formed upon fast cooling (>5°C/min), and the β′ form predominated at lower cooling rates (<1°C/min). An α to β′ transition took place upon storage. Fractal dimensions (D _f ) obtained by microscopy and image analysis showed no dependence on the degree of supercooling since D _f remained constant (∼1.89) at crystallization temperatures of 5, 22, and 27°C. Crystallization at 22°C at 1°C/min and 15°C/min yielded D _f values of 1.98 and 1.93, respectively. Differences in microstructure were observed, and changes in particle properties increased the parameter λ at higher degrees of supercooling.     

A kinetic study on isothermal crystallization of palm oil by solid fat content measurements

Isothermal crystallization of plam oil was studied by means of differential scanning calorimetry (DSC) as well as by nuclear magnetic resonance spectrometry to monitor its solid fat content (SFC). The temperature of crystallization (Tc) varied from 0 to 30°C, depending on the method used. The plot of %SFC vs. time at 25°C was sigmoidal in shape. However, at lower temperatures, two consecutive curves were clearly visible. Results from DSC experiments showed the following interesting features. At each Tc, the crystals produced were of different compositions. From 0 to 8°C, the thermogram showed three peaks, with the first two peaks (I and II) sharp, and the third (III) rather broad. At elevated temperatures up to 20°C, peak II disappeared totally while peak III tended to shift toward peak I. Above 20°C, both peaks shifted downward to longer times. Peak I continued to be broadened, and then suddenly disappeared at Tc above 24°C. The melting thermograms of the crystals obtained above and below this cut-off point were distinctly different. Kinetic studies on isothermal crystallization based on the data of SFC measurements showed that the data fit well into the Avrami-Erofeev equation with n=3 over the first 70% of the crystallization.     

Effect of surface nucleation on isothermal crystallization kinetics: Theory, simulation and experiment

Surface nucleation of poly(l/d-lactide) at the interface with aluminum was studied by performing isothermal DSC analysis of amorphous samples of varying thickness between 100 °C and 130 °C. To ensure complete wetting of the aluminum surface, a hot melt laminating process was used to prepare the samples. Theoretical aspects of surface crystallization kinetics were explored and the resulting model was compared with the results of Monte-Carlo simulations. Three stages of surface crystallization were identified depending on the growth geometry: (1) impingement-free growth, (2) increasingly laterally-constrained transverse growth, and (3) interstitial growth. By fitting the Monte-Carlo simulation to the experimental half-times of crystallization the surface nucleation concentration and the bulk nucleation rate was estimated at 4 different temperatures. It was found that both surface nucleation concentration and the bulk nucleation concentration decrease with increasing crystallization temperature.