Using USAXS to predict the under-tempered chocolate microstructure.

Chocolate is a manufactured product enjoyed worldwide. Over the years, manufacturers have learned how to appeal to humans using this rich-fat food that arouses all the senses. Good quality chocolate is recognized by its smoothness, a slow melt in the mouth, and a snap when bitten, and described as well-tempered. This work compares dark chocolate samples manufactured to obtain under- and well-tempered chocolate, where under-tempered does not show all the physical properties desired by consumers. The microstructure was studied using the ultra small angle X-ray scattering (USAXS) technique, complemented by small and wide angle X-ray scattering to identify the polymorphs. It was observed that under- and well-tempered chocolates exhibited differences in the q-region ~ 2 × 10-5 Å-1 < q < ~1.5 × 10-4 Å-1, which correspond to spatial length scales from 32 µm to 3.2 µm. The differences are manifested in the value of the mass fractal dimension, D, obtained when the USAXS data were fitted using the Unified Fit model (Irena software). The characteristic length scale at which these differences were observed falls in length scales detected by humans in the oral cavity. This work proposes that a D = 2.1 characterizes an under-tempered 70% dark chocolate while a D = 2.3 characterizes a well-tempered 70% dark chocolate. This work also presents a simple model that describes the disintegration of those aggregates formed by the basic scatter units for under- and well-tempered chocolate. The model proposes that aggregates formed in under-tempered chocolate persist after the bulk chocolate has melted, which can be perceived as grittiness. However, the model proposes that the aggregates for well-tempered chocolate melt at the same or lower temperatures than the bulk chocolate melting temperature; hence no grittiness is perceived. The model is supported by the observation that the heat of transition for the under-tempered chocolate is smaller than that of the well-tempered case.

Using the USAXS technique to reveal the fat globule and casein micelle structures of bovine dairy products.

Cows' milk is a commodity used worldwide to make many dairy products. We have used the ultra small angle X-ray scattering (USAXS) technique to reveal the fat globule and casein micelle structures of some dairy products. USAXS covers the q-range 5 × 10-4 Å-1 < q < 10-1 Å-1, thereby allowing the study of micron-scale structures present in those dairy products. We measured the USAXS intensity, Iq, as a function of the scattering vector magnitude, q, for samples of skim milk, non-homogenized whole milk, homogenized whole milk, half and half and heavy cream, at two temperatures, 7 °C and 45 °C. The data collected from the scattering experiments were fitted using the Unified fit model run under the IRENA software from the Advanced Photon Source, Argonne National Laboratory (Illinois, USA). The fittings were carried out when the data were plotted as log[I(q)] vs log[q]. We observed a combination of linear regions (LRs) and knees. Two parameters of interest were obtained from the fittings, a radius of gyration, Rg, and a Porod exponent, P. Unified fit allowed us to fit up to four structural levels. One of the knees was centered at q ≈ 8 × 10-3 Å-1 for all samples measured at 7 °C, but vanished at 45 °C. Two LRs were identified as being either due to casein micelles (CMs) or to fat globules (FGs). The porod exponent obtained from these LRs allowed us to describe the surface morphology of CMs and FGs. Two of the Rg values gave a rough estimate of the FGs and CMs sizes. FGs were identified for samples of homogenized whole milk, half and half and heavy cream in the q-region 2 × 10-4 < q < 8 × 10-4 Å-1. We found that, in the absence of chymosin, or changes in pH, CaCl2 concentration or temperature changes, skim milk and non-homogenized whole milk displayed a Porod exponent that indicated a behavior characteristic of aggregation. Using computer simulations, we found that, seemingly, bovine CMs spontaneously formed approximately 1-dimensional aggregates possibly analogous to swollen randomly branched polymers.

Small and ultra-small angle neutron scattering studies of commercial milk.

Milk and milk products are an essential part of global nutrition and the world-wide food industry. Studies of milk components using scattering techniques are well documented in the literature. However, those studies focused on the q scattering wavevector region 10 - 3 < q < 2 Å - 1 . This manuscript presents scattering results in the region 3 × 10 - 5 < q < 2 × 10 - 2 Å - 1 , a region that allows the simultaneous study of fat globules and proteins found in commercial food-grade milk. The small and ultra-small angle neutron scattering (SANS and USANS) measurements show that a model based on the Schulz distribution function using uniform spheres was a reasonable choice to successfully fit the scattering features below q = 0.007 Å - 1 . Contrast measurements using D2O on whole milk were carried out to distinguish fat from protein signals. Casein micelles were found to have mean diameters of 96 ± 10 nm with 33% polydispersity. The average scattering length density of the micelles varied from -0.04 × 10-6 Å -2 in homogenized, pasteurized commercial milk to 2.8 ×10-6 Å -2 with 50% dilution by D2O, with a match point of 43 ± 3%, as seen in previous studies. It was found that the average diameter of fat globules in homogenized whole milk was 0.47 ± 0.04 µm with a polydispersity of 45 ± 5%, and a volume fraction of 0.034 ± 0.002 when the scattering length density is fixed at 0.20 × 10-6 Å -2. These USANS measurements provide an important foundation as similar techniques are employed to study cheese varieties and cheese formation.

Molecular Insights into the Eutectic Tripalmitin/Tristearin Binary System.

A molecular interpretation of the eutectic behavior of a binary mixture of tristearin (SSS) and tripalmitin (PPP) triglycerides was formulated using computer simulations and experimental techniques (calorimetry and X-ray scattering). A eutectic composition was identified using both experimental and computer simulation techniques at a composition of 70% PPP and 30% SSS, in agreement with previous findings in the literature. The decrease in the melting temperature at the eutectic composition can be ascribed to an interplay between enthalpic and entropic effects. In particular, a lower global melting enthalpy at the eutectic composition was detected here, caused by a less efficient packing of the triglycerides in the crystal. On the other hand, a higher crystalline disorder is reflected in a lower change in the entropy of melting. The simultaneous decrease in global enthalpy and entropy has a contrasting effect on the melting temperature, with a slight melting point depression found in both experiment and simulations, resulting from a combination of enthalpic and entropic factors. Computer simulations showed, in fact, that the eutectic effect can be ascribed to the reduction of crystalline order when SSS molecules are incorporated into the PPP crystal structure. This decrease of the crystalline order is due to the protrusion of SSS end-chains (last three carbons of each alkyl chain) into the interlamellar space between adjacent lamella. These end-chains disturb the orderly stacking of the lamella, as evidenced by low-density regions in the interlamellar space. Thus, the greater disorder of the last atoms of the SSS alkyl chains is consequently due to the greater conformational freedom. At molecular level, in fact, the conformational freedom of terminal atoms of SSS surrounded by shorter PPP molecules is larger than the conformational freedom of longer SSS in the neighborhood of short PPP.

Characterization of the nanoscale structure of milk fat.

The nanoscale structure of milk fat (MF) crystal networks is extensively described for the first time through the characterization of milk fat-crystalline nanoplatelets (MF-CNPs). Removing oil by washing with cold isobutanol and breaking-down crystal aggregates by controlled homogenization allowed for the extraction and visualization of individual MF-CNPs that are mainly composed of high melting triacylglycerols (TAGs). By image analysis, the length and width of MF-CNPs were measured (600 nm × 200 nm-900 nm × 300 nm). Using small-angle X-ray scattering (SAXS), crystalline domain size, (i.e., thickness of MF-CNPs), was determined (27 nm (d001)). Through interpretation of ultra-small-angle X-ray scattering (USAXS) patterns of MF using Unified Fit and Guinier-Porod models, structural properties of MF-CNPs (smooth surfaces) and MF-CNP aggregations were characterized (RLCA aggregation of MF-CNPs to form larger structures that present diffused surfaces). Elucidation of MF-CNPs provides a new dimension of analysis for describing MF crystal networks and opens-up opportunities for modifying MF properties through nanoengineering.

Prevention of oil migration in palm mid fraction and palm olein using a stabilizer rich in behenic acid.

Oil loss is a common problem in high oil volume fraction fat-structured foods, where trans fatty acids are absent and saturated fatty acids minimized. Usually a stabilizer "fat", rich in behenic and stearic acids is added to these products. This stabilizer significantly reduces oil migration and loss; however, the mechanism by which it exerts its effect is not known. Here we study the structure and oil loss characteristics of blends of palm mid fraction (PMF) and palm olein (PO) mixed with a commercial stabilizer of fully hydrogenated cottonseed and rapeseed oils added at 1.25, 3.5, and 7% (w/w) levels. The addition of as little as 1.25% stabilizer reduced oil migration from 13% to ~2% for the 80:20 PMF/PO blend and from 29% to ~3.5% for 20:80 PMF/PO blend after 14days of storage at 20°C, as judged from a filter paper assay. Polarized light microscopy results demonstrated that addition of stabilizer decreased fat crystal size and lead to the formation of a denser network with smaller pores. The rate of crystallization, determined by monitoring increases in solid fat content as function of time by pulsed NMR, increased upon addition of the stabilizer as indicated by decreases in the half-time of crystallization. Addition of stabilizer lead to an increase in the mechanical strength of the material, as indicated by increases in breaking force. Differential scanning calorimetry showed the formation of a new fraction, intermediate between the high melting fractions in both PMF and PO and the stabilizer upon incorporation of the stabilizer in the blends. We hypothesize that the stabilizer helps the formation of this new fraction which has a higher nucleation rate and thus forms a denser network of small crystals with an improved ability to bind and retain oil.

Ultra small angle x-ray scattering in complex mixtures of triacylglycerols.

Ultra-small angle x-ray scattering (USAXS) has been used to elucidate, in situ, the aggregation structure of unsheared model edible oils. Each system comprised one or two solid lipids and a combination of liquid lipids. The 3D nano- to micro-structures of each system were characterized. The length scale investigated, using the Bonse-Hart camera at beamline ID-15D at the Advanced Photon Source, ANL, ranged from 300 Å-10 µm. Using the Unified Fit model, level-1 analysis showed that the scatterers were 2D objects with either a smooth, a rough, or a diffuse surface. These 2D objects had an average radius of gyration Rg1 between 200-1500 Å. Level-2 analysis displayed a slope between -1 and -2. Use of the Guinier-Porod model gave s ≈ 1 thus showing that it was cylinders (TAGwoods) aggregating with fractal dimension 1 ≤ D2 ≤ 2. D2 = 1 is consistent with 1D structures formed from TAGwoods, while D2 = 2 implies that the TAGwoods had formed structures characteristic of diffusion or reaction limited cluster-cluster aggregation (DLCA/RLCA). These aggregates exhibited radii of gyration, Rg2, between 2500 and 6500 Å. Level-3 analyses showed diffuse surfaces, for most of the systems. These interpretations are in accord with theoretical models which studied crystalline nano-platelets (CNPs) coated with nano-scale layers arising from phase separation at the CNP surfaces. These layers could be due to either liquid-liquid phase separation with the CNPs coated, uniformly or non-uniformly, by a diffuse layer of TAGs, or solid-liquid phase separation with the CNPs coated by a rough layer of crystallites.A fundamental understanding of the self-organizing structures arising in these systems helps advance the characterization of fat crystal networks from nanometres to micrometres. This research can be used to design novel fat structures that use healthier fats via nano- and meso-scale structural engineering.

Aggregation in complex triacylglycerol oils: coarse-grained models, nanophase separation, and predicted x-ray intensities.

Triacylglycerols (TAGs) are biologically important molecules which form crystalline nanoplatelets (CNPs) and, ultimately, fat crystal networks in edible oils. Characterizing the self-assembled hierarchies of these networks is important to understanding their functionality and oil binding capacity. We have modelled CNPs in multicomponent oils and studied their aggregation. The oil comprises (a) a liquid component, and (b) components which phase separately on a nano-scale (nano-phase separation) to coat the surfaces of the CNPs impenetrably, either isotropically or anisotropically, with either liquid-like coatings or crystallites, forming a coating of thickness ?. We modelled three cases: (i) liquid?liquid nano-phase separation, (ii) solid?liquid nano-phase separation, with CNPs coated isotropically, and (iii) CNPs coated anisotropically. The models were applied to mixes of tristearin and triolein with fully hydrogenated canola oil, shea butter with high oleic sunflower oil, and cotton seed oil. We performed Monte Carlo simulations, computed structure functions and concluded: (1) three regimes arose: (a) thin coating regime, Δ < 0.0701 u (b) transition regime, 0.0701 u ≤ Δ ≤ 0.0916 u and (c) thick coating regime, Δ > 0.0916 u. (arbitrary units, u) (2) The thin coating regime exhibits 1D TAGwoods, which aggregate, via DLCA/RLCA, into fractal structures which are uniformly distributed in space. (3) In the thick coating regime, for an isotropic coating, TAGwoods are not formed and coated CNPs will not aggregate but will be uniformly distributed in space. For anisotropic coating, TAGwoods can be formed and might form 1D strings but will not form DLCA/RLCA clusters. (4) The regimes are, approximately: thin coating, 0 < Δ < 7.0 nm transition regime, 7.0 < Δ < 9.2 nm and thick coating, Δ > 9.2 nm (5) The minimum minority TAG concentration required to undergo nano-phase separation is, approximately, 0.29% (thin coatings) and 0.94% (thick coatings). Minority components can have substantial effects upon aggregation for concentrations less than 1%.

Oil binding capacities of triacylglycerol crystalline nanoplatelets: nanoscale models of tristearin solids in liquid triolein.

Polycrystalline particles composed of triacylglycerol (TAG) molecules, and their networks, in anhydrous TAG oils find extensive use as edible oils in the food industry. Although modelling studies of TAG systems, have been carried out, none have attempted to address a problem of central concern to food science and technology: the "oil binding capacity" of a system of such edible oils. Crystalline nanoparticles (CNPs) have recently been identified as the fundamental components of solid fats in oils. Oil binding capacity is an important concept regarding the ability of fats particles to retain oil, and the ability of these CNPs to bind oil is important in designing healthy foods. We have carried out atomic scale molecular dynamics computer simulations to understand the behavior of a triacylglycerol oil (triolein) in nanoscale confinements between tristearin CNPs. We define a nanoscale oil binding capacity function by utilizing the average oil number density, 〈Φ(d)〉, between two CNPs as a function of their separation, d. We modelled pure tristearin CNPs as well as tristearin CNPs in which the surfaces are covered with an interface comprising soft permanent coatings. Their surfaces are "hard" and "soft" respectively. We found that for a pair of hard-surface tristearin CNPs a distance d apart, (i) triolein exhibits number density, and therefore density, oscillations as a function of d, and (ii) the average number density between two such CNPs decreases as d decreases, viz. the oil binding capacity is lowered. When a soft layer of oil covers the CNP surfaces, we found that the oscillations are smeared out and that the average number density between the two CNPs remained approximately constant as d decreased indicating a high oil binding capacity. Our results might have identified important nanoscale aspects to aid in healthy food design.

Nanoscale characteristics of triacylglycerol oils: phase separation and binding energies of two-component oils to crystalline nanoplatelets.

Fats are elastoplastic materials with a defined yield stress and flow behavior and the plasticity of a fat is central to its functionality. This plasticity is given by a complex tribological interplay between a crystalline phase structured as crystalline nanoplatelets (CNPs) and nanoplatelet aggregates and the liquid oil phase. Oil can be trapped within microscopic pores within the fat crystal network by capillary action, but it is believed that a significant amount of oil can be trapped by adsorption onto crystalline surfaces. This, however, remains to be proven. Further, the structural basis for the solid-liquid interaction remains a mystery. In this work, we demonstrate that the triglyceride liquid structure plays a key role in oil binding and that this binding could potentially be modulated by judicious engineering of liquid triglyceride structure. The enhancement of oil binding is central to many current developments in this area since an improvement in the health characteristics of fat and fat-structured food products entails a reduction in the amount of crystalline triacylglycerols (TAGs) and a relative increase in the amount of liquid TAGs. Excessive amounts of unbound, free oil, will lead to losses in functionality of this important food component. Engineering fats for enhanced oil binding capacity is thus central to the design of more healthy food products. To begin to address this, we modelled the interaction of triacylglycerol oils, triolein (OOO), 1,2-olein elaidin (OOE) and 1,2-elaidin olein (EEO) with a model crystalline nanoplatelet composed of tristearin in an undefined polymorphic form. The surface of the CNP in contact with the oil was assumed to be planar. We considered pure OOO and mixtures of OOO + OOE and OOO + EEO with 80% OOO. The last two cases were taken as approximations to high oleic sunflower oil (HOSO). The intent was to investigate whether phase separation on a nanoscale took place. We defined an "oil binding capacity" parameter, B(Q,Q'), relating a state Q to a reference state Q'. We used atomic scale molecular dynamics in the NVT ensemble and computed averages over 1-5 ns. We found that the probability of the OOE phase separating into a layer on the surface of the CNP compared to being retained randomly in an OOO + OOE mix were approximately equal. However, we found that it was probable that the EEO component of an OOO + EEO mix would phase separate and coat the surface of the CNP. These results suggest a mechanism whereby many-component oils undergo phase separation on a nanoscale so as to create a transition oil region between the surface of the CNP and the bulk major oil component (OOO in the case considered here) so as to create the appropriate oil binding capacity for the use to which it is put.