Development and validation of an ultrasonic shear wave reflection technique to monitor the crystallization of cocoa butter

Fat crystallization plays a critical role in determining the macroscopic properties of fat containing food products, such as chocolate, butter and margarine. However, most of the currently used techniques to monitor fat crystallization are off-line techniques, while on-line techniques could have considerable financial benefits by avoiding expensive rework or disposal of out-of-specification products. The aim of this PhD research was therefore to study the potential of an advanced ultrasonic shear wave reflection (USWR) technique based on low intensity ultrasound which has the potential to measure the crystallization behavior on-line. This reflection set-up was chosen to avoid problems with the high attenuation of fats and shear waves were used to obtain information about the microstructure of fats. The proposed technique was developed under both static (chapter 3) and dynamic (stirring) conditions (chapter 6) using cocoa butter as a substrate. The validation of both techniques (chapter 5 and 8) was conducted by varying process conditions and adding minor components to cocoa butter. Limonene and lecithin were chosen as minor components and their effects on the crystallization behavior of cocoa butter were also studied in depth by conventional techniques (chapter 4 and 7). In chapter 3, an USWR technique to monitor the isothermal crystallization process of cocoa butter under static conditions was developed. The custom-built experimental set-up basically consists of a cylindrical sample holder, which is open at both sides, with the bottom side connected to a thick plexiglass plate. A shear wave transducer is positioned at the bottom of the plexiglass plate. The remarkable oscillatory damped response in the shear wave reflection coefficient (swRC) as function of the crystallization time could be explained by constructive and destructive interference of a first reflection at the boundary between the plexiglass delay line and the crystallized cocoa butter and a second reflection occurring at the interface between crystallized and liquid substance. This hypothesis was supported by the excitation frequency dependence of the oscillations. An inverse model was developed based on the oscillatory damped response including four parameters: tind (induction time), K (crystallization rate constant), vs2 (shear ultrasonic velocity) and as2 (shear ultrasonic attenuation coefficient). Before this technique was validated, the effect of limonene on the crystallization behavior of cocoa butter was first studied with conventional techniques (chapter 4). Differential scanning calorimetry (DSC) and X-ray diffraction measurements revealed that the effect of limonene on the crystallization kinetics was temperature dependent. This could be explained by a two-sided effect: the addition of limonene resulted in a reduction of the amount of unstable crystals and an acceleration of polymorphic transitions. At 17°C, the crystallization process was accelerated due to the acceleration of the formation of more stable polymorphic forms, while there were insufficient a crystals for an a mediated b’ nucleation at 20°C, resulting in a slower crystallization process. In chapter 5, both the temperature effect and the effect of limonene in different concentrations on the isothermal crystallization of cocoa butter were studied by way of the USWR technique and the associated inverse model. Subsequently, the USWR parameters were compared to results of conventional techniques to monitor fat crystallization (DSC, microscopy). The study showed that tind and K provide information on the kinetics of the microstructure development, as the results of the USWR technique were much more comparable to microscopy measurements providing information about the developed microstructure than to DSC experiments which are known to measure the primary crystallization. The parameter vs2 is related with the equilibrium SFC (solid fat content), while as2 is both influenced by the SFC and the organization of the crystals in the network, yielding information about the microstructure of the crystallized samples. In chapter 6, the monitoring technique was adapted for dynamic conditions by placing an overhead stirrer above the sample holder. Under dynamic conditions, the typical pattern of the swRC exhibited a monotonic decline instead of the oscillating pattern and also the frequency dependence disappeared. This drastic change could be explained by the fact that in this case a homogeneous sample was measured, without the formation of additional layering and surfaces, and hence without interferences. Therefore, the inverse model has also been adjusted to be able to fit this monotonic decline, resulting in two parameters: t1 and Kini, indicating the start and initial rate constant of the microstructural development respectively. Only the first 30 minutes were modeled as a much faster stick release between the plexiglass and the sample was observed under dynamic conditions, making a further measurement impossible. In chapter 7, the effect of different lecithins on cocoa butter crystallization was studied with conventional techniques (DSC, rheology and a simple dynamic experiment). It was however very difficult to infer general conclusions as the effect of lecithin depended on the concentration and the matrix to which they were added. Furthermore, the observed effects depended on the measuring technique used to study the crystallization, whereby only the dynamic experiment could detect a difference between the different lecithins. Therefore in chapter 8, the effect of different lecithins was investigated with the more advanced dynamic USWR technique. To investigate the sensitivity of this technique, the stirring rate was first varied, and it was observed that a higher stirring rate resulted in a lower t1 and higher Kini, which is consistent with a faster microstructural development. Furthermore, one sunflower lecithin demonstrated a lower t1 and a much faster release between the plexiglass and the sample, suggesting that certain differences in microstructural development could be detected by the dynamic USWR technique, while no differences were observed by static DSC experiments measuring the primary crystallization. In conclusion, the proposed USWR technique shows definite promise to monitor the fat crystallization process. Its major advantages are the non-destructive character, the information about the microstructural development and the ability to measure under dynamic conditions.nrpages: 205status: publishe