Optimization of cocoa butter analog synthesis variables using neural networks and genetic algorithm

Area of Science:

• Food Science and Technology
• Biocatalysis
• Supercritical Fluid Technology

Background:

• Cocoa butter is a valuable fat with specific physical properties.
• Camel hump fat and tristearin are potential raw materials for cocoa butter alternatives.
• Enzymatic interesterification offers a sustainable method for fat modification.

Purpose of the Study:

• To develop a cocoa butter analog using camel hump fat and tristearin.
• To optimize enzymatic interesterification conditions using advanced modeling techniques.
• To achieve a triacylglycerol (TAG) distribution similar to natural cocoa butter.

Main Methods:

• Enzymatic interesterification of camel hump fat and tristearin in supercritical carbon dioxide (SC-CO2).
• Use of immobilized Thermomyces lanuginosus lipase (Lipozyme TL IM) as a biocatalyst.
• Optimization of process variables (pressure, temperature, substrate ratio, water content, time) using Response Surface Methodology, Neural Networks, and Genetic Algorithms.

Main Results:

• A general regression neural network model accurately predicted TAG distribution based on process parameters.
• Genetic algorithm optimization identified optimal conditions for producing a cocoa butter analog.
• Optimum conditions: 10 MPa pressure, 40°C temperature, 0.6:1 SSS/CHF ratio, 13% water content, and 4.5 hours incubation time.

Conclusions:

• Enzymatic interesterification in SC-CO2 is an effective method for producing cocoa butter analogs.
• Advanced optimization techniques (neural networks, genetic algorithms) are crucial for achieving desired fat structures.
• The developed cocoa butter analog closely mimics the properties of natural cocoa butter under optimized conditions.