Technology innovates and improves chocolate production. Advanced data analytics, algorithm generation, artificial intelligence, and 3D printing help develop innovative products and optimize the many steps of the supply chain.
From the cultivation of cocoa beans to the point of sale, new technologies combined with the Internet Of Things and Artificial Intelligence improve chocolate production. By analysing the myriads of data collected across the supply chain and benefiting from the benefits of machine learning, the chocolate industry develops new products, optimizes ingredient choice, improves processes, reduces waste and energy consumption.
The Pods Harvesting
Chocolate production begins with the cultivation of cocoa plants, which is still spread over many small family farms and some large plantations. In nature, cocoa trees can be more than twenty meters high, but in cultivation they are kept within five meters to facilitate controls and harvesting. The first fruits appear five years after sowing, the maximum yield is between the twentieth and thirtieth year and progressively decreases in the following decade. Cocoa pods (the fruit) grow by cauliflory, a phenomenon where flowers and fruits grow directly from the older, woody parts of a plant.
Each tree produces 20 to 50 pods, ripening takes 4 to 7 months, harvesting takes place twice a year. Each pod contains 20 to 40 beans (seeds) wrapped in a white, sticky, acidic flesh, mostly composed of water and carbohydrates. The quality of the beans depends on the territory, the variety cultivated and the possibility of inspecting the trees for a timely identification of anomalies. Smallholders rely on the work of skilled operators, while large plantations use drones that capture high-resolution images. Data processing allows for proper interventions in presence of parasitic or infectious diseases or signs of stress. Technologies associated with drones are multispectral and hyperspectral imaging or Infrared imaging.
The first detect changes in reflectance and transmittance. They differ in the amount of bands captured, the spectral resolution, and the electromagnetic radiation spectra in each band. In Infrared imaging, one or more thermal cameras detect the Infrared radiation from the plant, converting it into an electrical signal, and processing the information into a visible image representing the surface temperature distribution of the framed plants and caboose. Plants emit long-wave Infrared radiation, a portion of the Infrared spectrum furthest from visible light; the amount they emit varies based on the tissues ability to dissipate the energy absorbed from the sun. The emission is influenced by the tissues’ health, their metabolic activity, stomatal function, and water status of the plant.
Abnormal appearances indicate that a plant is suffering. Artificial intelligence tools and data analysis enable the utilization of satellite imagery, soil analysis and climate data to accurately assess cocoa plantations, improve growing conditions, predict yields, and select beans based on genetic traits that determine their aromatic potential. Research is also focused on the development of systems for robotic collection of caboose. In some areas of the fruit and vegetable sector, robotisation is already a reality; in the cocoa sector, however, its implementation is hindered by the irregular shape of plants and soils. The prototypes developed so far recognize and collect ripe pods using artificial vision systems, optical sensors, cameras, cutting systems and robotic arms equipped with harvesting grippers. A perfected robotic harvesting would guarantee faster intervention times with the same result, as well as a reduction in waste and in risks for operators as a result of adverse weather conditions or hazardous activities such as the use of machete or plant protection products.
Cocoa Beans Cleaning, Fermentation and Drying
Another interesting area of research concerns the mechanized opening of the pod and the extraction of seeds. Pods must be opened within three days of harvest. Pod breakers can be fed manually, with hoppers or powered conveyor belts. They typically use either lateral or longitudinal crushing (with shear, compression and impact) as well as lateral or horizontal cutting with sharp blades. The seeds are extracted by shaking the pod. A sieve separates the beans from other fragments. The beans covered with pulp residues are fermented for about a week in large wooden crates or under banana leaves protected from direct light. Yeasts first ferment sugars in the pulp to alcohol, then anaerobic bacteria turn alcohol into acetic acid, water, and carbon dioxide.
During this process, the residual pulp softens and liquefies, the beans’ colour changes, astringency is reduced, and the precursors of the flavours associated with chocolate are developed. The progressive increase in temperature to 50°C inhibits or prevents the germination of most seeds. Fermented beans retain 50% moisture and in traditional processing they are sun-dried for a maximum of two weeks. Drying slows down the fermentation process and prevents the growth of mold and unwanted acidic flavours, which could negatively impact chocolate quality. Hot-air drying is commonly used in large-scale processing. The drying room consists of a burner and a fan to force warm air upward through a thin layer of cocoa beans, typically 8 to 10 cm thick.
The beans are periodically raked to ensure uniform drying. The heated air transfers heat and removes moisture through convection. The drying room where the beans are moved is under slight vacuum conditions to effectively dissipate moisture. At the end of the process, the dried beans are transferred from the dryer into a bin using an auger. The following steps include beans calibration, quality-based classification, and packaging in breathable jute bags which prevent mold. An additional precaution includes storing and transporting the jute bags in ventilated areas. By leveraging sensors, IoT devices, and real-time data collection, manufacturers monitor temperature and humidity during these processes. Data processing provides suggestions on optimal fermentation and drying conditions and times.
Roasting
Prior to roasting, beans are cleaned and sorted to remove foreign materials, broken or deteriorated seeds and reduce moisture to improve its quality and flavour. Whole beans are roasted in rotating drum systems; the roasting parameters (temperature, time and heat transfer method) depend on the origins of the cocoa beans and are adjusted to develop the desired product. The most commonly used temperature range is 110-130°C with times ranging from 30 minutes to two hours. These conditions promote the Maillard reaction involving the reduction of sugars and amino compounds. In its intermediate stages, new volatile organic compounds (low molecular weight) are formed, while in its advanced stages, high molecular weight melanoidins are formed.
Traditional roasting is a slower process than convection, because the heat is transferred from the hot drum surface to the beans, whereas in convection transfer, the heat is transferred by the movement of hot air which surrounds the beans. The center temperature of the beans is lower than the surface temperature. The next step is the decortication that separates the cocoa beans’ edible nib from its outer shell peel before the cocoa beans are cracked. Drying cocoa beans is an alternative technique. It is preceded by decortication which involves pre-roasting with infrared lamps.
The shell dries out, becomes brittle and is separated from the inner kernel using an air jet. Fluidized bed roasting offers high particle mobility and uniform temperatures. The heat exchange with air occurs mainly by convection. The downward force of gravity on the particles to be roasted is counteracted by the upward force of the hot air flow. An emerging technology is the use of superheated steam. Experiments to date confirm that roasting improves the texture, color, aroma and microstructural properties of roasted grains. Conventional roasting can damage cellular structures like the cytoplasm and break cell organelles. Superheated steam technology results in looser cellular structures and smoother internal formations, leading to better qualities in roasted beans.
Cocoa Liquor
Cocoa liquor is produced by grinding roasted cocoa nibs in granite mills or knife mills, creating a thick, fluid paste or liquid that becomes increasingly fluid during the milling process. The heat due to friction melts the fat out of the grain and allows cocoa butter to be squeezed out using four different methods: mechanical press, hydraulic press, solvent extraction or supercritical fluid (carbon dioxide). This cocoa mass is then pressed to obtain cocoa butter and a solid residue (cocoa cake) that is finely ground to create cocoa powder that is classified with “ordinary” (with around 20 to 24% fat content) or defatted (with around 10 to 12% fat content).
Cocoa powder can be treated with an alkaline solution to neutralize its natural acidity, resulting in a smoother, less bitter flavour, a darker colour and improved solubility in water. Additionally, the fat is filtered out and deodorized under vacuum to improve its rheology. At room temperature, fat’s hardness is influenced by its saturated fat content and the crystalline lattice structure. Its main constituents are: 1,3-dipalmitoyl-2-oleoyl glycerol (POP), 1-palmitoyl-2-oleoyl-3 stearoyl glycerol (POS), and 1,3-distearoyl-2-oleoyl glycerol (SOS). Their proportions determine the properties of the product.
Asian cocoa butter generally has higher stearic acid and lower palmitic acid content than African cocoa butter. South American cocoa butter contains more unsaturated fats, making it softer. Cocoa butter has at least six distinct crystalline forms, each characterized by a specific melting temperature. Form I melts at 17.3°C, Form II at 13.3°C, Form III at 25.5°C, Form IV 27, 3°C, Form V at 33.8°C and Form VI at 36.3°C. Form V is described as shiny and smooth, and it melts quickly in the mouth. Tempering is utilized in the chocolate industry to obtain this texture.
Refining, Conching and Tempering
Refining reduces particle size through grinding to create a more homogeneous texture. The mass repeatedly passes through a battery of ever closer rotating steel cylinders. In conching, the conch stirs and mixes the chocolate mass at 50°C, creating a finer, smoother texture, reducing moisture and harsh flavours, though sometimes at the cost of positive aroma compounds. In the final step, the chocolate recipe can be completed by adding cocoa butter, vanilla and soy lecithin as an emulsifier. During tempering, the tempering machine mixes the product and heats it to 40°C. Then the melted chocolate is cooled to a temperature around 28°C and reheated to a working temperature of 32°C, to eliminate unstable crystals and encourage the formation of the desired Form V, providing the product being processed – and consequently the future chocolate – a smooth, glossy, and homogeneous finish.
The chocolate will have better resistance to breakage and be less susceptible to changes due to temperature and humidity changes or long-term storage. In properly tempered chocolate, the cocoa butter crystals regularly cross-link and homogeneously trap the other components; if the structure is not perfect, consumers would immediately notice an inhomogeneity. It would increase the likelihood of the formation of a white coating on the chocolate surface, typical of fat bloom. Improperly tempered chocolate is chewy, chalky, grainy and opaque, especially on the inside. An alternative technique allowing for the desired crystal structure to form consists in introducing Form V nuclei into molten chocolate and controlling the temperature during cooling.
Advanced automatic tempering systems are equipped with automated temperature sensors and regulators, alongside automatic flow and cutting controls, optimized for the various recipes to be produced. The system adjusts processing based on changing viscosity and progressive crystal formation. These solutions use the application of Computational Fluid Dynamics algorithms to solve complex fluid flow, heat transfer, and mass transfer problems simultaneously, and predict the resulting structure of the finished product based on detailed velocity, shear, temperature, and pressure profiles within the system. Plant manufacturers also focused on the design of agitators: to achieve the desired Form V crystal structure in molten chocolate, the agitators are designed to promote a downward axial movement and create a central vortex ensuring uniform temperature and distribution of these Form V crystals in the molten chocolate.
Moulding and Packaging
The processing ends with molten chocolate being poured into the molds and then solidified in a cooling tunnel before being removed from the molds and packaged. Optimal storage involves keeping chocolate in the dark, at temperatures between 15 – 18°C, and maintaining a relative humidity below 65%. The shelf-life for properly stored dark chocolate generally is around eighteen months, one year for milk chocolate, and eight months for white chocolate. The most significant innovation is the use of 3D printing for creating diverse textures and custom shapes in the last stages of production. Using a 3D printer, molten chocolate is deposited in layers; this process minimizes material wastes, because only the necessary amount of chocolate is used.
Again, careful control of temperature, nozzle mechanics, motion systems and extrusion algorithms is essential, because ingredient proportions significantly impact the material flow properties (rheology) during extrusion, directly affecting the final product quality. The near-instantaneous solidification of the deposited layer is vital to support the weight of the next layer. A cold printing table with a cold-water circulation system is essential. To prevent condensation, the printing table is also equipped with a cold airflow.
Crucial parameters when designing packaging include ensuring the welding temperature is optimal for sealing packages and designing sustainable packaging. Newer packaging machines are incorporating energy-generating systems, to create significant energy savings compared to traditional solutions. The machines used in this section of the plants are equipped with systems that collect consumption data and functional parameters and predict machine failures and wear in good time.
Artificial Intelligence
The chocolate industry has been one of the first to use Artificial Intelligence, initially in quality control and the development of new products, and later on for supply chain management and traceability. A first application concerns quality control during production and on the finished product. By combining the Internet Of Things, sensors and interconnections between machines, the Internet and Artificial Intelligence, the most relevant parameters are continuously monitored, and the system is modified in real time to ensure consistent quality.
Chemical analysis and flavour profiling help to accurately identify the aromatic compounds and off flavours present in the source beans, hypothesising how the different processing techniques will influence the taste of the finished product. Collected data can also be used to create innovative and customizable aromatic profiles. Research and development use the same mechanisms to identify specific flavours for the individual origins of cocoa and to intervene on fermentation and roasting to enhance flavours also taking into account the preferences of the market or, in more rare cases, of individual customers who ask for personalised recipes with specific ingredients, degree of sweetness and textures.
Artificial Intelligence, machine learning, and blockchain technologies are also the strengths for logistics and traceability applied to the chocolate industry. Predictive analysis helps predict demand and consequently the amount of raw materials to be purchased, ensuring product availability at the best possible price, optimizing inventory, helping to plan production, avoiding overproduction or stock outs, reducing waste and ensuring products are available when customers need them. With artificial intelligence and real-time monitoring, companies can monitor every step of the supply chain, from material procurement to delivery by tracking and sharing product information at every stage of the product’s life.
