The Impact of grinding intensity on particle properties and rheology of dark chocolate
The following is a summary of the stated research mentioned above. The content summarized here, including the figures and tables, all belong to the researchers (unless otherwise indicated). The summary attempts to stay as close to the original paper as much as possible with some adjustments in regards to jargon, length, or to focus on bean to bar aspects.
Introduction
Molten (liquid) chocolate is a multiphase system. In regards to chocolate, each phase is as a different ingredient or constituent. For instance, with dark chocolate there is the cocoa butter phase, the cocoa solid phase, the sugar solid phase. Chocolate is a mixture of all these various phases. Cocoa butter is the continuous phase, meaning other phases (cocoa solids, sugar, milk powder) are dispersed within it. The rheology (flow) of this concentrated suspension is important when it comes to moulding and coating (Afoakwa, Paterson, & Fowler, 2007).
The rheology of liquid chocolate is impacted by many factors:
The volume-based amount of suspended particles (in this case, the cocoa solids, sugar, milk solids) obviously plays an important role in the flow of chocolate (Beckett, 2000). For instance, if there are higher amounts of cocoa and sugar solids than cocoa butter, the mixture will be thicker and not flow as well than if there was a higher amount of cocoa butter relative to the non-cocoa butter particles.
The particle size distribution (the range of sizes), particle shape and surface, which is impacted by the stress intensity applied during the refining process. The goal of grinding chocolate is not to have all the chocolate particles 100% uniform, but to have a range of sizes (but with the largest size still small enough to not be detected by the tongue as gritty).
Amount and type of surface active compounds added during manufacturing (Arnold et al., 2014; Schantz & Rohm, 2005). Lecithin (a surfactant) is an example of such a compound used in chocolate.
Figure 1. Schematic illustration of the conditions applied during roller refining experiments.
For larger manufacturers, a 3, 4, or 5 roller refining is commonly used. This is where individual particles are forced to pass through a series of gaps between the rollers, where the gaps decrease in width from one roller to the next (Ziegler & Hogg, 2017). However, agitated ball mills are also an important tool used to refine chocolate. Here, particles randomly arrive at the stressing sites between the grinding balls (Alamprese, Datei, & Semeraro, 2007). In both systems (roller refining and ball mill), the intensity of the applied stresses can be controlled and adjusted. In roller refining, adjustments can be made to the rotational speed of the rollers and the size of the gaps between them. In ball mills, the ball size, agitation velocity, and milling time can be adjusted.
The aim of this study was to vary the conditions (applied stresses) in both the roller refiner and ball mill in order to analyze the particle morphology (shape, size) and how it relates to the flow properties of the chocolate.
Materials and Methods
Roller refining and conching
The mixture for the ball mill was 70% cocoa components, while the mixture for the roller refiner was 68.1% cocoa components. No lecithin was added.
Roller refining was conducted using an 80S three roll refiner (Exakt Advanced Technologies GmbH, Norderstedt, Germany). Melted cocoa mass was mixed with sugar and then refined. The gaps between 1st and 2nd roller were set to 145, 40, or 15 um. The gaps between the 2nd and 3rd roller were set to 70, 15, and 5 um. This was conducted at three different speeds (200, 400, or 600 rpm) (Fig. 1). This gave a total of 9 samples for analyses.
Conching was conducted using a UMC5 mixer (Stephan Food Service Equipment GmbH, Hameln, Germany) at 50 degrees Celsius for 10 minutes of dry conching before the cocoa butter was added, and then agitated for another 15 minutes.
Ball Milling
A MR6 agitated ball mill was filled to 80% by 4 and 5 mm diameter chromium steel balls. The temperature of the mill was kept at 45 degrees Celsius. To obtain a similar particle size to the samples that had gone through the roller refiner, the masses were pumped for either 20, 60, or 90 minutes.
Particle Characterisation
The refined material was diluted 2:1 in sunflower oil and measured with a 395 micrometer (Mitutoyo Corporation, Kawasaki, Japan). The particle size obtained was interpreted as X80,3 diameter.
After conching, particle size distribution (PSD) was analyzed using Helos KR laser diffraction spectrometer (Sympatec GmbH, Clausthal-Zellerfeld, Germany). The diffraction signal was converted into particle size distribution using the Fraunhofer model, and the volume density distribution based diameters x10,3, x50,3, and x90,3 were calculated.
Chocolate viscosity
Viscosity was measured using a HAAKE MARS 50 rheometer (Thermo Fischer Scientific Inc. Karlsruhe, Germany) with the chocolate at 40 degrees Celsius, with varying shear rates.
Results and Discussion
Effects of grinding intensity on particle characteristics
Figure 2. Particle size distributions of chocolate mass subjected to roller refining (A: 200 rpm at 3rd roller; B: 400 rpm at 3rd roller; C: 600 rpm at 3rd roller) and ball milling (D). Full lines, A-C: 2nd gap adjusted to 70 um (roller refiner); D: 20 min. residence time (ball mill). Broken lines, A-C: 2nd gap adjusted to 15 um (roller refiner)’ D: 60 min. residence time (ball mill). Dotted lines, A-C: 2nd gap adjusted to 5 um (roller refiner); D: 90 min. residence time (ball mill).
Table 2 summarizes the particle diameter and PSD (particle size distribution) spans. Than span refers to the range of particle size. The closer the SPD span number is to zero, the more uniform the particles are in size/shape relative to each other. A smaller gap between the rollers resulted in a smaller particle size, but a higher rotational speed of the rollers also contributed to a smaller particle size. The effect of the rotational speed was more pronounced at larger gaps (70 & 15 um), and it diminished when the gap width was set to 5 um.
Figure 2 shows the density distribution, and how particle size decreased with each successive roller passage. The high number of large particles (approx. 100 um) after the first passage can be attributed to large sugar particles. Higher shear by way of a higher (faster) rotational speed leads to a narrow particle size distribution, also evident by the span The span refers to a value which reflects how wide or narrow the range is between the largest and smallest particles. That is, if the span is closer to zero, then the granularity of the particles is narrow and more uniform.
After ball milling, the particle size distribution span diminished with time (see Table 2). The span value was lower for the samples that were ball milled than for the roller refined (Table 2, bottom of second column), which is characteristic of this communition (AKA grinding) technique.
Figure 3. Circularity distributions of particles in chocolate mass subjected to roller refining (A: 200 rpm at 3rd roller; B 400 rpm at 3rd roller; C: 600 rpm at 3rd roller) and ball milling (D). Full lines, A-C: 2nd gap adjusted to 70 um (roller refiner); D: 20 min. residence time (ball mill). Broken lines, A-C: 2nd gap adjusted to 15 um (roller refiner); D: 90 min. residence time (ball mill).
The density distribution for particle circularity (how close in shape they are to a sphere) as affected by grinding intensity are shown in Figure 3. In the roller refiner, with each subsequent passage at a particular roller speed, circularity slightly increased as oblong particles. The effect of gap width diminishes with increasing rotational speed. Circularity is higher for particles in the ball mill samples, along with the fact that particle size distributions are narrower. The effects of rotational speed and gap adjustment are also evident from the mean circularity values (Table 2).
Figure 4 depicts the morphology of the particles for the roller refiner (different gap widths at 400 rmp) and ball mill (different milling times of 30, 60, and 90 minutes). The smaller gap width and longer time in the ball mill resulted in a reduction in the number of coarse particles. For the roller-refined chocolate, the particle sizes were about 8.43 (70 um gap), 7.78 um (15 um gap), or 7.0 um (5 um gap). For the ball-milled chocolate, the particle sizes were about 7.9 um (20 minutes), 6.99 um (60 minutes), and 6.23 um (90 minutes).
Figure 4. Scanning electron microscopy images of chocolate particles after rolling refining at 400 rpm (top row; from left to right: 2nd gap adjusted to 70, 15, 5, um; scale bar length: 200 um) or after at ball milling (middle row; from left to right: residence time was 20, 60, or 90 min.; scale bar length: 200 um). The third row gives details on particle shape after roller refining at 400 rpm and 15 um gap at different resolutions (scale bar length, see pictures).
Sugar particles in the image appear with smooth surfaces with sharp edges. The cocoa particles have a rough irregular surface with honey-comb shaped cavities with diameters of a few micrometers (see Figure 4, bottom row). Another important aspect to take from Figure 4 is comparing the shape of the particles from roller refined (top row) to ball-milled particles (middle row). The shape of the particles from the roller-refiner are more oblong shape. This difference in shape leads to a higher specific surface and impacts the rheology and texture of the final product (Tan & Balasubramanian, 2017).
Chocolate viscosity
Viscosity measurements were taken after all samples were milled in their respective equipment, and conched. Therefore, all samples had a similar composition in regards to solids and fat content (see Table 1). Therefore, any differences can be related to particle size and shape, as well as particle packing density. If the chocolate is
Figure 5. Square-root transformed flow curves of chocolate subjected to roller refining at 200 rpm at different width of the 2nd gap (black, 70 um; grey 15 um, white, 5um). The Casson fitting functions are indicated by dotted lines.
Figure 5 plots the samples that were roller-refined at 200 rpm (3rd roller) with different gaps (black at 70 um, grey at 15 um, white at 5 um). Just as with other samples the Casson Yield Stress increased with refining intensity (whether it be smaller gaps in the roll refiners or longer time in the ball mill). The Casson Yield Stress essentially measures the viscosity by informing how much stress is needed to start flow (since chocolate is a non-Newtonian fluid and requires stress/shear to get it to flow/move). Why? The increase in the Yield Stress (a reflection of how thick or viscous the chocolate becomes) can be attributed to the fact that with less coarse particles, packing of the small particles increases and reduces flow (Bolenz & Manske, 2013). Interactions between particles within the chocolate is greater when particle size is low and particle number is high (Saputro et al. 2017). Yield stress is important when it comes to chocolate moulding or coating, so adjustments to equipment and processing is possible to improve this. Although the yield stress was within the same magnitude for the roller refined samples, the viscosity of chocolate that was ball-milled was higher. This can be explained by a possible less effective moisture removal during refining. This results in the formation of aggregates that immobilise fat (Sautro et al., 2017), but also the difference in how much fat was released in relation to cocoa solids (which is usually greater when particles are coarser) (Afoakwa, Paterson, Fowler, & Vieira, 2009). The more free fat available (the continuous phase) results in a lower visocisty (more fluid chocolate).
Conclusion
The work here demonstrates that during refining, both the intensity of the stresses and grinding technology (equipment type), significantly impact particle size distribution and surface properties (shape) of the sugar and cocoa particles. By targeting the various processes a manufacturer can better influence the rheological (flow) properties of the chocolate.
