Why molten chocolate flows as it does, what conching and tempering physically change, and where bloom comes from: the measurements behind the rules a maker is taught.
20 studies2002 to 2024Each checked against its published abstract
What these studies add up to
How chocolate flows is set by three things a maker controls: how much fat there is, how the particles are sized, and how much emulsifier is added. In one systematic study fat had the largest effect, then particle size, then lecithin, and each changed what the others did.
It is the spread of particle sizes that matters, not only the largest. Small particles filling the gaps between large ones free fat to flow, which is why two chocolates of the same fineness and fat content can flow differently, and why a roller-refined and a ball-milled chocolate differ.
Conching does two separate jobs. Physically, it shifts the point at which a dense powder-in-fat mixture stops being a jammed solid. Chemically, it makes no new key odorants; it changes the amounts, losing some volatiles and concentrating others.
Missing temper in either direction has a measured cost: over-tempered chocolate was harder, stickier and duller, and under-tempered chocolate bloomed.
Milk fat slows the crystal change associated with fat bloom, and warm, fluctuating storage speeds bloom up.
Conching trades aroma for flow. It lowered every odorant measured along with the viscosity, more so when hotter, and in one factory conche nothing measurable changed after twelve of twenty hours.
Cooling and release are tunable: in a pilot tunnel chocolate left the mould soonest at an intermediate cooling rate, and temperature, contact time and humidity each changed how firmly it stuck.
Aerating, printing and heat-proofing have each been measured in the laboratory: the gas chosen changes the bubbles and the taste, and ethylcellulose gave chocolate that held its shape at 40°C.
Conching made no new aroma compounds; it changed the amounts
Counet and colleagues, 2002. Journal of Agricultural and Food Chemistry. doi:10.1021/jf0114177
Question
Which odorants matter most in dark chocolate, and what does conching do to them?
How
The volatile fractions of dark chocolates were analysed by gas chromatography with a human assessor at the outlet, before and after conching.
Finding
Thirty-three potent odorants were found in the fraction studied; three aldehydes had a strong chocolate smell and several pyrazines carried cocoa, praline, nutty and coffee notes. No new key odorant was formed during conching, but some compounds, branched pyrazines among them, increased significantly, while most of the aldehydes were lost by evaporation.
Does not establish: That longer conching is better or worse. It removes some desirable volatiles along with the unwanted ones.
Limits: The abstract does not give the number of chocolates or the conching conditions.
Can the particle size distribution be arranged so that a lower-fat chocolate still flows and still melts in the mouth?
How
Model chocolates, dispersions of sugar in fat, were made with different particle size distributions and fat contents, and their viscosity, hardness and melting were measured.
Finding
Optimising the distribution so that particles pack more closely, while reducing fat to a critical 22% by weight, lowered the viscosity of the melt and the hardness of the solid. Samples with the optimised distribution also collapsed faster in a laboratory test standing in for melting in the mouth.
Does not establish: That a real chocolate at 22% fat would be acceptable to eat or to process.
Limits: Sugar-in-fat models with no cocoa solids or milk, and an in vitro stand-in for eating.
How does the structure of bubble-containing chocolate made with different gases relate to its sensory properties?
How
4 types of bubble-containing chocolate were produced using carbon dioxide, nitrogen, nitrous oxide and argon. Gas hold-up was determined from density, bubble size distribution from X-ray microtomography, and sensory properties by a nonexpert panel of 20 panelists using descriptive analysis.
Finding
The chocolates fell into 2 groups. Those made with carbon dioxide and nitrous oxide had distinctly higher gas hold-up and larger bubbles than those made with argon and nitrogen. Chocolates made with argon and nitrogen were perceived as harder, less aerated, slower to melt, creamier and higher in overall flavour intensity.
Does not establish: That aerated chocolate tastes more strongly of chocolate. The chocolates with larger bubbles and more gas were not the ones rated highest for overall flavour intensity.
Limits: A nonexpert panel of modest size and one chocolate base. The abstract gives no bubble dimensions.
How do particle size, fat content and lecithin each change the flow of molten dark chocolate?
How
Dark chocolates were made at four particle sizes (90% of particles finer than 18, 25, 35 or 50 micrometres), three fat contents (25, 30 and 35%) and two lecithin levels (0.3 and 0.5%), and measured on a rheometer.
Finding
All three factors significantly changed every flow property measured, and they interacted. Larger particles lowered plastic viscosity and yield value, most strongly at 25% fat and 0.3% lecithin and less as fat and lecithin rose. Fat accounted for the most variation, then particle size, then lecithin.
Does not establish: A target particle size or fat content. The study maps the trade-offs; coarser chocolate flows more easily and is also the one a tongue can feel.
Limits: One recipe family made in one laboratory. The ranking is across the ranges tested, which stop at 0.5% lecithin.
What do under-tempering and over-tempering each do to the texture and appearance of dark chocolate?
How
Dark chocolates at four particle sizes were brought to optimal temper, over-temper and under-temper, and their hardness, stickiness, colour, gloss and microstructure were compared.
Finding
Over-tempering significantly increased hardness and stickiness, reduced gloss and darkened the surface. Under-tempering induced fat bloom, with whitening of the surface and of the interior near it, and damaged texture, colour and gloss. Finer particles gave harder, darker chocolate at every temper.
Does not establish: The temperatures that give optimal temper for another chocolate. The study shows the consequences of missing it in either direction.
Limits: One recipe family, with temper states set in a laboratory.
How does replacing part of the cocoa butter with milk fat affect the surface of chocolate and the development of bloom?
How
Chocolates with and without milk fat were stored at a constant 26 °C or cycled between 26 and 29 °C, and their surfaces and fat crystals were followed over time.
Finding
Milk fat reduced large surface crystals and surface roughness, lowered the initial solid fat content, and slowed both the rate of whitening and the change from Form V to Form VI. Cycling the temperature accelerated crystal growth compared with holding it steady. Bloom began only at particular points on the surface, whatever the fat or the storage.
Does not establish: A shelf life. It shows direction and mechanism, not how long a bar lasts.
Limits: Storage a little below melting temperature, which is harsher than a cool cupboard.
How do processing conditions during moulding and cooling affect the ease of demoulding chocolate?
How
Laboratory study in which adhesion of chocolate to a polycarbonate mould surface was measured as the force required to separate a flat mould surface from the solidified chocolate sample, under varied temperature, contact time and relative humidity. No sample size is given.
Finding
Temperature, contact time and the relative humidity of the surrounding environment had a significant impact on chocolate crystallisation and solidification and on adhesion to the mould surface. The authors state that the findings confirmed observations made during commercial chocolate manufacturing.
Does not establish: Which settings to use. The abstract reports that the factors matter, without stating the conditions that gave the easiest release.
Limits: A flat test surface, not a shaped mould, and the abstract gives no figures or directions of effect.
Is there a better way to measure the yield stress of molten chocolate than fitting the Casson model?
How
Chocolates differing in fat content and in the amount and type of emulsifier were measured by oscillatory rheology, taking the stress at the end of the linear viscoelastic region as the yield stress, and compared with Casson fits.
Finding
The Casson model fitted the data well between shear rates of 5 and 60 per second but could not resolve real differences between samples below 5 per second. The oscillatory measure was sensitive to fat content and to emulsifier amount and type, and separated chocolates the Casson fit did not. The abstract notes that the industry body recommends against the Casson model, though it is still widely used.
Does not establish: That a Casson figure on a specification sheet is wrong. It shows what such a figure cannot tell apart.
Limits: A methods paper on a set of laboratory chocolates.
How do sugar, cocoa particles and lecithin affect the way cocoa butter crystallises, with and without seed crystals?
How
Model chocolate systems with and without each ingredient were pre-crystallised by seeding or left unseeded, and crystal growth was followed by confocal microscopy and calorimetry.
Finding
Every ingredient and both techniques had a large effect on how fast the fat crystallised and on the resulting structure. Seeded samples formed many nucleation sites and a network grew rapidly. Unseeded samples were uneven, with large spherical crystals in some regions. Lecithin shortened the time before crystallisation began in seeded samples and, in the absence of seeds, itself acted as the nucleation site.
Does not establish: How much seed to use, or that the differences seen survive in a finished chocolate.
Limits: Model systems under a microscope, not moulded bars.
How does the fat content during grinding affect particle sizes and flow, in roller refining and in ball milling?
How
Milk chocolates were made at pilot scale by both routes at different fat contents during grinding, and their particle size distributions and flow were measured.
Finding
Fat content during grinding changed the particle size distribution in both processes. Roller refining at higher fat gave two populations of particle size, in which small particles fill the gaps between large ones and free fat for flow. Ball-milled samples had narrower distributions lacking fine particles, and higher viscosity at medium and high shear rates, though sometimes lower viscosity at low shear.
Does not establish: That one route makes better chocolate. The comparison is of flow, and ball milling also removes less water and fewer volatiles, as the authors note.
Limits: Pilot scale, milk chocolate, one set of machines.
How do βVI-seeding and conventional pre-crystallisation, at different degrees of temper, affect the structure and storage stability of dark chocolate?
How
Dark chocolate was pre-crystallised by βVI-seeding or by conventional pre-crystallisation at various degrees of temper. Structure was characterised by traction tests and DSC melting curves; fat bloom was followed by DigiEye imaging and migration gravimetrically. The abstract gives no sample size or scale.
Finding
The two processes generated significantly different structures and storage stability. Well-tempered βVI-seeding gave a dense and homogeneous structure directly after solidification, which was optimal for retarding fat bloom and fat migration. Too high a structure density gave heterogeneous structures with reduced ability to withstand fat bloom. A lower structure density resisted moisture migration best.
Does not establish: That more seed or more temper is always better. Over-dense structures resisted bloom less well in this study.
Limits: No figures are given in the abstract, and storage conditions are not stated there. One dark chocolate.
Can chocolate that resists deformation above 40 °C be made by adding ethylcellulose dissolved in ethanol and evaporating the ethanol?
How
Laboratory study. A 20% ethylcellulose in ethanol solution was mixed with molten chocolate, the ethanol evaporated, and the chocolate incubated at 40 °C for 2 h and tested for hardness. Ethylcellulose viscosities of 4, 10, 20, 22 and 45 cP and concentrations from 1.0 to 2.2% were tested in different chocolates.
Finding
Milk chocolate containing 1.9% ethylcellulose had a hardness of 26.0 N at 40 °C, whereas the control was too soft to be tested. White and dark chocolates had hardnesses of 29.5 and 10.5 N. Hardness depended on formulation and ethylcellulose concentration and was independent of ethylcellulose viscosity.
Does not establish: That the product eats like ordinary chocolate or could be sold as chocolate. Neither eating quality nor legal composition was assessed.
Limits: Hardness after two hours at one temperature. Taste, texture in the mouth and residual ethanol are not reported in the abstract, and the authors say the mechanism was not determined.
Which fabrication settings allow a melt-extrusion 3D printer to build complex objects from chocolate?
How
Engineering study describing a melt extrusion 3D printer built from open source components and testing the ability of an extruded chocolate fibre to span distances without collapsing, while adjusting movement speeds, extrusion rates and cooling rates. Laboratory scale; no sample size is given.
Finding
Spanning distance was unaffected by movement speeds from 300 to 700 mm/min. The optimal extrusion rate was 10–20% leaner. Spanning distance improved when air was directed across the printing part, lowering the air temperature by approximately 3.5 °C.
Does not establish: That printed chocolate matches moulded chocolate in gloss, snap or shelf life. Only geometric performance was tested.
Limits: One printer design. The abstract does not state the chocolate used or whether the printed pieces were tempered, glossy or stable on storage.
Physically, what does conching do to turn a crumbly mixture into flowing chocolate?
How
The flow of a simplified chocolate formulation was studied through conching, as an example of mixing a powder into a liquid at the highest possible solids content.
Finding
Mechanical energy and the staged addition of surfactant together shifted the volume fraction at which the mixture jams, raising the maximum solid content that can still flow.
Does not establish: Anything about conching time or temperature for a real chocolate, or about the flavour changes conching also causes.
Limits: A simplified formulation, studied as physics. It is about flow, not flavour.
Can a minor lipid added to cocoa butter direct it into the right crystal form without the usual tempering procedure?
How
Various minor lipid components were added to refined cocoa butter and to melted commercial chocolate, which were then cooled rapidly to 20 °C without stirring, and structure, gloss and strength were measured.
Finding
Adding saturated phosphatidylcholine or phosphatidylethanolamine at 0.1% by weight accelerated crystallisation, stabilised the desirable Form V and gave chocolate with microstructure, gloss and mechanical strength comparable to a commercial tempered chocolate.
Does not establish: That this can replace tempering in production, or that a maker may lawfully do it.
Limits: Laboratory samples. The abstract does not report storage stability, flavour, or whether the additive is permitted in chocolate anywhere.
How does the local convective heat flux in an air cooling tunnel affect solidification, contraction and detachment of moulded chocolate from the mould wall?
How
Pilot-scale cooling tunnel trials on dark chocolate with cooling air temperatures of 2, 12 and 18 °C and mean air velocities of 1.5, 3.5 and 6.0 m/s. Heat transfer coefficients came from a CFD model, and crystallisation and detachment were followed by damping of ultrasound waves through the filled moulds.
Finding
Convective heat transfer from the bottom of the mould increased in the flow direction because of a recirculating zone of cooling air beneath the mould. Depending on the intensity and homogeneity of heat transfer, the structure density of the chocolate could be increased. For typical 125 g chocolate plate formats, detachment from the mould wall had an optimum at an apparent heat flux of 550 W/m², where time to detachment was at a minimum.
Does not establish: That colder or faster air is always better in a cooling tunnel. Detachment time passed through a minimum at an intermediate heat flux.
Limits: One mould geometry and well-tempered dark chocolate. Heat transfer values are modelled, not measured directly.
How do aroma-active volatiles and flow properties of the plastic mass change during conching of dark chocolate?
How
Different conched plastic masses of dark chocolate were characterised with a closed cavity rheometer, and five aroma-active volatiles were quantified in free cocoa butter by stable isotope dilution analysis and gas chromatography-mass spectrometry. The abstract does not state conche type, scale or batch number.
Finding
Free cocoa butter increased during conching and the complex viscosity of the plastic mass decreased with conching time. Concentrations of all five volatiles decreased with conching duration to varying degrees: acetic acid most, by about 60%, and linalool least, up to 26%. A lower polarity or boiling point was linked to a stronger decrease.
Does not establish: That conching removes only unwanted acids. The desirable odorants measured here fell as well.
Limits: Volatiles were measured in the free fat only. Five compounds were followed, including tetramethylpyrazine and 2-phenylethanol, which are usually counted as desirable.
In one factory conche, nothing measured changed after 12 hours of a 20-hour run
Valverde-Ayllon and colleagues, 2022. International Journal of Food Science and Technology. doi:10.1111/ijfs.16000
Question
Can the refining-conching time of dark chocolate in a Macintyre refiner/conche be reduced without changing the product?
How
Industrial case study in a Peruvian medium-scale chocolate company. Particle size (by micrometer and image analysis), plastic viscosity and yield stress were determined every 2 h during a total refining-conching time of 20 h. A trained panel compared chocolates made at 12 h and 20 h. The abstract does not state the number of batches.
Finding
After 12 h, no significant differences were found in particle size or flow properties. The trained panel found no significant differences between chocolate from 12 h and from 20 h. This amounted to a reduction of 8 h in refining-conching time.
Does not establish: That 12 hours is enough in any conche or melanger, or for other recipes. The result belongs to this machine and this chocolate.
Limits: One company, one machine type and one dark chocolate. Aroma volatiles were not measured, and the abstract's statistical wording for the sensory test is ambiguous.
How do conching temperature, shear direction and residue from previous batches on the conche wall affect odorants and viscosity during plastic conching of dark chocolate?
How
Time-resolved trials of plastic conching with varying conching temperature, shear direction, and with or without residue from previous trials (pre-charge) on the vessel wall. Six selected odorants were quantified in the fat and particle phases. The abstract does not state the scale or the number of runs.
Finding
Odorant concentrations decreased particularly at elevated conching temperature, by up to 78.0% in the fat phase. The highest concentrations of desired odorants were mostly found after conching without pre-charge. Odorants accumulated increasingly in the fat phase (up to 91.7%) with decreasing polarity. Temperature and absence of pre-charge had the largest effect on rheology, giving the lowest and highest complex viscosity respectively.
Does not establish: That longer or hotter conching makes better chocolate. The study found that retaining desired odorants and reducing viscosity were inversely related at elevated temperature.
Limits: Six odorants stand in for aroma as a whole, and no sensory test is reported in the abstract. Results are for the plastic phase of conching only.
What are the bubble features and hardness of aerated chocolate, with and without sugar, made by injecting nitrogen or carbon dioxide at different pressures?
How
Laboratory study. Sugar was replaced with inulin:maltodextrin mixtures at three ratios, and nitrogen and carbon dioxide were injected into these and a sugar-containing control at pressures of 3, 4.5 and 6 bar. Samples were imaged by X-ray computed tomography and segmented by image processing.
Finding
The bubble total volume (10440 ± 9206 mm3) and average diameter (1.30 ± 0.10 mm) of the sugar-containing control were larger than those of the sugar-free samples. Sugar-free aerated samples had lower hardness than the corresponding unaerated ones, while the reverse held for the control.
Does not establish: That aeration always softens chocolate. In the sugar-containing control the aerated sample was not softer than the unaerated one.
Limits: Much of the paper concerns the image segmentation method. The spread on the bubble volume is nearly as large as the mean.
What yield stress or viscosity a given job needs. The studies measure how the properties move, not the targets for moulding, enrobing or panning, which remain trade practice.
How far the common Casson figures on specification sheets can be trusted at low shear, where one study found the model cannot tell chocolates apart.
Whether additives that direct crystallisation could replace tempering outside the laboratory.
How any of this is perceived. Only a few studies connect flow or crystal structure to what a taster reports.
These are questions the studies read for this page leave unanswered. They are not a claim that nobody has answered them.
We would like to set two Google Analytics cookies to count page views, and to count them in our own cookie-free measurement as well. They do not identify you and nothing is shared with advertisers. Decline and the site works exactly the same — see cookies for what each one does.