Can we predict ganache texture from formulation?

A ganache may contain only chocolate, cream, glucose, invert sugar and butter. Ingredient ratios give a rough guide, since more chocolate often makes a ganache firmer and more cream often makes it softer. They do not describe texture.

One formulation can be firm at 16 °C and soft at 24 °C. Its structure also changes during the hours after mixing. Cocoa butter and milk fat melt differently, cocoa particles add structure, and glucose syrup changes the water phase.

The aim is to estimate ganache firmness from formulation, temperature and crystallization history before making a batch.

The result is a Ganache Firmness Index, or GFI. The index is a testable estimate based on formulation and process. It does not replace a texture analyzer, and it does not represent a universal sensory score.

What the model estimates

Texture includes several properties. A ganache can be firm and brittle, or soft and elastic. It can also measure as hard while having a weak internal structure. One number cannot describe all these properties, so GFI has a narrow definition.

$$GFI = \text{ordinal estimate of ganache firmness above 0 °C}$$

The index runs from 0 to 100. It is intended to rank practical states from very soft toward increasingly firm and cuttable structures.

Ordinal means that GFI ranks states from softer to firmer. A GFI of 70 is not 70 N, 70 Pa or any other instrument unit. The calculation estimates an internal structural state and maps it to a scale from 0 to 100.

Step 1: start with composition

The simplest model is a mass balance.

The mass balance separates each ingredient into water, cocoa butter, milk fat and cocoa solids. It also tracks soluble sugars, other solids, gelatin and liquid fats when present.

Component composition gives more information than a chocolate to cream ratio. Two chocolates with the same cocoa percentage can contain different amounts of fat, sugar and cocoa solids. White chocolate contains cocoa butter without cocoa particles, while cream contains water and milk fat.

For an ingredient $i$ with mass $m_i$ and component fraction $x_{ij}$, component $j$ is simply:

$$m_j=\sum_i m_i x_{ij}$$

The mass balance describes composition only.

Composition from mass balance

Change chocolate or cream to update component fractions. Temperature is not part of this calculation.

Guanaja 70%500 gCream 35%360 g

Cocoa butter20.8%

Cocoa solids14.2%

Water21.7%

Composition at 10°C20.8% cocoa butterComposition stays fixed

Composition at 25°C20.8% cocoa butterComposition stays fixed

LimitPhysical state is still unknown. Mass balance cannot say how much cocoa butter is solid.

The module shows the first limit. Cocoa butter content is identical at 10 °C and 25 °C, although texture changes.

A composition model tells us how much fat exists. We now need to know how much of that fat is solid.

Step 2: replace total fat with solid fat

McGill and Hartel studied ganache structure directly and found that crystalline fat content was strongly related to firmness and resistance to deformation. Increasing insoluble cocoa material also increased hardness, while soluble sugar and milk solids shifted the system in the opposite direction (McGill and Hartel, 2018).

The first physical variable is solid fat fraction, rather than total fat.

For cocoa butter:

$$F_{CB,solid}(T)=F_{CB}\times SFC_{CB}(T)$$

where $SFC$ is solid fat content at temperature $T$.

Milk fat is treated separately:

$$F_{MF,solid}(T)=F_{MF}\times SFC_{MF}(T)$$

The first estimate of solid fat is:

$$F_{solid}(T)=F_{CB,solid}(T)+F_{MF,solid}(T)$$

The calculation now responds to temperature. Much of cocoa butter can be solid at low temperature, but its solid fraction falls near its melting range. Composition stays constant while physical state changes.

The calculation uses representative SFC curves, not product specific DSC measurements. Botanical origin, tempering, crystal form and milk fat composition can shift these curves. SFC is therefore an estimate with uncertainty.

Step 3: solid fat is not immediately a network

Time also changes structure because crystallization is gradual. A ganache at 17 °C after five minutes does not have the same structure after 24 hours, even when formulation and evaluation temperature stay constant.

The primary cocoa butter crystallization term uses the Avrami equation.

$$X_c(t)=1-\exp(-kt^n)$$

where:

• $X_c$ is crystallized fraction;
• $k$ is a rate parameter;
• $n$ describes the shape of the crystallization process;
• $t$ is time.

Castro-Alayo and coauthors published cocoa butter crystallization data from 15 to 19 °C (Castro-Alayo et al., 2022). The calculation reconstructs $k$ from the reported Avrami exponent and half time, then interpolates within that temperature range.

Results outside 15 to 19 °C are marked as extrapolated. A calculation at 4 °C is possible, but the source did not measure a parameter at 4 °C.

A second empirical term represents slower organization over several hours. It rises toward a fixed limit at a rate that depends on temperature. The term is part of calibration and does not come from the Avrami study.

The implementation combines both terms as:

$$M_{fat}=X_c(t,T)\times A(t,T)$$

where $M_{fat}$ is fat network maturity and $A$ is the empirical aging term. The two factors keep published cocoa butter kinetics separate from unvalidated secondary aging.

Effects of temperature and time

Hold maturation fixed. Change only evaluation temperature.

Evaluation temperatureMaturation time

Evaluation temperature20°C

Fixed: 24 h maturation at 17°C

GFI71

Solid fat11.9%

Crystal maturity99.8%

Connectivity proxy0.039

Interpretationcrystal maturity. Temperature changes solid fat immediately. Time changes estimated crystal maturity. crystal maturity.

Change evaluation temperature while maturation stays fixed. Then change maturation time while evaluation temperature stays fixed. Solid fat responds to temperature, while crystal maturity responds to time.

Step 4: a solid phase still needs connectivity

Even solid fat does not automatically produce the same mechanical structure.

Ganache is a concentrated mixture of fat, water, sugar and particles. During cooling, its hot oil in water emulsion develops a partly continuous fat structure (Saglio et al., 2018). Solid fat amount and mechanical structure are therefore different variables.

$$\text{solid fat amount}\neq\text{effective mechanical network}$$

The calculation uses a fat connectivity proxy.

$$C_{fat}=f(F_{solid},M_{fat},\phi_{fat},\text{interfacial state})$$

The proxy is not a measured property. It multiplies solid fat fraction, crystal maturity and total fat volume fraction. A cocoa particle term then reduces the result, because particles can limit contact between fat droplets. The proxy prevents every gram of solid fat from receiving the same mechanical effect.

Step 5: separate the effects of cocoa particles

A simple model could use one rule.

$$\text{more cocoa solids}\rightarrow\text{more firmness}$$

The direction is partly supported by direct ganache data. McGill and Hartel found that increasing insoluble cocoa material relative to soluble components increased hardness and resistance to deformation.

Merachli and coauthors found that cocoa fibres can change the water phase and droplet interfaces. They can limit partial coalescence of fat droplets and form a crowded particle structure at high concentration (Merachli et al., 2021). Cocoa particles can therefore have two effects.

1. Particle crowding can strengthen the matrix.
2. Changes at droplet interfaces can reduce fat connectivity.

The calculation keeps particle crowding separate from fat connectivity. The crowding term rises faster as particle volume approaches a maximum packing fraction.

$$J(\phi)=\left(1-\frac{\phi}{\phi_{max}}\right)^{-2}-1$$

The equation is not a measured ganache law. It represents nonlinear crowding as particles occupy more volume.

Step 6: account for water and dissolved solids

A ganache contains water and dissolved sucrose, glucose syrup, invert sugar, dextrose, lactose and sorbitol.

Aqueous water mass fraction provides a first softening term.

$$W_a=\frac{m_{aqueous\ water}}{m_{total}}$$

The term counts water carried by ingredients after subtracting their dry matter. It does not represent thermodynamic “free water” or water activity. More aqueous water dilutes particles and dissolved solids, and it lowers the fraction of structural fat.

Water fraction alone does not describe the liquid phase. A sucrose solution at 40% solids has a different viscosity from one at 60%. Glucose, fructose and sucrose also have different viscosities at equal concentration.

Telis and coauthors measured carbohydrate solution viscosity across several concentrations and temperatures (Telis et al., 2007). The calculation interpolates the logarithm of viscosity because viscosity rises nonlinearly with concentration.

At 20 °C, sucrose solution viscosity rises from about 2 mPa s near 20% solids to tens of mPa s near 60% solids.

The calculation is:

$$\eta_{aq}=f(T,C,\text{sugar profile})$$

where $C$ is dissolved solids concentration.

The calculation reports aqueous viscosity but does not include it in GFI. Published work supports its role in emulsion behaviour and molecular movement, but it does not provide a direct relation between solution viscosity and ganache firmness across formulations.

Step 7: account for glucose syrup DE

“Glucose” in a pastry formula often means glucose syrup. Glucose syrups differ in molecular composition.

Dextrose equivalent, or DE, can estimate average molecular size. Rong, Sillick and Gregson describe the relation between DE and average molecular weight (Rong et al., 2009).

The relation used is:

$$M_n\approx\frac{18000}{DE}$$

as a molecular weight estimate for colligative calculations.

Below 0 °C, the number of dissolved molecules affects freezing point depression and amount of unfrozen water. DE60 and DE40 syrup therefore cannot be treated as equal at the same dry mass. Above 0 °C, no direct DE multiplier is added to GFI because available data do not support one.

Step 8: include the gelatin network in Namelaka

A dark ganache receives most of its modeled structure from fat and particles. Namelaka also contains gelatin, which forms a separate gel that melts when heated.

The gelatin term has two parts.

$$G=P(c_{gelatin},T,C_{sugar})\times M(t,T_{maturation})$$

$P$ estimates gel potential from gelatin concentration, temperature and sugar concentration. $M$ estimates maturation from 0 to 1 using time and maturation temperature.

Published gelatin studies used higher concentrations than typical Namelaka recipes. The gelatin term is therefore extrapolated and has lower confidence. Its purpose is to avoid assigning all cold structure in Namelaka to cocoa butter.

Step 9: firmness and structural quality are not the same variable

Neuwirth and coauthors measured ganache texture and rheology after different processing conditions. Short emulsification produced a higher firmness reading, but it also reduced measures of viscoelastic structure (Neuwirth et al., 2024).

Treating measured firmness as structural quality would give this relation.

$$\text{more measured firmness}=\text{better structural network}$$

it would reward the wrong mechanism.

GFI therefore remains separate from process quality. Emulsification time can lower confidence or produce a warning, but it does not multiply firmness. GFI estimates firmness only, not overall ganache quality.

Step 10: combine the mechanical contributions

The final calculation uses five state variables.

• fat connectivity, $C_{fat}$;
• cocoa particle crowding, $J_p$;
• gelatin network, $G$;
• aqueous water mass fraction, $W_a$;
• liquid fat fraction, $F_l$.

Above 0 °C, they form a combined value $L$.

$$L= 7C_{fat} +0.8J_p +0.45G -W_a -0.8F_l$$

The five coefficients are calibration values, not constants from published papers. They combine effects that have supported directions.

• A stronger crystalline fat network should increase GFI.
• Liquid fat and aqueous water should decrease GFI.
• Cocoa particle crowding and gelatin should increase GFI.

The weights were set by hand. They cannot establish effect size without texture measurements from matched formulations.

Step 11: turn latent structure into GFI

The combined value $L$ has no physical unit. A logistic function maps it to a scale from 0 to 100.

$$GFI=\frac{100}{1+\exp[-5(L-0.35)]}$$

The logistic function places lower and upper limits on the index. It does not add evidence or turn GFI into a physical measurement.

The calculation follows this order.

$$\text{composition} \rightarrow \text{physical state} \rightarrow \text{network state} \rightarrow L \rightarrow GFI$$

Future calibration may relate $L$ to force, yield stress or another measured property. GFI can remain a separate ordinal display.

Step 12: compare controlled changes

GFI is more suitable for comparing a controlled formulation change than for claiming that an absolute score is correct. For example, replacing some cream with cocoa butter at constant mass tests whether GFI moves in the expected direction.

Replace cream with cocoa butter

Total recipe mass stays constant. Each gram of cocoa butter replaces one gram of cream.

Reset

Cream replaced60 g

300 g cream + 60 g pure cocoa butter

Reference360 g cream1000 g total

→

Modified300 g cream + 60 g cocoa butter1000 g total

Solid fat+2.5 pt

Water fraction-3.5 pt

GFI+14

less cream water+more cocoa butter→changed solid fat network

InterpretationDirection has stronger support than exact GFI difference. Exact values require matched texture measurements.

Tests used thousands of random formulations above 0 °C to check directions such as:

$$\text{additional cocoa butter}\Rightarrow GFI\uparrow$$

and:

$$\text{cocoa butter replaced by liquid fat}\Rightarrow GFI\downarrow$$

under the tested constraints.

The tests check calculation behaviour. They do not show how large each effect will be in a real ganache. Experimental calibration is required.

Limits of GFI

GFI does not estimate force, $G'$, $G''$, yield stress or fracture behaviour. It also does not estimate sensory melt rate, droplet size or cocoa butter crystal forms. Product specific SFC curves, water activity, shelf life and whipped foam structure remain outside its scope.

Some properties need separate outputs rather than one firmness index. For example, the calculation reports aqueous viscosity but does not include it in GFI because matched ganache data do not support that relation.

Data needed for calibration

Calibration requires formulations with measured composition, temperature, process and texture.

$$(\text{formulation},T,t,\text{process}) \rightarrow (\text{firmness},G',G'',\tau_y,\eta)$$

Matched measurements would allow the calibration values to be fitted instead of set by hand. Published physical equations should remain separate from fitted values, so calibration does not change the meaning of each variable.

Conclusion

Ingredient ratio is a useful kitchen rule, but it cannot describe changes caused by temperature, time or process. GFI adds solid fat, crystal maturity, fat connectivity, cocoa particles, aqueous water and gelatin in sequence.

GFI remains an ordinal estimate. Its equations include published data, derived terms and values set by hand. Better ganache measurements can replace those fitted values without changing the physical variables used in the calculation.

References

• McGill, J., and Hartel, R. W. Characterizing the structural properties of ganache. Journal of Food Science, 2018.
• Saglio, A. et al. Ganache structure and phase behaviour during cooling. International Journal of Gastronomy and Food Science, 2018.
• Merachli, F. et al. Role of cocoa fibres in ganache-like concentrated emulsions. LWT, 2021.
• Castro-Alayo, E. M. et al. Cocoa-butter crystallization kinetics and Avrami modelling. Foods, 2022.
• Telis, V. R. N. et al. Viscosity of aqueous carbohydrate solutions at different temperatures and concentrations. International Journal of Food Properties, 2007.
• Rong, Y., Sillick, M., and Gregson, C. M. Determination of dextrose equivalent and molecular-weight relationships in glucose syrups. Journal of Food Science, 2009.
• Neuwirth, C. et al. Effect of formulation and processing on ganache texture, rheology and thermal properties. Journal of Food Engineering, 2024.
• Valrhona, professional recipe collection.
• Cacao Barry, professional recipe collection.
• Linnolat, professional products and inspiration recipes.