The Viscoelastic Fingerprint of Emulsifier Blends: A Texture Prediction Method Using SSL/CSL/GMS Ternary Networks as a Model

The texture of baked goods-softness, elasticity, chewiness-is a core determinant of consumer acceptance, but traditionally it can only be assessed through post-baking sensory evaluation or texture analyzer puncture tests, resulting in long feedback cycles and high costs. This article proposes a texture prediction method based on the "viscoelastic fingerprint" of emulsifier blends: using SSL, CSL, and GMS ternary networks as a model system, rheological parameters (storage modulus G', loss modulus G'', loss tangent tanδ, yield stress) of the blended emulsifier are measured in model interfaces and model dough, establishing a quantitative mapping relationship between the "viscoelastic fingerprint" and the final texture of baked goods. The core value of this method lies in advancing texture prediction from "post-baking testing" to the "formulation design stage," providing baking enterprises with a rapid, quantifiable, and iterative texture regulation tool.

In baking product development, formulators face a structural dilemma: texture can only be accurately evaluated after baking is complete.

This means:

• Each formulation adjustment requires a complete cycle of "ingredient mixing → mixing → fermentation → baking → cooling → testing," taking at least 2-4 hours.
• If texture is unsatisfactory, the formulation must be readjusted and the complete cycle repeated.
• Optimizing a formulation may require dozens of iterations, taking days or even weeks.

This "time lag" severely constrains formulation development efficiency. What formulators need is a method that can predict final texture before baking.

Emulsifier blends are one of the core factors determining baked goods texture. If "fingerprint information" related to final texture can be extracted from the rheological characteristics of emulsifier blends, early texture prediction becomes possible.

1. The Biological Significance of Viscoelasticity

The texture of baked goods is essentially the mechanical response of their internal network structure under external forces. This response exhibits typical viscoelastic characteristics:

• Elastic component: Derived from gluten networks, starch granules, and emulsifier crystalline networks. The elastic component determines the product's "springiness" and "support."
• Viscous component: Derived from the aqueous phase, lipids, and interfacial films. The viscous component determines the product's "softness" and "lubricity."

Texture quality is essentially a balance between elastic and viscous components. This balance can be quantitatively described using rheological parameters.

2. From "Single Parameter" to "Fingerprint Spectrum"

Traditional rheological analysis typically focuses on a single parameter (such as G' or tanδ). But texture is a multi-dimensional perception that a single parameter cannot fully describe.

The concept of "viscoelastic fingerprint" is: constructing a multi-dimensional spectrum reflecting structural characteristics through the combination of multiple rheological parameters (G', G'', tanδ, yield stress, creep recovery rate). This spectrum, like a "fingerprint," can uniquely identify the structural state of a given emulsifier blend.

1. Rheological Contributions of Each Component

2. "Viscoelastic Fingerprint" Characteristics of Ternary Blends

At different blend ratios, the SSL/CSL/GMS ternary system exhibits different rheological fingerprints:

Key finding: There is a strong correlation between tanδ (loss tangent) and sensory softness. Higher tanδ indicates a greater proportion of viscous components, making the product softer; lower tanδ indicates a greater proportion of elastic components, making the product firmer and more springy.

1. Construction of the Mapping Model

Through regression analysis of extensive formulation data, the following mapping relationships can be established:

Softness (sensory score) ≈ f(tanδ, G')

• When tanδ is in the 0.20-0.25 range, softness scores are highest.
• When G' is in the 5000-7000 Pa range, the balance between support and softness is optimal.

Springiness (sensory score) ≈ f(G', yield stress)

• Higher G' and greater yield stress correspond to stronger springiness.
• However, when yield stress is too high (>300 Pa), springiness actually decreases (brittleness increases).

Chewiness (sensory score) ≈ f(tanδ, G'')

• When tanδ is in the 0.25-0.30 range, chewiness is most comfortable.
• When G'' is too high, chewiness tends toward "gluey"; when G'' is too low, chewiness tends toward "dry and hard."

2. Verification of Prediction Accuracy

In a soft toast system, using ternary blended emulsifiers as variables, the results of "viscoelastic fingerprint texture prediction" were compared with "actual texture analyzer measurements":

Conclusion: The deviation between viscoelastic fingerprint-predicted texture parameters and measured values is within ±5%, demonstrating high prediction accuracy.

1. Sample Preparation

• Melt-blend SSL, CSL, and GMS at designed ratios (60-70°C), cool to prepare emulsifier premix.
• Add the premix to model dough (flour + water + salt + yeast) and mix uniformly.

2. Rheological Testing

• Use a controlled-stress rheometer (such as TA Instruments DHR series).
• Test modes: oscillatory frequency sweep (0.1-10 Hz) + creep recovery test.
• Record parameters: G', G'', tanδ, yield stress, creep recovery rate.

3. Fingerprint Extraction and Texture Prediction

• Input rheological parameters into the mapping model.
• Output predicted softness, springiness, and chewiness scores.
• Adjust blend ratios based on prediction results.

4. Verification and Iteration

• Bake the adjusted formulation and verify predictions using a texture analyzer.
• Feed measured data back into the mapping model to continuously optimize prediction accuracy.

1. Model applicability: This method has been well-validated in soft bread and toast systems; mapping relationships in hard bread, cake, and other systems may require recalibration.
2. Flour variability: Protein content and quality of different flours affect the rheological contribution of gluten networks; independent mapping models should be established for different flours.
3. Process variables: Mixing time, fermentation conditions, and baking temperature alter the final state of emulsifier networks; process conditions must be kept consistent during testing.
4. Cannot fully replace sensory evaluation: Viscoelastic fingerprints predict "instrumental texture," which still differs from "sensory texture." The method's value lies in rapid screening and trend judgment; final sensory evaluation remains irreplaceable.

The essence of the "viscoelastic fingerprint" method is to advance texture prediction in baked goods from "endpoint testing" to "process control." The SSL/CSL/GMS ternary network, as a model system, demonstrates the quantitative relationship between the rheological characteristics of emulsifier blends and final texture.

For formulation engineers, the value of this method lies in: being able to "see" the texture trajectory of the product before baking. It cannot replace sensory evaluation, but it can substantially reduce ineffective baking iterations, advancing formulation optimization from "trial and error" to "design."