One thing that almost all formulated products have in common is that they are non-Newtonian. Paints, inks, pharmaceuticals, lotions, creams, dressings, and many more do not have a single fixed viscosity. Instead, their viscosity changes as a function of shear rate. In practical terms, this means that the way a product flows depends on how it is being used. A formulation at rest on a shelf behaves very differently from that same formulation being pumped, sprayed, injected, or rubbed.

Because of this, viscosity is not a single number that can be optimized in isolation. It is a function of shear rate, typically represented as a flow curve of viscosity versus shear rate. That curve ultimately determines whether a product performs as intended. If it is wrong, the product will fail in use, regardless of how well it performs under a single test condition.

Viscosity is More Than One Number

Most formulated systems are shear-thinning, meaning their viscosity decreases as shear rate increases. This is what allows a paint to spread smoothly under a brush or a lotion to glide across skin. Some systems exhibit shear-thickening behavior, where viscosity increases under stress, as seen in materials like cornstarch suspensions. In both cases, the key point is the same: the act of using a product changes its viscosity. Rubbing, pumping, spraying, and injecting are not just external actions applied to a formulation, they actively alter its internal flow behavior.

The Flow Curve and Real-World Performance

Different regions of the flow curve correspond to different stages of a product's lifecycle. At low shear rates, the formulation is effectively at rest, and properties like sedimentation, sagging, and long-term stability dominate. As shear rate increases, behavior relevant to processing begins to matter, such as pouring and pumping. At higher shear rates, application-driven properties take over, including spreadability, sprayability, injectability, and lubrication.

A formulation that performs well in one region but poorly in another will fail in practice. The problem is not matching a single viscosity target, but ensuring that the entire curve aligns with how the product is actually used. In formulation design, the objective is not to optimize viscosity at a point, but to engineer the full flow curve.

Why Designing Rheology Is Difficult

Achieving this is difficult because rheology is impacted by phase behavior, micro/mesostructure and interactions amongst formulation ingredients. Surfactants, polymers, proteins, etc. can build up high viscosities based on their phase behavior, microstructure under various formulation conditions (pH, temperature, salt etc.) either independently or with interactions with formulation ingredients. These interactions are nonlinear and highly context dependent. An ingredient that thickens a system in one formulation may have a completely different effect on another.

As a result, it is very complex to predict rheology from first principles or intuition alone, even by experienced formulators for complex multi-component formulations. The standard workflow remains iterative: propose a formulation, measure its behavior, adjust, and repeat. This trial-and-error process is slow, expensive, and fundamentally limiting.

Even though rheology is central to product performance, it is often optimized inefficiently. Teams may converge on acceptable behavior after enough iterations, but exploring the full space of possible formulations is rarely feasible. At the same time, the shape of the flow curve itself can be a source of competitive advantage. Two products with similar compositions can behave very differently in use, and that difference is often encoded in their rheology.

Predicting the Flow Curve

If formulators could anticipate the rheological behavior of a formulation before running an experiment, the development process would change fundamentally. Instead of testing blindly, they could prioritize candidates that are likely to match their desired flow profile and avoid those that are not. This would reduce the number of experiments required and accelerate convergence toward a final product.

This is the problem FastFormulator was built to address. Rather than treating rheology as an isolated property, we approached it as a function of formulation composition and underlying chemistry. We built up rheological understanding progressively, starting from simpler systems and extending to complex, multi-component formulations across different industries. This includes polymers, surfactants, emulsions, and complete products, with data designed to capture not only individual behavior but also interactions between ingredients.

Using this foundation, we developed chemistry-aware machine learning models trained on a large, proprietary dataset of real formulations. These models are not general-purpose language models, but systems designed specifically to learn relationships between chemical structure, formulation composition, and physical properties. From this, we built what can be thought of as digital rheometers. Given a formulation, the model can predict its viscosity as a function of shear rate, producing a full flow curve before any physical experiment is run.

In practice, this allows formulators to evaluate and compare candidate formulations based on their expected rheological behavior. Instead of discovering mismatches late in the process, they can identify them upfront. Within our platform, ChemNexus, users can also define a target flow curve and search the formulation space for candidates that are likely to match it.

This shifts formulation development from an iterative, trial-and-error process to a more directed and efficient search problem.

Examples Across Applications

Polymer Chemistry Variation

Polymer based rheology modifiers play a critical role in building up the viscosity and viscosity vs shear rate behavior. This is dictated by their chemical structure, molecular weight, and MW polydispersity. The following highlights the ability of the ChemNexus Platform to adequately predict viscosity vs shear rate (comparison shown with experimental data) across a range of polymer and biopolymer rheology modifier chemistries:

Industrial Multi-Component Formulations

Although highly useful to be able to adequately predict the viscosity vs flow curve behavior in polymeric and biopolymer rheology modifiers, by itself it's not particularly useful for formulators. Instead, what is required is the effective optimization of the viscosity vs flow rate behavior in complex multi-component industrial formulations. The rheology behavior of rheology modifiers themselves can be drastically impacted in formulation as other formulation components can work synergistically or antagonistically to impact overall viscosity and flow curve behavior and certain components themselves can dominate the behavior as well based on microstructure and interactions.

In food systems such as mayonnaise, the flow curve determines pourability, stability, and texture. In personal care products like lotions and cleansers, it governs spreadability and sensory experience. In industrial systems such as drilling fluids, performance depends on maintaining specific viscosity profiles under varying conditions. In each case, the underlying challenge is the same: designing a formulation whose rheological behavior matches its intended use across the full range of shear conditions.

The following examples highlight the ability of the ChemNexus Platform to accurately predict the viscosity flow curve behavior (comparison with experimental data shown) across a range of different industrial formulations—foods, cosmetics and oil drilling fluids:

Conclusion

Rheology is not a secondary property that can be tuned at the end of development. It is central to how products function, and it must be considered as a design variable from the beginning. Because that behavior depends on microstructure evolution and complex, multi-component interactions, it cannot be efficiently optimized through intuition and trial and error alone.

The ability to predict and design for the full flow curve represents a fundamental shift in how formulations are developed, moving from reactive experimentation toward informed, data-driven design. In formulation design, the goal is not to match a single viscosity value, but to engineer how a product flows under every condition it will encounter.

References

• Larson, R.G. The Structure and Rheology of Complex Fluids. Oxford University Press.
• Davies, A.; Amin, S. Rheology of Cosmetic Products: Surfactant Mesophases, Foams and Emulsions. Journal of Cosmetic Science, 71, 481–496 (2020).
• Gentile, L; Amin, S. Rheology Primer for Nanoparticle Scientists.
• Macosko, C. Rheology: Principles, Measurements, and Applications.
• Barnes, H. A Handbook of Elementary Rheology.