SOLCALC helps formulators build a predictive digital twin of their drug formulation. Thus, decisions like polymer or solvent choice can be taken within weeks without extensive trial-an-error experiments. But until now, using SOLCALC software required thermodynamic expertise, making reliable parameter fitting a task for specialists.
With our latest update, that changes fundamentally.
We have now integrated novel parameter fitting and interaction screening algorithms directly into SOLCALC. This algorithm is connected to a large, curated, and continuously growing experimental database. You can think of it as a universal interaction brain that evolves as more validated data becomes available. The result is a workflow that is faster and more robust, while preserving the key advantage of physics-based modeling: mechanistic interpretability.
From manual fitting to a guided physics-based formulation workflow
SOLCALC is built on PC-SAFT (Perturbed-Chain Statistical Associating Fluid Theory), a physics-based thermodynamic model used to predict how small molecules behave in complex systems like amorphous solid dispersions (ASDs). This software is further enhanced with submodels (e.g. for predicting kinetic effects), delivering ultimate predictive depth.
For years, our modeling strategy has been consistent: calibrate the entire system against a large, carefully curated database. This means the model is trained simultaneously on thousands of experimental data points, building a comprehensive interaction landscape. Currently, our database includes over 2,500 verified curated solubility points compiled from literature sources and validated internal measurements across many APIs and solvents. Importantly, we only use validated data as training data for our interaction brain. Integrating unverified data can damage prediction quality.
The result is a generalized, unique model that captures essential intermolecular interactions, such as size, polarity, and hydrogen bonding, which are critical for formulation decisions. The physics-based modeling approach provides two fundamental advantages:
- It requires no empirical binary parameters. The predictability for unknown solvents or polymers comes from first principles.
- The model operates within strict thermodynamic boundaries, ensuring all outputs are scientifically plausible and inherently reliable. It cannot generate non-physical “garbage” predictions.
Until now, however, leveraging this powerful, physics-based big data strategy for a new molecule was a manual, specialist task. Our SOLCALC update removes that manual layer, creating a direct, living link between validated experimental data and formulation predictions. Instead of isolated fitting files, SOLCALC now supports a coherent fitting framework that can be refined in a controlled way as new high-confidence data are added.
What changes the SOLCALC update for users
A) Fitting a new molecule is now easier and much harder to do wrong
Users open a dedicated graphical interface, enter five experimental solubility points in different solvents, and click a button to run the fit and obtain the PC-SAFT parameters. This workflow becomes harder to misuse, because users no longer need to manually manipulate parameters in ways that can destabilize the model.
This matters because SOLCALC users are usually not thermodynamic specialists (see SOLCALC training blog). Standardizing the big data fitting procedure removes common failure modes that previously caused either crashes or silent quality loss. It also makes results reproducible across teams, i.e. the same experimental inputs lead to the same parameter set, independent of who performs the fit.
That reproducibility improves prediction quality in practical ways. The algorithm explores multiple starting points and searches for a robust optimum, reducing the risk of false fits. And because fitting is executed in the context of the broader interaction landscape, parameters become more consistent across molecules and solvents.
Once parameters are fitted, you can immediately start with screening polymers or solvents for spray drying processes. SOLCALC can now run standardized interaction and phase-diagram screenings directly in the interface, eliminating manual Excel-based screening workflows and reducing the time from experimental input to formulation decision.
In daily formulation work, this translates into:
- More stable extrapolation beyond measured conditions
- Higher confidence when screening polymers or solvent mixtures
- Better comparability of results across projects, teams, and sites
B) Companies can start building their own in-house databases digital twin
The update also strengthens how SOLCALC fits into real organizational workflows. Data and parameters can be managed consistently across colleagues: early-stage colleagues do the compound fitting when they first work with the molecule, another runs screenings, and teams do not lose context when projects move from early screening to late-stage optimization. The result is continuity, less redundant work, fewer project restarts, and clearer traceability of what has already been done.
This update makes it easier for organizations to build company-specific internal databases and digital twins. The SOLCALC base database provides a validated public foundation. On top of that, companies can develop their own internal interaction landscapes based on their own measurements for internal use, creating a long-term asset that reflects their molecules, their strategy, and their accumulated knowledge, precisely tailored to the companies’ proprietary chemical space.
Invitation to Connect
The SOLCALC update, which embeds this guided big data fitting directly into your workflow, is available now for all license holders.
This tool is instrumental in the digital design of drug formulations, providing a robust platform for comprehensive formulation analysis. To see a live demonstration of the new guided fitting process and discuss how it can be integrated into your specific workflows, please contact us to schedule a personalized walkthrough.
