PC-SAFT (Perturbed Chain Statistical Associating Fluid Theory) is a physics-based thermodynamic model that describes how molecules interact, dissolve, and stabilize in complex pharmaceutical systems. Unlike empirical trial-and-error testing, PC-SAFT uses physics-based equations to describe solubility, phase behavior, miscibility, and stability in pharmaceutical systems.

Pharmaceutical development often faces challenges like:

• Poor solubility of APIs
• Instability of amorphous solid dispersions (ASDs)
• Long drying processes with residual solvents

PC-SAFT helps predict these behaviors upfront. This means developers can select the right excipients, optimize processing conditions, and ensure stability before running extensive experiments.

For ASDs, stability is determined by a delicate interplay of thermodynamics and kinetics. PC-SAFT can:

• Predict miscibility of drug and polymer
• Calculate glass transition temperatures (Tg)
• Model molecular mobility as a function of humidity and drug load
• Anticipate crystallization risks and shelf-life windows

This enables formulation scientists to focus only on viable ASD candidates, saving months of experimentation.

Yes. Secondary drying of spray-dried materials is often slow and unpredictable. With PC-SAFT, we can:

• Simulate solvent evaporation under different temperature, pressure, and nitrogen sweep conditions
• Predict how assisting solvents (e.g., water, methanol) accelerate drying
• Generate drying curves that reduce process times by up to 50% without compromising product stability

Not entirely. Experiments remain essential for validation and regulatory purposes. However, PC-SAFT drastically reduces the number of experiments needed by eliminating unpromising options and focusing only on optimal conditions.

The in-silico model requires key training data to characterize the net strength of a drug molecule’s intermolecular interactions with its environment. By combining this minimal experimental input with a rigorous physical framework, PC-SAFT predicts molecular properties and behaviors with very little experimental effort.

Machine learning algorithms typically need thousands of data points because they must learn physical principles from data. In contrast, PC-SAFT already contains physicochemical laws in its structure. This means it requires only a handful of data points, making it particularly powerful for early-stage drug discovery, when experimental data is scarce.

The training set usually needs just five solubility measurements in different organic solvents. These are sufficient to determine the strength of key intermolecular interactions such as hydrogen bonding, van der Waals forces, and polarity effects.

Yes. Even for weakly crystallizing molecules, where solubility data are difficult to obtain, PC-SAFT can be trained with alternative thermophysical properties such as dynamic vapor sorption data or partitioning coefficients.

Yes. In fact, we observe that highly complex molecules are often described particularly well by PC-SAFT, as their diverse interaction patterns are captured in detail by the model.

The model captures a broad spectrum of forces, including:

• van der Waals interactions
• hydrogen bonding
• polarity effects
• ionic and charge interactions

This allows a realistic description of how drugs, polymers, and solvents behave together.

Very important question: No. The chemical structure is not required. Predictions are made from experimental thermophysical data, which means PC-SAFT can be applied even when the detailed structure is not available.

The model is highly flexible. Possible inputs include:

• crystalline or amorphous solubilities
• miscibility data
• partition coefficients
• sorption data
• osmotic or density data

We aim to use data already available from early-stage screenings, minimizing additional experimental work.

PC-SAFT can predict both thermodynamic and kinetic properties, including:

• Equilibria: crystalline solubility, amorphous solubility, sorption
• Metastability: crystallization risks, driving forces for phase changes
• Kinetics: crystallization onset times, dissolution rates

This enables applications such as excipient screening, ASD design, humidity risk assessments, shelf-life prediction, solvent screening, and dissolution modeling.

A typical phase diagram can be computed in just a few moments. This efficiency enables full formulation screenings within practical, economic timeframes — far faster than computationally intensive molecular simulations.

Yes. The number of components is not limited. PC-SAFT handles binary systems (drug + excipient) as well as highly complex formulations with many excipients. We have successfully modeled systems with up to eight excipients without major issues.

Virtually all relevant pharmaceutical materials:

• Solvents and co-solvents
• Drugs and APIs
• Polymers
• Sugars
• Amino acids
• Lipids
• Gases

Our predictions have been benchmarked against real-world data and consistently show high agreement with experimental results. For example, solvent removal kinetics and Tg profiles predicted by PC-SAFT have matched lab drying curves and crystallization behavior in shelf-life studies.

Typically:

• Basic solubility data (as few as five experimental points)
• Physical constants of API and excipients (molar mass, density, vapor pressure)
• Processing conditions (temperature, pressure, solvent type)

This minimal input enables robust predictions across a wide range of scenarios.

Regulatory agencies increasingly expect a mechanistic understanding of stability and process design. PC-SAFT provides this mechanistic foundation:

• Explaining why a formulation is stable
• Justifying why certain drying conditions are chosen
• Supporting shelf-life projections with predictive modeling

This strengthens CMC documentation and speeds up approvals.

• Time savings: Cut months off development timelines
• Cost efficiency: Reduce material and labor waste from failed experiments
• Reduced risk: Identify crystallization or instability risks before scale-up
• Faster scale-up: Transfer predictive insights directly to manufacturing conditions

We offer:

• Consulting projects for specific APIs or formulations
• Collaborative development with pharma partners
• Customized modeling packages for internal R&D teams

👉 Contact us