Every pharmaceutical formulator knows this painful reality: your drug powder comes out of the spray dryer, but it still needs secondary drying to remove residual solvents.
It slows down your workflow and can compromise product stability.
Two recent studies by Kerkhoff et al. (2025) and Kerkhoff et al. (2025) show how amofor uses PC-SAFT-based predictive modeling to reduce secondary drying times in amorphous solid dispersions (ASDs). This approach shortens the process while maintaining product stability and integrity.
These insights are already being used by leading pharmaceutical companies to optimize drying processes, reduce stability risks, and meet ICH solvent limits.
Here we show how it works and explain why it matters for your formulations.
The Bottleneck in Secondary Drying
Every spray-dried material contains traces of solvents. Reducing these residual levels is essential for patient safety and regulatory compliance. For instance, dichloromethane (DCM), a common solvent, must be reduced below 600 parts per million (ppm), according to ICH guideline for residual solvents.
However, achieving those low values is often a slow and inefficient process. Even with high temperatures and a strong vacuum, manufacturers still need two or more days to reach compliance targets. This is because many solvent molecules have a large hydrodynamic radius. They’re bulky and diffuse slowly through the dense matrix of an ASD. They get trapped, like threads in a spaghetti pile.
Multiple strategies exist to accelerate the process. For example, introducing a nitrogen sweep lowers the solvent’s partial pressure, while adding a second solvent, such as water, can accelerate diffusion and evaporation. This may sound counterintuitive because you are moisturizing your material. However, this isparticularly significant because it substantially improves drying speed. Without these strategies, prolonged drying slows production, overburdens equipment, and increases the risk of API crystallization.
What Physical Modeling Achieves in Practice
By combining our extensive knowledge of thermodynamics with physical modelling, amofor can help pharmaceutical companies dramatically reduce secondary drying times. Our validated predictive method delivers clear and accurate results:
- 50-75% shorter drying times: Clients regularly cut their drying duration from several days to less than one, significantly increasing production throughput.
- Reduction in residual solvents: Typical solvent levels around 50,000 ppm can be quickly and predictably reduced below regulatory limits (e.g., 600 ppm DCM).
- Reliable accuracy: Our models’ predictions closely match real-world drying experiments. Formulators can confidently rely on these insights, streamlining their operations and regulatory filings.
Step-by-Step: How Our Predictive Process Works
Step 1: Use PC-SAFT – The Science Behind the Speed
We apply our thermodynamic model, PC-SAFT (Perturbed-Chain Statistical Associating Fluid Theory), to simulate how solvent molecules behave within the ASD matrix.
With just five solubility measurements, we map out an API’s “intermolecular interaction fingerprint”. We model how it bonds, repels, or aligns with solvents, polymers, and excipients. This deep understanding enables us to predict solvent evaporation under various drying conditions without the need for trial-and-error experimentation.
Step 2: Creating Virtual Process Maps with Mechanistic Understanding
Based on molecular-level insights and your spray dryer settings, we generate a virtual process map. This map simulates how drying time responds to various variables, including temperature, pressure, and nitrogen sweep rates.
But they go further than surface-level trends: they reflect the underlying physics. By modeling pressure-dependencies, solvent diffusion, and evaporation kinetics, we provide a mechanistic understanding of the drying process. This is critical not only for process optimization but also for regulatory filings and submissions. Authorities want to know why a specific drying time is justified, not just that it works somehow. Our simulation combines thermodynamic equilibrium data and complementary fundamentals to explain how fast and why solvent removal occurs over time.
Importantly, these maps also uncover which solvent mixtures must be added to accelerate the kinetics of drying. Our water-assisted drying approach showed that under controlled humidity, ethanol residues were fully removed within 1500 minutes from PVPVA-based ASDs. Similar behavior was also observed in another studywhen using other solvents like methanol as assisting solvents.
Conventional drying under vacuum couldn’t achieve that. The added vapors increase molecular diffusion, allowing solvents to escape more quickly. This insight, grounded in physical modeling, leads to faster, safer drying and better outcomes.
Step 3: Clear, Actionable Client Outputs
Every client receives a complete package ready for direct use:
- Drying curves: Solvent concentration over time, easy to validate against experimental data.
- Wet glass transition temperature (Tg) as a function of time: Helps assess crystallization risk under drying conditions.
- Virtual DOEs: Drying time vs. temperature and pressure, identifying fast vs. inefficient settings.
Mechanical details, such as agitation speed, are not included, as they typically have a negligible impact in this setting. We focus on the primary drivers: temperature, pressure, and gas-phase solvent content.
All results are delivered in a visual, accessible format. We guide each client through the predictions, highlighting the variations, and show them how to apply the findings to inform process decisions, validation, or regulatory documentation.
A Proven Method in Action
We recently supported a pharmaceutical client scaling up a spray-dried ASD. They asked us to predict secondary drying times for a new batch and told us they’d run parallel experiments to validate the results.
Using their process parameters, including temperature, pressure, and nitrogen sweep, we simulated the expected drying times and provided a complete drying profile, including predicted solvent levels and Tg behavior over time.
Three weeks later, the client returned with the lab results. Our predictions matched perfectly. No adjustments needed. Next time, they said, they’ll skip the experiments and go straight to modeling.
These examples underscore our broader mission at amofor: solving real-world formulation challenges with rigorous, predictive modeling.
Ready to Optimize Your Secondary Drying Process?
Connect with our experts to discover how predictive modeling can enhance your formulation process, reduce drying times, and ensure process understanding.