Epsilon is a next-generation thermodynamic model for electrolytes

Problem: Electrolytes are difficult

At the heart of any process simulator lies the physical property engine that must deliver reliable predictions of e.g. phase equilibria and thermal properties to enable robust process design. Reasonably good predictive models exist for systems that do not involve reactive mixtures with electrolytes but unfortunately, the predictive capabilities of electrolyte models lag decades behind their conventional counterparts and require careful validation.

It is often difficult to know a priori if electrolytes will become important in a process, as they are typically present in trace amounts and may appear unexpectedly. However, accurate modeling of electrolytes can be critical in guiding e.g. material choice, process water recycling, chemical dosage, and instrumentation, to avoid expensive errors in process design and plant construction.

Solution: Epsilon

Epsilon is our vision for a next-generation reactive electrolyte thermodynamic model that provides accurate predictions even in the absence of experimental data. Combined with robust and fast algorithms for chemical and phase equilibria, Epsilon provides reliable predictions also for liquid-liquid equilibria with electrolytes – without going to the lab!

The principle behind Epsilon is to preserve detail from fundamental “0-parameter” science all the way to a truly predictive electrolyte equation of state (EoS). The foundational work was a prize-winning PhD thesis by one of the founders, which has since been improved using quantum chemistry and molecular simulation to develop Epsilon.

A new default

Epsilon will fully integrate with leading process simulation tools and is applicable to all systems irrespective of solvents, ions, T, P, phases etc. Epsilon is easy to use and runs fast, implementing modern computational methods and software practices, while the preservation of scientific detail makes results easy to interpret for both chemists and engineers.

The Epsilon methodology is backwards compatible with conventional thermodynamic models, including SRK, PR, CPA, and PC-SAFT. This makes it possible to use already known pure and binary interaction parameters with Epsilon and benchmark against existing flowsheets. Using Epsilon, you can then gradually increase the level of detail of your simulation to include effects of hydrogen-bonding, electrolytes, and reactions, allowing you to estimate their impact on your process without having to change simulation environment or unit operations.

The combination of versatility and backwards compatibility means you can always choose Epsilon. Whether working on a new process with no experimental data or looking for a second opinion, you never have to disregard electrolytes again.

With the customizable Epsilon-AI, Epsilon integrates with your engineering workflow by including your own data and experience and fine-tunes parameters for specific flowsheets using advanced automation algorithms. Epsilon-AI will not only provide the best possible predictions for electrolyte systems, but may replace existing non-electrolyte models to give equivalent or better predictions, that will improve over time by learning from your experience, data and operating range. Epsilon-AI also features uncertainty propagation using artificial intelligence and your experimental data to build greater transparency and confidence in your simulation results.