Academic Paper

Predicting Mass & Oxygen Transfer in Bioreactors

A Mechanistic Approach for Predicting Mass Transfer in Bioreactors

Chemical Engineering Science (2021)

John A.Thomas, Xiaoming Liu, Brian DeVincentis, Helen Hua, Grace Yao, Michael C. Borys, Kathryn Aron, Girish Pendse

Oxygen transfer in bioreactors is governed by complex fluid mechanics, which cause challenges when it comes to scaling up production of biologic drugs in biopharmaceutical manufacturing processes. In this paper, M-Star and Bristol Myers Squibb validate a mechanistic transport model to predict oxygen transfer rates within stirred tank bioreactors at multiple scales.

During many biopharmaceutical manufacturing processes, biologic drugs are produced by living cells growing within stirred tank bioreactors. These living organisms require a continuous and uniform supply of sparged oxygen in order to maintain optimal bioreactor operation—and appropriate cell growth, metabolism, and protein quality.

Typically, these biologic processes are first developed and optimized in benchtop bioreactors with small operating volumes. The challenge arises when manufacturers try to scale up from bench-top bioreactors to production-scale bioreactors with volumes measured in thousands to tens-of-thousands of liters. This is because the overall oxygen transfer rate to the fluid is a nonlinear convolution of the blend time, energy dissipation rate, gas injection rate, and bubble size. These complexities make predictive mathematical modeling slow and difficult.

In this work we present a framework for building time-dependent, bubble-resolved, two-phase models for investigating the real-time blending and mass and oxygen transfer in bioreactors. This method allows us to generate engineering predictions faster and with fewer modeling assumptions than RANS/population balance approaches in a multi-CPU environment. There were multiple order-of-magnitude improvements in computational speed when solving the fluid transport equations, further bolstered by the use of GPUs over a CPU cluster. The results were in good agreement with experimental data with no model reparameterization between scales or operating conditions.

In the full paper, we:

  • present relevant conservation laws that govern fluid, bubble, and species transport within an agitated tank.
  • solve transport equations using the Lattice Boltzmann method and Newton’s second law.
  • validate the model by comparing predictions to experiments across multiple reactor sizes and operating conditions.

Read the full paper to learn more about the development of the physics-based approach and see the industry practical results.