CFD-aided design of a fluidised bed bioreactor for bone tissue engineering

Introduction
The aim of this project was to perform CFD analysis of liquid-solid fluidisation to inform the design and development of a fluidised bed bioreactor for the culture and growth of bone cells seeded into porous bioceramic particles. The need for artificial bone graft material, particularly for large defects which cannot be filled with autologous bone, is driving research into synthetic alternatives1,2. The project utilises calcium phosphate particles (4 – 10 mm diameter), with large (500 - 1500 m diameter) interconnected pores. The fluidisation of the particles was modelled to inform the sizing of the bioreactor, to find a balance between quality of fluidisation and the volume of media required to be pumped.

Methods
The liquid phase was modelled to represent phosphate-buffered saline at 37 oC. The solid phase was modelled as 4 mm glass spheres with a density of 2500 kgm-3. The fluidisation was modelled in columns of 25, 35 and 80 mm diameter, all 400 mm long. A k- turbulence model was used to govern the flow of the liquid phase, a dense discrete phase model was used for the solid phase. The flow of liquid was incrementally increased from 0.01 ms-1 to 0.042, 0.06 and 0.10 ms-1 at 10, 30 and 50 s respectively. These velocities cover the range theoretically required to fluidise the particles.

Results
The steady state depth of the bed increased for each column size when the flowrate was stepped from 0.06 to 0.10 ms-1. Figure 1 shows the behaviour (volume fraction) of the solid phase as the velocity of the liquid is increased between these two values for the 25 and 35 mm columns. It can be seen that as the velocity increases, the bed of solid particles is temporarily pushed up the column as a slug before returning to a bed of greater depth than was present initially. The steady state depth of the bed in the 80 mm column increased by a similar amount, but no slugging behaviour occurred.

Discussion and Conclusions
The results of the simulations identified factors critical to the successful design of a fluidised bed bioreactor, i.e. the column diameter. A relatively narrow column can result in a significant slugging response that is not present in a larger column. This slugging effect would potentially have a negative effect on cells grown on the solid phase due to the increased shear stresses involved. It could be said that a wider column is preferable, however a balance must be found between the degree of fluidisation required, the dimensions of the column to house the bed and the cost of operation. As the diameter of the column is increased the amount of slugging decreases but the volume of media and size of pump required to maintain the same flowrate increase significantly.

Fig 1: Volume fraction of the solid phase as the velocity is increased, at t = 50 s, from 0.06 to 0.10 ms-1. Images represent a 2D plane through the middle of the reactor. a) 25 mm column, b) 35 mm column
These findings demonstrate the application of CFD modelling to enable the successful design and development of a fluidised bed bioreactor for the culture and growth of bone cells seeded on to porous bioceramic particles.

References
1 Boyan B. et al. 2003. Bone Graft Substitutes: ASTM International 2Stevens MM. 2008. Biomaterials for bone tissue engineering. Materials Today 11(5):18-25.

Acknowledgement
This work was funded by EPSRC Grant EP/I015922/1.

Affiliations
*Department of Chemical Engineering, University of Bath, Bath, BA2 7AY
** Department of Mechanical Engineering, University of Bath, Bath, BA2 7AY