Elsevier

Chemical Engineering Science

Volume 53, Issue 5, 6 February 1998, Pages 891-898
Chemical Engineering Science

On the mechanism of microbial cell disruption in high-pressure homogenisation

https://doi.org/10.1016/S0009-2509(97)00414-4Get rights and content

Abstract

The tensions produced in the wall of a rigid, thin-walled, liquid-filled sphere as it moves with an axisymmetric straining flow are examined. This problem has not been previously addressed. A generalised correlation for the maximum wall tension, expressed in dimensionless form as a Weber number (We), is developed in terms of the acceleration number (Ac) and Reynolds number (Re) of the straining flow. At low Reynolds number We is dominated by viscous forces, while inertial forces due to internal pressure gradients caused by sphere acceleration dominate at higher Re. The generalised correlation has been used to examine the case of a typical yeast cell (a thin-walled, liquid-filled sphere) passing through a typical high-pressure homogeniser (a straining-flow device). At 56 MPa homogenising pressure, a 6 μm yeast cell experiences tensions in the inertially dominated regime (Re=100). The correlation gives We=0.206, corresponding to a maximum wall tension of 8 Nm-1. This is equivalent to an applied compressive force of 150 μN and compares favourably with the force required to break yeast cells under compressive micromanipulation (40–90 μN). Inertial forces may therefore be an important and previously unrecognised mechanism of microbial cell disruption during high-pressure homogenisation. Further work is required to examine the likelihood of cell deformation in the high-strain-rate short-residence-time environment of the homogeniser, and the effect that such deformation may have on the contribution of inertial forces to disruption.

Introduction

High-pressure homogenisation remains the method of choice for the breakage of microbial cells in large-scale bioprocessing (Middelberg, 1995). Simplistically, it operates by raising the pressure of a cell suspension to approximately 1000 bar, and then releases this through a specially designed valve assembly. The cells experience a range of forces, and disrupt through their interaction with the fluid and the solid walls of the valve assembly. This simplistic description hides the complexity of the cell-fluid–solid interactions occurring in the valve assembly. Numerous mechanisms of cell disruption have been proposed, including turbulence, cavitation, impact with solid surfaces, and viscous shear (Middelberg, 1995, Save et al., 1997). Recent studies provide evidence supporting cavitation and viscous shear as the primary mechanisms of cell disruption during high-pressure homogenisation.

By constructing a purpose-built apparatus, Harrison (1990), Save et al. (1997) were able to showthat a cavitating ball valve causes yeast-cell disruption. Cavitation is a well-known cause of disruption, for example during sonication of cell suspensions (Middelberg, 1995). The likelihood of cavitation for a given geometrical system may be correlated using the cavitation number, σ:σ=P−PV(ρfu22)where P is the local static pressure, PV is the vapour pressure of the fluid, ρf is the fluid density, and u is the local velocity. In venturi meters, the cavitation inception number is less than 0.15 with up to 2.3% v/v air included, at throat Reynolds numbers as high as 106 (Hammitt, 1996). Recent information on the fluid mechanics of high-pressure homogeniser valves suggests that local static pressure remains greater than 1 atm for cell-disruption (CD) valves because of high frictional losses across the valve face (Kleinig and Middelberg, 1996). The minimum cavitation number thus occurs at the discharge, and for a typical fluid velocity of 200 m s-1 equals 3.6 for water at 40°C. This is significantly greater than the cavitation inception number for venturi meters, which approximate the axisymmetric flow field in a CD homogeniser valve. Kleinig and Middelberg (1996) also examined square-edged valves, which showed negative pressures near the valve entrance. This suggests that square-edge valves operating at low pressure with large gaps will cavitate near the inlet. However, these valves are used for fat-globule dispersion and find little application for cell-disruption. Thus, while the possibility of cavitation in a homogeniser cannot be ruled out, the evidence suggests it is very unlikely in cell disruption homogenisers that use a standard CD valve.

Shamlou et al. (1995) have addressed cell disruption in terms of a viscous shear mechanism. The “gross” strain rate in a homogeniser inlet, α, was defined by Eq. (2),α=urwhere r is the radius of the valve rod (a characteristic length (2.4 mm) over which fluid acceleration occurs). By replacing the characteristic length with the mean diameter of a yeast cell (5 μm), a force of approximately 540 μN was calculated to act on the cell. This exceeds the typical force of 40–90 μN required to break a yeast cell under compression (Roberts et al., 1994). However, Shamlou et al. (1995) assumed a Trouton ratio of 1000 in their calculations. Whilst the Trouton ratio may exceed several thousand for non-Newtonian fluids such as polymer melts, the value for a Newtonian fluid is 3. A dilute suspension of undisrupted cells is likely to be very nearly Newtonian. If a Trouton ratio of 3 is used in the calculations, a viscous force of less than 2 μN is expected, which is considerably lower than the reported compressive force required for breakage. The assumptions inherent in this analysis therefore fail to prove a viscous-shear mechanism for cell disruption.

It is clear that the mechanism of cell disruption during homogenisation remains unproved. In this study we examine the mechanism of homogenisation by considering a rigid, thin-walled, liquid-filled sphere (i.e., an idealised rigid cell) moving with an axisymmetric straining flow that approximates the inlet region of a high-pressure homogeniser valve. The homogeniser flow fields are defined using a previous study (Kleinig and Middelberg, 1996). Numerical simulation is used to determine flow fields around the sphere and tensions within the sphere’s wall. A critical Weber number for sphere rupture is defined based on the maximum von Mises tension within the wall. Using typical parameters for a cell in a homogeniser, it is shown that inertial force is the same order as that measured by compression tests. This suggests that inertial forces may be an important and previously unidentified mechanism for cell disruption in homogenisation.

Section snippets

Problem definition and dimensional analysis

This study estimates the tensions that are produced in the wall of a rigid, thin-walled, liquid-filled sphere as it moves with a flow field. The flow field used in this study represents flow through a constricted orifice, and is a combination of a steady uniform flow and a straining flow. It approximates the flow in the inlet region of a high-pressure homogeniser. Maximum wall tension, Tmax, will depend on the following parameters:Tmax=fn(us,rs,ρf,ρs,α,μ).With seven independent parameters and

Problem formulation and solution methods

The previous section outlines the restrictions and scope of the study. Additional assumptions and restrictions are necessary to fully define the problem being solved, and are outlined in this section. Determination of the correlation We=fn(Re,Ac) was then achieved in two stages. First, the external flow field around the rigid sphere was determined for various Ac and Re. Maximum wall tension (and hence We) was then calculated from a force balance for each flow condition. These two steps are also

Results and discussion

The numerical code was verified by comparing results with published data for flow past a single stationary sphere in uniform and straining flow (Table 1). Comparisons are made with data taken from Magnaudet et al. (1995) for the separation angle and total drag coefficient. Results agree to within approximately 2%. Excellent comparisons were also observed with pressure profiles at the sphere surface (results not shown).

Calculations were then conducted for moving spheres in axisymmetric straining

References (19)

There are more references available in the full text version of this article.

Cited by (0)

View full text