Hi Chris, I do not if I an expert,but I can help you.

Your assumption is correct. In principle,  the mass , or density does not have influence over  a falling body. Galileo demonstrated this principle. But this is valid in vacuum. In the presence of a fluid, the particle or a entire body will feel a drag. the drag force is  proportional to the square of falling velocity:  Fd= 0.5*Cd*rho*U²*A  where, Cd is the drag coefficient, rho then  density, U the velocity  of body, and A it  project area normal to the flux. For a sphere this area is equal to a circle area. The drag acts against the particle displacement due its weight Fp = rho*V*G, where V is the volume or particle, and G the gravity field acceleration ( 100g in you case). Cd is not a constant. It is function of Reynolds number at least ( there are other cases that Cd is function of other non-dimensional groups)  Re= rho*U*d/mu   where d is the  diameter or a equivalent scale of  particle, and mu the viscosity of the fluid and it behavior is complex Hence, as the RBC moves down the plasma must rise and a two-phase flow takes place. Particles to down, and  fluid to Up. In  this situation, the  velocity must be the relative velocity : Ur=(Ud-Uf)  (d=particles, f=fluid).  The effective viscosity will changes due to  changes in particles concentration. Bigger particles have a  bigger area, with a bigger drag, but it apparent weight is bigger also. after some time the drag nulls the weight and the acceleration is null, hence,  the inertia retain the particles in movement with a constant velocity, this velocity is called, terminal velocity. The re are other effect the shape drag: the wake  after the particle creates a low pressure field, which pulls the particle against its displacement. So, beyond the  density difference, the size and the shape will impact in the particle displacement. The fluid properties as its density ( we have a little buoyancy force) and viscosity are important too. if the particle can change  them shapes, as bubbles or drops, the drag is affected. The Uranium 235 production is made based on a little difference of densities between U238 e U235  (only 3 neutrons!!!) in ultra centrifuges.

I hope I have helped you!

Hi Ricardo

Thank you for your answer. I think I understand your general principle but what I particularly don't understand is the interaction between the bigger and smaller particles so that the smaller ones (platelets) are  displaced upwards against the gravitational force so that they become less concentrated in the part of the fluid occupied by the large particles (RBCs) than in the supernatant where there are no large particles.

It has been suggested by a friend that the physics behind Brownian Motion may be a contributary part. A platelet's volume is approx 9 femtolitre. I assume that the viscosity/relative density of the plasma will have a part to play in determining the  influence on the brownian motion effect but is the platelet in the order of magnitude for this to be relevant?

 PS for reference the average values for density in g/ml, voume in femtolitres and viscosity in milliPa-seconds are as follows, though there is a spread of sizes

• Plasma  density 1.025 - and viscosity 1.5-1.7
• Red cell density 1.125 volume 75-95 
• Platelet density 1.04 to 1.08 with a mean of 1.65  volume 7.5-11

Settling velocity will be different for the particles of different size. You may need to apply higher gs (g is the gravitational acceleration (m/s2)). Other option is, that you may be able run a low g centrifuge for longer time (this option is painful - you will need to rerun the centrifuge multiple times and the small particles should be in the middle of the tube; and large particles at the bottom).

V=(2/9)*(ρp - ρf)*g*R2/μ

where V is the flow settling velocity (m/s) for a sphere of radius R (vertically downwards if ρp > ρf, upwards if ρp 

 Hi Viral

I can answer some of your points in a moment but one burning question is whether the platelets are small enough to the physics behind Brownian Motion  to keep them suspended bearing in mind they are close to their density is close to the density of plasma.

Although the platelets have a range in size I am not trying to separate them by size - rather to keep as many of them as possible in the supernatant plasma, whatever their size. The Red cells and most of the white cells are both much larger and are denser and are easily removed  while  leaving the platelets suspended.

Interestingly there is virtually no concentration gradient of platelets in the supernatant, and they are spread virtually evenly, though after 20 mins a small one does appear, but at the cost of a decrease in concentration as they are spun down to sit in the buffy layer (mostly white cells) that sits on the Red Cell layer

I think I do understand why smaller particles are being pushed up.  With the 10:1 volume difference and the greater particles being slightly greater the force exerted from the mass difference is far greater for the greater particles.  Even if you assume that the density is the same, if they were to have the same shape, the greater particles would exert a force that is 1,000 times the force exerted downward by the smaller particles.As if a particle was of radius r and the other was 10(r), the ratio of the volume of both is 1,000 (the cubic power).  You may also need to review your definitions for the concentration and how the variation is computed after the centrifugation. In the bottom, where the larger particles are, you may recompute your concentration for reference exclusive of the volume of the greater particles and see what is there in terms of concentration of platelets in the plasma fraction. Now, I'm not in that area of blood and plasma so that I have to ask the question, along with the merits of your curious mind, Is there anything specific that you want to accomplish?

 Hi Yuri

I follow your suggestion but in this case the platelets are being displaced from the remaining plasma. As a non engineer I would assume that the conc of platelets within the plasma remaining between the RBC s would not change rather than being depleted by a large %age.

Perhaps there are electrostatic forces or tthere is a Brownian Motion effect? 

PS the volume is a tenth, not the radius!

Geometrical detail is necessary to answer this question because small particles move with swirl flow.

100 gs is 100 gs.  It may have something to do with deformation too. Under such force, the plasma having platelets will be squeezed out off the void space.  Upon removal of the acceleration the fluid will be coming back through the "free path" of the spaces so that the majority will be fluid. Another 2 cents.

I am very much interested in this problem. Just two observations, do not feel I am any close to being an expert on this: (a) this phenomenon may have very little to do with gravity. In a centrifuge inertial forces act on particles pushing the way from the axis of rotation. Larger particles with crowd out the smaller ones, pushing them upwards, simply because their terminal (falling) velocity is hihger - ok, in that sense gravity matters, (b)  plasma is a non-newtonian fluid, so some "creeping" motion may weigh in. Rgds. Jerzy Klimkowski

Just see how large and small particles behave in a cyclone

Bingo. I had a pleasure to mess up one cyclone design by way of ignoring the difference in terminal velocities' issue. Since I no longer work in dust pollution control prefered to keep my opinion to myself. 

By the bye, is Bosch GmbH working on two-phase injection for Diesel engines? From a water mist extinguisher designer's point of view, you should be doing just that. :-)

Greetings form rainy Warsaw. 

I work on cavitation & cavitation erosion for injector at Bosch.

Greetings from rainy Bursa :-)