Assuming that the solution input is flat on the bottom and that the impeller is a pumper type; then simply as close to the bottom as possible. The biggest concern with the position is over-mixing to obtain solution transfer. At 3L/hr you will likely need higher rpm's than otherwise needed just to keep sufficient vacuum to facilitate liquid flow.

You appear to be building a pilot plant sized circuit and I have built many of them. I would suggest putting the impeller a little above the bottom, say 5 mm or so. The issue is really the balance between pumping and mixing. To deal with that, your design should allow you to raise or lower the impeller off the floor of the mixer box and the mixer drive should be variable speed. If mixing is insufficient, raise the mixing speed. If that results in too much pumping (that is you start sucking air into the mixer box), then raise the impeller away from the inlet to reduce pumping. If you have to lower the mixing speed and pumping is insufficient, lower the impeller closer to the hole.

The issue with small mixer-settlers is in my experience really not one of over-mixing. Impeller tip speed and shear are the critical factor in producing the very small droplets that result in poor phase separation performance. Do the calculations and you can see that with the small diameter impeller in your cell, you can easily operate at 1000 rpm and tip speeds do not exceed the recommended values. And, you should be able to operate in the range of 600-800 rpm with this size unit. The more important concern is that balance between mixing and pumping and the fact that you don't want to start sucking air into the mixer as that can promote emulsion formation.

If your impeller diameter is large enough, and if you have not already done so, you can consider using impellers with curved blades rather than the straight radial slots of blades. These also reduce shear at the impeller tip.

The backswept curved blade top shrouded impeller (R 320 style) should give you good results at low tip speeds. Keeping air out of the box is very important so you should check the upstream weir boxes are flooded and vortexing is prevented in the pump mix box using egg crate baffles or such like.

There is a great topic and answers!

I think it's important to characterize the couple "agitator-mixer". Number of power and pumping, for example. The turbine must not have to be too big for having high rpm and keeping emulsion.
This is important to take heed viscosity and density of different phases.