This process description is missing an essential piece of information, the diameter of the Rushton impeller.  However, if the impeller is more than about 25 mm in diameter the problem is clear with either 6,000 rpm or 10,000 rpm.  The mixing intensity is off the charts at those speeds, but the impeller tip speed may be needed for the particle size reduction.
 
The splashing fluid and big fluctuations in torque are simple to explain.  Air is being drawn into the region of the impeller.  When the impeller attempts to pump air instead of liquid, the power requirement (torque) drops and the pumping is reduced.  When the pumping is reduced, the air is not drawn into the liquid and the power and pumping increase, resulting in another cycle of more air being drawn in.  This cycle of high flow / low flow, high torque / low torque will continue as long as air can be drawn into the impeller.
 
The two most obvious solutions to the problem are put a lid on the vessel or reduce the width of the impeller blades, or do both.
 
If you put a lid on the vessel and completely fill it with liquid, you can no longer draw air into the liquid and change the flow to the impeller.  To limit air entry and avoid splashing, a seal around the shaft may be needed.  (I presume that this impeller is being driven by a shaft.  Since I have never heard of a magnetic drive operating at 10,000 rpm.)  Even if you don't have a pressure seal, a lip seal may be sufficient.  Although getting any seal to work at 10,000 rpm or increasing the speed from 6,000 rpm may be difficult.  You may with to try overfilling the vessel, with liquid covering the lid and no seal, just a small clearance around the shaft.
 
The other option is to make the blades on the impeller narrower.  A true Rushton impeller has six blades, with a blade width (height, not thickness) of about 1/5 of the impeller diameter.  If all you are trying to do is break agglomerates (only fragile particles can be broken by a mixing impeller), the pumping characteristics of the impeller are probably more than adequate at 10,000 rpm and the tip speed will determine the ultimate particle size.  Start by reducing the blade width to 1/2 of the present width.  If you still experience problems, remove all of the blade extending above the impeller disk.  Reducing pumping above the impeller by having blades only on the bottom of the disk will reduce the ability of the impeller to draw air into the mixture.  As an alternate method of reducing the blade width, you may taper the blades from full or half width at the disk to a narrow width at the tip.  (You may have an impeller that looks a little like a cowboy spur.)  What ever you do, be careful to make all of the blades the same shape and width, otherwise you will have serious vibration problems due to imbalance.


The answers by this expert are based on the best available interpretation of the information provided.  The consequences of the application of this information are the responsibility of the user.  If clarification is needed, please submit a further question.