Not possible by the route you are using.  The best route is a bottom up process. Read: V R Vemula, V Lagishetty, S Lingala Solubility enhancement techniques International Journal of Pharmaceutical Sciences Review and Research Volume 5, Issue 1, November – December 2010; Article-007, 41 – 51 

Himanshu: This is copied from Drug Development and Industrial Pharmacy 41(1) · November 2013 "Media milling process optimization for manufacture of drug nanoparticles using design of experiments (DOE)"  Vijaykumar Nekkanti,  Ashwani Marwah, Raviraj Pillai.  If you bothered to read the article you will find that submicron particles were produced ( 'particle size' ~ 350 nm - indeed nowhere does it state whether this is a mean (and an assumed volume mean), median, or mode) and not nano (< 100 nm). Further, the imaginary part of the RI is not stated for the conversion to size from intensity-angle (only a real part of the RI of 1.5 is stated; is this reasonable?).  

Please do not plagiarize and try to understand that the difference between micro and nano is the difference between an elephant and a ladybug (Eric Drexler).

Himanshu: simply the ISO, ASTM, and EU definitions of nano refer to objects in the ~ 1 to 100 nm size range.  You can measure the distance to the sun in nanometers but it doesn't make this nanoscale.  I also have specific criticisms of the paper some of which I outlined in my first disorganized answer above.

Finally, a last point - the term 'comminution limit' should be researched.  This is my last contribution to this post.

A note to readers of this post: Himanshu Bhatt has removed his plagiarized comments and text from this thread making my answers to his comments stand in isolation.  This is completely unethical. I assume too that it is he that has downvoted all my answers as he continues to do on any thread.

For completeness, here are his 2 posts that fit in between my 3 first answers above.  Note that his first post was copied and pasted, without attribution, from the article I named in my second answer.

Himanshu Bhatt · BITS Pilani

Industrially applicable Method: Just check where you missed the process parameter.

The nanosuspension is prepared by dispersing drug in an aqueous solution containing a specified amounts of sodium lauryl sulphate and hydroxyl propyl methyl cellulose using an Ultra Turrax stirrer for 3-5 min at 10000 rpm. The homogenous suspension is then milled using a media mill. The milling chamber is charged with the milling media. The milling media is comprised of 0.2 mm yttrium stabilized zirconium beads. The milling operation is performed in a re-circulation mode with the suspension fed at a predetermined rate. The mill and pump speeds are operated according to the statistical design. The suspension is flowed axially through the milling chamber where the shear forces is generated during impaction of the milling media with the drug provided the energy input to fracture the drug crystals into nanometer-sized particles. The temperature inside the milling chamber is controlled by circulating cooling water through the outer jacket. After milling, the suspension is collected and stored at a temperature

 Dear Alan sir,

Thank you very much for you valuable answer for question,

What you said in the first answer was Bottom up approach is for academic purposes not for Industries,i have gone through the process it doesn't give reproducible results and it gives less yield.

Currently am working on media milling process it can give reproducible results and proper yeild,but the problem is am not achieving required particle size .i need D50 around 1 micron in range.

what  Himanshu Bhatt  sir said is right regarding the bottom up approach its not feasible in Industries.sir i preformed DOE trials as per the above reference article. 

according to the DOE results were not good as we required but required particle size D90 is 2.2 micron 

For media milling does bead size will effects on the milling parameters how to select the optimum milling speed and does  Drug input particle size will also  effets

Please give me suggestions to do 

Shanker - you have been very patient in this, so I will answer you, even though I stated above that I would post no further.  I also note that someone has gone to all of my recent posts on RG and, whatever the answer, has downvoted them.  This will be reported to the RG administrators who will be able to analyze the electronic signatures and see if there is a consistent violator.  Back to the reply, if you try a top down approach you get into 2 issues:

• The amount of energy required to form new surface becomes enormous and you can mill for hours without effect as smaller particles tend to flex and bend and not fracture - simply you are dispersing agglomerates and not creating new surface. If they do fracture then they can recombine again through van der Waals attraction (although this can be compensated for somewhat with huge stabilizer additions (ionic surfactants, polyvalent ions, or organic polymers)
• A limit is reached called the 'comminution limit' below which it impossible to size reduce any further for the reasons cited above.  Simple calculations in this area indicate that for many materials this limit is around 0.5 - 1 um or so - higher for some (polystyrene is calculated at some 2800 um and NaCl at 32 um!)  and much lower for others (e.g. diamond: 0.04 um).  See the paper by J. E. Field , M. Farhat, S. M. Walley" Comminution limit (CL) of particles and possible implications for pumped storage reservoirs"  J. Mater. Sci. (2014) 49:3780–3784 DOI 10.1007/s10853-014-8089-3.  Note too that bead mills are less than 2% efficient at best (in tems of converting energy to cleavage).  Most of the energy is converted into heat, light, and noise.

More on the comminution limit in:

October 6th, 2016 Controlling powder bulk density by optimizing particle size and shape distribution

https://www.brainshark.com/malvern/vu?pi=271602698&b=1&tx=120486&c1=22

and: October 2nd, 2013 The PSD, zeta potential, and rheology of ceramic materials and ceramic slurries/suspensions

https://www.brainshark.com/malvern/vu?pi=97328853&text=M021507

What can you do to get the best from your system?  In general, smaller beads tend to give a smaller final size although one should be careful with examining for contamination (e.g. zirconium) in the final product.  Stainless steel beads are never advised - blackening of white materials and Fe/Cr contamination is then obvious to see.  Contamination happens with ceramic beads too but just not visible.  Some high shear homogenizers can be used which avoid ball milling totally. Many other solutions were given in the paper I indicated above in my first reply in this thread (and not unsurprisingly was downvoted) - this is a good read.  You need to optimize the rpm to ensure that you have cascading in the mill - higher rates simply cause the particle to rotate with the wall - in my day an acoustic device could be used that would optimize the rpm for maximum noise (and hence maximum collisions).  Appropriate additives can be used (in the case of inorganic oxides then the optimum concentration of phosphate is usual) but be aware that, for any ionic additive, too much is as bad as too little.  The optimum can be found empirically in vial tests as can be seen in:

Nov 11th, 2008. Dispersion and nanotechnology http://tinyurl.com/hpywsge

I await the downvote to my answer.  Contact me privately for more information and assistance avoiding any public and childish downvoting.

 Dear Alan Sir,

Thank you very much for your valuable reply

I hope it would be helpful to my research work.Sir will contact you for future assitance