Yes the silica is very important. At physiological pH values, silica is weakly acidic (as its isoelectric point is pH ~2), giving it a weakly negative surface charge (Si-O- as opposed to Si-OH). In the presence of high concentrations of chaotropic salts (salts which disrupt and denature the structure of nucleic acids and destabilise hydrogen bonds), cations are able to form a stable layer around the negatively charged silica surface, effectively giving it an effectively positive surface charge. DNA is relatively stable in the presence of chaotropic salts and retains its negative charge in aqueous solution - via the phosphate groups present in the DNA ‘backbone’. Consequently, DNA can bind effectively via a ‘cation bridge’ mechanism to the now effectively positively-charged silica surface. The DNA out-competes other molecules to bind to the silica surface and they are left in solution. The newly formed DNA-silica complex can be washed in either a salt solution or an ethanol/water mixture to remove impurities, which aren’t as tightly bound as the DNA. The purified DNA can then be eluted from the silica using a buffer containing low salt concentrations, or water. Due to the difference in surface charges of magnetite and silica at physiological pH values, the interaction between the DNA and the surface will be much more different, and typically the magnetite will not elute the DNA with very high efficiency.

Ben...that's perfect. I can see a clearer picture now. One more thing....will this also prevent the non specific binding of proteins on the particles along with the targeted DNA? Will the purity of extracted DNA be more on using silica based particles.?

Yes it is much more specific towards DNA binding only because the chaotropic conditions are much more suitable for DNA than other species such as proteins.

I am not an expert on DNA or biology but I would assume to increase the purity of the DNA you could elute it from the surface, and repeat the binding and washing steps to remove impurities that may remain.