You can make an estimate using the plasmon peak shift of the bulk solution. For that you have to know the: average size of the nanoparticle, the shell of the 11-MUA layer around the NP and the 'spherical' size of your antibody (~150 KDalton) approximately 6 to 8 nm.

With this you can calculate the footprint of each antibody and the theoretical maximum antibodies on the surface of each particle. Now you have to do a scattering simulation (or discrete dipole approximation DDscat) where you simulate the NP as sphere of the metal (using the Johnson and Christy Optical constance of noble metals http://journals.aps.org/prb/abstract/10.1103/PhysRevB.6.4370), the 11-MUA as layer with guessed refractive index of 1.5 and then the layer of antibodies with a refractive index of 1.5.

I have more citations in my thesis but have to find those.

Ok dear...i have calculated the no of 11-MUA attached on gold nanoparticles. I have attached antibody to 11-MUA coated Gold nanoparticles and then i centrifuged.. but how i will estimate the free antibody present the solution(supernatent) at nanogram level..i have tried bradford reagent but it is not sensitive at nano gram level...so which reagent or process will be needed for free antibody estimation in solution without any fluorescent tagging?

Roti-quant is somewhat more sensitive, but best will be a fluorescent or RIA to quantify the content.