Maybe there is a problem with too high electrolyte concentration (addition of 10 % NaCl) leading to shielding of electrostatic repulsion.

Thanks for your reply, but the aggregation is appear before salt is added.

So the salt might not the problem

You may apply  AuNPs stabilisation using 0.05% Tween prior to Antibody addition

Thanks for Satish reply, I have tried that after you mentioned.

but I failed.

I just read some papers and by using sonication the problem is reduced. But I discover the color tend to be light purple after sonication instead of red.

My gold volume is O.D.=1 20 ul and antibody is 0.1mg/ml 1.2ul.

I just guess is that when the antibody concentration is higher, than antibody will appeal to others because the ion strength is not high enough. Means antibody is sensitive to the environment change.

I added PBS dilute to 1/20 fold but condition still exist.

Still suffering.

Refer Konstantin Sokolov's paper (Univ of Texas) for Antibody conjugated to Gold nanoparticles. 

I try to optimize the conjugate procedure by replace the gold solution with borate buffer pH 10.0 (9.0 and 8.0 are compared), it's huge reduce the aggregation, but still not the optimization. Cause the conjuagte flow the membrane slow. But weak color still can be seen on control line.

It's so disturbing.

But I change another monoantibody, this phenomenon could easily be solved with this one.

Try this gold conjugation kit, it will reduce the aggregation problem:

www.nanodiscoveryinc.com