I don't know what your samples are but in my experience cleaning surfaces is more likely to damage the gold coat. It may be easier to remove the coat and start again. Is the Au-coat really needed?, so much work is now able to be done uncoated, either at variable pressure, or by using low kV. Can you let us know what your material is?

In case you fancy having a go at removing the coat entirely - here is a method.

http://palaeo-electronica.org/content/issue-2-2012-technical-articles/234-removing-sem-coating

Dear Sabouri,

If your samples are smooth than you can clean with a soft cloth, because, the adhesion strength of the gold coating is not so high. 

1) If you sample is smooth and well polished, just swipe with a tissue paper,

    and do ultasonic cleaning.

2) If you sample is rough, just by abrasion on a fine sand paper a coupleof times,

    till you see you original sample surface, then ultrasonically clean it in some laboratory solvent like acetone oar IPA.

3) If the coating is sticking hard, the only way out is by grinding and polishing.

Dear Sirs,Thank you all.

The sample is a commpercially pure Aluminum, coated by Plasma Electrolytic Oxidation process (a porous oxide layer on Aluminum).

Already, the coated sample used for corrosion studies around 1000 h. Then, the corroded sample was characterized by SEM. At the moment, we need to do the SEM studies again for further morphological and chemical information on that sample. However, the new SEM characterization showed that the surface of the sample is polluted and need to be cleaned. Preparing a new sample is also time consuming and thus we have to do the new SEM studies on that sample, again.    

In fact, we need to remove the pollutants without damaging the Gold-sputtered layer that may also causes damage to the coating.

Dear Sabouri,

Thank you for explanation. You can apply ultrasonication to remove the pollutions from your coating.