Assumption: The area of your RE is significantly smaller than that of your WE and CE. If true, no problem here. If untrue, issues may arise due to current also flowing through RE, even if it is modified.

In a three-electrode system, the assumption is that the WE and CE are made from a material that is ideally polarisable (gold is good for this purpose), while the RE is made from a material that is ideally non-polarisable (Ag/AgCl is the default now, I believe, rather than Hg/calomel).

Now, if you use an ideally polarisable material for the RE, the current passing through it will affect its potential and throw it out of equilibrium, resulting in potential shifts (scalable with the applied potential!) on your CVs.

Luckily, there is an easy solution for this issue: just cover the conducting (i.e. the gold paths, which are not protected from contact with your solution by an insulating layer) areas of your RE with Ag/AgCl paste and you are ready for measurements.

A good source (personally tested) of such inks is ALS: https://www.als-japan.com/1391.html

You might also be interested in knowing that they offer a commercial 3-in-1 and 4-in-1 (dual working electrode, in the form of an interdigitated electrode system) electrode slips, using gold as the conducting metal, such as the ones here: https://www.als-japan.com/1379.html

Thank you. But Putting a RE of Ag/Agcl in a microfludic chip is a very difficult task. That's why I am using all the electrode with same material. Any suggestion for that?

Christian's suggestion above is a good solution, since a layer of Ag should deposit well on Au. This layer can be used as-is, as a pseudoreference electrode (will need to be callibrated for potential float from time to time, using a reference redox couple, whose standard redox potential is known, e.g. potassium ferrocyanide or ferrocene) or can be converted to AgCl via oxidation in dilute HCl.