I will offer some suggestions based on some of the other replies.
1: It is not a good idea to use 100% aqueous mobile phase on a C-18 column unless the column is designed to allow this. Can you tell us what column you
use? Instead of starting at 100% aqueous buffer, can you start at 5% acetonitrile/95% buffer?
2: First clean the column by flushing with 100% acetonitrile after you have flushed out all buffer salts with water.
Now do two test runs with your gradient using only water and acetonitrile and run the gradient with no injection.
If the peak appears, either your water or you HPLC system is contaminated.
If this blank is clean, inject pure water and run a gradient with water and acetonitrile.
If this blank is clean, add only one of the buffer components to your mobile phase and repeat the gradient.
Test each of the buffer components in this manner to find the one causing the peak.
Good luck and keep us posted on your results!