Utilizing a standard that is “off” is a problem. I have always assumed that my refractometer solution is accurate. You can check to see after calibrating to 35 ppt whether your ro/di calibrates to near zero. If there is a large discrepancy then you know something is wrong. I believe it was Randy Holmes-Farley that said it was best to calibrate at 35 ppt.
Of course you should also make sure your clean and dry your refractometer with each use. And definitely do not scratch the refractometer glass