This is obviously an odd one so a little context, essentially I'm knee deep in researching how to essentially replicate VRAM failiure without causing catastrophic failiure. Yes it's weird and borderline amusing, I'm not a repair technician but a research engineer (I also do a bunch of PCB design and troubleshooting). I couldn't think of a better place to ask since I've been a fan of Northridge fix's videos for a long time and I'm sure many of you are excellent repair techs, so you could probably give real helpful comments knowing what typically causes failiures which might help me go the other way. 

My goal is to end up with permanent visual artifacts without causing catastrophic failiure on a donor motherboard - i do however understand that stressing failed VRAM can lead to catastrophic failiure - I am trying my best to avoid any damage to the rest of the board. My current plan is to start heating a VRAM chip with a narrow nozzle focusing closely on the IC, starting at the TJmax and going in increments of 5c up while the machine is running a VRAM stress test, this will aid me in causing the VRAM chip to naturally heat up aiding the heat transfer as well as stressing it which should quickly show artifacts or failiure so I don't introduce external heat for any longer than needed. 

I have the option to inject voltages or introduce shorts but I think I can get the result I'm after through heat transfer, as opposed to probing the powered board.

I wanted to ask what you thought of this plan and if you had any other ideas. I am absolutely expecting some good jokes too, so don't dissapoint me.