jsonnet

So just to clarify, with typical HDDs the space is divided into physical containers, called sectors, if you want to immediately delete a file or even the hole disk you just override each sector with random data. Easy.
But Flash storage on the other hand is divided into blocks, which can wear out and are controlled by a flash controller, which provides an HDD-like interface to your system. Its job is to map sectors to blocks in a dynamic table, which allows the controller to spread the usage on different blocks, even if the operations are always on the same sectors. Which means, if you “overwrite” a sector, usually a new cell gets to be written on and the old one persists.
That's why in practice, flash storage is substantially larger (especially quality ones) than advertised, what allows for degrading blocks over time without losing storage space. All this is dealt with by the flash controller (invisible to user and system)!
BUT:
assume you are overwriting sector 0 typically doesn't change block 0, which means the system cannot see the data anymore. assume you are overwriting all sectors once, it does not affect all blocks, as the real size is larger. assume you are overwriting all sectors more than once, could work but might also not. => So if someone would take the flash media apart and insert their own controller and read block by block, he can definitely (potentially) recover the (all) data.
 
Of course there are solutions for secure deletion, but they typically require controller support and work by encrypting all cells, store keys in dedicated cells and if a file gets “deleted”, it does this by erasing the key cell. But in a nutshell this solution is far from being perfect and there are still no standards to date.
 
---------------------------------------
FYI: There is no intention in me sounding nerdy or anything like that, I just wanted to give a brief overview on how HDDs and SSDs differ in file deletion and the problem of Flash memory when wanting to securely delete data.