Actually, 50% of surface has already been read.
"Reading successfull but null" (light blue colour square) are sectors that have been read but they don't contain any data (00 filled).
It is still necessary to read 50% of surface, which is relative to the faulty head.
Try to read surface on head 1 at slowest UDMA 33 speed, if still gives problem try to decrease head block reading (example at 32 or even at 1).
If still have instable reading, you might try in PIO mode strarting from largest block, decreasing block reading may help reducing instable head because it reads sectors slowly.
If still doesn't read (or if is too slow) then you've to swap head assembly in clean room.
P.S.
If head reads but at very slow speed, you might jump all those 00 sectors, so that you will save time and reduce head stress.
To do that you should go to File Explorer in MRT so that it will access to filesystem and select just needed files/folders and clone sectors relative only to those data.
P.S.
You might folllow this video for best filesystem reading (mrt is not very good on showing folder structure) by using UFS Explorer software which can interact with MRT
https://youtu.be/NV3WY3WBqsI?si=5x0tBGdc7bdVA6_W&t=30Demo version is enough
https://www.ufsexplorer.com/