On a normal hard disk installation of most any EFI-based OS, you'll have, at a minimum, one FAT EFI System Partition (ESP) and one partition for the OS itself. The ESP holds a boot loader for the OS, possibly along with files to support the boot loader (fonts, configuration files, drivers, etc.), and possibly even the OS's kernel. The OS partition holds more-or-less the same OS files you'd find on a BIOS-based installation of the same OS. Depending on the OS, you might have additional partitions, too -- data partitions, a swap partition, etc.
There can be exceptions to this rule, particularly for installation media or emergency disks. For instance, you could put the whole OS in the ESP. Also, most EFIs are happy to boot from partitions that are not ESPs, so you could just have one big non-ESP FAT partition, as you've got. This can work fine for an emergency disk, but I wouldn't recommend setting up a regular OS installation in this way; I'd use a separate ESP and OS partition.
Note that a standard EFI can read FAT, but cannot read NTFS, ext2/3/4fs, HFS+, or any other filesystem. (Apple's EFI can read HFS+, and so can read its boot loader from a Mac OS X root partition rather than from the ESP, but Apple's EFI is the exception rather than the rule. A few EFIs also have ISO-9660 filesystem drivers -- but again, they're exceptions to the rule.) Because FAT is the only filesystem that's guaranteed to be readable by EFI, an attempt to build a boot disk that does not include a FAT partition is doomed to failure, except of course when used on those unusual EFIs that support additional filesystems.
I can't provide a procedure to set up a Windows emergency disk to use separate EFI and Windows partitions, since I'm more of a Linux person than a Windows person. Unless you run into a specific problem with your approach, though, I'd just stick with it; at least you know it works.
From Resolving the dreaded Windows 8 0xC0000225 error :
I've had this happen to me twice, all of a sudden Windows 8 refuses to
boot and gives an error code of 0xC0000225 (or something) and anything
you do won't fix it. The problem is that since for whatever reason
Windows thinks there's no bootloader, it refuses to boot. Period. You
can't even go into the recovery environment, which is my biggest
complaint at Microsoft.
In any case, there's one way I've found to resolve this. If you have
access to another computer, take out the affected computer's primary
drive and find a way to mount it in the system (internally,
externally, whatever). Then do the following:
- Open command prompt as an administrator
- Type in "diskpart"
- Type in "list disk". Find out which disk the affected drive is.
- Type in "select disk #", where # is the affected drive's number
- Type in "list partition", find the partition number of the system partition (it's usually 100MB, 200MB, or 300MB), then type in "select
partition #", where # is the system partition's number.
- Type in "assign letter=z", assuming you don't have a Z: drive.
- Exit out of diskpart by pressing CTRL+C
- Type in BCDBoot
[Drive letter of affected drive's Windows partition]:\Windows /S Z: /F UEFI
So if the affected drive's
Windows partition is say G:\, you would type in BCDBoot G:\Windows /S
Z: /F UEFI
It should fix the bootloader. If you can get into a recovery
environment on the affected computer, then this should work as well.
Best Answer
Format the USB flash drive as FAT32.
Then use RMPrepUSB to make the FAT32 USB flash drive bootable to use it to install Windows 8. AFAIK, only RMPrepUSB supports FAT32 and all other (Microsoft USB/DVD Download Tool/Rufus/etc.) support only NTFS.