The answer that Dennis gave is absolutely correct, but you should also know that in a way your question is nonsense. You need to remember that all any file contains is a series of binary 1s and 0s.
Now some files we agree to interpret by grouping those bits in groups of 8 and assigning certain agreed-upon logical associations to those 8-bit groups. One set of group we have decided will represent letters, while other groups are interpreted as representing numbers or punctuation. These files we call "text" files because we have all agreed to interpret the 1s and 0s in it in a particular way - as text.
Now there are other files (that we arbitrarily call "binary" files) where we have agreed to not use the 8-bit groups to represent letters or numbers. In fact, sometimes we don't event group the bits in groups of 8 anymore. Sometimes we group them with 16, 32, or even 128 bits in a group. There is nothing inherently different between these files and what we call text files. The only distinction is how we interpret what is in them.
Now you asked about a "hex" file. This is just another way of grouping and interpretting the 1s and 0s. Here, the bits are first grouped by 4s and then each of the 16 possible combinations is interpreted as a number - note that is this interpretation the letters "A" through "F"; get redefined as numbers. The resulting number interpretations are then grouped further into groups of 2, 4 or 8 numbers. But again, the only thing that distinguishes this file from the others is how we have agreed to interpret it.
This is why the answer Dennis gave is true. You can use the binary write function to write any data into any file. Of course if the 1s and 0s you write don't match the assumptions of the program that is going to be opening the file, the write will do more harm than good.
Mike...
