If they do not require corrections, one of those 2x chinese relay lenses (like the amscope) should be fine, or perhaps better a nikon 2.5x cf relay. The nikon ones are very nice, and available used cheaper than the 2x. In either case you will not need any intermediate optics, just straight to the sensor.
The eyepiece is the intermediate optic, whether the set up is afocal or if you use a photo tube with an eyepiece installed as a relay lens. -----
The principle difference is that in an afocal system, the tube length is automatically set to the optical tube length of the system and if you want a representative image of the view you are seeing through the microscope, and your sensor size is about the same as the f.o.v., then you are best to use the properly corrected eyepiece that you view through. If you are not getting a perfectly plan and corrected view through the eyepieces, then the photo image will likely be worse. A photo eyepiece used in that situation would have to be a) perfectly corrected for the objective or objective/telan lens combination and b)magnify the same amount as the viewing eyepiece.
With an APS-C sensor, which is approx. equivalent in size to a 20mm f.o.v., and a 2X eyepiece used afocally to photograph a f.o.v. that has been viewed with 10X eyepieces, you will get a huge f.o.v. but also magnification that is reduced by 5X. In order to present the image on a screen that roughly equates to what the eye sees at a 20mm f.o.v., the image will have to be cropped enough to compensate for the reduced magnification. There will be no gain in resolution by virtue of magnifying the original image less. Further , the likelihood of the 2X eyepiece to have been corrected for the optical tube length of the system is low.
When using an eyepiece as a relay lens in a photo tube, the situation becomes more difficult because photo tube specifications are seldom the same between any two manufacturers. Thus the reason for the extreme variability between the marked magnification on the eyepiece and what it provides with a tube length that is not original. For a photo eyepiece to perform up to needed specifications in terms of magnification and corrections when used outside of it's intended application it must be tested. Many do not make a claim of the intended tube length( Olympus does) but most use a different tube length than the optical tube length, so in order for the corrections to be accurate, the originally called for parameters need to be adhered to.
You can't really tell if an eyepiece has no chromatic corrections, unless you test it. All eyepieces do to some degree, unless the system they are from is precisely designed to not require chromatic corrections and then the precise engineered photo tube length would need to be adhered to as well. Eyepieces that are corrected, have varying corrections usually. Two over compensated eyepieces not necessarily will be over compensated to the same degree and some objectives require under compensation, so assuming that any given eyepiece will work such and such a way in a photo system , is the primary reason there are so many failures when diy'ing a photo system.
It really is a question of testing the entirety of the system , if possible finding one that is in use and works well, or adhering to the manufacturer's specifications for the set up.