Just as an example, 2540 lines per inch (really, line pairs), is 100 lp/mm.

This is certainly within the realm of high-resolution document/copy/litho film and lenses, so currently-available materials would be able to support the line gratings. In fact, in the text, the author states "photographic copies of these (gratings) were used for the experimental work."

As for creating gratings of the required frequency, this is more difficult. Commercially, you'd want to look for custom "Ronchi Rulings", but these are not inexpensive ($1000 per grating for 4x4" chrome-on-glass at Edmund).

A crazy idea I've had floating around in my head for a while would be to get 600- or 900- line-pairs per inch films on a 3600 DPI imagesetter (3-dots black/3-clear or 2-black/2-clear, respectively, or maybe 2-black/3-clear to account for dot gain [720 lp/in]), which you could then make reasonable-ratio reductions (in the range of 2x to 5x, depending on what your imagesetter films are) using a 4x5 camera or enlarger onto litho film. If you're using a white light source, best to use an APO enlarging lens. Can't guarantee the quality you'd get, but I can't see any other way to get there.