My friend was just an engineer. This is not a firm I referred to.
I read the article, that must have huge leakage when used in the photo showing the guy controlling the unit. I don't suggest doing similar unless you make your own sealed RF chamber.
I suggest something like using a microwave oven, most have rf coming out of the side wall vertically polarized. Cut a thin vertical slit in the wall opposite the source power. Have the paper feed thru this slit into the oven Have it wrap around a vertically oriented low loss dielectric tube (you'd have to install that), then back out the same slit opening. The RF will dry it as it moves towards and away from the source. To reduce leakage out this slit (from crosspole energy), add an extension to the slit (essentially a piece of flat waveguide, say 5-20 inches long (the longer the better for leakage suppression), as wide as your vertical slit, and narrow enough to fit some Magnetically resonant absorber inside and have the paper come thru.
I've done a similar thing winside a microwave oven whereby I had a waveguide to coax transition inside sitting atop a water filled tupperware. A coax attached to the waveguide transtion funnelled rf out the modified door (I added a bulkhead N-f to N-f adapter to the door to get the rf out). We didn't have any arcing problems and now had a cheap variable power RF source.
kch