For a circular star image, the amount of light passing through does not increase proportionally to the slit width so the penalty is not as large as expected. Having said that,I would expect your scope optics to be seeing limited on axis so for good throughput I would aim for a slit width equivalent to around 3 arcsec for typical uk seeing. (I use a 35um slit with my C11 at f10). Is your scope f8? in that case 23um might be better for you and give higher resolution. You will get a bit of vignetting in the spectrograph at f8 though which will lose you a bit of light compared with my setup at f10. (The spectrograph optics are f8 but you have to allow extra for the angle between the undispersed and dispersed beams so an f8 beam will not quite fit through the optics)
Robin