I did try to run gradient from 100% toluene to 100% methylethylketone using 150x4.6mm column packed with microporous (3nm pore size, 5micron particles Develosil) bare silica and evaporative light scattering detector to separate polymer blend containing several polymethacrylate homopolymers (Flow rate 1mL/min, injetion volume 50microlitre, concentration 1g/L). It was just to reproduce the results obtained long time ago by T.M Mourey (J. Chromatogr A 357, 101 (1986)).
One of the problems occured here was the "ghost" peak which alway appeared at the beginning of the BLANK gradient (gradient profile was monitored with density detector). Certainly, this peak was far behind the peak attributed to solvent peak, if any. In other words, elution volume was much higher than void volume.
Another problem was very poor resolution, peak broadening and tailing. For example, I could hardly get baseline separation of polyethylmethacrylate (PEMA) from polymethylmethacrylate (PMMA).