The efficiency of liquid chromatography separations could
be strongly
improved by changing the current packed bed columns by a bundle of
parallel capillary tubes. In practice, however, the polydispersity
effect, which emanates from the inevitable small differences in capillary
diameter, completely ruins this potential. The concept of diffusional
bridging, introducing a diffusive cross talk between adjacent capillaries,
has recently been proposed to resolve this. The present contribution
provides the first experimental proof for this concept and quantitatively
validates its underlying theory. This has been accomplished by measuring
the dispersion of a fluorescent tracer in 8 different microfluidic
channels with different degrees of polydispersity and diffusional
bridging. The observed degree of dispersion reduction agrees very
well with the theoretical predictions, hence opening the road to the
use of this theory to design a new family of chromatographic beds,
potentially offering unprecedented performance.