Résumé : The unparalleled theoretical performance of an ideal vector vortex coronagraph makes it one of the most promising technologies for directly imaging exoplanets with a future, off-axis space telescope. However, the image contrast required for observing the light reflected from Earth-sized planets ($\sim10^{-10}$) has yet to be demonstrated in a laboratory setting. With recent advances in the manufacturing of liquid crystal vector vortex waveplates as well as system-level performance improvements on our testbeds, we have achieved raw contrast of 1.57$\times10^{-9}$ and 5.86$\times10^{-9}$ in 10% and 20% optical bandwidths, respectively, averaged over 3-10$\lambda/D$ separations on one side of the pseudo-star. The former represents a factor of 10 improvement over the previously reported performance. We will show experimental comparisons of the contrast achieved and a function of spectral bandwidth. We provide estimates of the limiting error terms and discuss the improvements needed to close the gap in contrast performance required for future exoplanet imaging space telescopes.