A New Scientist article (with a somewhat amusing title) Fracking Hell investigates this exact question in some detail. The difference is made by an analogue to the process of fracking.
Instead of the water use in human-engineered process of fracking breaking the underlying rocks up to release the methane for commercial purposes. The considerably larger impact of an asteroid does the same thing in a massively uncontrolled manner - in sedimentary rocks, it pulverises the sedimentary rock, releasing the gases that were trapped in it, such as sulphur from evaporites in the vicinity of the Chixculub impact; other gases can include hydrocarbons, halocarbons (salt + hydrocarbons), chlorine to name a few.
The amount of gases released from within the sedimentary rocks were enough to cause mass extinctions (in an essentially extraterrestrial version of what is believed to have occurred during the Siberian Traps event).
In crystalline rocks (such as igneous and metamorphic rocks), according to the article, the rocks are far more stable and do not release such amounts of climate changing gases than sedimentary strata in an impact event. Although these rocks can contain gas bubbles etc, they have been found to relatively 'inert; in terms of gaseous release compared to sedimentary rocks, such as the 100km diameter Manicouagan crater in Canada, an event not associated with a mass extinction.
Essentially, the impacts can be devastating for either scenario, but the impact itself is not the only cause of mass extinctions, it appears that a large factor of the mass extinctions is due to the climatic change that results from the released gases uncontrollably 'fracked' out of the sedimentary rocks. Of course, if an impact is large enough, it could pulverise several layer of rock unlocking deep pockets of gases.