Isoprene is 2-Methyl-1,3-butadiene, photochemically a highly reactive organic compound, owing to the two reactive olefinic groups (double carbon-carbon bonds) in its structure:
It’s also of interest in air chemistry because it’s biogenic; the major source of emissions of isoprene into the atmosphere is hardwood trees and similar plants. Softwoods emit terpenes, the most common of which is alpha pinene, which is what turpentine is made of. Terpenes themselves are biologically synthesized from isoprene, but the terpene emitters don’t emit much isoprene.
The terpenes are even more reactive than isoprene, but that reactivity actually tends to reduce their ozone forming potential, for two reasons. The first is that they are so reactive that they react very rapidly with ozone itself. The second is that, after reaction, especially with ozone, the high molecular weight products that are formed tend to have a low enough vapor pressure that they form aerosol particles. Some amount of naturally forming haze comes from the reaction of terpenes with ozone, and it’s often said that the Smokey Mountains got their name from this pre-industrial haze.
One of the olefinic bonds in isoprene also has a methyl group associated with it, and this alkylation alters the electron distribution of the double bond and makes it more reactive. So that bond tends to be the primary reaction site, with the result that the product formed (after the usual smog reaction hugger mugger) is preferred. The usual product of that reaction site is methyl vinyl ketone, while the other site reactions give methacrolein, an aldehyde with a double bond in it.
Both MVK and methacrolein have interesting further reaction sequences. MVK will further react to give methyl glyoxal, which is also a product of the photooxidation of alkylated benzenes like toluene and xylene. Methyl glyoxal photolyzes like crazy, and it turned out to be the powerful radical source in the toluene system when we were earlier studying toluene.
Methacrolein primarily reacts via hydroxyl (OH) grabbing onto its aldehydic hydrogen, which give a peroxyacyl like compound. Faithful readers will be reminded of peroxyacetyl nitrate which was the first peroxyacyl nitrate (PAN) to be studied in smog chemistry, and it remains the most important. But the methacrolein analog to PAN turns out to be temporarily important in the photooxidation of isoprene. It serves as a temporary sink for both reactive radicals and nitrogen dioxide.
The nitrogen oxide sinks due to the PAN analog plus PAN itself cause an isoprene-NOx system to run out of nitrogen oxides relatively early in a smog chamber experiment, sometimes before the isoprene itself is gone. If that happens, the ozone may begin to decline (owing to the exhaustion of the NOx). But if the temperature continues to increase, as it does in smog chambers that don’t control the temperature (as is the case for outdoor chambers), then the thermal decomposition of the PANs will bleed both NOx and reactive radicals back into the system and the ozone will begin to climb again.
The first outdoor smog chamber experiments showed just such a phenomenon, resulting is odd looking “double peak” experiments. By selecting the appropriate value for the thermal decay of the “methacropan” and by paying attention to the radical-radical reactions that determined PAN decay generally, I was able to fit those experiments using gas phase chemistry while everyone else was suggesting that it was all due to “chamber effects.” Given a broad enough definition of “chamber effect” that was certainly true, but the actual effect that was responsible was simply the increasing temperature as the day progressed, a phenomenon that isn’t confined to smog chambers.
A Google search on “isoprene photochemistry” gives 45,000 hits. “Isoprene photochemistry Killus” gives only 200. Using Google Scholar gives 1740 and 84 respectively. That’s either 0.5% or 5%, not all that big a chunk either way. But for a brief while in 1983-4, I knew more about the photochemistry of isoprene than anyone else on the face of the planet.