You know what I don't miss about chemistry after years in the drug industry? Big, long, multi-step syntheses. Oh, we'll gear up to do eight- and ten- and thirteen-steppers here, even though some of those steps are just things like hydrolyzing methyl esters, stuff that blindfolded grannies should be able to do. But what I'm happy to leave the mighty academic natural product synthetic schemes behind, the ones where step fourteen finds you just getting warmed up.

As I've mentioned here before, I did that kind of thing in graduate school, and I swear it's scarred me for life. I pulled the plug on my total synthesis at step 27, about six steps short of the end (this is, if everything had worked perfectly, obese chance.) I've never regretted it. The benefits of getting out of grad school are huge, spacious, and well-appointed compared to the benefits of being able to say that I finished my natural product. Any of my readers in grad school, take note.

Long linear sequences are a slog. You have to start them in the largest buckets you can find, because you're never, ever going to have enough material. Now, we do large scale work in the drug industry, yes indeed, but that's because we intend to finish on large scale. If you're going to do six-week toxicity testing, you'd better have a fine keg of material on hand before you start. But those academic syntheses need huge amounts at the beginning in order to have anything at all by the time they finish. You work until you can't handle or characterize the stuff any more, then you trudge back down the mountain and start porting the loads back up the trail.

An example: I got to the point where I needed to take an optical rotation on the material from about step 25 or so. For those outside the field, this is an analytical technique that involves shining polarized light through a solution of your compound. If it's not an even mix of left-handed and right-handed isomers, that is to say, if there's some chiral character to the sample, the light will rotate. The degree of rotation can be used as an indicator of compound purity - I'm tempted to add "if you're a fool." They're not the most reliable numbers in the world, because some things just don't make the light twist much. And in those cases, a small amount of an impurity that rotates light like crazy will throw everything off. It's happened more than once.

Well, in my case, I loaded a half milligram or so of my precious stuff into the smallest polarimeter tube we had and jammed it into the machine. Hmm, I thought, a rotation of 0.00 degrees. A singular result, since I knew for certain that the molecule had six pure chiral centers. So I went back upstairs and loaded the whole batch into the tube, walking very carefully down the hall with this investment of several months of my life held in both sweaty hands. This time I got a specific rotation of about 1.2 degrees, which means that all those chiral carbons were roughly canceling each other out. Did I believe that number? Not at all! Did I put it in my dissertation? You bet! Gotta have a number, you know.

And that's how you work - purifying things through increasingly tinier columns, collecting them in slowly shrinking vials, running all the instruments for longer and longer with the gain turned up higher and higher, trying to prove that it's really still in there and really still what it's supposed to be. Then it's back to the buckets. Never again!