About this Author
DBL%20Hendrix%20small.png College chemistry, 1983

Derek Lowe The 2002 Model

Dbl%20new%20portrait%20B%26W.png After 10 years of blogging. . .

Derek Lowe, an Arkansan by birth, got his BA from Hendrix College and his PhD in organic chemistry from Duke before spending time in Germany on a Humboldt Fellowship on his post-doc. He's worked for several major pharmaceutical companies since 1989 on drug discovery projects against schizophrenia, Alzheimer's, diabetes, osteoporosis and other diseases. To contact Derek email him directly: Twitter: Dereklowe

Chemistry and Drug Data: Drugbank
Chempedia Lab
Synthetic Pages
Organic Chemistry Portal
Not Voodoo

Chemistry and Pharma Blogs:
Org Prep Daily
The Haystack
A New Merck, Reviewed
Liberal Arts Chemistry
Electron Pusher
All Things Metathesis
C&E News Blogs
Chemiotics II
Chemical Space
Noel O'Blog
In Vivo Blog
Terra Sigilatta
BBSRC/Douglas Kell
Realizations in Biostatistics
ChemSpider Blog
Organic Chem - Education & Industry
Pharma Strategy Blog
No Name No Slogan
Practical Fragments
The Curious Wavefunction
Natural Product Man
Fragment Literature
Chemistry World Blog
Synthetic Nature
Chemistry Blog
Synthesizing Ideas
Eye on FDA
Chemical Forums
Symyx Blog
Sceptical Chymist
Lamentations on Chemistry
Computational Organic Chemistry
Mining Drugs
Henry Rzepa

Science Blogs and News:
Bad Science
The Loom
Uncertain Principles
Fierce Biotech
Blogs for Industry
Omics! Omics!
Young Female Scientist
Notional Slurry
Nobel Intent
SciTech Daily
Science Blog
Gene Expression (I)
Gene Expression (II)
Adventures in Ethics and Science
Transterrestrial Musings
Slashdot Science
Cosmic Variance
Biology News Net

Medical Blogs
DB's Medical Rants
Science-Based Medicine
Respectful Insolence
Diabetes Mine

Economics and Business
Marginal Revolution
The Volokh Conspiracy
Knowledge Problem

Politics / Current Events
Virginia Postrel
Belmont Club
Mickey Kaus

Belles Lettres
Uncouth Reflections
Arts and Letters Daily
In the Pipeline: Don't miss Derek Lowe's excellent commentary on drug discovery and the pharma industry in general at In the Pipeline

In the Pipeline

« Exciting Nonsense Wins Another One | Main | Unacceptable »

December 3, 2007

The Big and the Little

Email This Entry

Posted by Derek

I was talking about vacuum distillation and recrystallization the other day, and several people pointed out in the comments that large-scale chemistry still relies on these techniques. That’s absolutely right, especially recrystallization. It’s all a question of scale.

These two are bulk techniques – they work out fine on reasonable amounts, but they’re very difficult to run on a microscale with conventional techniques. Distilling a kilo of something is just about as much work as distilling ten grams of it – but distilling ten milligrams, now that’s something I wouldn’t want to be in charge of. Going the other way, distilling ten kilos starts to take you into a another different world, and one of the main reasons operates across the entire scale: surface to volume ratio.

In a distillation from a really large flask, you have to find efficient ways to heat the thing, because just sticking it into a really, really big heating mantle or oil bath gets to be problematic. The surface area of the flask is going up as a square, and that’s what you’re depending on to transfer to heat to the inner volume. But that volume’s going up as a cube. In a 100-mL flask, no part of the contents is more than two or three centimeters from the wall, whereas in a 100-liter system the commute from the edge has grown to something pretty substantial.

Your heating problem is also a mixing problem, since those technologies don’t scale smoothly, either. In a 100-mL flask you can drop a good-sized magnetic stir bar in and whip the solution around smartly with the spinning magnet of a regular stir plate. A proportional stir bar for a 100-liter flask would be a real brick, liable to crash right through the walls of the flask, and you’d need some sort of diesel-powered stirring plate to spin the thing. Needless to say, there are plenty of heating, mixing, jkl and distilling methods for the industrial scale – the existence of gas stations is a testimony to that – but they don’t look much like what people like me use to make twenty milligrams of a test compound.

So much for the big stuff, now take it down to the ten-mg scale. The area-to-volume problem is now reversed. You can’t buy a proportionally sized distillation head, because you’d need a deranged artist of a glassblower to make one. Your tiny volume of solution will just spread out and coat the insides of the smallest distillation rig available. You’re working far within the error and loss of a normal distillation, and your sample will disappear into this gap and never be seen again. I can imagine some sort of microscale rig made out of glass capillaries, although I’ve certainly never seen such a thing. (Surface tension would surely start to become an issue with its operation). Microfluidics is a hot research area, but as far as I know they’ve yet to move on to distillation. I hope someone gives it a shot.

Now consider chromatography. Ten milligrams is plenty of material to work with on an HPLC system. (And if you’re just interested in analysis, and not isolating preparative amounts at the end, ten milligrams becomes a mountainous heap). But running an HPLC on 100 grams is pretty much out of the question. Running even a normal column on that scale isn’t much fun, and when you head up to ten kilos it becomes a major undertaking that you’d do all kinds of things to avoid. (Like, say, spending a week or two trying out recrystallization conditions). The amount of solvent become really substantial, as does the expense and trouble of handling it. It’s not that chromatography doesn’t get done on large scale, it’s just that it gets done only after better alternatives have been completely ruled out.

Recrystallization goes up and down the scale a bit better than these other two techniques. It’s tricky to do on a small scale, but if you have a good solvent combination to form the right kind of crystals you can recrystallize a ten milligram sample if you absolutely have to. One problem with trying to use the technique on that scale is that it generally takes a lot of messing around with the conditions to get a good system, and if you only have ten mgs you probably can’t get away with that. I’d much rather run that sample down an HPLC, though, believe me.

Crystals are much more fun when you’re making a few grams, where you don’t have to worry about every single bit stuck to the sides of the glassware. And the folks working on larger scale just love recrystallization more than anything. It’s true that you have to heat things up at the start, but the heating doesn’t have to be done as critically as in an actual reaction, since you’re generally just trying to get things to dissolve. And once everything has cooled back down and the new crystals have fallen out of solution, it’s just a filtration and wash, and that’s something that can be done well even on a gigantic scale.

Comments (17) + TrackBacks (0) | Category: Life in the Drug Labs


1. psi*psi on December 3, 2007 10:43 PM writes...

If anyone else has made and recrystallized your compound...the solvent is listed in Beilstein. Makes things a little easier.

Permalink to Comment

2. As You Lean on December 3, 2007 10:52 PM writes...

I always find it interesting on lab cleanup days to open the cabinet with the ginormous glassware and later open the drawer with the absolutely tiny glassware and marvel at the scale of it all.

Permalink to Comment

3. milkshake on December 3, 2007 11:13 PM writes...

I heard of scaling up column chromatography in process for making prostanoids - it was a highly potent compoundto so they were only making 10kg batches of it. To save on the non-recyclable solvent mix they run the column only halfway through, in a disposable polyethylene pipe. Then they sampled the column by using a drill, to find the exact position of the zone of purified material. They cut out the zone of purified material and extracted it from silica.

This cutting-edge process column technique gives me shudders. Sometimes the humongous protecting groups like trityl are good - to make things crystalline and UV active.

Permalink to Comment

4. Lucas on December 4, 2007 12:34 AM writes...

At my previous^n company, I was sent to one of our manufacturing facilities in Europe. While there, they walked me through the few non-sterile manufacturing areas. One of the products was so potent (sub-microgram/day) that an entire batch was a few grams, and easy to purify on a column.

And the sucker was still huge. Like, 3 meters long.

Gave me an appreciation for the difficulties of working on scale, it did.

Permalink to Comment

5. flaming crucible on December 4, 2007 2:03 AM writes...

On scale, in the pharma industry re-cryst is not just a matter of heating to dissove and cooling to colect the crystals. You may need to filter the hot solution, you most certainly will want to seed the solution and may even need to operate a complex heating and coolin system to obtain the correct polymorph. And that can be a real "*ç" .

Permalink to Comment

6. processchemist on December 4, 2007 6:44 AM writes...

The trouble with large scale chromatography is that you need special equipment... on kilolab scale is a bit simpler (I used to run columns with 12 Kg of silica gel). Anyway Simulated Moving Bed chromatography is an industrial reality (escitalopram is manufactured this way).

Derek's post underline the existence not only of scale up problems, but also of *scale down* issues...

Permalink to Comment

7. tyrosine on December 4, 2007 11:34 AM writes...

Preparative gas chromatography can be viewed as a type of small-scale distillation. A lot of highly strained fused-polycyclic cycloalkanes are isolated in this manner.

Permalink to Comment

8. milo on December 4, 2007 11:57 AM writes...

Using a megabore GC column and a good temperature gradient makes it really easy to trap enough material for a good set of 2D NMR experiments.

There are two main issues, IMO, with recrystallization: 1: often the solvents needed pose real containment and disposal issues, like DCM and 2: not everything produced in industry is a solid. Now, it is often possible to distill solids fractionally, if the mp if below say 100 C... but that has its own issues.

I think everyone should experience large scale (10-15 kg) chemistry at least once in their life, prefereably before they leave graduate school. It is amazing how humbled people become when they realize they cannot easily run a column.

Permalink to Comment

9. Jose on December 4, 2007 12:01 PM writes...

Along the lines of Milkshake's comment, there is a lab scale technique along the same lines, using silica and flexible nylon tubing, at the end you slice it like a sausage. I've thought about trying it, but never got over the energy barrier to set one up.

Permalink to Comment

10. Milo on December 4, 2007 12:50 PM writes...


I used that a lot in grad school, works like a charm and is really easy!

Permalink to Comment

11. Bunsen Honeydew on December 4, 2007 1:23 PM writes...

Didn't Novartis in Basel set up something like a 20 metre high column for purifying its discodermolide intermediates? I toured the site and I seem to remember someone telling me about it. It's too bad that the compound was too toxic. Alas, disco truly is dead.

Permalink to Comment

12. .... on December 4, 2007 4:46 PM writes...


I don't get this DCC technique. How do you find your compound if it isn't brightly colored?

Permalink to Comment

13. MTK on December 4, 2007 5:14 PM writes...

In process chemistry you usually make things many times over and progressively larger scale, so taking some time to find some recrystallization conditions is well worth the effort.

After some practice, it doesn't take any time at all to find reasonable crystallization conditions. A day, maybe two tops. Now finding conditions that give you maximum purity with maximum recovery, in the proper polymorph, crystal form, and particle size...that's a different story. BTW, if you've ever had to collect 80 kgs of something and realized the particles you've made are too small and clog the filter, you soon appreciate how important it is to control crystal growth vs. nucleation.

We had a small group of people that had it down, however. I would give them some compound and within a month or so they'd have a complete protocol worked out. The great thing was that this protocol would be completely scalable and predictable. Dissolve at X temp in Y liters/kg of your dissolving solvent. Add Z liters of your anti-solvent at a certain rate while stirring at a defined speed. Add your seed crystals at time point N and cool at M degrees/min then hold at P degrees for Q hours. Filter, dry, voila! Your API in the desired polymorph with the desired particle size. The first time you see it work on scale, it's a thing of beauty. We once had a project that required about a 24h filtration period on manufacturing scale. The crystallization group got on it and before you know it that baby was down to 2h. That might not mean to much to a lot of us, but if your the plant manager, who replaced the old manager in a cost cutting move, that means a lot.

Permalink to Comment

14. PDchem on December 5, 2007 12:04 AM writes...

@ 13:

You just took the words right out of my mouth, very well put! I work in process development and I also learned to appreciate the importance of controlling particle sizes in crystallizations and their impact on filtration times very quickly. As a rule of thumb we say that when it does not filtrate in the lab over a type 3 glass sintered funnel within 30 seconds, you are bound to run into trouble in the plant with long filtration times.

Permalink to Comment

15. flaming crucible on December 5, 2007 12:13 AM writes...

That was me. The column already existed, it was built to purify phospholipids. Actually it is regular use as it's companion on the other side of the Rhine

Permalink to Comment

16. process on December 5, 2007 4:23 AM writes...

@13 & 14

And what about crystal dimensions and efficiency in cake washings? Crystal dimensions can improve dramatically not only filtration time but also drying time, residual solvents profile and so on...

Permalink to Comment

17. bcpmoon on December 5, 2007 7:10 AM writes...

Perhaps this is comforting for the chromatography aficionados: Chromatography is used in the sugar industry and that means serious business.

Permalink to Comment


Remember Me?


Email this entry to:

Your email address:

Message (optional):

The Last Post
The GSK Layoffs Continue, By Proxy
The Move is Nigh
Another Alzheimer's IPO
Cutbacks at C&E News
Sanofi Pays to Get Back Into Oncology
An Irresponsible Statement About Curing Cancer
Oliver Sacks on Turning Back to Chemistry