Wednesday, March 5, 2014

Process Wednesday: photooxidation to get industrial scale artemisinin

Credit: OPRD, Turconi et al.
From the pages of Organic Process Research and Development [1], a fascinating article by workers at Sanofi-Aventis talking about their industrial route to the anti-malarial artemisinin. They decided to use a photooxidation to perform the semi-synthesis of the material and here, they talk a little about the challenges:
The utility of the Aubry reaction allows the generation of singlet oxygen, and thus the Schenck ene reaction, while avoiding the need for photochemical equipment. While in principle highly attractive, photochemical reactions are rarely practiced in industry25 and there is even less precedence for the large-scale application of the Schenck ene reaction with photochemically generated singlet oxygen.  
This is easily understood, as the hurdles to the implementation of a large-scale Schenck ene reaction are significant. Apart from the lack of experience and equipment, the choice of permissible solvents is very limited, the resulting products are inherently unsafe hydroperoxides, and the chemical engineering aspects such as mixing, gas transfer, and light transmission are unusually complex on scale.
More on the chemical engineering challenges:
On the basis of these results, a dedicated pilot unit was set up at Sanofi facility, Neuville (France). We selected a semibatch mode concept with a recirculation loop, which is possible as artemisinin is very stable under the reaction conditions and it offers optimal conditions versus energy consumption and the transformation rate. 
In addition to the other chemical engineering challenges, critical parameters included construction materials that minimize the loss of light (optimizing the quantum photonic yield) and the choice of a lamp with the optimal spectral distribution of emitted spectrum (medium pressure mercury/gallium lamp). The unit must also ensure a good turbulence of the recirculating fluid and allow a good gas−liquid transfer, while maintaining the internal temperature at −10 °C.
The first batch was piloted at 50 kg scale and now they're expected to run 60 tons a year, at a 370 kg batch size. That's 162 batches (?), which indicates a pretty impressive cycle time, although these numbers tend to get fuzzy around the edges (i.e. assuming running the plant at 100% capacity, which may not happen, etc.)

Of course, you have to read the Experimental Section* to get the full flavor of this:
Step 3: Photooxidation of Mixed Anhydride (DHAEMC) to Artemisinin. To the solution of DHAEMC (CJ's note: they started with 600 kg of artemisinic acid, 2 steps previous) were added 2570 kg dichloromethane and 300 g tetraphenylporphyrin (TPP) before the solution was exposed to light irradiation using photoreactors containing mercury vapour lamps and ambient air bubbling at about −10 to −15 °C. In the beginning of the irradiation 132 kg trifluoroacetic acid was added to the reaction mixture. The reaction was monitored by HPLC. As soon as the reaction was completed, the solution was treated twice with 720 L aqueous solution of sodium bicarbonate and was subsequently washed with 1440 kg water. The washed organic phase was treated with 30 kg activated charcoal and filtered. Before the crystallisation step,
the ADT24h of the reaction mixture was checked.
I don't have the professional capacity to comment on this intelligently, other than to say that it's a really impressive feat, in my opinion. I think it's apparent that the Sanofi coworkers not only had to design (and did!) robust enough chemistry to get it done, they had to design reactors to make this happen. Finally, it should be noted that Sanofi-Aventis was working with the Gates Foundation under a "no profit, no loss" model for anti-malarials. This seems like a pretty clear example of the good that industrial chemistry can do.

1. Turconi, J.; Griolet, F.; Guevel, R.; Oddon, G.; Villa, R.; Geatti, A.; Hvala, M.; Rossen, K.; Göller, R.; Burgard, A. "Semisynthetic Artemisinin, the Chemical Path to Industrial Production." Org. Process. Res. Dev. ASAP DOI: 10.1021/op4003196

*OPRD's experimental sections are the best, because it's always "To the reactor was added 14 quintillion liters of solvent..." 


  1. This work is a tour de force - Sanofi has done a brilliant job on a really important project. I saw this presented a couple of years ago at a Gordon - truly impressive stuff. My hat is off to the Sanofi Process team - Chapeau!

  2. Really a great piece of work!

  3. "...photochemical reactions are rarely practiced in industry..."

    From the perspective of someone thinking that all chemistry happens in a kettle reactor, yes that is true, but they've been used in industrial coatings for decades and their use is growing. Sometimes your reactor is long, thin and wide, and maybe even moving.

    1. Some days I do more web handling work than chemistry, but photochemistry on coatings is huge because it's "clean". At least I like it because of that.

  4. I always get a chuckle out of reading these things, as I try to compare my own little microgram scale sodium bicarb/water washes with "720 L aqueous solution of sodium bicarbonate and was subsequently washed with 1440 kg water". I like to imagine a giant doing organic chemistry and carrying out these reactions.

  5. I recently heard Peter Seeberger speak about this work, which was originally developed in HIS lab (not Sanofi). The paper does mention this but only in passing. His latest work by which he makes the actual drug entities, all in flow, will be published shortly and dwarf this work allowing for the production of the compounds at a fraction of the cost. Sanofi is not thrilled about this.

    I also wanted to comment on the Gates/Sanofi partnership. The Gates foundation purchases the compounds "at cost" which is determined by Sanofi. The cost is typically 2-3 times higher than that of other vendors so one would ask - why wouldn't the foundation buy them from a cheaper source and help more people? The answer might come from the fact that the Gates foundation is a "significant share owner" of Sanofi, therefore any profit Sanofi makes is returned back to the foundation. The overlap of corporation and philanthropy is not always as well intentioned as depicted.

    1. Simply not true! Sanofi shared their results at Artemisinin Conferences and filed patents a few years before Seeberger's work. Sanofi is displeased about Max Planck Institute filing patents on what is essentially their chemistry applied to flow.

      You are also quite wrong about the Gates/ Sanofi partnership. Gates funded initial research into development of fermentation and chemical transformation of fermentation products to artemisinin. Sanofi benefited from this research but also needed to develop the chemistry described in this publication. Gates does NOT purchase any compounds - they support research. The Sanofi prices are actually comparable to natural sources.

      Sanofi is not blameless in the lack of increased access to anti-malarial treatments devolving from semi-synthetic artemisinin but it is ignorant to accuse them of pirating Seeberger's work and to suggest that Gates somehow profits from this!

  6. @10:01 - I have a hard time understanding how anyone can belittle the terrific work of their chemistry colleagues. Critique is fine, but this comment is completely unprofessional and unwarranted. If a paper comes our improving on the synthesis, then great - we should always strive to improve, but an improvement would in no way detract from the good work that has come before. And I would like to see how the new chemistry is then implemented at scale - a synthesis is not a process.

    Really, your comment is way out of line.