I love a happy story

Also from this week's C&EN, the story of a first scale-up attempt: 

The article on continuous processing (C&EN, April 29, page 28) recalls my first experience with the concept. It was in 1988, shortly after I joined DuPont. We had a Grignard reagent to be reacted with an α-chloroketone to make a chlorohydrin intermediate to DuP 860, an antifungal drug candidate. The addition generated so much heat, on scale it couldn’t be kept at the temperature range required for stability over a reasonable reaction period. Instead we fried our first attempt and had to work day and night in the lab running 12 L reactors to make the amount required. 

My engineer, Mark Lauritsen, considered our resources at our pilot plant at Chambers Works in New Jersey and asked me for a 4 L jacketed glass reactor. We set up a continuous process in the midst of a bay with two large feed tanks (Grignard reagent and α-chloroketone) leading into the glass reactor and an overflow line leading out to a 100 gal (378.5 L) quench tank. The tiny glass reactor, dwarfed by the other fixed vessels in the bay, seemed so out of place. But once it started up, the continuous process worked perfectly and we made our delivery! Such a logical and simple solution, and my first experience with continuous processing. 

Jaan Pesti
Yardley, Pennsylvania

That's one of the cool things about flow, i.e. the reaction vessel doesn't have to be very big at all to do a big job.  

Great post by th'Gaussling

Mettler-Toledo couldn't possibly get a better recommendation for its reaction calorimeters than this post by th' Gaussling: 

There’s a good argument for a chemist to conduct RC experiments as well. A trained synthesis chemist is qualified to conduct chemical reactions within their organization. That includes sourcing raw materials, handling them, running the reaction, and safely cleaning up the equipment afterwards. But interpreting RC1 data has a large physical chemistry component. In my experience, run of the mill inorganic/organic synthesis people may have seen PChem as an obstacle rather than a focus in their college education. Their skill set is in instrumental analysis like NMR and chromatography, mechanisms, and reaction chemistry. I would recommend having a PhD chemist with a focus on thermo in a leadership role when calorimetry is a key part of a busy process safety environment. 

Safety data can be collected and archived all day long. The crucial and often tricky part is how to develop best practices from the data. I would offer that this is inherently a cross-disciplinary problem. Calorimetric data from reaction chemistry can be collected readily, especially with the diverse and excellent instrumentation available today. Adiabatic temperature rise, ΔTad, can be determined by a chemist, but it’s the engineers who understand how the equipment may respond to a given heat release. A smooth and efficient technology transfer from lab to plant happens when good communication skills are used. Yes, SOP’s must be in place for consistency and safety. But the positive effect of individuals who have good social skills and are prone to volunteering information cannot be underestimated.

Good stuff, with lots of relevant details around purchasing and setting one up.  

Process Wednesday: couple of short items edition

Couple of things: 

• Via Phil Baran's Twitter feed, this Army report in Organic Process Research and Development about a "greener" synthesis of methyl nitroacetate.
• A cGMP continuous flow synthesis of an oncology drug candidate by Lilly workers (C&EN article by Beth Halford, Lilly report in Science)

Enjoy! 

Process Wednesday: "seasoning"?

In an article titled "Industrial Suzuki chemistry" in the "Special Publication" section of the September 2016 issue of Speciality Chemicals (always a good read), the authors (John C. Parks and Eric L. Williams of Albemarle) mention an interesting technique I have not heard of yet:

Suzuki reactions are prone to catalyst poisoning. This means that it is wise to clean reaction equipment diligently and maintain a dedicated reactor for the Suzuki step. For very high value chemicals, Albemarle will run a reactor seasoning batch before starting the campaign. 

A seasoning batch is typically the desired Suzuki reaction run at one-tenth the normal concentration. If the seasoning batch runs to ~90% completion, we generally consider the reactor seasoned. 

I've never heard of this, has anyone else? I'm not quite sure what they mean by "seasoned." This usually makes me think of "seasoning" a cast-iron pan, i.e. polymerizing a protective oil coating on the surface. In this case, I suspect that it is a test to see whether or not there are catalyst poisons (sulfur compounds, etc) on the reactor wall? It'd be interesting to know if there was a difference between different reactor materials (glass, stainless steel, Hastelloy.) 

Process Wednesday: reactions with galvanized drums

I have been remiss in not tracking Scientific Update's process chemistry blog over the last couple of months. They have been posting with some regularity and adding some content from recent presentations at conferences that they run. Here's a nice bit of safety-related information written up by Will Watson: 

At the recent 'Scale Up of Chemical Processes Conference' in Baveno, Italy, Wim Dermaut from Agfa presented some case studies where problems have occurred as a result of reactive chemicals present in waste streams. In the first case study a drum containing a waste stream including diethylhydroxylamine exploded. The most likely cause was caustic present in the empty drum which initially raised the temperature of the waste to a temperature that interaction with Zn led to the runaway reaction (the waste drum was galvanized). 

Interestingly the supplier of diethylhydroxylamine maintained that it was safe as they only saw an endotherm in the DSC, whereas Agfa’s own studies showed an event with an onset at 105°C with a very rapid decomposition. The key difference was that the supplier had carried out their DSC run with an open cell. 

This is the first time that I've heard of a galvanized drum causing a runaway reaction, but I have heard of cases where zinc contamination comes from a galvanized drum. 

Process Wednesday: huh, that's a lot of "NEVER"s

Via a random Googling, I happened upon Dow's "Dow Answer Center" with a few helpful suggestions for handling acrylic acid:

• NEVER thaw frozen acrylic acid with steam!
• NEVER remove acrylic acid liquid from a partially thawed container!
• NEVER allow an incompatible material (peroxide, peroxide former, inhibitor or oxygen scavenger, strong bases, mineral acids, or iron oxide/rust) to come in contact with GAA or Acrylate Esters!
• NEVER sparge Nitrogen through MEHQ inhibited GAA or Acrylate Esters!
• NEVER blanket MEHQ inhibited GAA or Acrylate Ester tanks with Nitrogen!
• NEVER store MEHQ inhibited Acrylate Esters above 104 F (40 C)!
• NEVER store MEHQ inhibited GAA above 90 F (32C)! 
• NEVER deadhead pumps containing MEHQ inhibited GAA or Acrylate Esters!
• NEVER use steam to heat Acrylate Esters or GAA in storage!
• NEVER store MEHQ inhibited GAA or Acrylate Esters beyond the 1 year shelf life! Re-inhibiting with MEHQ and sparging with Oxygen will not extend the life!

While some of these are certainly understandable (gee, thawing acrylic acid with steam seems like a terrible idea), some of them (don't deadhead a pump with acrylic acid in it) is obvious only with some hindsight (if you're not used to dealing with pumps in the plant, anyway.) 

It's funny to me how safety literature is mostly of the "polite suggestions" variety, with strong words ("always", "never") reserved for material that has some pretty remarkable consequences. Considering that a random Googling of acrylate accidents reveals quite a few of them (including a plant explosion in Georgia that injured 6 and killed 1), following the 10 "NEVER"s seems like a good idea. 

Process Wednesday: cameras in process chemistry?

Thanks to Jyllian, I am reminded that Organic Process Research and Development has come out with its occasional (and wonderful) "Safety Notables" article [1]. Here's an interesting summary of an unusual set of experiments from a not-typical sector of research:

DSC Analysis of Liquid Sodium-Silica Reaction. 

Sodium-cooled fast reactors, used in the nuclear power industry, are typically built with steel-lined concrete. In the case of a structural failure, the liquid sodium coolant could potentially leak out and react with the concrete. This article (J. Therm. Anal. Calorim. 2015, 121, 45) looks at kinetic studies performed on the reaction of sodium with silica (SiO2) using differential scanning calorimetry. Samples of the reaction were run in open stainless-steel crucibles at different heating rates using a DSC within an argon glovebox. From the DSC data, five reaction stages were separated out using statistical deconvolution, and kinetic  parameters were calculated using the Kissinger and Ozawa methods. An interesting point about the experimental setup was the use of a videoscope with an optical fiber cable inserted into the sample crucible for viewing the sample as the reaction progresses. This allowed for a better understanding and characterization of the different reaction stages.

Certainly something I had not considered (that hot sodium and concrete could react), although it certainly seems reasonable.

It's also interesting to me that the article talks about using a "videoscope" to view the sample as the reaction progresses. I am intrigued by the possibility of sticking cameras in places we don't typically see them, i.e. inside a lab reactor or a plant reactor. I haven't heard about too many places where people are doing that, though - readers?

1. Brown, D.B.; Ironside, M.D.; Shaw, S.M. "Safety Notables: Information from the Literature." Org. Process. Res. Dev. DOI: 10.1021/acs.oprd.6b00013

Process Wednesday: gotta wait until the dryers are done

From Francis X. McConville's "The Pilot Plant Real Book" and its short section on dryers, a comment about dryer characteristics:

The properties of product from pilot drying equipment may be significantly different from that of product dried in laboratory vacuum tray dryers. This is particularly true of units that agitate the cake mechanically such as orbiting screw conical dryers. Particle attrition or agglomeration can result in major differences in particle size distribution, bulk density, compaction and flowability. These things in turn affect solubility, bioavailability, formulation, processing, packing and shipping. Therefore, it is not valid to base projected product properties on the results of tray-dried samples when different equipment will be used on scale-up. 

The behavior of a given product in different dryer types cannot be easily predicted. Bench or small pilot-sized test units are available for tumble or paddle dryers, but the dynamic similarity to large-scale equipment is poor. 

The best way to determine what the product will look like is by performing pilot studies in representative drying equipment. Sometimes the actual product characteristics will not be known until the first production batch comes out in the dryer. 

Just in case you thought you could predict the future in this sense, you cannot. Gosh, it is remarkable to me how much is not known in this business. 

Process Wednesday: the number of Suzukis done on process scale is less than I thought

From Organic Process Research and Development, a review by Elder and Teasdale [1]  on the difficultly in avoiding potentially genotoxic/reactive intermediates in later-stage syntheses. The authors reviewed the last decade of OPRD papers and found the following:

...The review covered 302 publications from this journal covering a 10-year period (2001−2010). The review data are provided in Tables S-1 to S-10 in the Supporting Information. (CJ's note: PDF) There was only one synthetic route (i.e., for netilmicin, published in 2002) that avoided the need to employ reactive intermediates during the entire synthesis....  

On the basis of this survey, we can conclude that the average number of steps required to synthesize each API was 5.9 and that the average number of reactive intermediates per synthetic route was 4.1. 

...In the second half of the decade (2006−2010; Figure 4), whilst these two classes of reactive intermediate (acid halide and alkyl halide) remained prevalent, they were joined by aromatic amines and Michael acceptors as the predominant classes. Aromatic amines (along with the related aromatic nitro precursors) have attained popularity because of the desire to introduce late-stage coupling reactions (e.g., Suzuki, Heck, etc.) into chemical schemes as part of green chemistry initiatives. Similarly, activated alkenes (or Michael acceptors) can readily undergo addition reactions with nucleophiles, one of the most attractive methods for the formation of C−C bonds, using mild reaction conditions. These findings are somewhat at odds with an earlier survey of good manufacturing practise (GMP) bulk reactions run in a research facility between 1985 and 2002, which showed that Michael addition reactions remained constant at about 7% of the total over that time period. However, this may be a vagary of the chemistries studied....

I am somewhat surprised at the relatively small number of boronic acids - I wonder if that number will either rise or fall for the 2010-2015 period? I wonder what future reactive intermediates will be seen?

1. Elder, D.P.; Teasdale, A. "Is Avoidance of Genotoxic Intermediates/Impurities Tenable for Complex, Multistep Syntheses?" Org. Process Res. Dev., Article ASAP DOI: 10.1021/op500346q

Process Wednesday: Synergistic effects from a combination of thiol/thiourea Silicycle products

Credit: Scheme and table taken from Wells et al.

I thought this OPRD ASAP [1] from a group at Janssen was pretty interesting, especially the notes about the effective combining of Silicycle-Thiol/Si-Thiolurea at removing 1400 ppm-1600 ppm palladium down to sub-10 ppm levels. The previous step was a tricky Suzuki with a cyclic boronic acid that seemingly required a reasonably high amount of palladium (see Scheme 9 above):

The results presented in Table 2 showed that Si-Thiourea was, in general, most effective at reducing the amount of Pd from product 1a; Si-Thiol was about 5 times less effective, whereas all others were ineffective at changing the residual palladium.  

It was interesting to find that Si-Thiourea was effective even at room temperature, reducing the residual Pd to 34 ppm, which was within specification for the initial scale-up campaign. However, an unexpected result was found by the synergistic effect of a combination of two Silicycle products, Si-Thiol and Si-Thiourea, when used in concert in EtOAc/MeOH at about 55 °C for 1−2 h, which was not apparent from the screening array. (emphasis CJ's) Their use in combination at 1:1 w/w (total 35−40% by weight relative to compound 1a) proved to be effective at removing residual Pd and decolorizing the product and consistently provided white to off-white solid product with less than 10 ppm of Pd.

 The authors don't discuss why there was synergy. Anyone have an idea?

1. Wells, K.M.*; Mehrman, S.J.; Abdel-Magid, A.F; Ferraro, C.; Scott, L.; Zhong, H.M.; Teleha, C.A.; Ballentine, S.; Li, X.; Russell, R.K.; Spink, J.M.; Diamond, C.; Youells, S.; Zhang, Y.; Tsay, F.-R.; Cesco-Cancia, S.; Manzo, S.M.; Beauchamp, D.A. "Synthesis of Mavatrep: A Potent Antagonist of Transient Receptor Potential Vanilloid-1." Org. Process Res. Dev., Article ASAP DOI: 10.1021/acs.oprd.5b00271

Process Wednesday: clogging in flow systems

A recent Organic Process Research and Development ASAP [1] from a group at Bayer in Germany was on the operation of a low-temperature (-35°C to -50°C) flow unit performing lithiation (using hexyllithium) of difluorobenzene to make difluorobenzaldehyde, among other things. There are, towards the end of the article, a very detailed couple of paragraphs about everyone's favorite flow chemistry problem, clogging: 

During commissioning of the LT (low temperature) unit, a number of problems were identified for various parts of the unit. The major problem spots that were faced at the beginning of the development are summarized in Figure 10. The hexyllithium filter (1) was frequently blocked by a jellylike buildup. Hexyllithium pumps (2) failed every few days due to the deposits in the pump heads and valves. Another problem relating to the pumps concerned the pulsation of flow caused by gas buildup. The problem of clogging in microreactors by gas bubbles is seldom discussed, but it seems to be an important issue, especially in multichannel devices. The hexyllithium precoolers (3) were frequently blocked due to the freezing of impurities if the temperature was too low. The serious problem for the continuous operation was clogging in the structure of the micromixer in the first reaction stage (4). Every 20−40 h of operation, the IKSM reactor became blocked due to the deposits of salts (5). After a few hundred hours of operation, fouling by polymer-like deposits in the residence time reactors occurred (6).  

Also the DMS precooler was susceptible to clogging when using an aged DMS−THF solution (7). Finally, the second reaction stage (8) and the downstream part (9) were regularly clogged by the salt deposits. The detailed discussion of the origin of these problems, precautions, and countermeasures is given in the Supporting Information.  

Generally speaking, the MRT-based  unit was significantly less robust than a classical stirred tank reactor setup. Figure 10 illustrates that in order to make MRT technically viable know-how has to be collected also with the auxiliary equipment. Pulsation-free operation, which is simple to achieve in the laboratory, became a challenge with the industrial equipment. Clogging was certainly the largest obstacle restricting a smooth and continuous operation. There were several causes of the clogging, namely, formation of solids due to moisture in feedstocks, impurities, formation of salt as byproduct, solid formation in the hot spots, and polymerization. 

Overall, this is no surprise. However, still, studies intended to overcome such obstacles are rare. Principally, it is well known that the design of the microreactor plays a key role in minimizing clogging, but so far, there is no design available which proves to be completely insensitive to any solid deposition. In a number of studies, use of a second, dispersed phase has been proposed as a measure to eliminate solid deposits. However, for the organolithium-based chemistry described in this paper, use of water in the Taylor flow regime is not possible due to the low temperatures.  

Therefore, the only way to free up clogging consisted of a classical approach to remove the deposits by purging after increasing the temperature. However, this method was time-consuming, and after being purged, residues of water needed to be removed to avoid formation of lithium hydroxide.  

Finally, it was important to clean the unit before the reactor was completely blocked. Also, the application of external forces, such as ultrasonic treatment, has been proposed. This method is hard to realize, though, when a reaction is performed in a set of reactors as described in this study. Therefore, a clean-in-place concept has been applied.
Technically, two parallel trains of reactors were installed. One line was in operation, while the second one was in cleaning mode. The reactor trains were supplied by one set of pumps and valves.

The Supporting Information has even more details on these issues.

The authors' conclusion is reasonably positive, though:

There are no easy solutions available to counter clogging or fouling. However, addressing salt formation or reducing side reactions by optimal temperature control can help to minimize its impact. Furthermore, clean-in-place solutions might help to increase the robustness of this technology. Based on these promising results, commercial units for low-temperature organometallic reactions have been designed.

I haven't really worked enough with microreactor systems to have a truly informed opinion on the subject, but it does seem to me (from the experience of coworkers and some perusal of the literature) that the handling of heterogeneous solutions that are prone to clogging is one of the top issues that seem to be unsolved.

1. Laue, S.; Haverkamp, V.; Mleczko, L.* "Experience with Scale-Up of Low-Temperature Organometallic Reactions in Continuous Flow." Org. Process Res. Dev., ASAP. DOI: 10.1021/acs.oprd.5b00183

Process Wednesday: Karl Fischer titration

From the second edition of "Practical Process Research and Development" by our mentor-by-literature Neal Anderson, a lovely little section on the determination of water content by Karl Fischer:

The Karl Fischer titration (or KF titration or simply KF) is the classic analytical method used to detect water, and is convenient to use. The basis of the Karl Fischer titration is the reaction of water with iodine and sulfur dioxide. In the early develop of this analytical technique, the solution containing water was titrated with a solution of I2 in benzene or MeOH until the I₂ color remained, providing a sharp, reproducible endpoint to determine the water content. Today of course benzene is avoided as a solvent for the laboratory and scale-up, and less toxic solvent or solvent mixtures are used. The accurate and rapid coulometric assay, which can detect down to 10 micrograms of water, is generally preferred, I₂ is generated elctrolytically at the cell anode, and the amount of water is determined by the current required for electrolytic oxidation of HI. 

My favorite thing about Karl Fischer is that it's a machine that spits out a number (assuming your molecules are compatible with the reagents, that is.) How much water is in there? You can find out, instead of the whole "I dunno, it's wet, I think." I highly recommend them. 

CSB animation on the DuPont La Porte deaths of 4 chemical operators

I've never worked in such a large facility, but I am always sorta kinda really terrified of the myriad venting systems that different plant facilities can have, including the tiny little one that I work at. Just because a pressure gauge is low does not necessarily mean the system is empty.

I'll have to read the draft report when it comes out thoroughly (PDF), but it seems like (as these things always are) a cascade of failures, from design on down to day-to-day operation at the plant. (Is anyone surprised (that in the CSB narrative, which may or not be the real thing) that part of this involved turnover between shifts?)

Here are my previous posts on this particular story.

UPDATE: Added the link to the PDF above, here's a Houston Chronicle story on the report from Lise Olsen, the reporter who wrote the Texas Monthly story linked above. 

Bonus Process Wednesday: Syrian mobile chemical reactors

I've been meaning to post on this July Wall Street Journal article* on what chemical weapons inspectors found in Syria:

Parked outside the bunker were mobile weapons-production facilities that sat on 18-wheelers. Inspectors would later liken the vehicles to Transformers toys because they looked so ordinary on the outside. As such, they would have been difficult to target from the air. 

“This wasn’t kitchen chemistry,” Mr. Smith recalls thinking. “It was a piece of quality engineering.” 

Syrian guards carried the empty aerial bombs out of the bunker and laid them in a row. The Syrians had stored them without chemicals inside. Mr. Cairns says the bombs contained two internal chambers separated by a thin membrane. When the bombs are filled with chemicals, activating them requires turning a crank attached to the back of the bomb, which rotates a rod inside, pierces the membrane and mixes the chemicals.

As someone who has done some 50-gallon chemistry back in the day, it doesn't surprise me that you could set up a mobile laboratory in a typical trailer. (Heck, they managed to stick a suite of reactors on a ship to neutralize those chemical weapons!) That said, it's probably not particularly comfortable for the operators nor would it be a particularly efficient means of production, I would think.

*Can't get to the article? Search for this headline: "Mission to Purge Syria of Chemical Weapons Comes Up Short"

UPDATE: A couple of nice articles about Scott Cairns, the Canadian chemist working for OPCW. 

Process Wednesday: more on static electricity

From a new-to-me book "Chemical Process Safety" [1], a nice summary of the different places that charge accumulation could result in electrostatic discharge in the plant: 

1. Contact and frictional charging: When two materials, with one being an insulator, are brought into contact, a charge separation occurs at the interface. If the two objects are then separated, some of the charges remain separated, giving the two materials opposite but equal charges.  

2. Double-layer charging: Charge separation occurs on microscopic scale in a liquid at any interface (solid-liquid, gas-liquid or liquid-liquid). As the liquid flows, it carries a charge and it leaves a charge of opposite sign on the other surface, for example, a pipe wall.  

3. Induction charging: This phenomenon is applicable only to materials that are electrically conductive. A person with insulated shoes, for example, may approach an overhead vessel that is positive charged (previously filled with positively charged solids). Electrons in the person's body (head, shoulders and arms) migrate toward the positive charge of the vessel, thus accumulating an equal quantity of positive charges on the opposite side of the body. This leaves the lower part of the body positively charged by induction. When a metal object is touched, there is transfer of the electrons, creating a spark.  

4. Charging by transport: When charged liquid droplets or solid particles settle on an isolated object, the object is charged. The transferred charge is a function of the object's capacitance and of the conductivities of the droplet, particle and interface. 

The myriad of places that static electricity can build up in the plant is not something that I've considered before; I tend to think of it as a phenomenon that only happens inside the reactor with certain solvents (hexane, for example) or during the transfer of various solvents from drums/totes to the reactor (or vice versa.) Lots to learn, I suppose.

1. Crowl, D.A.; Louvar, J.F. "Chemical Process Safety: Fundamentals with Applications." 3rd Edition, Pearson Education, 2013.

Process Wednesday: stress tests aren't just for the Treasury

Credit: Practical Process Research and Development

From our mentor-by-literature Neal Anderson's second edition of "Practical Process Research and Development", a good thought about stress tests of plant processes on page 416:

With stress tests reaction mixtures are purposely exposed to more extreme conditions to assess the potential impact. For instance, a key reagent may be added in portions to determine wither higher level of impurities will form due to micromixing. Another example of a stress test is shown in figure 15.1 (CJ's note: above). Merck researchers anticipated that by using n-BuLi as base, small amounts of excess base could lead to metalation of one aryl ring, leading to a bis-aldehyde. When two equivalents of n-BuLi was charged in a stress test, 10% of this side product was indeed formed. Further screening showed that EtMgBr could be used as a base, even in 20% excess, avoiding the need to titrate and control the addition of the n-BuLi solution. 

Every once in a while, you have this terrible "what if" dream where the horrible "what if" scenario is "what if the operator makes a math error and adds twice as much as they should of a key reagent?" Performing these sorts of stress tests (and yes, math errors definitely happen) are a good idea and an important part of developing a plant process.

This section also talks about identifying a "normal operating range" and a "proven acceptable range" for operating parameters - I think part of the issue is that the temptation is to make the proven acceptable range to be the norm, and then mistakes happen, and then you're into "here be dragons and impurities and reworks" land. 

Process Wednesday: double cone dryers

I've not had a ton of experience with double cone dryers, but I know that they are a common means of drying compound. They are certainly unique-looking - I presume that it took many years to come up with that shape.

McConville's "The Pilot Plant Real Book" mentions that the agitation is gentle and you get decent product homogeneity, which is good. I presume that you load up the product in one handway, turn on the vacuum (and the heat?) and allow it to slowly rotate and dry product. As the video shows, they don't rotate very fast. When you're done, you get your bag ready and open it up and let the product fall down? Dunno.

I am going to guess that cleaning one of these things is a bear, but again, not much personal experience with them. 

Process Wednesday: the dos and don'ts of scale-up

Credit: Dr. Clemens Brechtelsbauer

But hot filtrations are fun! (No, they're not.)

(Found via a random Googling of "minimum stir volume" a good list of "dos and don'ts" for scale-up included in a presentation by Dr. Clemens Bretchelsbauer of Imperial College, London. The list is credited to F.X. McConville, the author of the strongly recommended "The Pilot Plant Real Book.")

I'm gonna buy some goats for the plant

From NPR's "Goats and Soda" blog, a story of Nepalese accident investigation:

In Kathmandu, interested observers are waiting to see whether TIA will again take an approach that it used in 2007 to fix a technical problem with a Nepal Airlines Boeing 757. At that time, airline officials sacrificed two goats in front of the aircraft to appease Akash Bhairab, the Hindu sky god, and then declared the aircraft was ready to resume flying.

That'd cut a lot of my deviation investigations shorter.  

A neat ozone+UV paper pointed out by See Arr Oh

Check out this really interesting Science paper highlighted by See Arr Oh, one where the authors transform cyclohexane to adipic acid (among other transformations) using ozone and 300 nm UV light. Obviously, this is a pretty neat method and one that should be highlighted to the broader scientific community.

I think it's obvious to the authors that a batch reactor approach is/would be sufficient for the laboratory, but not for industrial applications to large-scale polymer needs. From the Supporting Information:

In the current photo irradiation process, liquid reactant was gradually converted to solid precipitate. At the late stage of reaction process, a small amount of liquid reactant/intermediates were trapped in the solid precipitate. It is therefore inherently difficult to reach more than 90% conversion. Nevertheless, the problem can be overcome by designing a dynamic flow reactor allowing regular removal of solid precipitated products from the bottom of the reactor when a large scale production of solid adipic acid is concerned.

In my favorite formulation of the moment, "can" is doing a lot of work in that last sentence there. (Presumably, you'd need a ozone flow unit (and those exist), combined with a photoirradiation flow unit -- who knows, it could be done.)

I presume the economics would not work out for adipic acid, but would probably work out for higher-value monomers. Who knows? Either way, definitely a very interesting method -- congrats to the Hwang group.