Skip to Content

The Dark Side

99% Yield? That, Friends, Is Deception

Here’s an attention-getting paper from Tomas Hudlicky (and his co-author Martina Wernerova), and I’d like to help it get some more. It begins:

One who has been reading the literature concerned with organic synthesis in recent years, be it methodology, catalysis, or total synthesis of natural products, may have noticed considerable inflation in the values reported for isolated product yields, ratios of diastereomers, and enantiomeric excess values. A comparison of papers published during the period 1955 to 1980 with those published between 1980 and 2005 reveals that those from the more recent period frequently report isolated product yields of reactions >95%. Such large values were rarely found in the older literature and are all but absent in Organic Syntheses, a journal that only publishes procedures that have been independently reproduced. . .

There, does that sound like the chemical literature you know? Just a bit? Hudlicky has tackled this issue before, and the reasons he advances for the problem remain the same: pressure to make your methods stand out (to the scientific community, to the journal editors, to the granting agencies), a decrease in scale in reactions (making accuracy and precision more difficult), and, finally, what he refers to as “deliberate adjustment”. That’s well put; the rest of us know it as fraud.
He identifies the mid-1980s as roughly the period when things really started to go to pieces, saying that most procedures in reputable journals before that era are reproducible by, as they say, one skilled in the art, while the numbers have been decreasing since then. And he puts some numbers on the problem, performing a series of test experiments with extremely careful weighing and analysis.
These confirm what every working organic chemist knows: the more manipulations, the more sample you lose. Filtration through a plug of silica gel, into one flask, can give you pretty much complete recovery. But if you cut fractions, you’re going to lose about 1%. And if you have to do a separation, even between two widely separated compounds on silica, you’re going to lose about 2%. So people who report a >98% yield after chromatography from a real-world crude mixture are kidding themselves. The same goes for extractions and other common methods. In general, every manipulation of a reaction is going to cost you 1 to 2% of your material, even with careful technique. Hudlicky again:

Given that most academic groups do not subject day-to-day reactions to serious optimization or matrix-optimization [6] as is done in industry, it is reasonable to assume that the vast majority of the reactions reported in the literature do not proceed with quantitative conversions. Such aspect would approximate our experiments with mixtures of pure compounds. Because a minimum of three operations (extraction, filtration, and evaporation) is required in working up most reactions, we conclude that yields higher than ca. 94% obtained by work-up and chromatography of crude reaction mixtures are likely unrealistic and erroneous in nature. Such values may arise as a direct consequence of not following correct protocols, which would be expected in the fast-paced academic environment. (An astute student of the organic literature may discover that this very author has been guilty of reporting yields in this range from time to time!)

He goes on to detail the limits of error in weighing, which depend greatly on the amount of sample and the size of the flask. (The smaller the sample-to-container ratio, the worse things get, as you’d figure). And he turns to analyzing mixures of diastereomers by NMR, LC, and the like. As it turns out, NMR is an excellent way to determine these up to about a ratio of 95:5 , but past that, things get tricky. And “past that” is just where a lot of papers go these days, with a precision that is often completely spurious.
Here’s the bottom line:

The conclusion drawn from this set of experiments points to the prevalence of serious discrepancies in the reporting of values for yields and ratios in the current literature. We have demonstrated that the facilities and equipment available in a typical academic laboratory are not adequate to support the accuracy of claims frequently made in the literature. . .The current practice of reporting unrealistically high isolated product yields and stereoisomer ratios creates serious problems in reproducibility and hence leads to diminished credibility of the authors.

He recommends a rigorous disclosure of the spread of product yields over multiple experiments, calibration of LC and GC apparatus, or (failing that) at least admitting that no such analysis has been done. (He also recommends getting rid of the concepts of diastereomeric and enantiomeric excess, in line with my fellow Arkansan Robert Gawley’s advice). But I think that these ideas, while perfectly reasonable, don’t get at the underlying problems – the inflationary pressure to produce more and more noteworthy results. Hudlicky’s rules should be adopted – but I fear that they might just push the self-deception (and outright fraud) into newer territories.
I’m glad he’s published this paper, though. Because everyone knows that this is a real problem – we complain about it, we joke about it, we mutter and we grit our teeth. But “officially”, in the published literature, it’s never mentioned. Let’s stop pretending, shall we?

66 comments on “99% Yield? That, Friends, Is Deception”

  1. john says:

    So this is something I’ve always felt strongly about. Yields are always published as a percent, no plus or minus. Well that tells me that you never repeated the experiment or don’t want to talk about the variance, why don’t journals start demanding a yield which has been calculated based on 3 different replicates of the experiment, at least. Give me some statistics, we all know yields are gonna vary day to day, heck from moon cycle to moon cycle given how tricky some reactions are. If you tell me 80 and I try it and get 50 I’m gonna call your work crap, if you tell me 80 but give me a range I can expect I am happier if I get something in that range, and am less likely to think you’re just trying to make your synthesis look better. Also it saves me time and money not using a route which isn’t suitable to what I’m trying to achieve.

  2. Brian says:

    99 % yield means you forgot to remove your magnetic stir bar from your reaction product or you still have residual metal in your product (like a nugget of Pb). A few times, I thought I had a crystallization nailed down, it was an inorganic salt that was entrained in my product that had precipitated. Having done analysis in my past, I look for water content by Karl-Fischer, a comparative wt/wt assay and a determination of residual inorganics before I would say something like I had >90 % yield. I approach things from a more rigorous approach.

  3. gyges says:

    Two points
    1) If he wanted to be read, why did he publish, perhaps more accurately, pseudopublish, in a closed source journal?
    2) “But I think that these ideas, while perfectly reasonable, don’t get at the underlying problems – the inflationary pressure to produce more and more noteworthy results.
    One of the underlying problems is the ignorant manner in which professional chemists regard one another. I’ve seen it time and time again, someone reports an excellent piece of work – mass balances, good analysis, sensible decisions at bifurcation points – the end of which has been to quote yields in the 60 – 70 per cent range. His/her reward has been some ignorant whispering campaigning about how it could’ve been done better etc … which is invariably utter and complete bollocks.
    To explain this culture look to the chemists with hair triggers who shoot from the hip and are not held to account.

  4. tuky tuky says:

    seeing a 99% yield always makes me giggle

  5. KC Nicolaou says:

    You’re all just jealous.

  6. Donough says:

    Different industry, similar story.
    Membranes have been touted for energy savings for a while. With membranes any number of polymer, zeolite, ceramic material or mixes can be used. In literature it is ‘normal’ to see selective comparisons to previous literature whereby the membranes are compared to obviously worse membrane (which no-one would use).
    In one field that pretty much ended when most of the membranes related were reviewed and compared to each other (200-300 membranes). The amount of novel membrane types being published suddenly dropped as the crap was cut. While this may seem anti-innovation, most of these novel membranes already has a good chance of being crap before testing.
    However not only are the researchers responsible, the publishers are responsible also. If I can identify crap results from the abstract, surely the reviewers who are supposed expects can also.

  7. number cruncher says:

    Very gratifying to see commentary like this appearing in the literature. I never understood why synthetic chemists are allowed to avoid reporting the sorts of statistical analyses without which papers in other fields aren’t even publishable.

  8. Nick K says:

    Bravo! At long last, someone is prepared to tell the truth about reported yields! It’s absolutely no coincidence that the reliability and trustworthiness of much of the organic literature went down at the same time as the yields went up. I rarely have problems with papers from the 50’s to the 70’s (yields 90%) often fails in my hands…

  9. milkshake says:

    I do get 99% occasionally, when all these factors consire: 1) I run a very clean transformation on a multigram scale and 2) I isolate the product by filtration and it is very insoluble (a common thing, in kinase field).
    Calculating the yield from 5 mg evap residue in a 50g flask is kind of joke – I think it would be better in case of small-scale reactions to provide the overall yield after several steps. (Then of course one has to consistently use the material from previous step – and not from a big batch sitting in a freezer that was made by someone else).
    Also, if you report 75% yield and someone else can reproduces your procedure with a 83% yield, it will not give you a bad reputation.

  10. Nick K says:

    An addendum to my previous comment: yield inflation may also have something to do with the recent decline in use of microanalysis. It’s far easier to claim a 99% yield with a clear conscience if you’re relying on HRMS rather than CHN.

  11. Will says:

    The greatest compliment anyone ever gave me on my chemistry was returning to grad school a few years after leaving and talking to current members of my PI’s lab and hearing that the stuff in my thesis was reproducible in the reported yields. They also told me the current students/postdocs couldn’t reproduce the work of my co-worker anywhere near the reported yields.
    Of course, that coworker published in JACS whereas I left with a masters after 5+ years…(cue miniscule violins, the bitterness went away after about 4 years)

  12. Anonymous says:

    There’ll be a few Nobel prize winners having to hand back their medals if this movement gets of the ground. We all know that in some groups a grad student’s life expectancy is determined by their yields. We also all know that there are some groups whose published work is extremely difficult to reproduce. What do you think the correlation is like ?

  13. mad says:

    Maybe they meant mass balance 🙂

  14. Jordan says:

    I had a chance to chat with Hudlicky a couple of years ago — he is a pretty cantankerous guy (lots of “kids these days” type remarks) but also very pragmatic and very dedicated to organic chemistry.

  15. partial agonist says:

    This is not an issue I ever got too worked up about since I pretty much take all reported yields with a grain of salt anyway. If it says 95%, I think 80%. If it says 85%, I think 70%. Most of us do that every day and have done it for so long, it’s just understood.
    Back in the day, my advisor wanted us to put the average yield for a transformation in the papers, not the best. That mindset is rare, it seems.

  16. Anonymous says:

    THud gave a few classroom lectures and handled problem sessions where I went to school.
    Much fun was had at the expense of us all.

  17. azetidine says:

    Hey #5, if I actually went back and tried to reproduce your chemistry, what percentage of the reactions would actually WORK as described, much less in the yields reported?

  18. molecular architect says:

    Reminds me of the work by a very prominent Nobel Prize Winner (HCB) who used to report “corrected yields” for organoboranes. The organoboranes were themselves rarely isolated and, as any organic chemist should know, can be converted to many different functional groups. His people would take known amounts of the end products, run them through the workup and calculate the % recovery. They then “corrected” the yield of the intermediates assuming 100% conversion to the end products. This was fully disclosed in the publications and “technically valid”, nonetheless, I always thought it was cheating since most chemists were interested in the final products not the organoborane intermediates.

  19. David Formerly Known as a Chemist says:

    This is the type of article that reaffirms what you already know, like those published scientific studies that concluded prayer has no effect on outcome of sick patients (really, see for example
    I was fortunate to never experience pressure to increase my reported yields, since I never worked on development of synthetic methodology. Development of new synthetic methods seems to be the papers where “yield inflation” would naturally be the most intense. Of course, working in process chemistry, there’s always the pressure to increase your yields, but for purely economic reasons.
    As bad as this problem is in the scientific literature, it doesn’t hold a candle to the irreproducibility of procedures in patents.

  20. Anonymous says:

    This seems like agruing about semantics. Do you really disregard a reaction or favor one method over the other because it is 85% vs. 95% yield?
    While the point Hudlicky brings up may apply to methodology papers, I don’t see where it matters that much in reagards to total synthesis. You cannot keep adding 5% to your yield in each step of a multistep synthesis, the solvents, impuritys, etc. eventually average out throughout the steps.

  21. processchemist says:

    Those who work on the process side know far to well how much “yeld” is a stressed word.
    Milkshake talks about “stone-like” kinase inhibitors. Yes, many of these compounds can precipitate as rocks from a reaction mixture, so you can have very high yelds. But high yelds of what? Of a product with an HPLC grade excessing 98.5%, with all impurities less than 0.5% by HPLC, total ashes under 1%, AND of the right crystalline form, with a particle size distrubution good for further processing (milling/sieving/pharma manufacturing)?
    And I totally agree about dropping enantiomeric and isomeric excess as parameters to evaluate the output of a reaction.

  22. Spiny Norman says:

    The problem starts with laboratory courses where all or part of the grade depends on yield x purity. QED, chumps.

  23. MTK says:

    I’m with partial agonist on this one. It doesn’t bother me that much really. I don’t take the numbers that seriously. To me 95% yield means it’s a good reaction.

  24. petros says:

    I think it started rather earlier than 1985, close to where Derek works, with it being the norm in JACS articles from a certain group.

  25. Bruce Hamilton says:

    Good subject to discuss. I’d also like to see the end of ee, it’s unnecessary with the widespread use of chiral HPLC, and ignores other impurities.
    I’d also like to see both achiral and chiral HPLC data for molecules claiming chiral purity. That alone will involve some extra effort, as a chiral system has to show it will separate all molecules of interest.
    I’ve run published chiral separations with enantiomers producing a single peak, suggesting the author did not confirm the chiral column was appropriate for the separation.
    I’m not so thrilled about the suggested calibration of analytical HPLC response factors, as it can be a major effort to make relevant “separate source” reference compounds. I’d just prefer evidence that potential impurities and precusors have been separated from the analyte by the submitted method.
    Too often, people assume that what’s loaded onto a chromatography system will be detected, which is only true for a few types ( eg TLC, Iatroscan (TLC-FID)), whereas the more common GC, Flash, and HPLC techniques obviously can only ever detect what is eluted. “xx% by chromatography” was introduced to describe such products, rather than just “xx%”
    I also agree with the concern in a comment above about the increasing use of MS instead of elemental analysis. It complements the composition data, rather than replacing it.

  26. Resveratrol Receptor says:

    My reactions create matter, muahahaha.

  27. iridium says:

    Even if the author is formally correct, I am with 23 on this one:
    ” To me 95% yield means it’s a good reaction. ”
    – If you are working in process development you will do your optimization on your substrate.
    – If you are a medicinal chemistr….you do not care about a difference in 10%yield, and you most likely are using anyway a slightly different substrate.
    Although it is obviously wrong, unethical and not educative to embellish the yields, I do not see the point of students wasting time calibrating everyday HPLC, calculating error of the balance, giving statistical analysis of the yield of their reactions…it is better they use their time for learning and thinking.
    The reaction conditions are anyway usually optimized on ONLY ONE substrate and if 10 entries are reported in a table, you can immagine that 9 of them are obtained under NOT optimized conditions.
    If there is not intentional cheating…to me 95% yield means it’s a good reaction.

  28. paul says:

    I want to see % conversion, %yield and %purity with some indication of reproducibility. This would give a good indication of how good the transformation is.

  29. provocateur says:

    So, you just proved ppl exaggerate!I do not need a paper to say this..thats why some procedures get popular and some don’t…let the market decide it because we will never get the truth!

  30. HK says:

    I’ve published a few synthetic papers where I report high yields and high ees, and I do things on small scale. I do show both chiral and achiral HPLC traces in my SIs. I will be the first to admit that it probably wasn’t statistically accurate, but when I weighed my scintillation vials after pumping down, there was that much material in there, and the NMRs were clean, so I reported it. I ran the reaction multiple times and got similar results, but I don’t report statistical error because I wasn’t rigorous about it.
    I don’t think it’s an issue. The papers got into their journals not because I got 90% instead of 80%. They got in because they were interesting mechanisms, or interesting extensions to reported reactions. As it was, we were more rigorous in our mechanistic analysis and kinetics studies etc etc than in the actual synthetic steps, because they weren’t the focus.
    If someone were to report a synthesis for the purposes of saying “this is practical, use this in industry” then you should definitely be way more rigorous in analyzing your reaction yields and other forms of efficiency, and perform everything in triplicate. The stuff I’m working on now, I’m scaling up and repeating multiple times so that I can get good readings on what the yields are – around 80%, and I’m very happy with that, on the scale I’m running it.
    I’ve heard about Hudlicky… I appreciate what he’s doing, but his ire should be directed at synthetic methods that tout efficiency at large scale but don’t put any effort into statistical analysis, not just at all methodology papers in general.

  31. Chrispy says:

    Anyone else ever hear of the Corey 1:2 inversion?
    That’s where 29% becomes 92%…

  32. anonymous says:

    Someone once told me that organic chemists are pathological liars. I tend to agree. The only yield I know of is isolated, in the flask, passed purity analysis. No prefixes like GC, NMR, HPLC are acceptable (I recall some papers from a very recent Noble winner reporting GC yields). And too often I hear in casual conversation the words conversion & yield being used as synonyms.

  33. frequent visitor says:

    In chromatography/ultrafiltration steps in protein purification, the best I ever expect is 95% at manufacturing scale with errors in the assay, and that is ELISA which is +/-15% with good operators of the assay. Getting 97-98% is the tops, and I have seen it on well developed ultrafiltration where the membranes were rinsed.
    It is a strange artifact, that often in good capture steps from mammalian cell culture (or human extracts) that you can obtain 110% yield (ELISA) at small scale, but in those steps I account for 120%ish recovered in the mass balance…and more often than not they turn out 90% yields at manufacturing scale…strange why the small scale does this. I have seen many attempts in the literature to explain this, but no has, at least to my satisfaction. Some of these papers date back to the 80’s, and I bet many people just stopped reporting this type of artifact…if anyone knows here I am all ears.
    Yes 99% is ridiculous…I don’t trust very many assay’s to +/- 1%

  34. frequent visitor says:

    sorry I should have put +/-15% ELISA in terms of accuracy.

  35. medchemist says:

    #18 followup. MolArchitect, your example was the first that came to mind. I worked in a well-regarded lab down the road from the one you cite. All yields from the borane lab were regarded as highly suspect (if not outright hilarious) and not achievable in practice as isolated product. A divisor of 2-3 was usually about right. I’ve always thought that that lab was where the problem started; a bad practice very visibly well-rewarded encouraged emulation.

  36. Anonymous says:

    I think this is much ado about not so much. Some of you act like a yield is a physical constant up there with Plank’s constant. If we run the reaction enough, we’ll know the yield to five places. Meh!
    The “hands” factor in this is so great, that the reported yield is merely a rough guide to the quality of a reaction, considered along with other data (like the author’s rep).
    Reporting accurate yields (with statistics) isn’t worth the trouble; however, reporting honest yields (what you actually got) is appreciated.

  37. processing says:

    The best compliment I ever got came from my own PhD advisor, many, many years after graduation. He said “No one will ever doubt your procedure or yields as your experimental section begins with…A reactor was charged with 3 kg of X followed by 6 L THF dried to 0.1% water content…and the solution yield was xx% …and the isolated yield after crystallization was yy%” etc. But not everyone has the luxury of working with the accuracy of multi-kg quantities, or working with highly skilled engineers. Or working in a team environment (quelle dommage!). Lots of cross checks make for better yield reporting, especially when there are millions of dollars at stake in a single batch.
    I read the literature yields with a HUGE grain of salt. And often, with a little careful analysis, I can improve upon them. The isolated yields are very rarely >90%,and that is only if the solution yield is close to quantitative (eg. Hydrogenation reax).
    The real key to honest yields are interdependent teams…but that won’t happen anytime soon in academia…

  38. Azo bob says:

    Another issue I see people not considering when calculating yields is whether they have produced a solvated crystal form. If I’m expecting 3.00 g of a 300 Da product and I get 2.90 g my yield is 97%, shame that I vac’d down from DCM and have a monosolvate so I should get 3.86 g, now my yield is 75%. Your sample can be dried on the high vac as long as you like but that DCM is still there. People seem to ignore this in their yields even when their NMRs, CHN and x-ray structures show what’s really there.

  39. I tried one of Mr KCN’s procedures, published in Angew. Chem., at that. The explosion was so violent that it took out a fume hood pane. Fortunately no-one was around when that happened.

  40. mmol says:

    THud is, as always, direct and provocative and as usual, right and hitting the nail with a sledgehammer. We all know the pressure here….referees, journal editors, even…egos…….. But the issues are more fundamental… academics who don’t (actally won’t) publish the “low yielding examples”, the reactions that don;t work as this somehow demeans what they have done…..there’s that ego again…..when of course knowing what doesn’t work, and it will be for a good reason, is what is most useful to potential users. I’d challenege editors to demand of methodology authors a range of examples, including some that do not work…….. and certainly some that do not work so well. The only yields i believ are those from process groups where you have to actually deliver what you claim……… Otherwise the error bars (and forget calibrating LCs etc…….life is too short and yields for one-off reactions are too unimportant) are, as someone once said, a unknown known; we all know that they are there but don’t recognise them. As for “total synthesis”, then we all have our stories, many of which could have been penned by Mr H C Anderson.

  41. bbooooooya says:

    I had the pleasure of dinner with Hudlicky once, and he is a breath of fresh air. Maybe a tad(?) cynical but, gosh, fantastically intelligent. Best thing to ever happen to Brock University.
    The bigger issue here isn’t just blatant lying about yields: if we know (and I think everyone does) that many lie about yields obtained, how do we know they’re not lying about other important matters in their reported procedures?
    I recall old RSC papers which would often include statements that “The Checkers found….”. Good times….good times.

  42. Tokamak says:

    There are 2 reactions I’ve run that I can honestly say give 99% yield. Pd/C hydrogenation where the Pd/C is simply filtered off and especially TMSCHN2 methyl ester formation. I can certainly believe that a reaction is >99% if you weigh your sm into a tared flask, dissolve it in your solvent, add the reagent, then rotovap it back down again and re-weigh the flask. No transfers = no loss.
    More on topic, I appreciate that my advisor’s mentality to yields on total synthesis papers is >50% is fine, don’t waste too much time optimizing things.

  43. cookingwithsolvents says:

    @36 “The “hands” factor in this is so great, that the reported yield is merely a rough guide to the quality of a reaction, considered along with other data (like the author’s rep)”
    This is 100% true. I’ve topped lit yields from time to time, which is always a nice surprise.
    @42 re: Pd/C
    Usually mine check in at 97-98%; maybe I should wash with more solvent.
    I’m happy to see this article though in my experience it applies more to o-chemists than inorganics (my field). I can say with certainty that the number the balance reads is the % that goes into my notebook and into my papers. I’ve had a few inorganic ones that came in at 99% but they were simple reactions and on a scale where loss of ca. 1g would have been loss of a full percent.

  44. watcher says:

    Just be real, honest, forthright, and then get on with it. If some people want to exaggerate, who really cares, except those who are less well known, who are then jealous. Forget it. Get over it, and go out for a walk in visible nature where you can see trees, sky, deer, birds…..

  45. Anonymous says:

    I think the main point is that A LOT of the new literature is really non consequential methodology.
    Example: Ligandless Pd couplings and diethyl zinc additions to aldehydes.
    I think the argument of many of these papers is an often incremental improvement. But if these yields are falsified then really what is the point of all this science, and (ahem) funding.
    Sometimes I am embarrassed to be a chemist. You’d think we would be working on some better stuff.
    +1 on lets show stuff that didn’t work so well, lets test the limits of this “new” methodology.

  46. GreedyCynicalSelfInterested says:

    People need to graduate. Period.

  47. Canageek says:

    I’ve been wondering lately what is a ‘good’ yield to be satisfied with. I’m an undergraduate chemistry student and I’m looking at the lab report that I just finished editing and my yields range from ~80% to ~30% (I spilled something part way through). I typically seem to get ~70% on a good day, and I’m pretty happy with this. We lose 1 mark if a yield is under 50%, which happens on a semi-regular basis. So at what point should I be happy with my yield, and what what point should I be kicking myself for not re-precipitating the filtrate?
    To the points about lying about yields: By 2nd year chemistry (When you move from general science to a major at my uni) this is already pretty deeply embedded, along with altering results to match what is expected. This seems more common among people who don’t want to be chemists at least, as they seem to find it to be typically less work to alter a few numbers than to actually do an outlier analysis or whatever. This seems to drop off each year, though this could just be because faking an NMR spectrum is a lot harder than a physical observation.

  48. RB Woodweird says:

    1. When I am looking for a way to turn A into B, I don’t really care much about the yield. As long as it is nonzero, I figure I can tinker with it to get what I need.
    2. I couldn’t really figure out a variance on yield for most of my reactions, because I hardly ever run them the same way twice. I have the insatiable urge to improve the conditions every time.

  49. bcpmoon says:

    Sorry, but a procedure with GC-, HPLC-, NMR-, …yield and a HRMS is completely useless. Its like taking a photo of a crowd when you should be taking portraits.

  50. GSDecay says:

    Agree with 40 – When will journal referees realise that a methodology paper that actually discloses the lower yields and poorer substrates is infinitely more useful than the sanitised stuff people mainly feel obliged to submit. If there are only 10 examples in the paper and all show 90%+ yield (all usually very uninteresting alkyl-substituted aromatic examples) with none lower, you immediately think “why no more diverse examples?” (I think we know why).
    When the referees send the message it is OK to include the instructive limitations on scope then maybe more folk will start to include more of the useful negative data in the manuscript leading to more people trusting the data and allowing greater take up the new reactions.
    Don’t forget, publishing the paper is only the start of creating the value and impact from the work, not the end as many seem to think.

  51. Nick K says:

    @50 GSDecay: Perceptive comment. Negative examples are at least as useful as positive examples in methodology.

  52. PMP says:

    The THud paper is a good start but, like all papers, it is just one side of the issue.
    – First, the mass recovery. This is vastly dependent on the type of compound. Typical total synthesis intermediates armored with greasy protecting groups (TBDPS etc) will fly through short columns with >98% mass recovery. On the other hand, polar heterocycles such as imidazoles will be retained on silica very tightly and in fact require specially functionalized silica for >80% recoveries. As such I would never make a statement that >94% yield would be a lie. The same applies to extractions – this is why people often do back-extractions to recover the compound lost in the aqueous phase! The maximum possible yield depends strongly on what you are making, and how. In analytical chemistry, “quantitative transfers” are somehow possible – why not in organic chemistry?
    – Second, the yield range. A typical methodology paper includes 10 to 20 examples with different substrates. One of the very reasons why this is asked is that it forces the people who develop the method to repeat their protocol at least 10-15 times. As such the tables do include the yield range in a way – especially these days when people have the habit of including substituent ranges such as Ph, 3-ClPh, 4-ClPh, 3-BrPh, 4-BrPh, 4-MePh, etc. Do we really need to repeat _each_ of these 3 times to make sure the yields are reproducible each time? What if you spill some – do you then repeat again, or do you include the spilled example in the range reported? Would this really advance the science? I don’t think so. It might be useful to repeat some of the examples – and this is done anyway when the reaction is being optimized – but repeating the example with R = Ph and R = 3-ClPh each three times does not add much value.
    The yields are also dependent on scale (for reasons too numerous to explain here). I would encourage everyone in method development to demonstrate at least one example with >10 mmol range (preferably 50-100 mmol at least). Sometimes the yields will actually increase in this process. I think this makes more sense than repeating each and every example at 0.1 mmol scale three times. I do not see any point in this exercise and I really do not understand why THud is insisting that this should be done.
    What makes more sense is the comment in the paper that the conversion (and the method for determining the conversion) should be reported. In practice this means using some sort of internal standard – I think no one should do any method development without an internal standard-based monitoring of the reaction.
    At the end of the day, it is the reputation of the scientist that is at stake. If other people are using your methodology and obtain results similar to what you have published, you should feel happy about yourself. I know there are cases where this is not true – but you could not have prevented them with these rules.

  53. London Chemist says:

    I remember the post-doc in my PhD group who did one reaction twice: the first time he got only 50% but about 12:1 diastereomeric ratio, the second time he got 80% but the dr was only about 5:1 (numbers are approx–it was a long time ago). Guess how it got reported in the subsequent paper?
    Yep! 80% yield, dr of 12:1…..
    The boss didn’t see anything wrong with that….
    I actually don’t mind the yields being off–its the literature reactions that DO NOT WORK at all, no matter how hard you try, that really annoy me.

  54. anon the II says:

    I don’t know if anybody is still reading this post but it’s important to reply to Nick K (# 51).
    The negative data is almost never as important as the positive data. Not even close.

  55. X-files says:

    I enjoyed reading this Hudlicky paper, which he disclosed his opinion on inflated yields. Our group had experience in applying an asymmetric reaction published in JACS, but we could never get it work. Later we discovered that we were not the only people who failed to perform this transformation. We did a lot of literature search on this reaction and discovered that while some people got it work while others couldn’t. I suspected that there were more failure out there but possibly people are not keen on reporting negative results. I managed to find the Ph.D. thesis of the original investigation online but it
    was not of much help. So exactly how many times each reaction is repeated in the original investigation, only god knows. So did the authors make up the results? I think not. However, I doubted if they have looked extensively into the effect of reagent purifies. I suppose why bother if they were getting very high yields and ee(s).
    By the way, I was told that KCN mentioned recently in a symposium that his group is close to completing the synthesis of maitotoxin. I wonder how he is going to quote the yield of his final step? Perhaps 99%?

  56. molecular architect says:

    @54: I strongly disagree with you. Negative data is valuable as it defines the limit of a method. Even more useful in med chem. An SAR profile which spans a wide range of activity offers far more insight that does a tight grouping of active compounds.

  57. CMCguy says:

    #54 anon the II I will take the counter position that often looking at what did not work or worked poorly relative to other substrates is of great value. Such data can be helpful in sorting out mechanisms and can also act as warning for people who do wish to attempt on similar substrates. It can lead to further modifications necessary to make reactions work where was not applicable before. I actually wish there was more published in each paper on what was found doesn’t work since majority of focus in literature is on only positive results. There have been a number of times I have talked to people after the fact and learned some analog I had struggled with was already known to not work.
    No one has mentioned above I think and it has been a topic of Derek’s before but in terms of yield and conditions I am always highly skeptical of what is in Patents after having poor experiences attempting to repeat a fair number.

  58. anon the II says:

    To you fans of negative data, call up Homeland Security and tell them you know where Bin Laden isn’t hiding. See how excited they get.

  59. Orthogon says:

    Think anyone has ever tared a 100 mL flask straight from the drawer or oven, but, after drying down under vacuum, backfilled with Ar the same flask containing their 100 mg of pure product, and then weighed that flask to determine their yield?
    (The density of argon is 1.78 g/L; the density of N2 is 1.25 g/L.)

  60. Nick K says:

    @Anon the II: Are you seriously comparing chemistry methodology with the hunt for Bin Laden? Here’s a question for you: you’re about to run a new reaction on a complex substrate. Would you feel more confident if the paper reporting the methodology had a wide range of examples, including a few failures, or merely half a dozen alkyl aromatics?

  61. Jose says:

    “I actually don’t mind the yields being off–its the literature reactions that DO NOT WORK at all, no matter how hard you try, that really annoy me.”
    KCN’s benzylic ox with IBX springs to mind. Zero, nada, zlich for any substrate. Last time I checked, no other report anywhere in the literature of that working. Thanks for wasting my time!

  62. Nick K says:

    @Anon the II: Off-topic, but more germane to your post: Do you think the US Army would like to know where they WON’T find IEDs in Afghanistan?

  63. DrAliP says:

    #61 has it right, it’s the time wasted that makes inaccurate reporting so damaging. I spent 18 months of my doctorate research proving to my supervisor that d.e.s reported in a paper of his were greatly exaggerated. Of course, the paper where we published the real values ended up in a journal that no-one reads, so as usual the liar did well out of it and the incentive to exaggerate remained.

  64. Tyrosine says:

    I think I know partly why yields in more recent years are inaccurate and inflated. In the 70s and 80s we used to routinely do reactions on a multi-gram scale. Hence mechanical losses, relative weighing errors etc. were smaller as a percentage of the material. Today academic research is mostly mg scale or less (to save on costs?) so the percentage error is greater.
    As many have said the accuracy of the yield quoted is not really that important. I’d be just as happy, happier, with a rough guide or even a range. Imagine published reactions were described as giving (e.g.) 40-90% yield, or 60-70% yield. Numbers like that tell you something about both the yield and the reliability of the reaction.
    I certainly don’t believe yields can be quoted with any accuracy greater than +/- 5%. But we all do it, just out of convention.

  65. chemadvocate says:

    Thanks for further drawing attention to this aspect of synthetic life. While doing my doctoral work I remember often wondering why no one bothered publishing the range of yields they obtained. In fact, I believe I was at a disadvantage because, as an aspiring process chemist, instead of defaulting to “>99%” I often vocalized my wondering of when the last time the hplc was serviced, the polarimeter calibrated or what the detection limit of our 600 MHz Oxford magnet actually was! I also gave far too much time to the consideration of the ratio of my flask mass to my expected yield. All of this was rewarded with fewer papers and no jobs because these things matter only to process chemists and those jobs had moved to China and India during my training years 🙂

  66. otherside says:

    Reactions are inherently inefficient. I am happy someone has drawn attention to the trend to overstate yields. However, after doing synthesis for many years, I supported med chemists in analysis and purification. Not so shocking is the similar trend to believe more target compound resulted from the reaction than what the data proved. Everyone could always find their ion in the noise of an FIA-MS results, for instance. Uv-only detection gave a false sense yields were higher, but with more sophisticated systems incorporating other detectors(MS, CLND, ELSD)revealing otherwise, chemists were disbelieving and sometimes angry. Chemists are attached to their own molecules, like they are children and want to believe the best outcome, 99% yield and 100% purity (by UV of course).

Comments are closed.