The OM Paper vs. Drinkel’s PhD Thesis

August 9th, 2013

ChemBark InvestigatesAs part of our investigation into the controversial paper published by Reto Dorta and coworkers in Organometallics, ChemBark contacted a source in Europe who was able to obtain a copy of the Ph.D. dissertation of the first-author of the paper, Dr. Emma Drinkel. Chapter 4 of the thesis carries the title “Synthesis, Structure and Catalytic Studies of Novel Palladium and Platinum Bissulfoxide Complexes”, and the chapter appears to describe the vast majority of the work reported in the publication in Organometallics.

The entire thesis is 174 pages long. ChemBark has made the editorial decision not to republish Drinkel’s thesis in its entirety, but rather, to provide a set of small excerpts that highlight important information, including a number of discrepancies with the paper in Organometallics. We are also republishing excerpts from the SI of the paper. I believe that this approach constitutes “fair use” with respect to copyright law, because (i) there is a time-sensitive need for the community to be informed about this important case, (ii) these excerpts represent a small fraction of the whole of the published works, and (iii) republication of these excerpts does essentially nothing to deprive Drinkel or ACS Publications of financial gain.

ChemBark’s excerpts from Chapter 4 of Emma Drinkel’s Ph.D. Thesis
ChemBark’s excerpts from the Supporting Information of the OM Paper

Drinkel’s thesis is dated “Zurich 2011”. In her curriculum vitae included at the end of the thesis, Drinkel reports her Ph.D. studies as having spanned “09.2007-09.2011”. Dr. Drinkel’s LinkedIn profile reports that she was at Zurich until December 2011, and she began work as a postdoc at Universidade Federal de Santa Catarina (Brazil) in July 2012. It is worth noting that the Organometallics paper was received by the journal on January 7, 2013—a full year after Drinkel departed from Zurich. This piece of information is interesting when one considers whether “just make up an elemental analysis” could mean “perform an elemental analysis” versus “fabricate the elemental analysis data”. Of course, the (arguably) ambiguous instruction could have been written many months prior to submission of the paper—while Drinkel was still in Zurich—or Drinkel could have carried all of her samples from Switzerland to Brazil.

A brief examination of the dissertation reveals that much of the information published in the supplementary file of the OM paper is identical to the information published in Chapter 4. This includes most of the characterization data and the prose used to describe the experiments. But a rapid comparison is hindered by what appears to be the root cause of the confusion between the main paper in OM and its corresponding supplemental file: Drinkel misnumbered some of the compounds in her thesis. The numbers in the discussion section of chapter 4 are shifted relative to the data reported for the same compounds in her experimental section. For example, compound 14 in the thesis’s experimental corresponds to compound 15 in Figure 13 from the thesis (pasted below). This compound is labeled 14 in Scheme 5 from the OM paper and 154 in the OM supporting information. The same problem goes for compound 16a/15a/15a/165a and others.

Figure 13 from the Thesis

Figure 13 from the Thesis


Scheme 5 from the Main Paper

Scheme 5 from the Main Paper


The numbering discrepancies in Drinkel’s thesis not only went unnoticed, they were exacerbated when the OM authors built their paper off of the chapter and decided to delete the label from compound 14 in Scheme 5. This is the compound associated with the now-infamous instruction to “just make up an elemental analysis”.

The written response from the editor-in-chief of Organometallics on the SI’s controversial statement regarding compound 14 included the following:

The author has explained to us that the statement pertains to a compound that was “downgraded” from something being isolated to a proposed intermediate. Hence, we have left the ASAP manuscript on the web for now. We are requiring that the author submit originals of the microanalysis data before putting the manuscript back in the print publication queue.

Indeed, there are no data for 14 written in the experimental section of Drinkel’s chapter 4—its preparation occurs as an intermediate in the preparation of 15 (using the numbering from Figure 13). With that said, the discussion section of chapter 4 mentions:

When 5a was treated with only 1 equivalent of AgBF4, unlike in the Pd case, the stable complex 14 was formed. No crystals could be grown to confirm the structure, but the 1H NMR spectrum of the complex shows the ligand is still symmetric. There is precedence for this type of chloro-bridged Pt dimers in the literature with phosphine ligands.

This statement from the thesis might appear to refute the claim in the letter that the authors could not isolate 14 and that it was simply a proposed intermediate, but the text of the main paper states that NMR was taken “in situ” after the first reaction. With that said, no NMR data are provided for compound 14 in the Supporting Information file, and an instruction is given to Emma (Drinkel) to insert these data. Perhaps the instruction to “insert” was given because the instructor already knew the data existed (based on what was written in the discussion section of the thesis)?

Beyond the problems associated with misnumbering, there are several discrepancies between the data reported in the thesis and the data reported in the SI of the Organometallics paper. All of the examples that I could find related to elemental analyses. Specifically:

SI-5b vs Thesis-5b

Compound 5b from the SI

Compound 5b from the SI


Compound 5b from the Thesis

Compound 5b from the Thesis


SI 11a vs Thesis 9a

Compound 11a from the SI

Compound 11a from the SI


Compound 9a from the Thesis

Compound 9a from the Thesis


SI 12 vs. Thesis 12

Compound 12 from the SI

Compound 12 from the SI


Compound 12 from the Thesis

Compound 12 from the Thesis

SI 165b vs. Thesis 15b

Compound 165b from the SI

Compound 165b from the SI


Compound 15b from the Thesis

Compound 15b from the Thesis


You can see that the authors chose to “count” different associated solvents when calculating the expected values for the elemental analyses, and they reported different observed results in the paper vs. the thesis for some compounds. Were these samples run multiple times? Since the original data have been demanded by the journal, I guess we’ll find out.


Stay tuned for continuing coverage…


Note: In the reporting of this story, we wanted to give both the first author of the paper (Emma Drinkel) and the corresponding author (Reto Dorta) the chance to comment on the discrepancies we found in the data prior to the publication of this post. ChemBark first attempted to contact Professor Dorta by e-mail on Tuesday night (St. Louis time) and received no response. Dorta also has yet to respond to a second message, sent Thursday afternoon, that sought comment on the discrepancies reported in this story. A message seeking comment was also sent to Dr. Drinkel, at the same time, through her Facebook account. Should either author respond to our requests for comment, the responses will be posted in their entirety.

76 Responses to “The OM Paper vs. Drinkel’s PhD Thesis”

  1. DS Says:

    Starting sentences with numbers, and no use of oxford comma? Disturbing to say the least.

  2. Vojsov Raison Says:

    Yes, there are discrepancies with the thesis, but there are also two other authors on this paper who could potentially have had another go at the EA for these compounds. Not possible to conclude anything from a comparison of thesis and SI without knowing what EA analyses were actually carried out. Saying that, assuming solvent of crystallisation without a crystal structure showing its presence is a bit of a fudge…

  3. number cruncher Says:

    As I stated on the other thread “It’s worth pointing out that some of the EAs in the SI (e.g. those of 10 and 12) have closely matching found and calculated values, but these don’t match what the calculated values SHOULD be for the given empirical formula. Coincidence? I can’t be bothered checking if the empirical formulae are correct, but someone might?” This is even more strange given the differences between the thesis and the SI

  4. Andy Robertson Says:

    DS, if you’re going to bring up the subject of Oxford commas, this thread will go wildly off topic and stay there until the damn things are consigned to the grammatical hell where they belong.

  5. Bob Says:

    It’s interesting to look at the measured EAs for compound 15b (from thesis) and 165b (from OM SI), i.e. the bottom two figures above. The “Found” measures for both C and H have the exact same digits after the decimal point (48.16, 4.59 and 47.16, 4.59). Too much coincidence. I suspect the following happened: the student actually measured the EAs and, like everyone else, fit them as best as possible in her thesis with some added solvent. When transferring the data from thesis to manuscript form for submission, one of the authors decided that publishing with fudged solvent molecules wouldn’t look as good and may even raise ire of the reviewers. So the analyses were re-calculated without solvent and the measured analyses were adjusted to fit better with the expected values. I believe that the student actually did obtain real EA data for the compounds above, but that these were manipulated prior to publication, which seems aligned with the comment in the OM SI. It’s even possible that the EAs from the thesis were included in the original OM SI submission, but a reviewer may have questioned these and they were “adjusted” after the review process. The reviewer may have also requested an EA for the intermediate 14. It’s likely that the incriminating comment was added at this stage, and the other EAs were tinkered with after the paper was reviewed. The paper may have then been accepted by the editor without the reviewers ever seeing the version that includes the “just make up the EA” comment.

  6. Serenity Says:

    Contacting a source for the thesis is quite easy – it is publicly available at the university library:

    There are also three different labs listed with the elemental analysis.

  7. A recent PhD Says:

    I agree Vojsov Raison, one cannot forget that
    1.- It’s perfectly common to synthesizes many times a the same complex when you{re doing your PhD thesis, therefore it’s completely normal to have several elemental analysis data for a giving compound as one has many different samples.
    2.- There are other authors who could’ve performed the reactions and obtain different data.

  8. ChemGuy Says:

    If only we could obtain the originally submitted version of the SI and Manuscript? They must be somewhere in the ACS manuscript database. However, I am pretty sure that this information will not be made public due to the confidential nature of peer review.

    Firstly, this would allow us to determine if that controversial statement was in the original SI.
    I think if the EA values were magically different between the two manuscripts, we would start to question where the new EA values came from.

  9. come on people Says:

    While synthesising the same complex a few times is common, getting repeated elemental analysis is not. The experimentals are the same in both the thesis and paper, so the differing elemental results presumably came from different attempts that were not reported?

    There is a huge amount of evidence pooled that there is a very large degree of dishonesty at play. The case has been made for misconduct, the authors of this paper must now unequivocally clear their names or plead guilty.

    As a third year phd student, who has papers held up trying to get reasonable elemental data, my question is how has this not been identified as a problem sooner? Highly important numbers with no evidence required and large margins of tolerance are going to be faked by the less than honest researchers. I think the chemistry field as a whole needs to take a long look in the mirror about the academic dishonesty which we all know happens, but refuse to talk about.

  10. A recent PhD Says:

    Dude, I don’t know what kind of chemistry “come on people” do, but if you do catalysis, each time you synthesize a new batch of catalyst you have to send it to EA before you can do anything with it.

  11. SpeedyGonzales Says:

    @A recent PhD
    I have never heard of such a thing. I have done my PhD and post-doc in two prominent groups that do reactions in catalysis, and whenever we remake catalysts that have already been subject to full characterization (including EA), we judge our success by the homogenity of the NMR spectra and the performance of the catalyst in whatever reaction is of interest. I can’t imagine my boss condoning the expense of doing repeated elemental analyses on a catalyst that has already passed EA, has the same NMR spectra, and gives expected results with regards to ee or regioselectivity.

  12. See Arr Oh Says:

    I love the Oxford comma [ducks]

    Excellent job, Paul. I’ll be interested to see how OM / Prof. Dorta handle their end of it.

  13. A recent PhD Says:

    Most metal complexes can have NMR silent impurities that’s why you should check for a good EA. EA is not really that expensive and it gives you always peace of mind. Maybe you use a frit that someone else used before and had some silica or charcoal and there you go, your complex will never be clean. I had to redo my catalysts many (many times) and trust me I have an EA for each and every single batch I used for catalysis. If I was to use it as an intermediate for the synthesis of a derivative then I didn’t have to do the EA. But those were the rules at my PhD lab and at my current postdoc lab as well.

  14. Just some chemist Says:

    Well, this certainly puts the light on things now. Paul, I think you just gave sufficient evidence that writing this paper was made into a circus by the PI and made a nightmare for all the authors by the PI. Based on what’s seen here, I’d say the EA data from the dissertation was the real data, and like Bob said, the data was transferred to the OM SI with the EA data manipulated (seemingly at Dorta’s request). While I wouldn’t go so far as to say this is outright fabrication, I do believe that fitting the data to solvent molecules (which appears to be common practice…? I’ve only done a couple EA’s in my career – I deal with x-rays and MS mostly in full chracterization) should have been reported instead of this “adjusted” data. It really does appear to be a request from Dorta to Drinkel to “adjust” the EA data for this compound.

  15. ChemGuy Says:

    @See Arr Oh

    Based on Prof. Dorta’s recent non-responses, his response will be not to respond, (Talk about professional). Perhaps he has already fled for some kind of safe haven?

    Since all the EAs are all over the place for the same compound, my prediction is retraction in OM. I think the ones with the solvents included are the actual experimental ones, then they did some creative accounting… I could be wrong but that is my guess.

    I agree, excellent job Paul!

  16. SpeedyGonzales Says:

    @A recent PhD
    Those are good points that I hadn’t thought of. I guess that the particular systems I work with are unlikely to have NMR silent impurities because of the nature of the oxidation state of the metals and reagents we use in the final step, and so that is presumably why we don’t have that rule. I can think of one case in a different lab at my school where it turned out that a bunch of silver oxide and triflate contaminants were actually doing the real chemistry in some elaborate system using an entirely different metal, so point taken :) I do however jealously guard my frits from my coworkers. Out of curiosity, do you ever find much variation from batch to batch of your catalyst preparations if things otherwise appear to go well?

  17. A recent PhD Says:


    When it comes to EA of different batches, usually the variations are withing 0.3%, which is I think normal given the fact that you might have different amounts of residual solvents etc, and we always submit a sample of the bulk, because I know of labs in which you sent for EA a sample that you extradry or extrapurify just to get the values right, but in my opinion that’s wrong as EA is a proof of homogeneity of a sample, and that’s why crystal structure is not enough. As you can only account for the crystal that you measured, but in reality you can have a completely different mixture.

    We sent our samples to a guy who always performs the analysis in duplicate and the differences between analysis is sometimes in the order of 0.1-0.2%. I did have a couple of odd batches when dealing with dirty frits and the solution was almost always start over (that’s why I’m well aware of the “dirty frit problem”) sadly in my former lab all glassware was common, so you had to be extra careful all the time.

  18. Nick Says:

    Have Prof. Dorta’s former or current university said anything about this? Surely he should at least be under investigation for this?

  19. Nick Says:

    Actually, has anyone informed them? They may still be unaware this is going on…

  20. SC SynthGuy Says:

    This reminds me of when I had to get elemental analyses for compounds. It was not unheard of for the heterocycles I was making to associate water, but my boss was always very hesitant to just fit the results to water or some other solvent. I had one case where I had 1:1 associated methanol (the sample was recrystallized from methanol). The EA fit rather well, but I still had to prove to my boss that it was not a fluke. I think we ended up getting a crystal structure that backed up the result. I’ve met others, however, who did not have nearly as high of a standard when it came to EA results.

  21. Old Dude Says:

    While the behavior in this incident is execrable, if you want to have nightmares about current practice in the life sciences, see the article by Fang, et al., in PNAS, 110(3), 1137. These authors did an analysis of 2047 retracted papers in the PubMed catalog up to May, 2013 and found that only about 20% of the articles were retracted because of error; more than TWO-THIRDS of the retractions were attributable to misconduct!!!! They also argue that the percentage of articles retracted because of fraud has increased roughly 10-fold since 1975. This analysis dealt with biomedical and life-science research, of course, where ethical problems abound.

  22. Less newt Says:

    @number cruncher: The empirical formula for 12-SI is correct. As you note, the calculated values are then incorrect; they should be: 46.5 % C, 3.29 % H, 2.85 % N.

    @Bob: If there are enough compounds with suspect EA data in the SI, a chi-squared test versus a uniform distribution can usually support allegations of fraud (e.g., ). The particular behavior you suggest would probably not be detectable this way, though, if the original EA data were real (and I see no reason to fit to solvent if you’re making up data).

  23. OldGuy Says:

    There’s a fair bit of “holier than though” discussion building up here. I think it’s best to assume, as a starting point and in the absence of detailed information, that your own way is not ALWAYS the right way. It depends on the specifics of the chemistry that you are doing, amongst other things.

    @come on people: You say, “While synthesising the same complex a few times is common, getting repeated elemental analysis is not.” and then “As a third year phd student, who has papers held up trying to get reasonable elemental data..”. How are you going about trying to get reasonable elemental data? I’d imagine you are sending off a sample for EA, seeing how it come back, repurifying, then sending again? Which would be getting repeated elemental analysis… Am I missing something?

    @Just some chemist: “I’ve only done a couple EA’s in my career – I deal with x-rays and MS mostly in full chracterization”. Do x-rays and MS confirm purity? I’m not criticising, just wondering.

    Plenty of criticism of including solvent to fit – but some chemical systems just don’t release solvent fully, without decomposition. My approach in some cases has been to avoid drying too much so as to end up with integral molecules of solvation, rather than residual amounts that just won’t go to zero. Horses for courses.

    I’m not suggesting there haven’t been some errors or bad decisions made by the authors, but I don’t think there is sufficient evidence to confirm any systematic misbehaviour at this point. Another author could have repurified samples, and sent off for EA, explaining the differences between thesis and paper, for example. Various other, less nefarious, possibilities are discussed in other threads.

    @Nick: I doubt there is any academic chemistry department that is not aware of this. At least any department with an internet connection!! All of my connections are well aware, and that spans a few continents.

  24. Dave Says:

    With regard to routine use of EA on bulk samples:
    I totally agree that combustion analysis is the best way to determine purity in conjunction with evidence of molecular composition (NMR, MS), and analysis should definitely be done on the bulk and not the triply-recrystallized material. One major problem with this technique is that many departments don’t have in-house EA services. At my PhD institution, we had in-house, with a great technician. Our air, moisture, light, and heat sensitive compounds would always pass as long as they were pure. But now I have to ship samples across the continent to be run by some person I’ve never met in a glovebox I’ve never vetted. Now all of a sudden my air sensitive samples analyze as decomposition products (hydrolysis, oxidation). I can’t trust the numbers I get from these places, so it totally eliminates the usefulness of EA as an analytical technique. If EA is going to be a hard requirement for publication, every major research university should have in-house service.
    BUT. This certainly doesn’t mean that people should be making up shit. Incidentally, the University of Zurich has in-house EA:

  25. kyujkyjuhgtfhuygt Says:

    In tomorrow’s post, Paul gets a restraining order against him by all the authors of this paper….

  26. Locale Says:

    The OM paper lists Dorta’s affiliation as University of Zurich, but he moved to University of Western Australia in December of 2011. I suspect the discrepancy is due to not having updated his electronic bio in the ACS system.

  27. Suspicious guy Says:

    The OM editor should require original EA data for ALL new compounds reported in the paper. If Dorta is honest he will comply quickly to save himself.

  28. z Says:

    While people are right in asking all of these questions, I find it disturbing that so many self-professed scientists are completely jumping to conclusions about what happened and who is responsible before the actual answers to these questions have come to light. There seem to be multiple reasonable explanations that lay varying amounts of blame on different people. Maybe we should wait until the evidence appears before lighting up torches and joining a witch hunt.

  29. Burn Says:

    Reasonable explainations, jumping to conclusions, evidence appears? Come on, you know we are a society that has to have all the answers right now. We love to crucify people before there is a trial. Best off, we love hiding behind the internet saying things that we never would say in person.

  30. A recent PhD Says:

    I couldn’t agree more with Z and Burn!

  31. tt Says:

    Great job Paul. It seems fairly clear what went on here based on comparing the SI and Thesis and the fact that the primary authors had moved on. The responsible move by the PI should have been to inform the editor/reviewers that they didn’t have an EA for the intermediate and then provide the EA’ for all the others like that reported in the thesis. That would leave it up to the editor to then decide whether to accept or reject. The PI could have then just had a current student crank out the compounds and EA, or submit the paper to a different journal that doesn’t mind reporting EAs with solvent.

    The sad fact is, that as a process chemist, we always refer to most papers as reporting “academic yields” because I have yet to find one example where the reported yield reflected what we observe on multigram scale (even with lots of optimization). The primary cause is that academics rarely report purity. In industry we adjust our yields based on purity from HPLC versus an analytically pure standard (QNMR, EA, KF, XRPD, HPLC on the standard). I wish all editors/reviewers were much more strict and at minimum required a statement of purity based on solid analytical methods. I guess that in some ways, nothing in this story really surprises me.

  32. CV Says:

    shows 8 more publications while working with Rota:

  33. EA Says:

    Emma’s thesis is public information, easily relocatable by searching

    I would think the data reported in her thesis are correct.

  34. Joey Walsh Says:

    tt, as a chemical engineer, we never expect things to scale. Very few large productions in chemical process industry are batch. You have to balance conversion with purification with costs with throughput. I don’t fault people reporting what they got in a small scale batch reaction, but one has to assume they are indeed honest.

    Kind of like sensor research: Everybody has the best sensor in the world when it is in a controlled environment. Once you get these devices into a complex matrix, selectivity and sensitivity go to pot. Lab scale reactions cannot assume to scale and rarely do. I personally wish they would report all side reactions (this is often impossible) because the products from these can be where trouble starts.

  35. tt Says:

    Joey Walsh.
    Absolutely understand that lab scale vs pilot is quite different, but I’m talking about simply replicating at gram scale in the process lab (even after DoE). Never observed an academic paper working as advertised (especially catalytic). It’s a very subtle dishonesty to report yields without a purity correction, but in the lit this is common practice

  36. Syas Says:

    As several have pointed out, there is a lot of stone casting going on without allowing the process to properly run its course. I wouldn’t fault the authors for not responding to a blogger since they are a dime a dozen although if the blogger used their professional name then that would be a slightly altered ball of wax. Regardless, I wouldn’t fault the author from only corresponding through the editors at least until the issue is resolved, especially if I were being contacted through my facebook page, lol.

  37. Dave Says:

    If anyone has not already seen it, please read John Gladysz’s recent comment on the original post
    (rather near the bottom at this point, best to do a search for ‘Gladysz’)

  38. Thomas Says:


    Well, what to say. PhD’s are doing somehow all the today scientific job, group leaders are running for funds and stay alive and finally Prof. are not willing to do “micromanagement” and to look after PhD’s.

    I had the chance to work in several top ranked uni. . In such competitive “industrial” field, work is mainly driven by paper, money and h factor. It is very difficult to do “nice and careful” science as it has to be. In fact dangerous for your career.

    I have seen and have been asked to make up data too,…,I nearly did not get my PhD because I very strongly refused to follow this unacceptable path and weird leadership… The words of the ombuds person on this above behavior were “If you want to have your thesis do exactly what your supervisors are asking you to do”…and basically “zip your mouth!”.

    Some of my PhD’s friends crossed the dark line,…, please…PhD’s this is your work you have to stand behind so fight for the top quality and only the top quality…it is your PhD not the one of your supervisor…

    I wish all the Best to Emma, as her story highlights black on white, how an ASAP comm. is reviewed and how Prof. and Co-authors are reading “carefully” a PhD manuscript. The scientific system has somehow reached a limit and maybe Emma’s story will contribute to decrease such “unacceptable” inquires from the leadership.

    I’m currently in a beautiful group in a less “top ranked university” where research is done to its purest level……

    As advice for future PhD’s: You should avoid industrial like science in top ranked uni…if you are in deep love with Science and Nature.


  39. Just some chemist Says:

    @ OldGuy: Typically, full characterization required for identification of the compound includes two types of NMR (for us organic chemists it’s typically 1H and 13C that are required) and either high resolution MS or getting an EA (IR, melting point if it’s a solid, and specific rotation if it’s chiral are also required). Purity by GC or HPLC helps too. In the case of a copper complex I made, NMR, the X-Ray structure, and high-res MS were enough. We went and got the EA for it too in case the reviewers requested it, but it wasn’t ever requested.

  40. Just some chemist Says:

    @OldGuy: Sorry, I completely forgot your question while answering it! X-ray can’t determine purity, but if you also take a high-res MS of the compound the crystal was grown from and it agrees with the X-ray structure as far as mass goes, combine that with NMR and you have a reasonable argument for purity of the compound these days.

  41. OldGuy Says:

    @Just some chemist. Thanks for the polite and constructive replies. I understand and agree with your responses – I know others would be more insistent on EA for everything. It’s often problematic with my chemistry because of the tendency to hang on hard to solvent, disordered solvent occupancy etc etc. So I’m entirely on board with you. It’s a matter of being as rigorous as possible, within reasonable bounds, and adapting to the chemistry you are dealing with – and then letting the publication process decide if you have done enough. It irks me when people take a black and white position on these matters without knowing the full story (which you didn’t do – you just got caught in the cross fire :).

  42. Joseph Merola Says:

    @Just some chemist. I can’t fully agree with high res Mass Spec showing purity. As with anything, the sampling is the important part. But, let’s pick an extreme, that the organometallic sample is contaminated with 10 or even 20% salt. The high res M.S. is not going to show that. As I posted in a different part of this blog, I think that EA is not needed for EVERY new compound, but for key compounds in a paper. We have just submitted a publication where the issue of cis/trans isomerism in a 4-coordinate complex was an issue. The x-ray structure was clearly of just one isomer. Screening multiple crystals from the same batch confirmed the same unit cell, but even there we can’t screen every single one. We used powder diffraction to show the bulk corresponded to the single crystal. Of course, this does NOT make the case for EA since both isomers would have the same EA – it is simply the case that one must be aware of the capabilities of both the analytical method and the sampling techniques that go along with the analytical method.

  43. Just some chemist Says:

    @ Joseph Merola:

    I completely agree sampling is important. The complexes I made typically stood up to chromatography, so I didn’t have to worry about salts (which I agree EA would be incredibly useful in detecting). I usually chose LIFDI for these complexes when looking at it by HRMS, and I came to swear by it for the types of complexes I was making.

    I can relate to your case of multiple isomers in a sample. I developed a heteroleptic complex that could have presented as two different isomers. It took a good amount of time to really figure out if the x-ray structures I saw (multiple batches made, multiple x-ray structures obtained) were reflective of the whole sample, and in the end NMR gave more insight as only one isomer was observed.

  44. Matt Says:

    @Thomas I’m sorry that that was your experience, good for you for not giving in. Such things may be distressingly common, but there are groups at top 10 ranked universities that have reputations for doing things right. You have to ask around. I know, I graduated from one last year. I had to get dozens of EAs on organics, and yield was reported on the mass of material that passed EA. It was a major pain, but you can trust my yields.

  45. Random chemist Says:

    According to the ACS at :

    “Policy Summary on Prior Publication

    Manuscripts submitted to Organometallics must, except for reviews, be based upon original research by the authors. The manuscript may be precluded from consideration for publication if it deals with information that has been made publicly available in print or electronic format. Such disclosure includes posting of conference presentations, posters, and preprints on Web sites. All such print or electronic materials, with the exception of short abstracts, must be itemized in the cover letter.

    An exception to these disclosure restrictions is the final version of a thesis filed as a publicly stated requirement for a degree and posted on the official Web site of the degree-granting institution.”

    Clearly, unless this thesis is posted on the OFFICIAL Web site OF THE degree-granting INSTITUTION, inclusion of any of it’s contents in the manuscript both violates ACS policy and is plaguirism of a prior (2011 copyright) work. The thesis is a single author document; the paper has other authors taking credit for the single author’s words. According to my University’s Dean of Graduate Studies, this legally constitutes plaguirism. I asked him personally in front of my peers and he was very stern and clear about it, indicating he could produce legal precedence if needed.

    I suppose since lots of people violate this rule and plagiarize from thesis, that makes it ok, eh?

  46. EA? Says:

    My experience with some commercial analytical labs is that they will sometimes try too hard to match the “anticipated numbers” provided by me on the lab’s sample submission form. I can’t say whether this is due to unconscious bias by the chemical analyst or whether it is a dishonest effort to give the me what the EA lab thinks I want, with aim of keeping me as a customer. I understand that having a rough idea of what the composition is likely to be is useful to an analytical lab because it enables the lab to run its measurements under conditions where accuracy is likely to be greatest. The protocols can differ for maximizing accuracy when looking for, say 30% of element X versus 0.5%.

    Still, after getting numbers that appeared too good too often, I switched commercial EA labs yet again and thereafter have only sent “anticipated values” that are intentionally somewhat off from what would correspond to a completely pure sample of the desired compound. Human nature being what it is, this seems like a good way to keep from being deceived by an analytical service, whether intentionally or otherwise. Keep in mind that while one reason to obtain EA data is to get a paper published, a more important one is to determine whether one indeed is working with clean compounds.

  47. OldGuy Says:

    @Random chemist. I think the exception statement would apply to 99% of theses, wouldn’t it? Unless you make the mistake of dumping it out as a “book” through one of those “publishers” you hear from every time a thesis is submitted.

  48. Random chemist Says:

    @OldGuy. Perhaps my campus is the exception then. They use some company called ProQuest to handle our theses, so nothing is hosted on our University’s websites. They warn us that it is very important for the thesis author to set and monitor a thesis embargo with ProQuest. They say publication of articles from the thesis contents must be completed before the embargo is ended. Since the thesis author owns the words, only the thesis author can adjust the embargo. ProQuest evidently won’t allow the thesis advisor to be involvedmin the embargo because the thesis advisor has no rights to the words in the thesis. We were told that the best option is to make sure all the work has been published before the thesis is submitted.

  49. OldGuy Says:

    @Random chemist. Very interesting – not something I was aware of, perhaps because I haven’t seen that much of the workings of the US (?). I don’t like the idea, but perhaps I don’t see the advantages from where I’m sitting. I can think of some disadvantages – but this is getting a bit off topic.

  50. Joseph Merola Says:

    @Random Chemist Electronic publication of Theses and Dissertations is something that was much more problematic when they were first introduced. It wasn’t so much an absolute issue of publication, but of access. UMI has long been a place where one could obtain a copy of a thesis or dissertation and I have ordered several from them over the years. But, that is obviously not a readily available source, but it was a prepublication. When anyone could put information on a website and that information became available to just about anyone, well that is when journals began to worry about the issue of prior publication. I know my university and a lot of others entered into discussions with ACS and other publishers to come to an agreement that would not preclude prior publication as a dissertation. Each ACS journal has its own policy and you published “Organometallics”‘ above. Interestingly, “Inorganic Chemistry” does not mention theses or dissertations in it’s policy. Note that all policies say that prior publication MAY jeopardize publication in their journal and that one should notify the editor of any form of prepublication when submitting an article. Elsevier has an agreement with Proquest that dissertation’s deposited there do NOT constitute prior publication.

    I certainly would abide by the opinion of the Dean of the Graduate School, but I could also make that case that, if publication by a third party is a university REQUIREMENT then that would fall under the heading of the University’s Official Website because that is the site attached to a dissertation requirement. Most universities also allow a delay of dissertation publication if there are IP issues to be considered. Again, the issue is a general one that each researcher needs to be aware of their particular university’s, journals, etc policies.

  51. Mike Says:

    Sadly, I have seen repeated examples of this kind of behaviour: faking of EA data by inorganic chemists (two academics at a middle-sized uni in eastern europe). At one point, almost the whole faculty (inc. head of school) new about it, but they did nothing (except politely telling the offenders not to do this anymore and forcing the whistleblower to shut up)! One of this people moved to industry (to a senior position), the other will shortly be promoted to senior professorship.

  52. John A. Gladysz Says:

    topic: Responsible Reporting of Solvated Comounds

    There has been due attention given to the microanalytical data in the Dorta manuscript and Drinkel thesis in this string.

    Personally, the first thing first caught my attention was that both documents report identical reactant quantities, identical reaction conditions and workups, and identical product quantities and yields for the four compounds highlighted in the ChemBark post. However, the products are represented as solvated in the thesis, and unsolvated in the manuscript.

    This of course cannot be, and is issue I want to focus on in this comment.

    First, I’m a stickler for reporting both mass (typically g or mg) and molar (mol or mmol) quantities of all products (just as one does for reactants), not just the yields or yields and masses. I recognize that yields/masses only is a common format today, but this was not case some decades ago. Having the molar quantities greatly helps in checking the yield data, something I always do as a reviewer when solvates are claimed.

    It is often the case that yields are incorrectly reported for solvates, and I made this mistake myself once (“Regiospecific and Stereospecific Reactions of Ph3C+ PF6– with Rhenium Alkyls (η-C5H5)Re(NO)(PPh3)(R); α vs ß Hydride Abstraction”, Kiel, W. A.; Lin, G.-Y.; Bodner, G. S.; Gladysz, J. A. J. Am. Chem. Soc. 1983, 105, 4958-4972. DOI: 10.1021/ja00353a020). Hence, ever since the mid-80s I have had a group handout on this topic with the example from the paper. The present version is on my research group website and pasted below:

    I’ve never sent this out for external review, but this post is a step in that direction. Subject to input, it may eventually become an appendix in the author guidelines of Organometallics.

    One point is that a solvated compound has a greater FORMULA weight than an unsolvated compound. This is the quantity that must be used in the yield calculation. The yield associated with a solvated product will always (for a given mass) be lower than that associated with an unsolvated product.

    Another point is that the presence of solvate molecules must be independently verified, with NMR being the obvious choice, but there are other options. I’ll refer readers to the handout for this. I’ll also comment that hydrates are the most difficult types of solvates to treat quantitatively, sometimes there is no perfect slam-dunk solution for them.

    Thus, with respect to the Dorta manuscript and Drinkel thesis, we will be focusing (apart from many other questions) on whether the reported procedures give solvated or unsolvated products (it cannot be both), and then whether the yields given are correct (we have done the calculations both ways, and also looked at the NMR spectra per the group handout).

    There’s more to say about a lot of related topics, but they depart from this string, so I’ll inquire with the ChemBark management.


    Solvates: Avoiding Common Errors
    (Gladysz group handout,

    Special care that must be taken in characterizing compounds that crystallize as solvates. An experimental procedure from a thesis is given below. The purpose of this handout is to critique several misleading features.

    [(η5-C5H5)Re(NO)(PPh3)(=CHCH3)]+ PF6– (25). A solution of 18 (0.314 g, 0.549 mmol) in CH2Cl2 (30 mL) was cooled to -78 °C. Then solid Ph3C+ PF6- 0.234 g, 0.604 mmol) was added with stirring. After 30 min, the cold bath was removed. After 1 h, hexanes (15 mL) was added. The solvent was removed by oil pump vacuum. The off-white powder was collected by filtration and washed with hexanes followed by several small portions of Et2O. The sample was dissolved in CHCl3/CH2Cl2, and Et2O was slowly added by vapor diffusion. After 2 days, greenish yellow leafs of 25 (0.311 g, 0.434 mmol, 79%) were collected by filtration, mp 165 °C dec. IR (cm–1, CH2Cl2): νNΟ 1720 s; 1H NMR (δ, CD2Cl2): 15.82 (ReCH, q, J = 8.0 Hz, 1H), 7.62-7.28 (phenyl, m, 15H), 5.94 (C5H5, s, 5H), 2.53 (CH3, d, J = 8.0 Hz, 3H); 13C NMR (ppm, acetone-d6): 310.7 (br s), 133.8 (d, J = 11.9 Hz), 133.0, 130.6 (d, J = 32.3 Hz), 130.1 (d, J = 10.9 Hz), 100.7 (C5H5), 44.0 (CH3). Anal. Calcd. for C25H24F6NOP2Re: C, 41.91; H, 3.38; N, 1.95; P, 8.63; Calcd. for 25•(CHCl3)0.5: C, 39.30; H, 3.16; N, 1.79; P, 7.93; Found: C, 39.59; H, 3.35; N, 1.73; P, 7.68.

    1. The isolation of “greenish yellow leafs” of the ethylidene complex “25″ is claimed. However, upon further reading, this material is noted (?) to be a CHCl3 solvate. Therefore, the final sentence should read:

    After 2 days, greenish yellow leafs of 25•(CHCl3)0.5 (0.311 g, 0.434 mmol, 79%) were collected by filtration, mp 165 °C dec.

    2. The yield is incorrectly calculated. The formula weight of 25 is ca. 716, whereas that of 25•(CHCl3)0.5 is 776. The former value was used to calculate the yield, but the latter material was isolated. Therefore the yield is erroneously high. Corrected version:

    After 2 days, greenish yellow leafs of 25•(CHCl3)0.5 (0.311 g, 0.401 mmol, 73%) were collected by filtration, mp 165 °C dec.

    3. The only datum supporting the existence of the solvate is a partial elemental analysis. There is no direct evidence, such as (1) observation of the solvate by NMR, or (2) chlorine analysis. A reader might suspect that the researcher hadn’t bothered to check for a solvate, submitted the sample for analysis, and then obtained an analysis that did not fit the formula. The analysis was then “retrofit” to a solvate. This is poor technique. Someone who wishes to repeat this work really has no idea of what was really isolated. The following protocol would prevent this from occurring.

    3.1 After obtaining crystals, run a NMR spectrum in a solvent other than the recrystallization solvent(s) to see if any solvate is present. If a solvate is present, integrate (or add an internal standard) to determine the amount. Some solvates will be difficult to detect by 1H NMR (e.g., CHCl3 in the presence of PPh3). In these cases, a 13C NMR assay could be used. The solvate signals should be included in the NMR data reported in the experimental.

    3.2 Determine how “tight” the solvate is. Does it come off in vacuo (drying pistol)? Is it retained after several weeks on the bench? This should be part of the experimental.

    3.3 If your solvate contains Cl, request a chlorine analysis. The solvent quantity determined by microanalysis should agree with that determined by NMR.

    3.4 Other points:

    3.4.1 If you have isolated a solvate, it may save you work to change your isolation conditions to obtain a non-solvated material. Diffusion recrystallization conditions often give solvates (and three-solvent systems such as in this example should be avoided).

    3.4.2 If you are using a non-solvate, but only the solvate has been reported in the literature (or vice-versa), then you are using a new, previously uncharacterized form of matter. There is no need to repeat solution spectra, but you must put your new isolation procedure in the literature, and establish the purity of your new material (microanalysis)

    3.4.3 Avoid at all times “fantasy solvates”. This is a compound that did not pass microanalysis, and for which a “solvate” was invented, with no independent supporting evidence, in order to make the microanalysis fit.

  53. anon Says:

    Solvated vs unsolvated complex – this is exactly the example that I remember from my previous-previous uni. One guy submitted a paper to Polyhedron with a solvated complex in the experimental part (but the fake EA was for the unsolvated one). One of the referees spotted the disrepancy, so the lead author resubmitted the corrected manuscript with a fake EA of the solvate (explaining that he repeated the analysis on a purified sample). Another example from the same uni: the manuscript about some inorganic complexes (with x-ray structures) was almost prepared for submission, just one of the collaborators was asked to run the NMR. He found out that the complex cocrystallized with HBr (stoichiometric amount of HBr seen on 1H-NMR) – which was neither present in the x-ray structure, nor in the fake EA. So the x-ray was re-solved with HBr in the unit cell, and the EA recalculated and everything got published. As far as I know, this practice ran for years. The problem is that there is no written evidence…

  54. notafake Says:

    There is a recent Swiss paper from Firmenich which, to me, appears to follow the rules outlined by John Gladysz above, at least to some extent:



    Excerpt from the SI: “The microcrystals of 2e contain CH2Cl2 (0.5 equiv) as solvent of crystallization, as indicated by the integration of the H NMR spectrum. EA: Calcd. for C36H54Cl2N2RuS2·0.5(CH2Cl2): C 55.25, H 6.99, N 3.53. Found: C 55.03, H 6.52, N 3.36.”

    No Cl analysis reported but at least they give the basis for inclusion of solvent.

  55. Nick Says:

    Re: John Gladysz’s post

    This seems like pretty much the perfect way to handle this kind of thing.

    I strongly suggest this is added to the Instructions for Authors for Organometallics, and also for JACS and Inorg Chem (EA doesn’t seem to be used nearly as regularly by organic chemists so might not be so relevant for JOC/Org Lett).

    Sadly, I suspect most solvents added to EA are just trying to account for random unidentified impurities…so people will keep using them as a “fudge” factor.

  56. bad wolf Says:

    Second Nick’s comment, and thanks to Prof. Gladysz for an especially informative post. Always appreciated!

  57. Rhenium Says:

    What seems more unbelievable is two of the solvates added.
    Two deuterated chloroforms?
    0.8 pentanes?


  58. bacon Says:

    Good lord, this is getting excessive. Keep things in perspective people:

    Does the absence of this EA value change their results? NO

    Are the intellectual merits of this work diminished? NO

    I honestly don’t care if they made up an EA value- everything else was sufficiently defended and the field ultimately benefits from their work.

    What concerns me the most is how much excitement surrounds proving there was intent to fabricate an EA value for a single compound. Are the benefits of confirming the correct syntax of this PI’s comment really worth discrediting their entire work and damaging their reputations?

  59. ChemGuy Says:


    I agree with the fact that this particular EA is not crucial. If it was not so important, then why didnt the authors simply not include anything there. This would have been a better decision on their part.

    However, fabricating any data is on a slippery slope to fabricating more data.

    The essence of the way the current science field works is that I trust your data, because you try to accurately report what you do. If all the sudden one piece seems fabricated, then I start to question everything.

    I think many other people who work in the field have a similar attitude.

  60. anon Says:

    Your word is all you have. If you make your data up, then your word is no longer sufficient, and neither is your data, so what do your publications have that anyone can care about? All they have is your ideas. Without trustworthy data, ideas are worth less than a dime a dozen.

    There aren’t too many things people can get righteous about in science, but making stuff up is one of them. The importance of the data that would have been faked isn’t relevant (and based on the letter from Organometallics, the “argument from language ambiguity” holds as much water as an empty spaghetti sieve). The importance is in the decision to tell one of your students to fake data, which makes everything else you do hard if not impossible to trust.

  61. KGM Says:

    Bacon (and anyone else with a similar view):

    I’m appalled by your comment. The issue here is not elemental analyses. It’s truth and integrity. If students are taught that it is ok to fudge EA results just to get a paper published then some fraction of them are going to think it’s ok for fudge and outright fabricate other data and results. If you can’t get acceptable EA results on your compound then you report that FACT. If you suspect solvate formation then you report the FACTS that support that suspicion. You don’t teach students (or accept doing it yourself) to just fabricate compositions to match expectations.

    0.8 pentanes? 3.0 H2O for a compound shown by x-fay to be a CH2Cl2 solvate and prepared under air/moisture free conditions? Start doing that and you are well on your way to fabricating nmr data, etc. Section 5.4.1 of the Organometallics Author Guidelines clearly spells out the rules for EA including those that do not meet the 0.4% rule. Follow these rules or don’t publish.

    Actually as I write this I am more than appalled by your comment, I am disgusted by it.

  62. Chemical JUSTICE Says:


    +1 and I share the same groud with You.

    Report the FACT, no adjusting and playing around.

    Better do not publish and do not make from science some Canterbury tale…

  63. schinderhannes Says:

    Just a question to the crowd: what is the correct way to handle the following case?
    You make 50 gms of x, it looks nice, is used in next reactions, but nmr is so so. you recrystallize 500mgs get great NMR, EA is still of, you column, recrystallize, sublime, pray, light a candle and dry in high vac for ages, EA is spot on.
    Hoiw do you report this. You should write it down in epic details shouldn´t you?
    How come I never ever read about it? Cant be so uncommon?

    I am writing this to say (and I am a convinced atheist!)
    Let him who is without sin cast the first stone.

  64. Joseph Merola Says:

    @schinderhannes I honestly have never lit a candle for EAs. As a long-term candle lighter for other causes, does it work?

  65. schinderhannes Says:

    You´re asking an atheist, remember?
    I put “light a candle” here as a substitute cause I felt admitting that I instead worship the invisible spagetti monster by going to the local Italian for pasta (and to get drunk) sounded too silly!

  66. Chemical JUSTICE Says:


    I will tell you my own experience during my career being undergrad, grad, -postdoc.

    1. There were numerous cases that after several columns and tedious recrystallizations EA for compound was not good to be published (though according to NMR the substance was pure). Yeah, it was pain in the …. , but after changing the solvent system for crystallization, boiling it with activated carbon (which absorbed most of impurities as well as some amount of purified compound), then columning it again, it was possible to get very good EA.
    Yes it was difficult and really discouraging to do it several times, but the outcome was always positive. No one can guarantee that EA should be perfect after first attempt (even if NMR is very good one).
    2. Also faced the cases when the starting substrate was not pure enough and it caused all the further steps to contain the same impurity.
    Try hard, change conditions: try different reaction solvents (DMF for instance very often causes gumming when used as a reaction solvent, while switching to acetone/acetonitrile for the same reaction resulted in no side reactions (no gumming, etc).
    We do not live in the Stone Age and there are some purification methods known for a long time
    More here:
    1.Purification of Laboratory Chemicals
    By W. L. F. Armarego, Christina Li Lin Chai
    2. Organic Solvents: Physical Properties and Methods of Purification, 4th Edition
    By John A. Riddick (Author), William B. Bunger (Author), Theodore K. Sakano (Author)
    3. A Microscale Approach to Organic Laboratory Techniques
    By Donald L. Pavia, Gary M. Lampman, George S. Kriz, Randall G. Engel

    I should admit that I am unaware with a trick using a candle (maybe will ask some of my colleagues).

    P.S. The question arises whether you want to have a lot of papers or you have less, but you are confident with a quality (quantity doesn’t guarantee quality, this statement also rules in everyday situations).
    Good luck in your research!

  67. schinderhannes Says:

    @ Chemical Justice and all the rest as well:

    Maybe I was unclear but you missed my point:

    If the EA is not from the bulk material you should not put it in the experimental as if it was and as so many people do!
    You should write “NMR of bulk was so so, recrystallized material had following NMR, EA was of for both, after lengthy purification (including indulgment of alcohol by investigator), we managed to get a tiny sample with correct EA”.

    This I have never seen in writing though!
    It certainly is a far smaller sin than to fabricate EAs by calcualting phantom solvents in or doing worse, but still, I wanted to point out that there is a fine line to hypocrisy here…..

  68. schinderhannes Says:

    @ Chemical Justice
    on a more personal note:
    I went thru all that stuff you mentioned as well. I am a Principal Research Scientist in industry.
    Most of my publications are patents nowadays.
    And boy am I glad I don´t need to run EAs any more (last one was 12 years ago…)
    Nowadys I “light cancles” for clinical candidates.

  69. A recent PhD Says:


    All information depends of the context.
    If you want to show purity of your sample by EA, you have to use the bulk.
    If you want to use EA as a characterization method then It’s ok to send the ultraextrapurified/sublimed/prayedfor sample, but then again, if you really have to go through all that trouble, you know you are not really making what you claim, or/and you obviously have an NMR silent impurity on your sample, which then renders your yields completely useless as you should always report the yield of clean product or specifically state that you’re giving a crude yield.
    To be honest, I’ve seen people praying/petting/nourishing/and even singing to reactions in order to “make them work”, of course I don’t think people would really take seriously a procedure reporting that in order to obtain white crystals you should do the recrystallization while listening to Metallica… maybe in biology? or should you do the recrystallization twice changing the music you listen to in order to determine the influence of such factor! That’s a paper I would read!

    But besides the humorous tips some of us have to make some reactions work, academic misconduct is a serious thing, and nothing to laugh about.

  70. schinderhannes Says:

    Once again and without any sarasm, side jokes or anything that might cause you to misunderstand me just because I´m a little loose (and a non native speaker):

    A question to the crowd:

    How often do you think that the yields reported are for bulk material and the EA is for extra purified material?

    Have you ever read such a comment similar to “Yield is for bulk, EA is not bulk but specially purified sample to prove identity?”

    Most importantly: do you consider this (not reporting diffenreces between bulk and analytical sample) to be acceptable bahavior?

  71. A recent PhD Says:

    I have seen papers where the say stuff such as
    “analytically pure samples where obtained by recrystallization in such solvent”
    I believe that reporting the EA of a “specially treated” sample without stating it in the publication, is misleading.

  72. schinderhannes Says:

    @ A recent PhD
    thanx for the answer!
    And while were at it I forgot to mention that I had a supervisor who routinely made us go this dark alley insiting it was ok and standard procedure performed all over the chemical universe…..

  73. A recent PhD Says:


    Sadly, at the moment I’m doing some reported ligands, and while they claim yields of 66% I barely managed to get 10% of product and you guessed it, it wasn’t clean… “a yellow precipitate turned out to be a brown oil” :/ but as you said, we all know that when trying to synthesize a reported compound, the yields you will get are never the same. I would like to think that it’s mainly due to my personal lack of familiarity with the reaction and not that the person in question lied about it. But writing a procedure that’s is not reproducible is technically lying.

  74. Dorta Paper Link Roundup | ChemBark Says:

    […] News, Analysis, and Commentary for the World of Chemistry & Chemical Research « The OM Paper vs. Drinkel’s PhD Thesis […]

  75. anon Says:

    For the record, an editorial review of this case has been published.

  76. logical rigor Says:

    Interesting that he had to pull several analyses and use NMR instead. Massive paper rewrite (should be a retraction…impossible to read with all the errata).

    Also, that they pass on the moral issue. Seems pretty obvious that his institution should fire him for advising a student to lie.

Leave a Reply