Basing your own work on something that doesn't work is incredibly frustrating and can lead to enormous amounts of wasted effort. It doesn't even have to be fraud, there are so many factors you often can't fully control, and reactions can depend on very subtle details or minor impurities.
My impression is that usually the informal communication about stuff "that everyone knows doesn't actually work" is far more efficient than in your case. But this is something the PI has to do, as a new PhD student won't be connected enough for this, and your PI seriously failed you there.
It would be nice if someone published that this method doesn't work, but that doesn't seem to be how this works. The amount of effort to actually demonstrate that it really doesn't work is so much higher than the reward.
In a healthy environment people should have been much more sceptical much earlier. At the latest when you saw potential manipulations in the NMR. I'm curious what kind of artifacts you saw there, did they just remove or add signals?
> the informal communication about stuff "that everyone knows doesn't actually work"
Informal communication, as an important part of the system of "science", seems very underappreciated in nearby threads.
Science in quotes because even subfields can be very diverse.
Often the corrective mechanism isn't retractions or demotion, it's the hallway gossip at conferences, the "don't believe it - he (high-profile PI) sees what he wants to see". And associated differential aging-out of relevance. There can be a lot of science system state that isn't captured by the short-term state of the research literature.
But regrettably, as the stories here of smashed careers and lives illustrate, it can be very far from "everyone" that "knows". And a big difference between someone "knowing", and that being well expressed in their mentorship and leadership.
Thank you very much for taking the time to respond. I would like to see the world in a better light, and I do try most of the time; but it's pretty difficult for me when it comes to academic misconduct.
> It doesn't even have to be fraud...
I absolutely agree -- we all make mistakes and scientists are no exceptions. In my case, I honestly believe that nobody except for the student who manipulated the said NMR spectra initially committed any fraud.
As for what my PI did (or didn't do for that matter), that's really up to interpretation. Even in the unlikely case that nobody had told the PI, in the 10 years prior to my arrival, that the reaction doesn't work as advertised, there really aren't any excuses for not responding to my e-mails and simply brushed it off when I had told him what the issue was face-to-face.
> ... there are so many factors you often can't fully control, and reactions can depend on very subtle details or minor impurities.
I also agree. I left out the technical details earlier, here are a few other things I haven't mentioned:
* I had friends and colleagues check my calculations.
* I had friends and colleagues check the analytical data of my substrates and catalysts.
* I borrowed the same catalyst that a coworker made for her own reactions, which was made recently then, and it didn't work for the reaction I was trying to reproduce.
* I hunted down previous batches of the same catalysts in the entire building, none of them worked for the reaction I was trying to reproduce. It is worth noting that the catalyst is very stable under ambient conditions.
* I used my own batch of catalyst on other types of reactions reported in the literature and it worked as expected.
* The reaction is not supposed to be water/light/oxygen sensitive. I did try the reaction with and without Schlenk conditions, with and without light excluded, and combinations of them. Nothing worked.
* At some point I even had a few coworkers looking over my shoulder to see if I was doing anything wrong.
* When I used 10 times the amount reported in both the relevant paper and the PhD thesis, the reaction profile I observed was then consistent with what was reported.
> I'm curious what kind of artifacts you saw there, did they just remove or add signals?
Those artifacts happened in multiple spectra, there were three main types:
* In proton spectra, signals were just removed without much effort made as in noise simulation at the baseline. In addition, the signals removed were not just solvent and water signals. This is back in the days when signal removal wasn't so prevalent and accessible in everyday spectrum-processing software. Signal-removal in synthetic chemistry should never be allowed in the first place.
* In carbon spectra, there were regions that looked like signal truncation at first glance, but were definitely signals that got edited out (~0.3 ppm wide regions) and replaced by a straight line.
* In carbon spectra, in my friend's (who was an NMR practitioner then) words, "it looks like someone has DRAWN A VERTICAL LINE IN BLACK AT [multiple regions in ppm]". For context, I sent her high-resolution images for comments without telling her what they were or the issues I was dealing with.
There were other kinds of artifacts that I was less sure about, such as inconsistent phasing across different parts of a spectrum that hinted at parts from different spectra were stitched together.
I should note that not all of these spectra were related to what I was doing, but the spectra relevant to the reaction I was trying to reproduce had all of the artifacts listed above.
> It would be nice if someone published that this method doesn't work, but that doesn't seem to be how this works. The amount of effort to actually demonstrate that it really doesn't work is so much higher than the reward.
I think at that point the effort has usually been made and it's fear that stops people from disclosing such misconducts. My then PI was not a typical scientist, and "powerful" is the first word that comes to mind to most people when describing him (before "brilliant", "charismatic", etc., which also apply to him). Even though I had decided that I didn't want to do chemistry anymore pretty quickly after that, I never had the courage to try to correct any of it for the following reasons:
* I wasn't sure how it would affect my ex-bosses.
* I wasn't sure how it would affect the careers of those who are associated with the group.
* I wasn't sure how it would affect the status of my fellowship. I had already decided that I wouldn't do chemistry anymore after my contract was up, but I didn't want the extra burden of having to explain to future employers about what happened.
My impression is that usually the informal communication about stuff "that everyone knows doesn't actually work" is far more efficient than in your case. But this is something the PI has to do, as a new PhD student won't be connected enough for this, and your PI seriously failed you there.
It would be nice if someone published that this method doesn't work, but that doesn't seem to be how this works. The amount of effort to actually demonstrate that it really doesn't work is so much higher than the reward.
In a healthy environment people should have been much more sceptical much earlier. At the latest when you saw potential manipulations in the NMR. I'm curious what kind of artifacts you saw there, did they just remove or add signals?