Sciencemadness Discussion Board

Cancer of science: the impact factor

kristofvagyok - 5-12-2012 at 14:26

Why I am writing on a home chemist forum about the impact factor? Where probably just a few people is going to publish an article and start to gain some IF. Easy, I am writing because of the problems what this type of cancer had caused to me.

At first: sorry for my English.

Second: nowdays in the scientific life impact factor will say how awesome you are, how many citing have your articles gained and several places they give money for your researches after this number. And because of this a few people publish anything for this.

Why is this a problem? I have opened a recent copy of the Tetrahedron Assymetry and found an article about making enantiopure BINOL from the racemic with and easy method. They said that I should dissolve the rac.BINOL in acetonitrile, add some S-proline, boil it for a while, filter, recrystallize and isolate the pure 99,88% ee. L or S BINOL in a 75% yield. All this illustrated with fancy crystallographic data and ect.

The only problem I have noticed that the 85% of the BINOL was still in the acetonitrile after the workup and didn't reacted. The second thing was that the filtrate was almost the pure proline, with just traces of the BINOL-Pro salt and after recrystallization it gave a poor 8% yield and it wasn't enantiopure even. Shit happens.

What is the problem? They usually lie about the yields, the 99% is never-ever true, not even the 80% and they often miss a few things from the articles what is hard to find out and it is often critical performing a synthesis (e.g.: how to recrytsallize, purify, ect.)

One other memorable experiment, my first preparative experience what was long ago:

My fist ever "had to make chemical" was 2-carboxyethyl-5-bromo-indole and we had 4-bromophenylhydrazine as a starting material. It looked like an easy experiment, a simple Fischer indole synetesis with some ethyl pyruvate. The first part went awesome, the hydrazone formed, crytsallized and I've got an almost white crystalline solid, but the ring forming caused some problem. It was described in a patent to use polyphosphoric acid and heat it to 100Celsius for a while. I've got a brown-black insoluble mass, similar to potting soil. NMR spectra was strange, because all aromatics have gone. The other strange was that in the patent they described the pure material as a brown solid... -according to my data it contained no indole, so they simply lied.
After a lot attept I had made the goal compound as a white-yellowish beautiful crystalline compound, but it was made by my method and not after the journals.

I hate the current scientific life, because if I want to work after a published journal or a patent I am sure that I won't end up with a thing that they described. And why is this? Everyone wants some IF and they publish a promising experiment immediately and write 99% to the yield what is simply not true.

So, dear sciencemadness members, especially the not so experienced part of the forum: If you find an awesome article what says they have got 99,9% yield with a fancy e.g.: triflate catalyst, then think it again and do not order anything expensive for your experiment, because I promise: it won't work.

sargent1015 - 5-12-2012 at 14:47

I'm going to have to agree with this. I was working up a purification of a simple polymer from a journal and their procedure was rather vague. Needless to say, I needed to create my own conditions with what little help they had given me. Plenty of failed attempts ensued before I got it right. A little extra ink on a piece of paper could have saved me hours of time.

497 - 5-12-2012 at 15:31

Maybe if the peers where actually performing the reaction to review it these problems wouldn't be so common?

As for patents, it makes perfect sense. They just want to give enough info to get approved, but not enough to be actually useful to anyone else.

smaerd - 5-12-2012 at 15:58

I never trust patents for that reason 497. As for the journal's and the yield lies, it's a sad truth. I'd rather know to expect a realistic 40% yield then be hoping for that 80% that's sort of 'promised' in publications. I think I read on this forum some large figure about Chinese scientists admitting to accepting bribes and lying in publications...I view science as a sort of romance between man and a search for knowledge. It's really upsetting to hear about all of these perversions. I smile when I read my experiment note-book at the failures and 15% yields, because I learned from them, but I kind of view journals as a method of preventing others from reinventing the wheel.

I feel like it would be very important to create a resource for practicing scientists, amateurs or professionals, to give ACTUAL results of experiments with descriptive conditions. Kind of like an open-forum such as this. There seems to be a good deal of resources for researchers to collaborate, and home-grown computational chemistry data gathering, but still nothing like this.

Ultimately as stated above by other posters it seems like a lot of journals leave out critical details, and imply that many experimental conditions are implicit and see no need to be explicit. Which is a huge mistake, not everyone who is reading an article to make a given compound as a piece of a different area of research is seasoned with a particular reaction and if they were they probably wouldn't be reading the article after-all. I don't know how many journals I've read where there is no mention of using an inert atmosphere for reactions where it is critical or similar conditions. Sure it may be implied, or to some common lab wisdom, but leaving out little things like that can cost other fellow scientists a lot of time... Science is supposed to be a gift economy, sure it's contents may be esoteric to the majority of people, but forcing fellow's to read between the lines isn't beneficial, and isn't helping anyone out.

Maybe that's partially why a lot of these bunk articles get published. Is peer reviewers can't be bothered to recreate the experiments and decypher their supposed conditions. Just an idea.

Kristoff, I have done some basic research about fischer indole synthesis and it does seem like a lot of them promise easy work-up and high yields but actually give as you described the classic "oh look my flask turned completely black".

Or the classic "The product was isolated with column chromatography in 95% yield", what kind of eluents were used, or do I have to run 10 TLC's to find out something that might work myself? Not a big gripe but still, it's not helping anyone to leave information like that out.

stoichiometric_steve - 6-12-2012 at 01:22

As a rule of thumb, avoid chinese, pakistani and indian papers. Most of those i've read coming from these countries were complete bollocks. You can't really blame them, though - exaggeration is a cultural thing with them, at least in india/pakistan.

Endimion17 - 6-12-2012 at 05:53

Quote: Originally posted by stoichiometric_steve  
As a rule of thumb, avoid chinese, pakistani and indian papers. Most of those i've read coming from these countries were complete bollocks. You can't really blame them, though - exaggeration is a cultural thing with them, at least in india/pakistan.

Reminds me of their exaggerated, servile e-mails. Damn that creeps me out. :o

So, patents, eh? I learned long time ago that just because something is patented, it doesn't mean it actually works.
It just means it's patented. :)

kristofvagyok - 6-12-2012 at 11:31

Quote: Originally posted by Endimion17  
So, patents, eh? I learned long time ago that just because something is patented, it doesn't mean it actually works.
It just means it's patented. :)

There are several patented applications about "free energy generating devices" :D

To avoid chinese, indian and like that journals is a good point. Trust German, Russian and USA papers, especially the old ones. Long ago (before 1970-1960 the scientific publication was based on successful and reproducible experiments).

Also I would like to make a little advertisement for where they reproduce the recipes and they publish them just if they have worked. Also the comments and the "small letter info" is also written down.

Magpie - 6-12-2012 at 20:31

Len1, who is a physics (IIRC) professor at a university in Australia, often commented on the dishonesty and worthlessness in technical papers. He felt that his peers were driven to publish at all costs. I can remember from my college days long ago a professor telling us that the rule was: "publish or perish."

I have learned through my home experiments to look at all technical papers and patents with a jaundiced eye - at least those published since say 1965 or so.

But even procedures published in the early 1900's can be found worthless. One in particular is a German procedure for making malonic acid from malic acid by Dessaignes. Recently I attempted to make sodium 2-naphthalene sulfonate using a procedure from Gattermann's 1937 book. It also was worthless, at least in my hands. It called for cooking the reactants for 6 hrs at 170-180C. By that time all of the naphthalene had evaporated. Luckily I noticed that benzylchloride1 had success using the procedure in Vogel. I tried that and now have more of the salt than I'll likely ever use.

[Edited on 7-12-2012 by Magpie]

DJF90 - 7-12-2012 at 00:46

I've done reactions where the claimed yield was never obtained (and significantly above what I achieved) but at the same time I know there are still honest scientists out there. A good indicator for me is to look at the paper and see if there are any mediocre yields mentioned; I'd find it hard to believe that every reaction they did gave an excellent yield. Fortunately not all groups are anal about publishing as much as they can; the research group I was part of would not publish until they had something realy good, and until there was sufficient quality data for it to be accepted into a good journal like Angewandte or Nature.

Practical interpretation may be another thing that causes problems with replicating results. Unless otherwise specified, run it under inert atmosphere (and in anhydrous solvent). This will cause you the least problems. I also find that on repeating an experiment several times (and knowing what to expect) makes higher yields more achievable. Getting a good yield on a first attempt is an act of god. I find 2-3 runs is what I need to be comfortable with a new reaction. I'll always do an attempt or two on small scale (10mmol or 1g, whichever is most practical) before scaling up. This way I can work out any niggles without wasting a whole bunch of material, and I also gain experience with the procedure at hand.

Also remember that the guys writing these papers have been running these experiments for a while (1yr + in some instances) prior to publication, and they'll only publish the best yields for each reaction. So if at first you don't succeed, try again.

turd - 8-12-2012 at 09:08

Concerning the subject of unrealistic yields, you might also be interested in:

Even manipulations aside [cheating, making 10 publications out of one, inter-group self-citings, trivial papers with 10 authors (one to hold the beaker, one to press the I/O-button,...), review articles, ...], the whole concept of 2-year impact factors is idiotic. It's only a measure of conformism with the current fad, i.e. of scientific worthlessness. Wherein lies the gained insight of the millionth compound characterized with method x or the millionth complex of metal y? Sure, this data should be published, but as an indicator of scientific relevance? Puh-leeze!

I know the ironic situation of two competing journals in a field, where the one with a much more rigid review process and distinctly higher quality publications has about half the 2-year impact factors of the other. I think that says it all.

DJF90 - 11-12-2012 at 05:47

It might also be worth drawing attention to this: I've just come across it while looking for something unrelated.

SM2 - 11-12-2012 at 07:19

One other tactic, especially when used by contemporary big pharma, is to try to cover AS broad a claim as they can, They will throw out a very wide net, and pull in any fish which could (under the patent) be claimable. Even if that info is worthless, synthetically. All the variations (you pick the best temperature), it is sometimes meant to confuse. When big bucks are at stake, things have to be arguably true, but they can like hundreds of conditions and let the casual observer pick one.

Mildronate - 11-12-2012 at 08:39

Actualy problem is deeper than impact facor, example: student forking in university research lab or something, profesor ask results and dont acept bad results, there is time deficite and so, student is not the main author, but done all work from A to Z, or other example here need results for reserch project and somebody lay about them, its not about impact factor. I never read patents.