Inorganics/compounds with no H or C in your sample won’t show on NMR. Even HPLC with NMR is not perfect, for the same reason, though JMC accepts this. Characterisation data of a compound that includes EA and a mpt are fundamentally useful, since after that a melting point will do. Not doing EA on a new compound is slightly sloppy, particularly if the compound’s destined for biological evaluation. Do you eat your cake with no icing, too?

But we’ve all done it. Readers of papers need to evaluate an experimental section critically. No EA -> can I trust it?

(I’m an inorganic/organometallic chemist) I love x-ray but, as we all know, that’s just representative of THAT crystal. That’s fine if you, say, grow a crystal of some minor product or decomp. but if it’s a major product you should spare enough for EA. HRMS is good but a lot of metal compounds don’t fly or blow up even w/ soft methods.

I guess I’d say one should use all the methods possible to characterize their material. Every project is different.

I have NO problem when the prep is for “crude” material (1H and 13C basically clean and in the SI) and a phrase like “analytically pure material may be obtained by crystallization from XXX”. This can be important if the compound is really soluble and it’s easier to make it 95%+ pure than purify it ’cause you loose 50% when you recrystallize it but you at least did it once. I also think that “one pot” reactions should at least have the spectra of each step of a typical reaction reported and/or in the SI. I guess I see the above examples of being intellectually polite to anyone following your preps more than some example of draconian/stickler/ attitudes.

Is part of the problem not having enough time to do all the things that you think should be done?

I study a class of coordination polymers that cannot be recrystallized. All characterizations are done as synthesized. I have to count on ALL of the starting materials and solvents being pure as well as dry, and the reaction rate is as slowly as I can possibly let it without being obsolete. Unfortunately, the chance of getting a nice single crystal is still pretty slim after all the precautions are taken, and NMR is incredibly difficult to do on these polymers. So CHN analysis becomes one of the most important information that we rely on to identify the product.

I guess the necessity of EA depends on the type of compound you’re working with. In my case, it’s just about the only thing that brings definitive conclusion.

Isn’t the acceptable alternative to EA doing Hi-Res MS? Even if one does not get in Hi-Res mode it would seem that some type of MS would be reported (although as cookingwithsolvents suggest not everything flys well). EA is still useful when qualifying a reference standard but can be a pain. I would find it hard to believe that IR would disappear unless have some other technique such as Raman, UV or NIR to characterize.

In real world pharma manufacturing NMR has limited use except for characterizational efforts whereas MP are routinely practiced to evaluate materials. I have seen too many chromatography compounds that where “pure by NMR & IR” to put much stock in the value of such testing.

I recently wrote up my undergraduate thesis, and my supervisor made me get EA and HRMS on *all* starting materials, intermediates and final products.

I’m wondering if I need to get EA for a paper I am just finishing now. I made a new metal complex and characterized it with X-ray, mass spec, and NMR. I tried to get EA but cant get them close enough to include in the paper. I’m going to give it one more shot with some crystals I recently grew but I doubt they will be any better. Seems like a huge waste of time to me..

I would still include it in the paper. This kind of oddity is exactly what would have some scientists scratching their heads for weeks if they needed your compound in the future.

In times past EA was needed to understand composition but that can now be done by exact mass. Although the technique was once considered the ultimate arbitor of purity, it has largely been supplanted by NMR, LCMS and/or HPLC. Although these methods are not fool proof, in this age of constantly trying to do more with less the latter methods offer a fairly high throughput way to reliably give gather information on the pruity of products arising from routine synthetic transformations on large numbers of substrates.

Still there is nothing like the exerimental rigor of EA coupled with the above mentioned methods to demostrate composition and purity of the material prepared. Crystallization is one of the most astetically satisfying parts of being a synthetic chemist but unfortunately one may not have the opportunity to employ it on a routine basis on final products when using high througput synthesis techniques.

Many times I’ve seen people running small columns and doing recrystallizaions of a few milligrams in order to get a “correct” EA of some compound. In my opinion, this is useless since the extra-pure material you get this way is not the same as the material you use to continue with the synthesis. In principle it would be necessary to purify all your material the same way, but then you don’t do it because that would mean losing too much of it.
The problem is that the elemental composition must be accurate within a range of, say, 0.4%. If it lies outside that range, you cannot include it in a paper.
Regarding the issue of trust, an elemental analysis is extremely easy to forge!

I think u are right, gone are those days (when I was doing my Ph.D.,) we used to spend lot of material for analysis purpose and even we did spend lot of material for pharmacological activities to study the toxicity (like LD50). But with easy availability of NMR (many varieties like 2D, COSEY, NOESY & other analytical methods), nowadays, I agree with u. And in my opinion, again it depends on the number of steps involved and the target compound (cost of raw materials)..

As I said before, I’m cool w/ the prep being for (e.g.) 85% crude yield with analytical samples being recrystallized from saturated pentane (or something similar). If the material was pure enough for what you did in the paper as-prepared then just SAY SO. If you needed analytically pure catalyst to get that one tough substrate to get to 95% ee but not for benzaldehyde then say so. No one is going to blink when they read that in the experimental.

HRMS is awesome for your compound but it is a method of purification if there is that little GC in front. I’ve also seen WAAAY more “strange” things happen because people were putting too much stock in HRMS than EA. Supposed “H-bonded dimers” (not the product at ALL), ACTUAL “H-bonded dimers”, no parent ion at all (esp. w/ tBu groups in ligands, etc), dirty/residual sh*t in the MS, hydrolysis back to starting material, etc etc. Now, I won’t propagate some of the things that are possible with EA but I DO know that some companies seem to always match the numbers you give them. . . . Sure, there are potential problems with everything.

Bottom line: if the compound is stable you should attempt EA in my line of work. If you consistently get (e.g.) low N or something you should note it in the experimental w/ a typical result or an average of three different batches if the compound is really important to the premise of your paper (30-step natural product synth’s aside…again, common sense and all). Just because we have the *luxury* of many fancy analytical methods being “routine” doesn’t excuse intellectual and scientific laziness when we get an unexpected result using a “routine” technique.

Just to give a better insight into the source of my strong feelings on the matter, here’s an incident from my PhD school (different lab): Final step “clean” by mutinuclear NMR, 2-D, etc. complete crap by EA. . .why? The fancy-pants organic molecule was binding LiCl on the extraction into CH2Cl2 and kept it in CDCl3. Crystallized on rotovaping the vial (not surprising, related compounds did, too). Was low enough ACTUAL yield that the extra mass from the LiCl gave 85% yield. X-ray had been done on a previous variation of the system so they didn’t do that one.

Good thing the person “wasted” 10mg on EA (of 150, probably, given the person and lab). That boss is a stickler for EA (and was even more after this incident). I’m guessing he’ll tell that story to new students till the end of time. . .

why? The fancy-pants organic molecule

I’m going to start calling all of my molecules ‘fancy pants’ when they misbehave. Thanks for giving me a mid-morning chuckle

So is it acceptable for me to just say in the paper that I tried EA several times and they were always off? I know exactly why they are off so I could add that to it.

Personally, I think that EA has become a waste of time, its just an older method.

-Eric T. http://www.jazdchemicals.com

If you know WHY they are off then by all means list a typical result and your reasons for why they are off w/ evidence. All of this probably only warrants a mention in the experimental section unless it’s central to your theme.

An somewhat common example from the inorganic world: one of your potential CVD single source precursors to HfO2 has low volatility and you get a lot of C in your thin films (incomplete “decomposition”). The EA of the precursor is also low in C, presumably also from the formation of Hf carbides, even with the O2 around.

I believe new compounds should be completely characterized. Now, if you have a few mg of priceless material trying to recrystallize a solid or measure bp (I work with lots of oily compounds) is not worth at all!
For me HRMS is the way to go.
I really don’t think IR is going to disappear.