As an undergraduate I used 1D 1H and 13C NMR to determine unknown structures, with either a given molecular formula or EI mass spectra limiting the number of possible options. The methods I came up with were well suited to test questions, but aren’t so good in a research setting. So, when I entered graduate school I had to mix things up, learning a new way of using NMR. This post will cover both methods. If you just want the research POV, jump to the next heading.

Building from the Ground Up

The only way to get good at reading NMR spectra is to worked through the bulk of your textbook, solving the practice problems [1]. Rote memorization will get you a bit of the way, but much of the learning curve lies in recognizing patterns in the peaks. High marks come from doing this fast. The NMR problems in an undergraduate course are falsifiable, and given enough time you can work them over until every solution is correct.

Most of these problems can also be solved with only the 1D proton and carbon spectra, even if you are introduced to more sophisticated techniques (ie. COESY, DEPT or HSQC). Even a three hour exam isn’t long enough to solve truly complex spectra, which limits what can be thrown at you [2].

The standard approach:

1. Using the given formula/mass , determine the unsaturation number. This tells you the number of double bonds and rings, and can be calculated from

U = (2*#carbon – #hydrogen – #halogens + #nitrogens + 2)/2

An aromatic ring is worth four (three double bonds and one ring), and all alkenes, ketones and cyclocarbons are one.

  1. Group the proton peaks into shift regions, according to the following table [3]. Memorize the shift of benzene, and use it to group aromatic peaks into “shielded” and “deshielded” camps.

Bear in mind that these are fuzzy ranges, as proton shifts vary depending on the neighbouring functional groups [4]. If no table is provided, memorize the broad ranges in advance [5]. Always pad non-aromatic ranges on the deshielded end. An example: hydrogens connected to ethers have 3.0-4.5 ppm shifts.

  1. Integrate proton peaks, if necessary. Start building functional group fragments, based on what you know so far (aldehydes, t-butyl group, etc.). Write everything down on scrap paper or in the margins.
  2. Determine coupling, and if possible coupling constants. If your professor is the type to get cute, learn how to build up to J4/ddddd with this simple technique. Use the coupling interactions to build more advanced functional group fragments (ethyl groups, cis-alkene, etc.).
  3. Turn to the carbon NMR and repeat step 2.
  4. Build a complete structure with the functional groups you think are in the structure.
  5. Test the structure, moving down through the list again. Does your compound have the same mass/formula as was given? Do your functional groups fit their expected shifts and integrations? Does the coupling match?
  6. If the structure fails for whatever reason, go back to step 6 and build it again. If it fails twice, go to the next problem and come back to it later.

If your problem comes with additional information (IR spectra, MS shows isotopes, etc), slide that in after step 5. 2D analysis isn’t normally required until peaks start to overlap, or you get into stereochemistry.

At the Research Level: What’s this peak?

The undergraduate approach builds from minimal information, but in synthesis a known structure has been transformed, and the structure of the proposed product needs to be confirmed. This shifts the emphasis to “key peaks,” which are only present before/after the proposed transformation, or are greatly shifted due to alteration of neighbouring functional groups.

The new approach, for pure compounds (ie. post column chromatography):

  1. At the start of the project assign the prominent peaks of your first starting material. To make this job easier you can search SDBS for simple structures, or use Scifinder to locate publications on the assignment of the more challenging molecules. Even if you’re working from a literature source obtain a full range of NMR spectra (1H, 13C, COESY, HSQC, HMBC, etc.) in your solvent of choice when time permits. This tends to come in handy later on.
  2. After the reaction select at least two key peaks, one from the starting material and one from the product. Verify their presence/absence. If you’re looking for a peak that has moved significantly, use HSQC to link the proton to its parent carbon. Carbon shifts are far less susceptible to changes in neighbouring functional groups.
  3. If the reaction proceeded as planned, check for impurities. Key peaks from relevant solvents are easy to memorize, and the spectra for your reagents can usually be found online at Sigmaaldrich or the SDBS. If the spectra is clean, double-check that you have all the tests you need. Then move on.
  4. When you have time, weeks or months later, assign each peak and build your experimental section. Completing several related structures at the same time limits the amount of repetitive work. If you have multiple rings you can use HMBC to group your proton peaks, provided the rings are linked by heteroatoms.

————————————————————————————————-

[1] Here’s a few databases if you’ve run out of all the interesting problems. (Link 1) (Link 2)

[2] Still study the more sophisticated techniques, just focus the bulk of your efforts at getting really good at 1D.

[3] More exacting shift tables can be found on Hans Reich’s website, but at the undergraduate level they’re counterproductive. You want your first assignment to be fuzzy.

[4] If this is a third or fourth year problem you’ll have to take spacial organization and anisotropy into effect in step 6.

[5] A trick to memorize the shifts, and other random pieces of data. Keep your shift table an arm’s reach away.  When you need it, pull the page over and read the shift, then put the page back. Pause a moment, then classify your peak.

Advertisements