How much substance do I need?
Open-Access: Complete characterization incl. 13C-spectrum and a set of 2D spectra
You need 600 ul of a 50 to 150 mM solution of your molecule in a deuterated solvent. For a typical organic molecule this is about 10-30 mg. As an example, these are good settings for a 50 mM sample on one of our 400 MHz spectrometers:
- 1H: 8 scans (1')
- 13C: 1024 scans (43')
- COSY-FAST: 1 scans per increment (1.5')
- HSQC-FAST: 2 scans per increment (3')
- HMBC-FAST: 2 scans per increment (3.5')
Total experiment time is 52 minutes, 80% of which is spent for the carbon spectrum. The 2D spectra can be kept short because at this concentration the generous signal-to-noise on the 1H channel allows us to use non-uniform sampling (NUS) of the second dimension. On the 500 MHz spectrometer also 512 scans might be enough for a decent 13C-spectrum at this concentration. Very often the 13C-spectrum is not even needed for complete characterization because one can read off the 13C-shifts from the indirect dimension of the HSQC and HMBC spectra.
With a 100mM sample the 13C spectrum takes only one quarter of the time (on a 400 MHz spectrometer ca. 10 minutes). For the 2Ds the higher concentration does not make a difference because we need the 2 scans per increment for the minimal phase cycle. The total experiment time is then 20 minutes.
NMR is a completely nondestructive method. If you can afford it it makes sense to use a bit more substance to save a lot of time. If you do not forget to collect your sample you will not loose anything.
I have only enough for a 20 mM sample. Can I still use the open-access?
Yes, but you have to invest a bit more time. The 13C spectrum needs to be recorded with 4096 scans now to achieve a result as good as for the 50 mM sample. This takes approximately 3 hours. COSY and HSQC can still be recorded with NUS and the minimal 1 and 2 scans per increment, respectively. For an HMBC at this concentration we recommend 8 scans per increment and no NUS anymore.
Wait! We have only half as much substance as in our 50 mM sample but the 13C spectrum takes four times as long? This does not sound right. Unfortunately, it is true. The signal-to noise of a FT NMR spectrum does only increase with the square-root of the number of scans. So for a 10 mM sample we would have to measure 12 hours on the same instrument. If you want a carbon spectrum of such a sample you have to submit it to the service.
But proton-detected experiments including COSY, HSQC and HMBC are still possible in the open-access even at relatively low concentrations. For a 10 mM sample we recommend ns=1, ns=2 and ns=16 or 32 for COSY, HSQC and HMBC, respectively. At 5 mM a good HMBC will take too long, but COSY and HSQC are no problem: we recommend ns=1 and ns=4 or 8, respectively.
I have only 1 umole of compound. Can I get a 13C spectrum and how long will it take?
In 600 ul of solvent this corresponds to a 1.7 mM sample. If you consider what we have said above it would take more than 2 weeks to get a good carbon spectrum on a 400 MHz spectrometer. But you can submit your sample to the service. We have a 600 MHz spectrometer equipped with a cryoprobe optimized for 13C. The 13C sensitivity of this instrument is ca. six times higher than the 13C sensitivity of our 400 MHz spectrometers. Even on this spectrometer it takes half a day to get a good carbon spectrum with this amount of sample.
We are close to the limit now. For most organic compounds we deal with at LOC 1 umole corresponds to a sample amount smaller than 1 milligram. At this scale even small absolute amounts of impurities, humidity etc. can significantly lower the effective concentration of the sample and make experiment times of over one day necessary.