25 us pulses are fine, especially for just normal 1D experiments. They are routinely used, for example when determining 90° pulse angles.

It could be a problem if you want to record 2D spectra though, as very long pulses can cause problems, for example due to relaxation, or high power required for decoupling 13C in HSQC experiments.

Using 3 mm tubes instead of 5 mm can help a bit.

Dear Lorenzo,

Janis Advice is perfect - to add one comment:

with 25 us pulse you will have a sensitivity which is three fold reduced compared to the normal pulse.

With the 3mm sample I'd expect that you will not have a substantially longer pulse than normal. Due to reduced volume the sensitivity is about 3 fold lower. So: I expect that a 3 mm tube with the short pulse will not give about the same sensitivity compared to 5mm tube with the long pulse.

Finally the question is: do you have a 3mm Spinner?

Alfred

Hello everyone and thank you for your help.

I've tried acquiring 1D spectra with long pulses and they came out fine, we also got some TOCSY spectra after the bruker technician gave us the go ahead with the long pulses and they also look fine.

Unfortunately, as Jani predicted, we aren't able to get HSQC spectra but we can do without them if we don't find a way around the salt.

We do not have 3mm spinners but I think I can get my hands on some tubes that are 5mm at the top and 3mm at the bottom to try with the 5mm spinners we have if you think that it's worth giving it a shot.

I've also tried diluting a sample from 600uL to 1.2uL and that helped a bit, pulse time went down from 22us to 17.5, we even got higher signals.

Dear Lorenzo,

I am VERY much wondering why HSQC is a problem.

• What I do to estimate what NMR time it takes in a 2D HSQC I do the following:
• First acquire a 1D 1H
• write this spectrum to a 2nd process number (e.g. wrp 2)
• Divide the spectrum in procno 2 by 100 (dc 0.01, mulc) to accomodate for 13C natural abundance
• superimpose this spectrum with the noise of the original 1D 1H spectrum.

Doing so you can estimate how many more scans it will take to have a signal substantially higher than noise.

• If you need to multiply the down-scaled spectrum by 8 and it took 32 scans for the 1D 1H you will need (at least) (8*8)*32*2 scans in total in the 2D.
• For a HSQC with 256 t1 increments this means NS has to be at least 16.

All this is of course ONLY true for small molecules where relaxation can be neglected.

In our experience matching and tuning is NOT highly affected on 13C by high salt....

Goog luck

Alfred

The issue with HSQC is that most common pulse programs are fairly complicated, so there is some relaxation taking place during the pulse program. This is amplified by both the length of the pulses increasing this relaxation, and the sensitivity lost due to long pulses.

An HSQC sequence requires 180° 13C pulses which have a fixed length when using adiabatic pulses. Therefore you can't make them longer, and would have to adjust the power instead. Since approximately three-fold pulse lengths are required, the power would need to be increased by a factor of nine to bring the pulse lengths back to the normal values. This is beyond what your typical liquid probes can handle.

One alternative would be to use the simpler HMQC experiments, which contain fewer pulses, so they suffer less from the long pulses, and most importantly they don't require 180° pulses on the X nucleus.

Thank you again for your help.

I can try with HMQC if you think that will help. I'm up for trying various experiments what we are worried about is that longer pulses and higher power levels could damage the instrument/probe. Do you think that's the case? If so is there a way to determine if an experiment is safe?

Dear Lorenzo,

it is absolutely not clear what type of molecules you try to observe in your sediments. If the line-width LW of the signals gets close to J-(CH)-couplings you will loose most/all signals in HSQC with transfer delay D, if you are interested in low-MW compounds with a "few" Hz of line-width relaxation is no issue.

The total delay D gets long for multiplicity edited types of HSQC with senstivity enhancement (mostly used nowadays) .... have you adjusted all your shaped/adiabatic pulses in HSQC? You also can try the Bruker sequence hsqcetgp ... this is very simple not using any long delays or shaped pulses.

Alfred