Dear Clemens Anklin

You must be talking about the IUPAC unified scale. Unfortunately, I don't have any deuterated solvent with those referencing substances (TMS, DSS or TSP), I usually use the solvent residual 1H signal for referencing.

What I did for the external referencing is :

1. Acquired 31P{1H} spectrum of the sample (locked and shimmed);

2. Acquired 31P{1H} spectrum of H3PO4 (Unlocked, unshimmed);

3. Calibrated the 31P signal of H3PO4 at 0 ppm and typed ''sr'' command

4. Returned to the experiment where my sample was run, typed ''sr'' and entered the number I obtained in step 3.

Can you please tell me if I did right ?

Many thanks

Dear Salem,

Using the SR from the H3PO4 spectrum is indeed the correct way to apply external referencing for your spectrum.

Dear Luc,

Should I update it from time to time (the SR value) or record the H3PO4 spectrum juste once and use the SR value obtained to reference all the spectra I record ?

Thank you for your feedback.

I am not sure how often you are running 31P spectra, but best practice is to run the standard as part of every sequence of samples. If the referencing does not need to be super accurate, I guess you could get away with measuring it once a month or so, providing that nothing changes in the hardware of the instrument (e.g. no probe changes or other impactful changes).

Dear Luc,

Thank you very much for your answer !

Dear Salem,

just a comment: I'd expect that the chemical shift of H3PO4 does depend strongly on e.g. pH of the sample (which also holds for organic solvents with traces of acid!)....but I do not have a better idea compared to what has been proposed by Luc and Clemens....

Cheers

Alfred

Dear Alfred,

The 0 ppm reference for 31P NMR is the undiluted 85% H3PO4. Its pH is therefore fixed and not a concern. This is also the reason it is an external reference only. For internal referencing, non reactive compounds such as PF6 could be used. However, the compound must at least match the ppm range you're interested in. PF6 has a strongly negative chemical shift. And, as Clemens indicated, Xi referencing is also a good option in most cases.

Salem Baroudi There are two ways you can do this: the easy way or the harder way.

The easy way is as Clemens Anklin

The harder way is how everyone used to do it, and many probably still do. After running your sample, you remove it and put your standard in (H3PO4 for 31P). Without changing the frequency of the lock (ie. the Field value; and this means you definitely don't use autolock); run a single scan experiment; set the H3PO4 peak to zero; take note of the SR value; return to your experimental spectrum; set the SR value to match that in the H3PO4.

* Clemens Anklin

• TMS or DSS/TSP
• residual protio solvent peak from deuterated solvent
• protio solvent peak from protonated solvent (eg. if you're reaction monitoring in the reaction solvent)

The idea is that it's reproducible.

** If the referenced spectrum isn't in expno 1, the expno must be specified as an argument of the command; eg. "xiref 42" to reference to the spectrum in expno 42.

Luc Alders Related to your answer starting with "I am not sure how often you are running 31P spectra..." I submit that it is imperative to do the calibration for every sample, not just periodically or even just with every sequence of samples.

This is because the magnetic field the sample is experiencing is directly related to the FIELD value of the lock channel and the magnet drift over time. So, whenever these change (such as the FIELD value may be when using the auto lock; and the magnet is always drifting), the magnetic field for the sample is changed. It is typically not a large change if using the same solvent under the same conditions, but it could be quite different, and for that reason you must do it each time.

Here are a couple of examples where it might be quite different:

• In a solvent like toluene-d8 or THF-d8, where there is more than one deterated chemical shift, if it is locked on the 'other' peak
• If you turn the lock and sweep off (eg. if you are monitoring a reaction in the protonated solvent), it will be dependent on whatever was the last deuterated solvent the spectrometer was locked on

You can try this at home. Run a spectrum of H3PO4, set the peak to 0 and record the SR. Change the FIELD value for the lock by eg. 200 units, and run the spectrum again.

This is why the Xi reference method is nice - it's so easy to do, so why not do it?

I have set up the xaup program on my spectrometers for any nucleus other than 1H & 13C to do xiref as a part of the automated processing. For this, I also made a slight modification to the xiref AU program to silence the message box at the end (commented out the last line starting with QUITMSG).

@stevens....I am defenitely taking your answer into consideration, thank you for those details. I was wondering if you can share the AU prog. you did ?

Thank you.

Salem Baroudi , the AU program for the referencing is "xiref", which is bundled with TopSpin. I just started with that one, saved it as "xirefs", and changed the last line from QUITMSG(qmsg) to QUIT (because in automation I don't want the message).

For automated processing, I take the standard program "proc_1d" (bundled with TopSpin) and replace the command SREF (in line 55) with XAU ("xirefs", "") and I name that "proc_1d_xiref".

Set the processing parameter AUNMP to "proc_1d_xiref", and then it will use this if you invoke the command xaup (instead of eg. efp).

Thank you Steven !