Hi, I am an undergraduate student who is doing a chemistry/computing project. I have chosen to write a simple C program that can import NMR data from commercial spectrometer software into a common format. The scope of my project is restricted to "1D Proton" as time is limited. I'm not sure if the code is going to get used for anything later (I doubt it) and I am not intending to do anything with my work other than get a good grade! I have a question that relates to the phase correction parameters present in Bruker and Varian datasets: are the 'rp'/'PHC0' & 'lp'/'PHC1' parameters correlated in some way? I understand from a Varian manual that 'rp' is the zero order correction value and that the frequency-dependent 'lp' value is applied to each point in the spectrum such that the "pivot point" is at the right-hand edge. I have been trying to work this out empirically by importing a Varian data set into our lab's copy of TopSpin 2.1 using the 'vconv' command. But my results don't seem to make any consistent sense, and I'm hoping that someone here can point me in the right direction. Most times the value of 'rp' is simply "inverted" when I look at 'PHC0'. So, rp=35 will become PHC0=-35. However, I cannot see the relationship for 'lp'/'PHC1'. Only when lp=0 does PHC1=0. Other times the numbers are wildly different to each other. Can anyone explain what I am seeing? Thanks a bunch if you can help, Jean asked Mar 19 '13 at 18:57 |
Hello Jean! There is no trivial way to answer it. At the moment I am also converting data from JEOL to Bruker and Varian/Agilent to Bruker. First - very important is, where is the pivot point for the zero order phase correction. One of the more or less intelligent algorithms sets it to highest point (Bruker). But sometimes it is more prudent to take a big signal next to the right border to do there the zero order adjustment. This is also the procedures by Agilent/exVarian. To set the two red marks on the right hand side to do first the zero order there. Then set the red marks on the left side and do then the first order now. After this, you also should control the base line of the signal, where you have done the zero order before. No you have to do a rephasing of zero on the right hand side signal, too! aph, aphx, aph0 are the automatism commands by Varian. apk, apkf, apkm, ... are these by Bruker. apk, apkf, apkm are different algorithms, some do iterations! First it depends mainly, where you have set your pivot point. This leads to a derivation of values! Also there are differences in the sampling process between Bruker and Varian. Bruker is alternating between real and imaginary very fast. Varian does it slower, but the two channel at the same time! Bruker has a group delay in the beginning of the FID! Also the binary structure may be different, ... so this also leads to differences! NMR Spectroscopy Explained by Neil E. Jacobsen - Wiley - ISBN 978-0-471- 73096-5 for better reading! Yours, Ulrich Haunz answered Mar 20 '13 at 08:19 Ulrich Haunz I have to disagree on saying that the values deviate depending on where you put the pivot point. The position of the pivot point is completely irrelevant to the application of the phase correction. The pivot point only helps finding the correct values. It is just a "tool" offered by the programs. - Pascal Fricke (Mar 24 '13 at 03:51) Furthermore, Bruker spectrometers nowadays also acquire the points simultaneously. Only very old Bruker spectrometers still acquire sequentially, or if you manually set it to do so. - Pascal Fricke (Mar 24 '13 at 03:52) Yes, Pascal is correct. The pivot point is simply a convenience for doing interactive phase correction. The actual phase constants stored with the data have the first order phase zero at one of the other end of the spectrum. - Kirk Marat (Mar 25 '13 at 07:43) Thank you Kirk & Pascal, I came to the same conclusion too by playing around with pivot points and looking at the final result in the parameter values. - Jean Davidson (Mar 30 '13 at 12:23) |
Hi Jean, as far as I know, PHC0 = -rp and PHC1 = -lp. I never had any problems converting the phase values that way. The pivot point is only used to determine the point in the spectrum whose "overall phase correction" (frequency-independent zero-order and frequency-dependent first-order) does not change while correcting the first-order phase. So when changing the zero-order value, the position of the pivot point does not have any influence at all. But when correcting the first order, the position of the pivot point marks the area where there is no change in the phase, so that you can correct the phase looking at a signal far away. After this second step, you phase correction is done. So just to explain the meaning of the pivot point, a possible way of correcting the phase of the spectrum would be: 1.) Change the zero-order phase looking at one signal. This one signal should preferably be at one end of the spectrum and it should be strong enough to be able to nicely phase it correctly. 2.) Once this signal is phased nicely (using zero-order only!) you set the pivot point manually directly on that signal. Now you never touch the zero-order correction again and only change the first order. You will see that no matter how much you change the first order, the chosen signal will not change its phase. This is because the program (TopSpin, VNMRJ, ...) automatically recalculates the zero-order phase so that the overall phase at the position of the pivot peak does not change. This behaviour can now be used to find the correct first-order value: Look at a signal at the other end of the spectrum (the further away from the pivot point, the better) and change the first-oder correction until the phase is okay. Then you are done. So ideally, a 1D-spectrum has two phase correction values that can be found in the way I just described. And which these values are is completely irrelevant to where you put the pivot point. They are always the same. answered Mar 24 '13 at 03:47 Pascal Fricke Thank you Pascal, this simple relationship works for the 1H spectra I have! Being an inquisitive person I also looked at a 13C{1H} spectrum I happen to have lying around as well. PHC0=-rp & PHC1=-lp does NOT work here! I am confused why this is so as it is just another 1D. Any insight here? - Jean Davidson (Mar 30 '13 at 12:21) |
Hi, I don't know exactly the answer but I remember is reported on a paper by Wishart in J. Biomol. NMR. I don't remember the correct reference, of course :) I think (1997) 10, 373-382 best marco answered Mar 26 '13 at 00:48 marcs Thank you for trying but I cannot get that publication without paying for it online ... and I am just a poor student! ;-) - Jean Davidson (Mar 30 '13 at 12:21) |