Pure shift methods can greatly improve the resolution of proton NMR spectra. However, current pure shift spectra show small periodic artefacts that prevent their use for studying dilute mixture components. A new technique, compatible with all current pure shift methods, is presented that suppresses such sidebands to arbitrary order, allowing ultraclean spectra to be obtained.

Pulse sequence diagram:
The new SAPPHIRE method gives essentially complete suppression of sidebands by manipulating the phase of the residual
modulation through small (ms) timing changes.

Choosing parameters in interferogram pure shift experiments using SAPPHIRE :

In any pure shift experiment using interferogram acquisition, the two main decisions needed are the digital resolution to aim for in the final spectrum, and how clean a spectrum is required. The better the resolution, and the cleaner the result required, the longer the experiment will take. Interferograms are constructed from M chunks worth of data each of duration 1/sw1, where sw1 is an even integer submultiple of the total spectral width sw3, so each data chunk contains sw3/sw1 complex data points. sw1 is typically chosen to be about twice the maximum width of a multiplet; this restricts the sideband artefacts that result from acquiring pure shift data in chunks to a few % of parent peaks. Increasing sw1 reduces artefacts still further, but increases the total experiment duration needed. The total number of complex data points acquired is M*sw3/sw1, giving a limiting digital resolution in the transformed spectrum of sw1/M Hz per point.
In SAPPHIRE experiments a small number N of such interferograms, acquired with slightly different timings, is averaged; for simplicity of programming, N is restricted to even values. Averaging N interferograms suppresses sidebands to order N–1; in normal circumstances N=4 is sufficient to give very clean spectra.

In the Topspin implementation the incremented timings are carried out in a 3D acquisition framework, where F3 is the direct dimension, F2 the SAPPHIRE dimension, and F1 the pure shift inerferogram dimension.

Parameters (with Topspin names in brackets) are therefore set as follows:

sw3 (3 SWH): as usual, sufficient to accommodate the full range of Larmor frequencies of interest in the direct dimension
sw2 (2 SWH): 2*N*sw1
sw1 (1 SWH): about twice the maximum width of a single multiplet; in the parent pure shift experiment this is subject
to the requirement that sw3/sw1 be an even integer, but SAPPHIRE adds the requirement that sw3/sw1
be an multiple of N
td3 (3 TD): at least 2*sw/sw1 + 2*GRPDLY, but best set to 8-16k so that individual FIDs can be Fourier transformed
and examined if necessary (e.g. to check shimming)
td2 (2 TD): N
td1 (1 TD): M + 1 (the extra 1 is to accommodate the small timing shifts used in SAPPHIRE)
ns (3 NS): as usual, the number of scans to be averaged in each individual acquisition; usually at least 2. The total
number of scans in the complete experiment is td2*td1*ns
d2 (3 D2): at least p16 + 2*d16 + 1/(*4*sw1), typically 1 ms or so greater

To suppress chunking sidebands to first order in a 500 MHz 1H spectrum, typical values would be
3 SWH=5000; 2 SWH =200; 1 SWH =50; 3 TD =16384; 2 TD =2; 1 TD =32; 3 NS =2; 3 D2=0.009
This would give a limiting digital resolution of 1.6 Hz per point

There is a typographical error on page 10189 of the paper; the expression "Min(...) ... Max(...)" should read "Max(...) ... Max(...)", as is used in the pulse sequence code.

Varian / Agilent
Pulse sequence:
sapphire.c (right click and chose 'save ... as')

pm_fidadd_v (right click and chose 'save ... as')
pm_pshift_v (right click and chose 'save ... as')

Pulse sequence:
sapphire_PSYCHE (right click and chose 'save ... as')
sapphire_ZS (right click and chose 'save ... as')

AU programs:
pm_fidadd (right click and chose 'save ... as')
pm_pshift (right click and chose 'save ... as')

Parameter Sets:
Raw data can be downloaded from the following DOI:
DOI: 10.15127/1.309229

N.B. 1. Topspin is not yet fully compatible with the floating point data acquisition used in Neo consoles, so in some versions of Topspin is may be necessary to convert interferogram pure shift data to integer form (e.g. with the Bruker AU programme sertoint.ptg) before processing with sapphire_pshift.

N.B. 2. The latest version of Topspin 4 has lost some backwards compatibility, and gives an error message when processing the original SAPPHIRE data obtainable from the DOI above. Removing the files audit.txt from the raw data restores normal function.

Moutzouri, P.; Chen, Y.; Foroozandeh, M.; Kiraly, P.; Phillips, A. R.; Coombes, S. R.; Nilsson, M.; Morris, G. A.
Chem. Commun. 2017, 53, 10188-10191
DOI: 10.1039/C7CC04423B