Settle in and get COSY!

In November 2014 Nanalysis was pleased to announce the expansion of their product line with the NMReady-60PRO.  With this we also extended our experiment library for the NMReady-60e.

We now include the *much-requested* homonuclear 2D experiment…..drum roll please…..the COSY!

COrrelation SpectroscopY (COSY) is one of the most commonly acquired 2D data sets.  Why?  Because, like most 2D experiments, it deconvolutes complex data.  A COSY can be useful for assigning resonances in a 1D spectrum because it literally shows you the proton spins that are coupled to each other.

I know, I know what you’re thinking here, and yes, I have talked a lot about multiplicities and how splitting patterns can be used to determine connectivity.  And they do! The beauty of 2D methods is that they make this information less ambiguous.  And, of course, less ambiguous = more obvious = less prone to misinterpretation = good call!

A COSY is obtained by stacking a series of 1D experiments that differ only in their ‘t1’ parameter. [1] What exactly do I mean?  Well, I mean that (in it’s simplest form) a COSY pulse program is simply:

1) interscan/relaxation delay
2) first pulse (p1)
3) resting time interval (t1) where the nuclear spins begin to relax
4) second pulse (p2)
5) data acquisition over a time (t2)

The 2D aspect to this data is introduced by performing successive Fourier transforms (FT), first on the rows of 1D stacked data measured at various t1’s.  Depending how short, or how long, the t1 time is the signal amplitude observed at t2 will vary.  Performing a second FT on the columns generates a contour plot, showing the hills (if J > 0) or valleys (if J < 0) of each peak formed.

To illustrate this let’s look at simple fragment – maybe an ethyl group taken from a zoomed in COSY of ethyl cinnamate.  I’ve shown both the 1D stacked-plot like display and the contour plot.

It doesn’t matter what data display we’re looking at; there are 4 peaks in both.  To simplify our data interpretation we can divide these peaks into 2 ‘types’: (1) on-diagonal and (2) off-diagonal or cross peaks.

The on-diagonal peaks are effectively just a 1D spectrum and can be largely ignored for 2D data treatment.  The 2 peaks on-diagonal show where the same signals align – the triplet with the triplet and the quartet with the quartet.

The cross peaks are where all the information is!  These cross peaks show the nuclear spins that are connected through magnetization transfer – i.e., the phenomenon that affords spin-spin coupling.  Because a COSY is a symmetric plot each spin correlation will be observed twice – once on either side of the diagonal.  From left to right across the top – we see quartet (on horizontal) aligned with triplet (on vertical) and triplet (on horizontal) aligned with quartet (on vertical).

So these cross peaks show us the correlation between bonding groups.  The CH2 is bonded to the CH3, in a vicinal (3-bond coupling) relationship.

So instead of manually looking at the spectrum like we do with a 1D a noting that a quartet has 3 neighbours i.e., is bonded to a methyl group (n + 1 = 4 so n = 3) and a triplet has 2 neighbours i.e., is bonded to a methylene (n + 1 = 3 so n = 2) we can see the correlations!

Okay, that makes sense!  Let’s expand our view to the whole ethyl cinnamate COSY – we see 8 peaks – 4 on-diagonal and 4 off-diagonal.  The 4 on-diagonal peaks don’t show connectivity so we can narrow our focus to the 4 cross peaks that result from magnetization transfer.  We highlighted the ethyl groups above (correlations shown in green), but we can also see the geminal (2-bond) couplings for the vinyl protons.  These are highlighted in blue – again, two signals for one bonded pair.


COSY’s typically show up to 4 bond J 1H-1H and come in all sorts of flavours.  Shown here is a COSY-90 because it has the highest signal-to-noise ratio (SNR).  If SNR is too high and the diagonal becomes overwhelming, a COSY-45 can be run to reduce this and make cross peaks more evident.  Double quantum filtered DQF-COSY can remove lower couplings so as to only see higher order transitions.[1]

Typically these experiments are done to determine 1H-1H correlations, but coupling between other nuclei can be determined using the same method……

[1]Silverstein, R, M.; Webster, F. X.; Kiemle, D. J.; “Spectrometer Identification of Organic Compounds” 7th Ed. John Wiley & Sons Inc: USA