termine the correct attenuation corresponding to a 90° pulse at 50 ms duration (see Exp. 7.1). Determine the phase difference between the hard and the soft pulse and take this into account for the actual phases used in the pulse program. You have to set:

td	32 k
sw	10 ppm
o1	on resonance of selected signal. If the software allows offsets for selective pulses, one can also put o1 in the middle of the ¹H NMR spectrum. However, the different phases of the selective pulses at different offsets must be determined.
p1	Gaussian shape, 50 ms length, transmitter attenuation corresponding to 90° excitation
p2	90° ¹H transmitter pulse
d1	2 s
d2	1/[2J(H,H)] − p1/2, typically 37 ms, calculated from J(H,H) = 8 Hz
ds	4
ns	16

5. Processing

Use standard ¹H processing as described in Experiment 3.1. Note that the signals of the coupling partners show the active coupling in antiphase.

6. Result

The figure shows the result obtained on an AMX-500 spectrometer. In a an expanded portion of the normal ¹H NMR spectrum is shown, in b H-12 was selected, giving the responses of both H-11 and (weakly) of H-13, and in c (d2 = 50 ms) H-15β was selected, giving the responses of H-15α, H-14, and H-16. Note that the coupling constant J(H-15β,H-16) can be measured selectively (4 Hz).

7. Comments

For the COSY part of the sequence exactly the same theory applies as given in Experiment 10.3. Note that the delay d2 determines the intensity of the “cross peak”. It may be necessary to perform the experiment twice, for example in order to identify spin coupling partners with both small and large spin-spin coupling constants. In the spectrum c, d2 was set to 50 ms.

The 90° Gaussian soft pulse used here can be replaced by many other types. You may try a 270° Gaussian, a half Gaussian with or without an additional purge pulse [3], or DANTE excitation (Exp. 7.4).