Minimizing the t 1 -noise when using an indirect 1 H high-resolution detection of unlabeled samples.

The most utilized through-space correlation 1 H-{X} methods with proton indirect detection use two consecutive transfers, 1 H → X and then X → 1 H, with the evolution time t1 in the middle. When the X isotope is not 100% naturally abundant (NA), only the signal of the protons close to these isotopes is modulated by the 1 H-X dipolar interactions. This signal is theoretically disentangled with phase-cycling from the un-modulated one. However, this separation is never perfect and it may lead to t1 -noise in case of isotopes with very small NA, such as 13 C or even worse 15 N. One way to reduce this t1 -noise is to minimize, 'purge', during t1 the un-modulated 1 H magnetization before trying to suppress it with phase-cycling. We analyze experimentally several sequences following the HORROR condition, which allow purging the 1 H transverse magnetization. The comparison is made at three spinning speeds, including very fast ones for 1 H resolution: 27.75, 55.5 and 111 kHz. We show (i) that the efficiency of this purging process increases with the spinning speed, and (ii) that the best recoupling sequences are the two simplest ones: XY and S1  = SR21 2 . We then compare the S/N that can be achieved with the two most used 1 H-{X} 2D methods, called D-HMQC and CP-CP. The only difference in between these two methods is that the transfers are done with either two π/2-pulses on X channel (D-HMQC), or two Cross-Polarization (CP) transfers (CP-CP). The first method, D-HMQC, is very robust and should be preferred when indirectly detecting nuclei with high NA. The second method, CP-CP, (i) requires experimental precautions to limit the t1 -noise, and (ii) is difficult to use with quadrupolar nuclei because the two CP transfers are then not efficient nor robust. However, CP-CP is presently the best method to indirectly detect isotopes with small NA, such as 13 C and 15 N.