The magnetogyric ratio of 1H is larger than any other nuclei. There is an increase in signal-to-noise ratio in HSQC/HMBC 1H detected experiments, compared to X detected 1D or 2D experiments (eg HETCOR)

It will also depend on the configuration of your NMR probe. Is the 1H coil or the X coil closer to your sample? If the 1H coil is closer, the 1H detected 2D experiments will be more sensitive, in addition to the points raised by Guilherme.

Dear Sirs,

This cannot be a sensitivity issue! I forgot to mention that I am dealing with highly concentrated samples (more than 40.000 ppm).

Thank you for your remarks.

At natural abundance, 13C is only 1% of all the carbons of your sample, so naturally, you only observe 1% of the concentration of your sample, but since the magnetogyric ratio of X nuclei is always lower than 1H, doing 1H detect experiments will always give you more signal than their X detected counterparts. Experiments such as HSQC (and HMBC) use the 1H magnetization, transfer it to the X nuclei, and then back to the 1H before detection. while X detected experiments start with the magnetization of the X nuclei directly. both HSQC and HMBC work by "enhancing" the detected signal related to that X nuclei

If sensitivity is excluded as an option (although 15N at natural abundance is very insensitive), it could also be that there is no power value defined in the parameter set for your pulses. For some reason, Topspin removed a while ago the power values for some of my saved parameter sets after an update. If there is no pulse, only noise is recorded. And Topspin does not give an error or warning if no power value is set when acquisition is started. The getprosol command should fix that (if power values are defined in the prosol table).

Dear Salem, for 15N direct observe experiments there are other factors to consider in addition to just sensitivity. The two major issues are heternuclear NOE and relaxation. Some nitrogen atoms can have very long relaxation times. In addition to acquiring a lot of scans you also have to use long relaxation delays. The other factor is heteronuclear NOE. 15N has a negative gyromagnetic ratio, therefor the maximum NOE can be -4. It is possible that the NOE is -1 and that leads to cancellation of the signal. For this reason it is often advised to use inverse gated decoupling (zgig) again with long d1's. Both 15N and 31P have a large shift ranges and the signal might not be where you expect it. If you have seen it in HSQC or HMBC you should know where to look.

The absence of signals in a direct X{¹H} NMR experiment (where X is a heteronucleus such as ¹³C, ¹⁵N, or ³¹P) while obtaining clear signals in 2D [¹H-X] HSQC/HMQC experiments can be explained by fundamental differences in sensitivity and experimental design. Direct detection of X nuclei suffers from intrinsic limitations including low natural abundance (e.g., 1.1% for ¹³C, 0.37% for ¹⁵N), low gyromagnetic ratios (resulting in weak signal intensity), and often long relaxation times (T₁) that require extended recycle delays. In contrast, HSQC/HMQC experiments overcome these limitations by indirectly detecting the X nucleus through polarization transfer from nearby protons, leveraging the much higher sensitivity of ¹H detection. This transfer process significantly enhances the observable signal from X nuclei, making them detectable even at low concentrations where direct observation would fail. Additionally, the direct X{¹H} experiment may suffer from practical issues such as improper pulse calibration for the X nucleus, insufficient ¹H decoupling during acquisition (leading to broadened signals), or incorrect spectral window settings that could cause signals to be missed. The 2D experiments inherently avoid many of these pitfalls by using ¹H excitation and carefully optimized transfer periods that select for only those X nuclei coupled to protons. For cases where direct X observation is necessary, improvements can be made by implementing ¹H decoupling, increasing scan numbers with appropriate relaxation delays, or using sensitivity-enhanced pulse sequences like INEPT. However, for routine analysis of X nuclei in protonated systems, HSQC/HMQC remains the preferred approach due to its superior sensitivity and reliability. The discrepancy you observe thus reflects the combined effects of fundamental NMR physics and experimental methodology rather than any error in sample preparation or instrumentation.