J-filters, low-pass

In its standard implementation, the suppression of the undesired one-bond correlations is far from complete using a one-step low-pass J filter HMBC pulse sequence. Effective modifications are needed to achieve this goal when there is a wide range of Vch values. There are several kinds of pulse... 

In 1983, Kogler et al. had proposed that the low-pass J filtering process can be repeated and improved several times with different intervals x, x2, in order to properly eliminate all 1Jch peaks when a wide range of one-bond coupling constants 1Jch is present in the molecule under study (Figure 4).15 After the low-pass J filter, as shown by Equation (8), the... 

For a one-step low-pass J filter, the one-bond responses /lpi are expected to be proportional to... 

Obviously, the one-step low-pass J filter produces the narrowest profile and is therefore the less efficient in removing direct correlations in HMBC spectra, while the five-step tuned low-pass J filter (f lps) offers the most broadband profile. For this filter, the intensity of the residual /ch signals remains below 0.11% for the range 125 < /ch < 225 Hz, which is remarkable. Interestingly, as long as the /ch couplings range of the molecule remains moderate (—60 Hz), the four- and five-step filters... 

Attenuation of the long-range magnetization after N-step low-pass J filters... 

Unavoidably, the efficiency of the different low-pass J filters in removing direct correlations induces attenuation of the signals associated with long-range couplings. For some molecules, the choice of the most suitable low-pass J filter may be first determined by the attenuation of those signals. Neglecting relaxation, pulse imperfections and offset effects, the... 

TABLE 2 Maximal residual transverse proton magnetization and maximal attenuation for nJCn coupling constants <8 Hz < 20 Hz and < 30 Hz for one-, two-, three-, four- and five-step low-pass J filters... 

For non-experienced users, which may prefer to use one single experimental parameter set suitable for all samples, or for being used in automatic mode, we found experimentally that the three-step low-pass J filter offers the best compromise between 1Jch suppression and "/ch attenuation. It provides excellent results for nearly all molecules, and only in very special cases a four- or a five-step low-pass J filter will be required. 

The efficiency of the various low-pass J filters may be appreciated by considering the HMBC spectra of 1,3-butadiynyl (ferf-butyl) diphenylsi-lane (Figure 6), which provides a stringent test because of the extensive range of 1Jch coupling constants present. 

Partial plots extracted from HMBC spectra recorded with the five different low-pass J filters are shown in Figure 7. Clearly, ambiguities may arise, particularly if accidental degeneracy in chemical shifts causes true long-range correlations and residual 1Jch correlations to overlap. 

The results shown in Figure 8 are in good agreement with the theoretical results shown in Table 2. The value of the 2/c2hi coupling is very large, 49 Hz, and the attenuation during the respective low-pass J filters is subsequent (left). It must be pointed out that relaxation has not been... 

Figure 8 Comparison of nJch intensity illustrated with ID rows taken from HMBC experiments recorded using a one-step low-pass J filter (denoted LP), a two-step LPJF (denoted LP2), a three-step LPJF (denoted LP3), a four-step LPJF (denoted LP4), and a five-step LPJF (denoted LP5) showing the nJch response of C-2 (left) and the nJCn response of C-3 (right) of the 1,3-butadiynyl (tert-butyl) diphenylsilane molecule dissolved in CDClj. The measured signal-to-noise (obtained using the sino macro) are indicated on the top of each peak.

Figure 9 Timing diagram of the BIRD-HMBC pulse sequence for the detection of nJch correlations, including an additional two-step low-pass J filter. Thin and thick bars represent 90° and 180° pulses, respectively. 13C180° pulses are replaced by 90°y — 180°x — 90°y composite pulses. <5 is set to 0.5/(Vch) and A is set to 0.5/("JCH). Phases are cycled as follows fa = y, y, —y, —y 4>j = x, —x fa — 8(x), 8(—x) fa = 4(x), 4(— x) ( rec = 2 (x, — x), 4(—x, x), 2(x, —x). Phases not shown are along the x-axis. Gradient pulses are represented by filled half-ellipses denoted by Gi-G3. They should be applied in the ratio 50 30 40.1.

The low-pass J filter efficiency obtained may be improved by implementing a second low-pass J filter element. Taking advantage of the double-difference principle, this two-step filter is expected to yield high... 

Figure 13 Timing diagram for the clean HMBC experiment with an initial second-order and terminal adiabatic low-pass 7-filter.42,43 The recommended delays for the filters are the same than for a third-order low-pass J filter. <5 and 8 are gradient delays, where 8 — <5 + accounts for the delay of the first point in the 13C dimension. The integral over each gradient pulse G, is H/2yc times the integral over gradient G2 in order to achieve coherence selection. The recommended phase cycle is c/)n = x, x, x, x 3 — 4(x), 4(y), 4( x), 4(—y) with the receiver phase c/)REC = x, x.

For those purposes, the authors used constant-time version of the sensitivity-enhanced HMBC sequence,79 combined with a two-step low-pass J filter. Constant-time experiments have no coupling structures in the carbon dimension making it easy to identify the centre of signals in... 

For D-HMBC experiments, we usually omit the low pass J-filter (the first 90° pulse for C nucleus in the HMBC pulse sequence) aiming to suppress the cross peaks due to the direct Jc-h correlation, but it can be implemented if desirable. In the D-HMBC spectra, the cross peaks between directly bonded C and H do not, in most cases, hinder the easy analysis of the spectra, because these cross peaks appear as singlets. On the contrary, these peaks even contribute to easy NMR spectral analysis when HMQC spectral data are not in hand. 

Kogler H, et al. Low-pass J filters. Suppression of neighbor peaks in heteronuclear relayed correlation spectra. J. Magnet. Reson. (1969) 1983 55 157-163. 

---