Quentin Reynard-Feytis (CEA Grenoble, France)
LinkedIn: @Quentin Reynard-Feytis
Abstract: Recent developments in MAS-DNP have dramatically enhanced the sensitivity of solid-state NMR, making it possible to perform increasingly complex experiments. Notably, this includes 13C–13C and 13C–15N correlation spectroscopy on powdered samples at natural isotopic abundance (NA), without the need for isotopic enrichment. Working with low-abundance nuclei also reduces dipolar truncation, thereby facilitating the observation of long-range polarization transfers. However, the potential of these techniques is often compromised by strong artefacts such as t₁-noise, which arises from instabilities during indirect evolution. Because t₁-noise is multiplicative, signals from abundant but uncorrelated nuclei can mask weaker cross-peaks, particularly problematic in NA samples where signal overlap is common.
In this study, we present a new approach to suppress t₁-noise in natural-abundance 13C–13C DQ-SQ correlation spectra. The method involves converting double-quantum (DQ) coherences into longitudinal two-spin order (zz-terms), followed by the application of a “zz-filter” to selectively remove magnetization from uncorrelated 13C spins. We describe the theoretical basis of the technique and demonstrate its application to both J-coupling and dipolar-based DQ-SQ experiments at NA. At 100 K, using a standard Bruker MAS-DNP system, we show that this filtering enables the clear identification of long-range cross-peaks previously obscured by noise. Furthermore, at 30 K on a helium-cooled MAS-DNP setup, where t₁-noise is typically more severe due to higher sensitivity, we observe substantial SNR improvements in the indirect dimension (up to 10×). These advances make it possible to acquire, for the first time, a reliable 13C–13C DQ-SQ spectrum at natural abundance and 30 K.
-
Hi Quentin, great work and nice presentation! Could you elaborate how zz-filter distinguish between “correlated” and “uncorrelated” nuclei at the quantum mechanical level? What happens to the overall sensitivity when you apply this filtering – is there a trade-off between noise suppression and signal intensity?
-
Hi Amit,
thank you very much !
before the DQ excitation block, the spins are prepared among the z-axis and from there you apply the DQ-excitation. At this stage, there is two possibilities:
– correlated spins: you generate DQCs (which are for instance DQy = I1xI2y + I2yI1x)
– uncorrelated spins: they cannot create DQCs so they stay along the z-axis
When we apply the zz-filter, we convert these terms such as:
– correlated spins: the first pulse will create I1zI2z terms from the DQy (cannot be completely converted)
– uncorrelated spins: the Iz magnetization is put in the x.y plan
The subsequent delay will diphase uncoupled spins’ Ix.y magnetization, but will preserve the I1zI2z terms that can only be formed from coupled spin pairs.
The second pulse will convert the I1zI2z terms back into DQCs, and the sequence keeps running 🙂
Let me know if that was clear or if you have any more questions !
-
update: I forgot to answer the 2nd part of your question.
Yes, there is a trade-off. For isotropic DQ-excitation (i.e., all the DQCs generated have the same phase in the DQ subspace), we loose a factor 2 when applying the zz-filter.
This is problematic, although we obtain a 5 to 12-fold noise reduction with the zz-filter, which leads to SNR improvements of ~2.5 to 6.
When facing relatively “low” t1-noise, the applicability of the zz-filter isn’t straightforward and there might be other solution more viable. (for instance Fred Perras paper 10.1016/j.jmr.2018.11.008 )
I hope this was clear !
Best,
Quentin
-
Thank you Quentin, for your detailed answer.
-
-
-
-
Dear Quentin,
Thank you vor the very nice presentation!
I was wondering if the efficiency of the zz-filter depends on the refocused lifetime of the coherences (T2′) and/or on the spinning frequency ?
-
Dear Chloé,
thank you very much for your comment !
You’re perfectly right, since we want the SQCs associated to uncoupled spins to diphase during the delay of the zz-filter, this needs to be done through various (Zeeman truncated) interactions: CSA, 1H-X couplings, 1H-1H couplings, etc…
Although all of these interactions are supposely averaged out by MAS, their still quite present in rather slow MAS speed regime (<10kHz). At higher spinning speeds, the time necessary to diphase the SQCs (which create the t1-noise) might be exceedingly high ! The T2' partially transcripts for how present these residual interactions influence the spectrum, so this is definitely linked ! A really long T2' might mean that the residual interactions are not strong enough to quickly diphase the SQCs, which can affect the zz-filter efficiency…
However:
– at natural abundance, the zz-terms lifetime is reeeaaaally long since they don't face spin diffusion issues, so in principle there is no limitation to extend the zz-filter delay.
– It is possible to add a DARR-field on the 1H channel, which reintroduces 1H-13C coupling, to speed up the decay of the SQCs during the zz-filter delay.
So in conclusion, yes it definitely matters and the stategy needs to be adapted, but one shouldn't face any major limitations….although this still needs to be proved 🙂
-
Thank you for the detailed answer
-
-
Leave a Reply