Correcting for magnetic field drift in magic-angle spinning NMR datasets
Graphical abstract
Introduction
Drift in the main magnetic field in magnetic resonance techniques is a well-known issue. If not corrected for, the drift of the B0 magnetic field results in broadening of spectral peaks and distortion of lineshapes. In magnetic resonance imaging (MRI), using spectral registration, each spectral average is fit to a reference scan in the time domain by adjusting both frequency and phase terms [1]. In solution NMR spectroscopy, the issue is usually solved by correcting the magnetic field while the measurement is recorded. This field lock is typically implemented by tracking the field drift by detecting deuterium and constantly adjusting the main field with the use of room temperature electromagnets [2], [3].
In magic angle spinning (MAS) NMR measurements, this solution can be impractical, due to the small sample volume, which results in low sensitivity of deuterium and an unstable lock. Additionally, many probes built for proton detected MAS NMR do not have a deuterium channel. An alternative approach for carbon detected spectra is to use the proton channel for a lock [4]. External locks can also be used to track the main field. In this case, a small sample (e.g. containing deuterium oxide in solution) is placed close to the MAS rotor in its own dedicated detection coil [5]. Such external lock systems are available from instrument manufacturers, but they do not entirely remove drift, since they are not detecting the field at the sample, and in addition, they are often not temperature controlled, resulting in a long equilibration time before the temperature sensitive D2O sample can be used. The result is that MAS NMR data is often acquired while the main field drifts. Even adjusting the linear drift compensation of the spectrometer, we often observe drift of up to 0.075 ppm after a 24 h measurement. For 13C-detected data, this is usually insignificant, since linewidths are generally greater than 0.3 ppm. However, proton-detected spectra of deuterated microcrystals can have linewidths below 0.05 ppm [6], [7], and even the less ideal preparations of membrane proteins can have linewidths of around 0.1 ppm [8].
In solution NMR, the lock is also important in order to maintain good water suppression. This is because the typically used water suppression schemes, such as presaturation methods [9], [10], [11] and WATERGATE [12], [13] are highly dependent on the carrier frequency. With a small drift, the water suppression can be severely compromised. This is not the case in cross polarization-based proton detected MAS NMR, where the water is suppressed by relatively strong saturation pulses [14], [15], [16]. It is therefore possible to correct for large drift in the field, provided the drift is known. Alternatively, the drift may be assumed to be linear if the acquisition time is made sufficiently short such that a linear correction is a good approximation of the actual drift.
Section snippets
Linear phase correction of the FID
We apply two principles to acquire the best resolution possible under conditions of a slow field drift. First, we acquire data in short blocks of about 24 h, or even shorter, such that the drift over this time will be approximately linear. Longer acquisitions can be divided into 24–36 h acquisitions by use of reduced phase cycles and non-uniform sampling. Next, we correct the data in the time domain assuming linear drift occurred during the measurement. This drift is determined from a one
Application of drift correction
While datasets may contain no drift at all, we have found that occasionally, data would need to be thrown away if correction were not used. For example, the rate of drift is different and unpredictable for several days after filling helium. Also, when recording data over the course of several weeks, the linear drift compensation of the magnet may not be sufficient, and drifts of up to hundreds of proton Hz are not uncommon. The linear component of the field drift during measurements can be
Conclusions
We present a simple method to correct for drift occurring while recording MAS NMR spectra in the absence of a lock. We demonstrate that the detrimental effect of a significant field drift can be minimized if the data is recorded in short blocks and, after acquisition, a linear phase ramp calculated from the drift during each experiment is applied to each spectrum across all dimensions. We show that drift correction leads to a significant increase in peak intensity and improved peak shapes. The
Declaration of Competing Interest
None.
Acknowledgments
We thank Karin Giller and Stefan Becker for preparation of the VDAC sample. We thank Wolfgang Bermel for his help in understanding the data structure of the binary file containing time domain data, the ‘ser’ file, stored by Bruker Topspin, and initial direction in implementing the script.
This research was supported by the Deutsche Forschungsgemeinschaft through grants SFB803 INST 186/794-3 project A04 and the Emmy Noether program grant AN1316/1-1.
References (23)
- et al.
A field-sweep/field-lock system for superconducting magnets–Application to high-field EPR
J. Magn. Reson.
(2006) - et al.
“FLIPSY” - A new solvent-suppression sequence for nonexchanging solutes offering improved integral accuracy relative to 1D NOESY
J. Magn. Reson. Ser. A
(1996) - et al.
Purge NMR: effective and easy solvent suppression
J. Magn. Reson.
(2005) - et al.
Pre-SAT180, a simple and effective method for residual water suppression
J. Magn. Reson.
(2008) - et al.
Gradient-tailored water suppression for H-1-N-15 Hsqc experiments optimized to retain full sensitivity
J. Magn. Reson. Ser. A
(1993) - et al.
High-performance solvent suppression for proton detected solid-state NMR
J. Magn. Reson.
(2008) - et al.
Crystallization of the human, mitochondrial voltage-dependent anion-selective channel in the presence of phospholipids
J. Struct. Biol.
(1999) - et al.
beta-Barrel mobility underlies closure of the voltage-dependent anion channel
Structure
(2012) - et al.
Analysis of non-uniformly sampled spectra with multi-dimensional decomposition
Prog. Nucl. Magn. Reson. Spectrosc.
(2011) - et al.
Frequency and phase drift correction of magnetic resonance spectroscopy data by spectral registration in the time domain
Magn. Reson. Med.
(2015)
High stability nuclear magnetic resonance spectrograph
Rev. Sci. Instrum.
Cited by (19)
Magic-angle spinning NMR structure of Opa60 in lipid bilayers
2024, Journal of Structural Biology: XRaman imaging: An indispensable technique to comprehend the functionalization of lignocellulosic material
2022, International Journal of Biological MacromoleculesCitation Excerpt :Drift in the magnetic field is another issue, which will cause peak broadening and lineshape distortion. This is especially the case when analyzing biological macromolecules, making the interpretation very complicated [24]. Thus, the indirect structure information obtained from NMR needs to be validated by other accredited techniques [25].
No-D quantitative <sup>1</sup>H Nuclear Magnetic Resonance spectroscopy method for the determination of ethanol in distilled spirits
2021, Microchemical JournalCitation Excerpt :This is to prevent the interference of the solvent peak with that of the analyte. Also, deuterated solvents are used to address the drift in the magnetic field which is apparent for long run times [30]. Gama et al. and Valim et al. used the no-D qNMR technique in their analyte solutions and both studies showed good analytical results [31,32].
Fast time-resolved NMR with non-uniform sampling
2020, Progress in Nuclear Magnetic Resonance SpectroscopyCitation Excerpt :Recently, methods based on the Radon transform have been proposed for processing serial datasets which have controlled non-stationarity induced by temperature [57–60], and these can be potentially used to deal with time-resolved data. A commonly met source of frequency non-stationarity is magnetic field drift, and this is often corrected with dedicated processing algorithms [61]. A significant improvement in temporal resolution can be obtained by combining uninterrupted NUS acquisition with a processing method based on selection of “time windows”, as originally presented in the work of Mayzel et al. [49].
<sup>1</sup>H-detected characterization of carbon–carbon networks in highly flexible protonated biomolecules using MAS NMR
2023, Journal of Biomolecular NMR