Dear all,
Many thanks to all who responded to my, in hindsight, rather obvious question regarding x-axis gradient strength measurement (when you’re in a rush, you don’t have time to stop and think), in particular Geoff Akien, Dan Plant, Mike Brown, Ryan McKay, and Clemens Anklin.
Two main ways to do it, exactly the same as the z-axis gradient, viz. either measure the diffusion constant of a sample of known diffusion constant (e.g. water or doped water) and calculate from the discrepancy what the real gradient is or use a phantom and acquire an image. Both methods are described in the 100/150/200 and More NMR Experiments series of books (explicitly for the latter, and for the former, the correction is Grad (real) = Grad (“used”) x SQRT(D measured/D literature). The diffusion constant method is described in the attached Bruker DOSY manual and also in the pdfs attached from Mike Brown from Bruker, which also describe the phantom method. Other methods are described in the book by Bill Price, NMR Studies of Translational Motion: Principles and Applications, though the two aforementioned methods are probably the primary methods.
Though the measurement of a diffusion coefficient seems to be considered the more accurate method, temperature calibration and the question of the actual value of the diffusion coefficient are likely sources of considerable error. The phantom image method can be prone to user interpretation errors due to uncertainties about where to measure due to unclear boundaries, but it is much, much quicker (as long as a phantom is on hand, at least for the z-axis). The disk phantom for the z-axis seems to be particularly bad with respect to unclear boundaries, possibly due to the factors suggested in the comments section of the below blog post concerning gradient uniformity and B1 homogeneity, though acquisition parameterization can also be a source of this (see Price’s book). However, for the x- and y-gradients, estimating boundaries seems not to be a problem and it seems to be really something to do perhaps with the disk phantom considering that sharp boundaries for the z-gradient can be obtained with other phantoms (see Mike Brown’s pdfs). I have never seen good examples of boundaries for the disk phantom, including my own sample, but have gotten clear boundaries now for the x- and y- gradients.
For either method you can say it is the same effort for the x- and y-gradients compared to the measurement of the z-axis gradient with no alterations. (Note, the distance really is the internal diameter of the tube, not the radius, for the phantom measurement of just water in an NMR tube in the case of x- and y-gradients.) We were in rush to finish an abstract for ISRM so before the replies came in we used the diffusion constant method and then checked it later by the image method (which I rather prefer). Both were in agreement.
The disk phantom for the z-gradient seems an odd choice (though it is simple in design). You need to make a disk of exact size (width and diameter), that fits into the NMR tube nicely but not too loose so that it sinks to the bottom under its own weight but stays where you place it (too tightly so the air cannot escape should not be a problem as a small hole can be drilled into the disk, http://u-of-o-nmr-facility.blogspot.com/2008/03/gradient-calibration-1d-mri.html, but hopefully air bubbles will nevertheless not remain trapped), and made of not-too-soft material that it is easily deformed or gets skewed when you push it down the tube… I actually did make up a phantom like this many years ago but have since lost it over the course of moving from place to place, but it seemed more effort than was worth it tbh. Fancier and more complicated phantoms can be made up with more effort (e.g., see the pdfs from Mike Brown) but which do not move and should be easier to position, though one can also just use a Shigemi tube (so the opposite to the disk configuration where the water is of fixed and known length rather than defining the boundaries of a phantom). Or use a plug to fill the bottom of the tube and a plug of lighter-than-water material to place on top of a short sample of water which should be easier in principle if not necessarily in practice to make up. But for x- and y-gradients, no inserts are actually required as the phantom can just be water in a tube where the limits are bounded by the tube - and the physical dimensions are probably much more accurately known to boot. So actually, measuring the strength of the x- and y-gradients is easier in terms of making up a sample than measuring the z-axis gradient. Sometimes the obvious can be elusive.
Regards,
K. Klika
Received on Sat Nov 10 2018 - 10:04:25 MST