Best coil design for homogeneus NMR RF pulse on liquid discs

In summary: To summarize, in my opinion, the coil you mention is not the best option for the application you are considering. I would recommend investigating other coil types or using a solids-NMR machine.
  • #1
_maxim_
105
6
Hi there,

we have a NMR spectrometer equipped with a standard bore 54mm cryomagnet working at 14.1 T (600 MHz Larmor Frequency) produced by Oxford.

NMRSystemSchematic.jpg


I would implement a special probe capable of generating a RF pulse and reading a NMR answer from a sample that is not standard for such experiments (typically a cylinder 40-60mm high x 5 or 10mm diameter in a standard glass tube) , but a "disc" of 1-2mm high only with a diameter of 5 or 10mm, contained into a glass tube of coarse.

Because of its lower height than the one of a probe coil, the disc introduces an enormous inhomogeneity and the linewidth is far from being a lorenzian line. There is no way to correct such a "bad magnetic field" by the well know "shimming" procedure.

I was wondering if there is a way to desing a perfect coil for this application, since the sample will be a disc of liquid substance. Of course, the coil circuitery must be complete with all the capacitors needed to implement a resonance circuit (Tuning and Matching) at a given frequency (i.e. 600 MHz):

img8.jpg


Obviously the RF coild must generate and read an electromagnetic field othogonal with respect the main field generated by the superconducting magnet.

Does anyone have experience on RF coil design?

Very quick intro in NMR
 
  • Like
Likes Nino MMP
Physics news on Phys.org
  • #2
I don't think you are going to like my reply. I think that 1) fabricating this coil is likely to be far beyond your abilities, and 2) in any case it's the wrong thing to do. I'll explain.
1. NMR coils may look simple but actually are exquisite works of inspired design and engineering. To list just three points,
a) The coil you mention is actually two wavelengths long, once you take into account the high dielectric constant of the water sample that normally fills it. Thus its properties are closer to a microwave traveling wave structure than to a lumped-element RLC tank circuit. The prediction of current flows, electric hot spots, to say nothing of matching it and expecting a uniform B1 field, is far from trivial.
b) The coil itself is typically made of a sandwich of materials carefully engineered to have zero bulk magnetic susceptibility, and materials for the support structure are similarly complex.
c) Designing the coil and structure to have very high Q is an art in itself.

2) In fact, a new coil will do nothing to improve the lineshape and width if it is due to B0 inhomogeneity. The problem is the sample shape. You can improve things by "padding" your sample, by embedding it into a standard sample tube that is filled with material of like magnetic susceptibility (assuming that your sample is liquid ). The glass interfaces will still cause problem. The best approach is to invest your time into collecting enough material to fill a standard sample tube so that it can be shimmed conventionally.

If your sample is solid, instead, then you need to get access to a proper solids-NMR machine that performs magic angle spinning. Machine your sample to fit into the special spinning sample holder. The strong spin-lattice coupling possessed by a solid sample makes any attempt to get a useful signal from a liquid NMR spectrometer virtually hopeless.
 
Last edited:
  • #3
Hi marcusl,

Thank you for your reply. I've read it carefully, and I would like to reconsider some points I've found interesting.

As a marginal note: I have been in the field of NMR from decades and I had to handle virtually every type of situation by acquiring multidimensional experiments in high resolution as well as in HRMAS or solid-state, with Room Temperature and cryogenic probes. I've dismounted, repaired and modified dozens of probes, so I know the implications of the matter.

That said, going more deeply into details, the goal of this project is create a probe capable of reading the NMR signal from a liquid sample located in several vertical positions different from the canonical (which is that one fixed and centered with respect to the main field B0). In order to minimize effects of the inhomogeneity of the magnetic field - whose linearity degrades after few cm from the center of the magnetic field - a sample in "disc shape" has been assumed to be the best: reduced height = less impact of the field homogeneity. That's why I was evaluating which kind of coil geometry could better match that form.

On the other hand, of course I can evaluate other possibilities (as spherical) if other forms were preferable. The point is that sample must be contained in a glass and "moved" along the z-axis of the main field to allow different field-dependent measures.
Obviously he problem of designing a resonant circuit for any different fields, or multiple resonant circuits (isolated) for each given field remain to solve.

Responding directly to your questions:

1a) The dielectric constant of the sample is mainly dominated by the solvent, which is 99% H2O
1b) I guess insulated copper will suffice at least in the beginning
1c) The resonant circuit will must mainly match the resonant frequency at that field, the Q-factor is not decisive since the measure will consists into a simple echo or cmpg-loop for T2 measures and the sample molarity can be arbitrarily high.

2a) I agree. Any suggestion here is welcome.

The challenge is very interesting (to me).
 
  • #4
So you are trying to do NMR on a liquid sample at different field strengths? And you want high B0 homogeneity so you can perform spectroscopy?
 
  • #5
Basically yes, this is the idea
 
  • #6
Your biggest challenge is to get a homogeneous B0 field. First, the flat interfaces of a thick disk (a wheel of cheese, if you will) perturb the field within in a way that can't be corrected by your shim fields, so you'll never get narrow lines. That's why NMR sample tubes are very long, ensuring that the ends are far from the central active spot. Stay with a conventional sample shape.

Second, the homogeneity of your magnet gets bad in a hurry away from the center. If you try to put your sample in a region where B0 is substantially different, the inhomogeneities will be too huge to shim away.

Third, your shim coils are carefully designed to operate well at the magnet center. Moving the sample up or down positions it where the shim fields are asymmetric--not good.

My suggestion is this: Change the field by dialing the Z0 shim up and down, since this changes the Larmor frequency without affecting homogeneity. If you need more range, charge or discharge your superconducting magnet coils a bit. It's the only way to change B0 without ruining the homogeneity. If you need more range still, put your sample into a variety of different NMR machines (i.e., 400, 500, 600 and 800 MHz).
 
  • #7
Ok: let say that the sample shape could be changed in order to obtain the best homogeneity reachable in a non-homogeneous field.

Has it to be spherical? Cylindrical? I can try to manage this requirement.
From this assumption, it arises the coil geometry.

The point is that we must measure T2's at different heights respect to the centered field, i.e. at different B0 -> Larmor frequency for the same sample, in a given magnet.

Doesn't matter if I have more than 10 fields available from 400 to 950 MHz (as it is my case) ... the goal is to perform these measures in the Oxford Magnet (14.1 T nominal field).

First of all, I guess I must map the field ...
 
  • #8
How big a shift are you looking for?

To address your question, spherical sample bulbs were used in the early days of NMR.

upload_2017-5-16_8-35-45.png

The glass sphere introduces asymmetric inhomogeneities caused by the stem on the top as well as the liquid/air interface at the meniscus, so performance is better with long cylindrical test tubes for reasons I mentioned above. Spherical bulbs are still available, but cylinders are used almost exclusively.

I wouldn't touch your coil. Your problem is B0, not B1, homogeneity.
 
  • #9
I came up with an idea: I could create a single chip containing the sample and the coil.

The liquid sample is collected in a suitable container (eg glass or other non-protonated material), closed, cylindrical and completely filled with the sample;
the coil is a simple solenoid wrapped around the sample container;
the chip "sample + coil" is then oriented in the direction normal to that of the field.

If the diameter of the coil (= sample) is quite small (2-4 mm), I expect B0's inomogeneity to be not so devastating. Even B1's homogeneity at this point would no longer be a problem.

upload_2017-5-18_11-1-56.png


The critical points I see in this solution are:

1) find the way to fully fill the cylinder without bubbles or meniscus;

2) find the way to mechanically move the chip along the vertical axis of the magnet using (sliding?) contacts to connect it to different RF compartments, depending on the specific resonance frequency of a given distance from the magnetic center;

For 2) I guess it's possible to use endless screws driven by stepper steers easily operated by an Arduino microcontroller.

Alternatively, wishing to use fixed positions rather then a moving chip, one could hypothesize a network of horizontal capillary tubes interconnected, with a battery of solenoids placed on each of them and connected to a separate RF circuits resonating at a given frequency.
The problem is to be able to select a single RF network among the many available, having only one output to handle ...

What do you think about?
 
  • #10
How much change in field (or frequency) are you seeking?
 
  • #11
marcusl said:
How much change in field (or frequency) are you seeking?

Ideally, I would like to acquire a T2 spectra in rather different fields respect to the main; for example, in the case of a magnet operating at 14.1 T (600 MHz 1H Larmor Frequency), 500-400-300-200-100-50-25 MHz might be interesting values.
More generally, each value is acceptable to perform a measurement; one just need to know at what field is being acquired.
Of course, initially even a single measure at a predetermined field is fine.
 
  • #12
I think your best bet is to move your sample to various spectrometers operating at their design frequencies.
 
  • #13
I see you point. Thank you anyway for devoting time to this discussion.
Max
 
  • #14
You're welcome. Good luck!
 
  • Like
Likes _maxim_

1. What is the purpose of using a coil design for NMR RF pulses on liquid discs?

The coil design is used to generate a homogeneous magnetic field that is necessary for the accurate detection and measurement of nuclear magnetic resonance (NMR) signals from liquid samples. It ensures that the magnetic field is uniform throughout the sample, allowing for precise analysis of the sample's properties.

2. What factors should be considered when designing a coil for NMR RF pulses?

When designing a coil for NMR RF pulses, the following factors should be considered: the sample size and shape, the frequency of the RF pulse, the desired homogeneity of the magnetic field, and the materials used to construct the coil (e.g. copper, silver, gold).

3. How does the coil design affect the homogeneity of the magnetic field?

The coil design, specifically the shape and size of the coil, can greatly impact the homogeneity of the magnetic field. A coil with a larger diameter and a smaller length will produce a more homogeneous field compared to a coil with a smaller diameter and a longer length.

4. Are there any limitations to using a coil design for NMR RF pulses?

Coil designs for NMR RF pulses have limitations in terms of the sample size and the desired homogeneity of the magnetic field. It may be challenging to achieve a completely homogeneous field for larger samples, which can result in less accurate measurements. Additionally, certain materials used in the coil may introduce noise or interference in the NMR signals.

5. How can the performance of a coil design be evaluated for NMR RF pulses?

The performance of a coil design for NMR RF pulses can be evaluated by measuring the homogeneity of the magnetic field, the signal-to-noise ratio, and the efficiency of the RF pulse in exciting the sample. These parameters can be measured using specialized equipment, such as a field mapping system or a probe head. Additionally, the NMR spectra obtained from the sample can also indicate the performance of the coil design.

Similar threads

Replies
4
Views
2K
Replies
4
Views
2K
Back
Top