A Advice for 100 MHz resonator

[BillKet]

Hello! I would like to apply an electric field of 100 V/cm, along the z-axis (defined by an externally applied DC field), with a frequency of 100 MHz at the location of an atomic cloud (everything is inside ultra high vacuum). We can assume that the atoms are located in a volume with a diameter of 50 microns (e.g. in an optical dipole trap). However, I need to minimize any induced magnetic field along the z-axis at 100 MHz as much as possible (ideally at the nano Gauss level or below). Does anyone have any advice on how to proceed? I am interested mainly in two things:

1. What is the best way to design such a setup mechanically, to have from the start a magnetic field as small as possible? Naively, I can imagine having 2 parallel plates, 1 mm apart, with 10 V applied on them at 100 MHz (part of a resonant RLC circuit). But I assume there might be smarter ways to design this (e.g. specific plates design to minimize edge effects?). Also, for example, I am not sure how parallel these plates can be to each other and relative to the externally applied z-field, or how I could measure this.

2. How can I measure such a magnetic field in practice, and what are the best ways to minimize any non-zero magnetic field actively? For example, if I knew the magnetic field at the center of the plates (where the atoms will be), I could place some coils above and below the capacitor and generate a 100 MHz magnetic field that would counteract any residual magnetic field along the z-direction. But in practice, I am not sure what would be the best way to do this. Also, I am not sure what device is the best to measure such a place in-situ and how well can I even use it. For example, how can I be sure that I am measuring the B-field along the z-direction I am interested in, and not along 0.01 degrees relative to the z-direction?

Any advice or reading recommendation would be greatly appreciated. Thank you!

 

[Baluncore]

For example, how can I be sure that I am measuring the B-field along the z-direction I am interested in, and not along 0.01 degrees relative to the z-direction?

You cannot. Perfection is the enemy of progress.

A uniform electric field can be generated between two mesh grids.
A uniform magnetic field can be generated inside a Helmholtz coil.
https://en.wikipedia.org/wiki/Helmholtz_coil

If you can measure the errors, trim coils and small grid electrodes can be used to correct near uniform fields.

 

[BillKet]

You cannot. Perfection is the enemy of progress.

A uniform electric field can be generated between two mesh grids.
A uniform magnetic field can be generated inside a Helmholtz coil.
https://en.wikipedia.org/wiki/Helmholtz_coil

If you can measure the errors, trim coils and small grid electrodes can be used to correct near uniform fields.

I was hoping for a more detailed reply... I know some stuff in general terms, but I am not sure how I would proceed in practice. For example "If you can measure the errors" - as I said in the original post, I don't know what is the best way to measure tiny RF magnetic fields in practice.

 

[DaveE]

what is the best way to measure tiny RF magnetic fields in practice.

SQUIDs are the most sensitive, I think.

 

[BillKet]

SQUIDs are the most sensitive, I think.

But I thought that the high RF electric field and the relatively high RF magnetic field in the x-y plane (assuming I won't align the SQUID perfectly parallel to that on the first attempt) would damage the SQUID.

 

[Baluncore]

I don't know what is the best way to measure tiny RF magnetic fields in practice.

Whatever you introduce into the field, to measure the field, will distort the field.

Turn challenges to your advantage. If the angle must be precise, then use the reason it must be precise to access the deviation in the angle.

 

[DaveE]

I am not sure how parallel these plates can be to each other and relative to the externally applied z-field, or how I could measure this.

This is a distant analogy, but this reminds me of the issues we had decades ago with perfect alignment of the mirrors in laser resonators. Our solution was an active servo system that adjusted for a peak in the response. Getting as close as possible to zero is an easier problem than measuring zero.

 

[DaveE]

Since you are creating (or can measure) the 100MHz source of the fields, I would try a lock-in amplifier for synchronous detection of the z-axis field, regardless of the sensor type. This can greatly reduce noise issues improving sensitivity.

I guess I'd try a simple coil or solenoid for the detector and focus on really good signal processing to see how well that works first. Then at least you'll understand a lot more about what needs to be improved.

Sensor directivity and alignment might be a difficult issue too.

No one's said it yet, so I'll ask. Can shielding be part of your solution? Sounds like not, IDK.

 

[BillKet]

Since you are creating (or can measure) the 100MHz source of the fields, I would try a lock-in amplifier for synchronous detection of the z-axis field, regardless of the sensor type. This can greatly reduce noise issues improving sensitivity.

I guess I'd try a simple coil or solenoid for the detector and focus on really good signal processing to see how well that works first. Then at least you'll understand a lot more about what needs to be improved.

Sensor directivity and alignment might be a difficult issue too. No one's said it yet, so I'll ask. Can shielding be part of your solution? Sounds like not, IDK.

I can do shielding if that is possible, but I am just not sure how. For example, if I use 2 parallel plates, I might have some induced currents in the surface of the plates, that would generate a magnetic field at the center if the plates are not perfectly parallel. I am not sure how I would even shield against that.

 

[berkeman]

I have been holding off replying so far in this thread. @BillKet do you have access to EM simulation software through your university and research project? Can you show us a 3D CAD drawing of your setup so far?

I can do shielding if that is possible, but I am just not sure how.

To allow a LF RF E-field like 100MHz and exclude the B-field, I was tempted early in your thread to suggested a shorted turn, but until I can see a scale drawing of your setup, it is hard to suggest anything, IMO.

 

[Malamala]

I have been holding off replying so far in this thread. @BillKet do you have access to EM simulation software through your university and research project? Can you show us a 3D CAD drawing of your setup so far?


To allow a LF RF E-field like 100MHz and exclude the B-field, I was tempted early in your thread to suggested a shorted turn, but until I can see a scale drawing of your setup, it is hard to suggest anything, IMO.

I don't have a design yet, I am in the design phase. I am thinking of a measurement, and I did all the math, but this kind of magnetic field is the main systematic. So before deciding if I should pursue this idea, I would like to know if it is even possible at all to achieve this. At this point, I am open to any design suggestions.

 

[tech99]

If you use a parallel plate transmission line and tune to its resonant frequency, the electric field max position will correspond to a zero, or node, in the magnetic field. Maybe this would be a starting point.