Laser interference fringes fading in and out of view

[PokemanDude]

TL;DR Summary

In my lab, the interference pattern from a mach zehnder interferometer fades in and out of view, and we can't determine the cause. How can we go about fixing this?

Hi! I'm working in a lab this summer doing quantum optics, and my team is working on a Mach-Zehnder interferometer setup. We got it well aligned and interference fringes appeared as expected. However, soon after building it, the fringes began to fade in and out of view without us changing anything about the setup, and we can't figure out why.

This is how it looks: there is a circular beam showing up on a screen with dark interference lines going across it. The interference lines fade to the point where the beam looks uniform as if there is no interference and then they'll come back into view. This happens quite often (about 10 times per minute) and it's random rather than periodic. As the fringes fade, they mostly don't move or change in size. Typically, the only thing that changes is their brightness, although occasionally the bright and dark fringes change places.

The only thing we know of that changes randomly over time is the laser intensity. It fluctuates a slight bit, but we found no correlation between the laser intensity and the visibility of the interference fringes.

What could be the cause of this issue, and how should we fix this?

 

[DaveE]

I'm betting on an issue with your laser coherence. Do you have a way to measure that, like a spectrum analyzer or a stable cavity?

 

[PokemanDude]

We don't have that equipment. What would you suggest?

 

[WernerQH]

The interference lines fade to the point where the beam looks uniform as if there is no interference and then they'll come back into view. This happens quite often (about 10 times per minute) and it's random rather than periodic.

That suggests a coherence time of the order of 6 seconds.
I think the beam splitter works as expected only if the input beam has a definite polarization. Have you checked it?

 

[sophiecentaur]

If the fringes change in peak amplitude then could one leg have varying loss? Do they shift laterally? I'm just thinking in basic interferometry terms.

 

[DaveE]

That suggests a coherence time of the order of 6 seconds.

Coherence time normally refers to the path length difference. 6 seconds would imply that an interferometer with a 5 light-second path difference (1.5e9 meters) would work OK, I doubt that for any laser. I think frequency stability is a better term for this.

 

[sophiecentaur]

I think frequency stability is a better term for this.

If it's stability then wouldn't the interference pattern march across the screen?

 

[DaveE]

If it's stability then wouldn't the interference pattern march across the screen?

Maybe. Maybe too quickly for a person to see. IDK, there are a lot of possibilities here.
For example, what if the laser switches modes, essentially instantly. like TEM₀₀ → TEM₁₀, adding new frequencies above the fundamental.

 

[sophiecentaur]

. Maybe too quickly for a person to see.

Then, wouldn't the fringes be just a blur?
I often read threads about the behaviour of laser beams in which people seem to draw parallel with the very basics of diffraction and interference. If you listen to be beats between interfering radio signals, you can expect any interference pattern to be racing across the country. I'm an old RF engineer but the same theory can often provide solutions to questions that the 'jet pilot' laser engineers come up with.

 

[DaveE]

Then, wouldn't the fringes be just a blur?
I often read threads about the behaviour of laser beams in which people seem to draw parallel with the very basics of diffraction and interference. If you listen to be beats between interfering radio signals, you can expect any interference pattern to be racing across the country. I'm an old RF engineer but the same theory can often provide solutions to questions that the 'jet pilot' laser engineers come up with.

Maybe, IDK. It could be quite complicated.
Like TEM₀₀ → TEM₁₀ → TEM₀₀ →...

I don't have a hypothesis, beyond "laser". This isn't based on the description, it's based on my experience that the laser is the most complex part and most likely to be the source of your problems.

Also based on the fact that many people that try to use lasers won't actually ante up to support them with an accompanying "laser lab" and skilled staff. For example, why don't we even know the laser type here (let alone mfr, model, etc.). I don't mean to be insulting, but for a laser guy (which I am NOT!) this question is similar to asking the surgery forum "I've cut open their chest and can see the heart, now what should I do."

BTW, back in the day, the really big and successful laser company I worked for made a lot of money with field service support (contracts mostly) for this sort of thing. OTOH, our stuff was expensive, and if the customer didn't appreciate the life cycle and down time costs, they wouldn't buy from us to begin with.

 

[sophiecentaur]

I don't have a hypothesis, beyond "laser".

But 'waves is waves' and a free space light beam would only have one mode (no?). I know that polarisation can be relevant so perhaps the splitter could be affecting it. We get this effect from reflections along an RF transmission path and the block of apartments and the local hill are hard to characterise.

 

[DaveE]

a free space light beam would only have one mode (no?).

No. That would depend on where they came from.
https://en.wikipedia.org/wiki/Transverse_mode

 

[sophiecentaur]

No. That would depend on where they came from.
https://en.wikipedia.org/wiki/Transverse_mode

IS that the link you want to give me? It makes the point that the other modes are in waveguides, fibres and other resonators. They are not sustained in free space. TEM is the free space mode (unless you can give me other information).

 

[DaveE]

IS that the link you want to give me? It makes the point that the other modes are in waveguides, fibres and other resonators. They are not sustained in free space. TEM is the free space mode (unless you can give me other information).

The modes are created in wave guides, but they don't magically vanish when they leave the laser resonator. Granted, they'll change as they propagate, but the different frequencies remain. This is why good interferometry requires a single mode source.

 

[sophiecentaur]

they don't magically vanish when they leave the laser resonator

Are they not evanescent modes? That's what happens around an RF antenna. The energy in the weird modes is just reactive. Is it possible that different rules apply? It's all Maxwell, isn't it? I'm sure @Andy Resnick could give us chapter and verse on this.

 

[sophiecentaur]

but the different frequencies remain.

You may have a different context for the word 'mode'. Modes of the same frequency can exist but only TEM will propagate in free space. I think you are referring to the different modal frequencies that can exist in a cavity. The only way their energy can get out will be when they are converted on exit into TEM modes.
I think that resolves our problem as it would explain why a "single mode" source can be important for interferometry.

 

[DaveE]

You may have a different context for the word 'mode'.

Probably.

Modes of the same frequency can exist

The different TEM_xx modes are at often at slightly different frequencies, especially the fundamental mode.*

I know this in practice because our largest Ion laser could tell you if it was operating in TEM₀₀ mode or if other modes were also present. The way we did this was to detect the beat frequency in the (extra-cavity) amplitude. (HF range, something like 10-30MHz, IIRC). You could then easily select the correct intracavity aperture to extinguish the higher order modes.

* BTW, not to be confused with the longitudinal modes, of course. You'll select only one of those with an etalon, or equivalent, first (if they are possible).

only TEM will propagate in free space.

And yet the sunlight still makes it all the way to my backyard.

 

[DaveE]

You can also buy an instrument to measure the beam profile parameter M² (Modemaster, etc.) which is typically located a significant distance away from the laser aperture. Yes, IRL, higher order beams do propagate far enough to mess up your experiment.

Back in the cowboy days of ion laser mfg. it was normal practice to hold a concave mirror in front of your high power laser (after yelling "mode check" so the other techs would look down/away) then you would expand the beam and shine it up on the ceiling or a wall many feet away. The tech, with a trained eye, could the see the difference between a gaussian profile and one with higher order modes.

 

[renormalize]

only TEM will propagate in free space.

And yet the sunlight still makes it all the way to my backyard.

By this statement, do you imply that sunlight reaching earth is not TEM; i.e., that the electric and magnetic field vectors of the light are not each perpendicular to the direction of propagation, as well as each other?

 

[DaveE]

By this statement, do you imply that sunlight reaching earth is not TEM; i.e., that the electric and magnetic field vectors of the light are not each perpendicular to the direction of propagation, as well as each other?

Nope, sorry. You are correct I think. I missed the nuance that non-TEM modes were under consideration, since no one had mentioned them before. The issue with (good) interferometry is the fundamental TEM₀₀ versus higher order TEM modes. In the laser world, it is exceedingly rare for people to talk about optical modes that aren't TEM, because as you've said, they aren't normally sustained.

 

[DaveE]

I suspect the semantic disconnect here is that when I refer to laser modes I mean TEM₀₀, TEM₁₀, TEM_01*, etc. while you were using a broader view, TE, TM, TEM, etc. Maybe I should have said mode number?

 

[sophiecentaur]

Maybe I should have said mode number?

It seems to me that the full Mode Number describes what goes on in the cavity, so yes. Different frequencies may emerge from different modes of resonance. the energy from each mode will match differently to a free space wave. it's precisely the same with a radio antenna which will have a bandwidth within which signals can emerge. It's possible to generate a pretty narrow band RF signal but I guess that could be a higher fractional bandwidth than that of a single mode laser.

But does our deviation contribute to the problem of 'fading' of the pattern? If the relative levels of what's produced in the different modes is drifting then the resultant polarisation could drift about and the beam splitter could be sensitive to that. I'd expect that the actual way in which the fringes behave could indicate what to look for.

 

[DaveE]

It seems to me that the full Mode Number describes what goes on in the cavity, so yes. Different frequencies may emerge from different modes of resonance. the energy from each mode will match differently to a free space wave. it's precisely the same with a radio antenna which will have a bandwidth within which signals can emerge. It's possible to generate a pretty narrow band RF signal but I guess that could be a higher fractional bandwidth than that of a single mode laser.

But does our deviation contribute to the problem of 'fading' of the pattern? If the relative levels of what's produced in the different modes is drifting then the resultant polarisation could drift about and the beam splitter could be sensitive to that. I'd expect that the actual way in which the fringes behave could indicate what to look for.

Yes, there are a lot of possibilities, and little actual information to go on. My experience is that polarization in lasers is usually maintained pretty well since the loss from a intra-cavity Brewster window (or equivalent, cavity geometry, etc.) is really large. But this all depends, of course, on the laser design. Undesirable TEM modes or longitudinal mode hopping are more common problems.

The analogy to radio antennas is good, but the antenna size wrt to wavelength is tiny compared to optical devices. The laser "antenna" is also a high Q oscillator so while many modes may be possible, there can be strong selection based on small losses that select a preferred oscillating mode.

 

[sophiecentaur]

The analogy to radio antennas is good, but the antenna size wrt to wavelength is tiny compared to optical devices.

Absolutely. There's a factor of >=10,000 difference between the wavelengths involved. I'm not using the word 'analogy' in theis discussion because it's not analogous; it's equivalence and, until you start to deal with individual photons, all the equations have to apply.
The problem you and I had was really just 'usage' of terms.

High order TEM modes or longitudinal mode hopping are more common problems.

I can see that. RF cavities are seldom (someone will correct me here, perhaps) more than a few wavelengths is size and the higher modes (overtones) are well spaced in frequency so monochromaticity problems are different.

 

[Andy Resnick]

TL;DR Summary: In my lab, the interference pattern from a mach zehnder interferometer fades in and out of view, and we can't determine the cause. How can we go about fixing this?

Hi! I'm working in a lab this summer doing quantum optics, and my team is working on a Mach-Zehnder interferometer setup. We got it well aligned and interference fringes appeared as expected. However, soon after building it, the fringes began to fade in and out of view without us changing anything about the setup, and we can't figure out why.

This is how it looks: there is a circular beam showing up on a screen with dark interference lines going across it. The interference lines fade to the point where the beam looks uniform as if there is no interference and then they'll come back into view. This happens quite often (about 10 times per minute) and it's random rather than periodic. As the fringes fade, they mostly don't move or change in size. Typically, the only thing that changes is their brightness, although occasionally the bright and dark fringes change places.

The only thing we know of that changes randomly over time is the laser intensity. It fluctuates a slight bit, but we found no correlation between the laser intensity and the visibility of the interference fringes.

What could be the cause of this issue, and how should we fix this?

(BF added)

First, the basics: what laser, what mirrors, what beamsplitter? is there a sample cell or other similar object located in one arm? If so, what happens to the fringe visibility if you remove it?

Does the fringe visibility across the entire spot gradually decrease and increase, or are there sudden drops and increases?

Can you describe the mechano-optics of your interferometer (how everything is aligned and then locked into place)?

Can you describe the room/enclosure that the interferometer is placed? (air currents, for example)

 

[sophiecentaur]

@Andy Resnick You don't seem to include lateral movement of the fringes in your list. Is that not 'a thing' in those interferometers? I'm only considering the very basics of interference.
But I knew you'd have relevant things to say about the problem.

 

[Andy Resnick]

@Andy Resnick You don't seem to include lateral movement of the fringes in your list. Is that not 'a thing' in those interferometers? I'm only considering the very basics of interference.
But I knew you'd have relevant things to say about the problem.

I was simply trying to reply to the OP, who did not include fringe drift in their description. Fringe movement would also be a diagnostic.

 

[Vanadium 50]

I was simply trying to reply to the OP,

Who, sadly, wandered off after message three.

 

[sophiecentaur]

Who, sadly, wandered off after message three.

PF has its own style which can put some people off. Nothing can be done about that - or even should it?

 

[DaveE]

PF has its own style which can put some people off. Nothing can be done about that - or even should it?

Nope. For some it's a community, for others just another reference source.
Either they got an answer and left or didn't get what they wanted and left. There's nothing wrong with that.

 

[PokemanDude]

We ended up writing a program to correct for the blurriness in the interference. We gave up trying to fix it haha.