Light coupling from optical fiber to glass?

[Mary09]

Hey everyone,
I am trying to couple light from an optical fiber connected to an LED (365nm) into a small glass capillary which is pulled on one end and should be used as a small light source. The idea is to use the glass as a light guide. Has anyone tips for efficient incoupling of the light?
I would really appreciate your help.

 

[tech99]

Not clear exactly about your set up - can you provide a sketch? Why are you using a capillary?

 

[DaveE]

I would guess you can just treat it like an optical fiber where the core is air and the cladding glass. There's lots of info on the web about fiber coupling.

 

[Mary09]

Many thanks for your answers! I have attached a sketch. The glass capillary is used to melt in a small metal wire which is then used as an electrode. The idea is to use the glass fiber both for insulation of the metal wire and for local illumination as we want to study photocatalyst reactions on a surface.
My question is how to position the optical fiber for the coupling?
-Directly on one side of the glass sheath? Therefore the diameters of the fiber (0,4mm) and sheath (0,25mm) would not match now. Maybe with a smaller fiber then? 0,4 is the smallest we have right now
-Illuminate both glass sheath and air with the bigger fiber? (Also directly place the fiber in physical contact with the capillary)
-Another idea would be to melt the glass end of the capillary and use it as a lens to collimate the light?

 

[tech99]

Today I tried shining a torch [flashlight] into a glass rod. It works really well as a light guide. Then I tried shining the torch into a glass tube. In this case the light is guided mostly through the glass rather than through the hollow centre. I tend to think that any light source placed against your capillary tube will shine effectively through the glass and emerge at the end. In order to illuminate the glass rather than the hollow centre, I suggest trying the 1 mm fibre butted against the end of the capillary.

 

[Mary09]

Yes the glass works very well for guiding the light. However as I saw today again, it seems unavoidable to have major losses while coupling in the light (about 90% or more of the initial power of the light source is lost). Why is that? Reflection losses should only be 4% ?

This is also the case in these two publications using a similar approach:
https://pubs.acs.org/doi/10.1021/acs.analchem.6b03706 (from e.g. 150 W to about 0,35 mW?)
https://pubs.acs.org/doi/10.1021/acs.analchem.9b03347

Butting the 1 mm fibre directly on the end of the capillary works for my purpose right now. I will continue that way but if anyone has another idea i am open to try.
I will check for capillaries with smaller inner diameter so that more light is coupled directly into the glass also...

 

[tech99]

I think most light enters to hollow capillary and is lost. Maybe use a solid glass rod tapered to the size of the capillary and having a cone-shaped depression made in its end made using a pin when the glass is soft. We need to convert the optical mode to a coaxial waveguide mode to suit the capillary. Incidentally, I have previously found that it is not possible to collapse a capillary by just heating, but in principle I think it needs a transition from solid to hollow glass waveguide. Are you going to use a laser or special light source or an LED?

 

[Mary09]

Unfortunately i cannot use completely solid glass as i need the capillary to melt in the metal wire. However i can melt quite a large part at the end of the capillary without major deformation but there is still a very small channel. The only option would be to open the capillary at some point and use a vakuum pump (this is the way we melt in the metal wires). The resulting glass ball at the end works also as a ball lens and i could reduce the losses a bit.
I will try with the pin to make a cone shape, how would you recommend to place the fiber then? butt-coupling directly into the cone?
I am not sure if i got your idea with the coaxial waveguide, additional shielding of the capillary (e.g. acrylate or another metal)?
An LED 365 nm is connected to the fiber.

 

[DaveE]

Unfortunately i cannot use completely solid glass as i need the capillary to melt in the metal wire. However i can melt quite a large part at the end of the capillary without major deformation but there is still a very small channel. The only option would be to open the capillary at some point and use a vakuum pump (this is the way we melt in the metal wires). The resulting glass ball at the end works also as a ball lens and i could reduce the losses a bit.
I will try with the pin to make a cone shape, how would you recommend to place the fiber then? butt-coupling directly into the cone?
I am not sure if i got your idea with the coaxial waveguide, additional shielding of the capillary (e.g. acrylate or another metal)?
An LED 365 nm is connected to the fiber.

Maybe fill the capillary with a high index fluid like immersion oil? You may need to do it in a vacuum to avoid trapped air. Messy and might not work, but it's an idea.
Maybe a runny flavor of clear epoxy?

 

[Vanadium 50]

Maybe it's just how you've drawn it, but It's the reduction of radius that concerns me. Aren't you trying to reduyce the phase space of the light? And doesn't Liouville's Theorem get in your way?

 

[Mary09]

I think most light enters to hollow capillary and is lost. Maybe use a solid glass rod tapered to the size of the capillary and having a cone-shaped depression made in its end made using a pin when the glass is soft. We need to convert the optical mode to a coaxial waveguide mode to suit the capillary. Incidentally, I have previously found that it is not possible to collapse a capillary by just heating, but in principle I think it needs a transition from solid to hollow glass waveguide. Are you going to use a laser or special light source or an LED?

I tried to make a cone shape at the end while melting the glass but was not successful in a quick approach as the glass cools down very fast and it tends to bend while heating....

Maybe fill the capillary with a high index fluid like immersion oil? You may need to do it in a vacuum to avoid trapped air. Messy and might not work, but it's an idea.
Maybe a runny flavor of clear epoxy?

with the clear epoxy was also one of my ideas, i filled a capillary and will check when its hardened.

Maybe it's just how you've drawn it, but It's the reduction of radius that concerns me. Aren't you trying to reduyce the phase space of the light? And doesn't Liouville's Theorem get in your way?

As far as i have understood, when the light enters the glass it propagates via total internal reflection and will exit at the tip. Its right that some light will be lost where the radius is reduced, but i also checked capillaries without any tapering and most of the light is lost already here with the coupling.

I have also found one paper where lenses are used and losses are significantly reduced. However i am not too familiar with the optics and alignment etc. also i think not everything is clearly explained here
https://www.scirp.org/pdf/JCC_2013112910444793.pdf

 

[Vanadium 50]

As drawn, the system seems to decrease etendue. How?

 

[Hyperfine]

Hey everyone,
I am trying to couple light from an optical fiber connected to an LED (365nm) into a small glass capillary which is pulled on one end and should be used as a small light source. The idea is to use the glass as a light guide. Has anyone tips for efficient incoupling of the light?
I would really appreciate your help.

Consider the option of using the optical fiber all the way to your sample. Assuming you could use such a fiber that is not clad, at least on the sample end, you might be able to plate the fiber tip with gold to act as the electrode.

 

[johnbbahm]

Hello, Decades ago, when I did this sort of thing for a living, NSG SELFOC has
small graded index lenses that attached to the end of fiber with index matching glue.
This (or any biconvex lens) would convert the diverging cone of light coming out of the end of the fiber,
into a focusing cone of light. By adjusting the lens distance, you could optimize the amount of light getting into the glass capillary.

 

[hutchphd]

I would guess you can just treat it like an optical fiber where the core is air and the cladding glass.

This makes no sense to me. The index of the "fiber" must be larger than the index of the cladding to produce total internal reflection.

Yes the glass works very well for guiding the light. However as I saw today again, it seems unavoidable to have major losses while coupling in the light (about 90% or more of the initial power of the light source is lost). Why is that? Reflection losses should only be 4% ?

How do you know this? The hollow part of the tube is problematic. Most of the coupling losses are going to be "acceptance angle" issues and not reflective losses at the entrant surface. If you fill the glass tube with epoxy and use it to glue on the fiber (or the LED!) you will do far better than 10%. Keep it simple.

 

[tech99]

I have made a sketch of a possible coupling arrangement.

 

[berkeman]

I have made a sketch of a possible coupling arrangement.

Would you just polish the ends of the FO cable and the glass, and affix them somehow mechanically? I wonder how the OP is cutting the FO cable to prepare it for the assembly...

 

[hutchphd]

I believe you could make a robust connection by stripping the cladding from the (polished) fiber end for 10mm. Make the glass tube hole as short as possible and put in low viscosity (UV cure?) epoxy with a syringe sufficient to fill the interstitial space between inner tube and fiber (try to match the glass index of refraction) and shove in the fiber. Bubbles are your enemy (you could put it in vacuo). Keep it simple and strong. Complicated schemes for light pipes never perform up to expectations and LEDs are now very bright sources.. (Twenty five years ago when I was designing bad light pipes they were much less so)
Worth a shot
EDIT: I just reread the OP and realized the incident light is is UV (365nm). That makes the clear glue a bit more challenging... the sentiment is still the same however.

 

[hutchphd]

One more crazy idea. Can you put an unclad fused silica fiber into the capillary tube as you draw it for the wire? Get the polished fiber as far forward as possible to where the hole disappears? Its crazy idea night I guess.

 

[Mary09]

Thank you all for your answers and ideas!

Consider the option of using the optical fiber all the way to your sample. Assuming you could use such a fiber that is not clad, at least on the sample end, you might be able to plate the fiber tip with gold to act as the electrode.

Yes, i think some people did that already where the inner core is the actual fiber and then coated with a metal which works as an ring electrode (then insulation again...). However the size of the fiber limits the size of the electrode and right now i think only around 100 µm in diameter is what i have seen so far...
Some companies also provide ver thin fiber cannulas for biological applications, which could be then covered with a metal etc.

Hello, Decades ago, when I did this sort of thing for a living, NSG SELFOC has
small graded index lenses that attached to the end of fiber with index matching glue.
This (or any biconvex lens) would convert the diverging cone of light coming out of the end of the fiber,
into a focusing cone of light. By adjusting the lens distance, you could optimize the amount of light getting into the glass capillary.
View attachment 318874

This is something i wanted to try. But i am not sure about gluing lenses or the capillary itself onto the fiber...i might damage the end of the fiber permanently when removing? i will check if i can sacrifice one or try first without glue.

This makes no sense to me. The index of the "fiber" must be larger than the index of the cladding to produce total internal reflection.

How do you know this? The hollow part of the tube is problematic. Most of the coupling losses are going to be "acceptance angle" issues and not reflective losses at the entrant surface. If you fill the glass tube with epoxy and use it to glue on the fiber (or the LED!) you will do far better than 10%. Keep it simple.

Maybe you are correct, but i thought the LED light is quite widespread so the exact angle may not be a big problem?

I have made a sketch of a possible coupling arrangement.
View attachment 318919

i made a holder to fix the fiber exactly like this on the glass wall, i will try and check the losses when i find some time...

Would you just polish the ends of the FO cable and the glass, and affix them somehow mechanically? I wonder how the OP is cutting the FO cable to prepare it for the assembly...

the fiber i am using has a metal sheath like in my drawing at the beginning, i am not cutting anything just put them directly onto each other.

One more crazy idea. Can you put an unclad fused silica fiber into the capillary tube as you draw it for the wire? Get the polished fiber as far forward as possible to where the hole disappears? Its crazy idea night I guess.

To your comment and the one before: the issue will be to get the epoxy inside, we have very thin syringes but they may clog... i will give that a try anyway also i will try to melt the glass further up.
How do you want to couple the light then into the fused silica fiber?

 

[Hyperfine]

Perhaps we could be more helpful if we knew some additional features of the experiments. For example:

Which is more critical, the light reaching the sample, or positioning of the electrode?

How precise and how reproducible does the apparatus need to be?

Does the probe need to be moved relative to the sample during an experimental run? If so, could it be easier to move the sample rather than the probe?

Regarding the electrode wire, is that to pick up signal and if so what kind of signal (DC, short pulse, rf)?

How does the end of the electrode away from the sample couple to other instrumentation? Does that end inhibit gluing a lens to the capillary?

Are you working in a vacuum chamber?

I would also suggest that you consider approaching the manufacturer of the fiber optics for assistance.

 

[hutchphd]

To your comment and the one before:

Let me reiterate my initial lack of care here. I think you may have trouble finding an epoxy that is transparent at 365nm. Many of them are definitely opaque and I have never really used one in the UV. Maybe they exist.

 

[DaveE]

Let me reiterate my initial lack of care here. I think you may have trouble finding an epoxy that is transparent at 365nm. Many of them are definitely opaque and I have never really used one in the UV. Maybe they exist.

Yes, it's a problem. But there are about 1 million different versions. A call to the manufacturers is in order here.

Also look into MasterBond EP30LV-1 it my be OK.
Epo-tek 305 maybe? Norland 63? Norland 88?

Note that I have no useful experience with this, just some quick mostly random search results.

 

[Mary09]

Perhaps we could be more helpful if we knew some additional features of the experiments. For example:

Which is more critical, the light reaching the sample, or positioning of the electrode?

How precise and how reproducible does the apparatus need to be?

Does the probe need to be moved relative to the sample during an experimental run? If so, could it be easier to move the sample rather than the probe?

Regarding the electrode wire, is that to pick up signal and if so what kind of signal (DC, short pulse, rf)?

How does the end of the electrode away from the sample couple to other instrumentation? Does that end inhibit gluing a lens to the capillary?

Are you working in a vacuum chamber?

I would also suggest that you consider approaching the manufacturer of the fiber optics for assistance.

You helped already a lot!

- as much light as possible should reach the sample, but also we need to position the electrode precisely. But it is possible only to move the sample and the electrode is static. Right now the adapter (fiber to glass capillary) i have works for my case and i can couple enough light to investigate the reaction. For me it is just the question if i really need such high power (comment #6) or if it is possible to optimize the coupling somehow

- The electrode will pick up a DC which is measured against a reference electrode

- The electrode wire will continue to the outside and will then be connected to a potentiostat. Like in drawing from #16 the wire can either come out somewhere at the front (i need to open the glass) or at the end of the capillary to the side when the fiber is place on the other side of the glass wall

- As these electrodes can be damaged and may need to be exchanged, i would prefer not to glue any lens there (lenses are quite expensive i suppose? or could there be an easy way to remove the glue?) so maybe design a specific holder for the fiber/lens with the capillary?

- i am not working inside a vacuum chamber

- i contacted the company and they suggested to use an aspherical or achromatic lens but not anymore

Yes, it's a problem. But there are about 1 million different versions. A call to the manufacturers is in order here.

Also look into MasterBond EP30LV-1 it my be OK.
Epo-tek 305 maybe? Norland 63? Norland 88?

Note that I have no useful experience with this, just some quick mostly random search results.

it should ideally not absorb light in this range you are right. I will check what we have and the ones you suggested, thank you!

 

[Hyperfine]

Thank you for those details. I must say that I am not a fan of the proposed capillary and its coupling to fiber optics. That reflects my biases based on my laboratory experience in a high-resolution LASER lab. Just so you know that.

I would approach the problem by separating the mounting of the detector wire and the method of delivery of light to the sample. Regarding the light, I would urge that you consider simply focusing the light onto the sample by means of one or more lenses. More on that below. I would then mount the electrode separately allowing precise positioning of that electrode as well as facile replacement if needed.

Back to the light. I do not know the properties of the LED you are using--divergence for example, so I cannot make specific recommendations. In general however, I think you should be able to devise a series of lenses that will allow you to tightly focus the light as needed. Some general comments on how i would approach that. Assuming you have sufficient space, I would use a standard optical rail to mount the components. Each individual component would be mounted on a stable mounting system that can be precisely manipulated in all three dimensions. With the proper lenses, that should give you the ability to deliver a tight focus spot on your sample as desired. It also allows easy placement of other optical components in the beam path. Potentially useful ones include a simple shutter to block the beam if needed (better than turning the LED off), or optical filters (neutral density perhaps).

A further comment regarding the manipulation of a hard to see light beam. At your wavelength, high quality business cards are a must have. The paper must be highly bleached white paper. Why you ask; because that paper fluoresces with good efficiency making it quite easy to see your beam spot.

 

[Tom.G]

or could there be an easy way to remove the glue?

Yes. Methylene Chloride dissolves epoxy, it also removes paint from metal, bark from trees, skin from bone. It is sold (diluted?) over-the-counter as paint stripper.

Full strength is available from makers of epoxy. Containers are a problem though, only all-metal or all-glass will hold the stuff, no gaskets on the cover, they dissolve. (I havent tried, but one of the Teflons might hold up.)

A further comment regarding the manipulation of a hard to see light beam. At your wavelength, high quality business cards are a must have. The paper must be highly bleached white paper. Why you ask; because that paper fluoresces with good efficiency making it quite easy to see your beam spot.

Ordinary copy paper or printer paper that claims "High Brightness" (94+%) has optical brighteners added that fluoresce under UV light. There is also "Fluorescein", a moderately common chemical that fluoresces under UV.

Another critical subject is the optical acceptance angles of the fiber and of the glass capillary. Those, along with the dispersion angle of the light source (LED) can severly limit the optical coupling efficiency. Note that these angles are often reported as a 'half angle', that is the limiting angle off the physical axis of the device/material beyond which incoming light is reflected rather than accepted (sometimes known as the 'critical angle').

Amazon has UV flashlights, 395nm, at low cost.

A dimensioned sketch of the optical system would help.

Cheers,
Tom