Yes, it appears that your beam diameter is way too wide, to couple reasonably well with the single mode fiber. Notice how this (fairly old!) paper describes one high efficiency coupling method, where a microlens is used for coupling. Its base is the same diameter as the fiber core:

http://digitalcommons.ohsu.edu/cgi/viewcontent.cgi?article=1231&context=etd

This tutorial makes the same point. The laser light has to focus on a spot about the same size as the fiber core. That discussion follows after the paper talks about fiber to fiber connections.

http://hexagon.physics.wisc.edu/research/technical%20info/newport_fiber_tutorial.pdf

Sorry, I can't rate up your answer. When I click the button, nothing happens. 

Thank you for the links. However, I would expect that, if I place an f=20 mm lens in front of a 5 mm wide Gaussian beam, the focal spot will have a diameter of ~ 5 um - just from the Newport's formula. I wonder if the mistake was to use a condenser lens...

Elena,

"I found an aspheric lens with f=20 mm, which is close to what I calculated. I aligned it in front of the laser, with flat side facing the laser".

 Aspheric lenses (and plano-convex spherical lenses) are designed to give minimum aberrations with the most strongly curved surface facing the collimated beam - in this case your laser.  The flat side should face the optical fibre.

With the lens reversed spherical aberration will be high, so you won't get a good match to the mode field of the fibre.

You may not get diffraction-limited performance from a moulded aspheric lens as these are often intended for less critical applications, but the spot size should be much less than 0.5 mm if used the correct way round.  A precision CNC machined and polished asphere may perform better, but will be more expensive.

You don't need the high numerical aperture offered by a precision asphere, so I would suggest an achromatic doublet.

e.g. https://www.thorlabs.de/thorproduct.cfm?partnumber=AC080-020-C-ML

http://www.edmundoptics.co.uk/optics/optical-lenses/achromatic-lenses/near-ir-nir-achromatic-lenses/45792/

As for the asphere, most strongly curved surface towards the laser, flatter surface towards the fibre (though the difference is less obvious for a typical achromat).

Alternatively a 10x microscope objective has a focal length around 16 mm, but should work reasonably well.  Likely to be more expensive than a simple achromat, but you may have one already to hand.  Expect around 90% of the coupling efficiency compared with 22 mm optimum focal length, less any additional losses due to surface reflection losses if the objective is not optimised for 1060 nm operation.

Good luck.

For a simple explanation of coupling from a source to a sink, see https://www.osapublishing.org/ao/abstract.cfm?uri=ao-23-23-4187.  (If you can't get the reprint, just search for the authors, Joyce and DeLoach, and title elsewhere -- it's out there for free.)  Matching the beams is the key so you need to consider the numerical aperture of the fiber and also of the beam once it has been transformed by the microscope objective.  The cited reference will enable you to roughly calculate the coupling efficiency for any pair of sources and sinks and will explain why you got so little coupling in your first attempt.  The reference is a bit dry but I think it is very clear.  Good luck!

One more thing ... one can think of coupling beams as matching the angular Fourier components of the source the the angular 'acceptance' angles of the sink (in this case, your fiber.)  I know that this isn't how most physics teachers present the topic and it probably isn't rigorous, but consider the following.  A beam with very high numerical aperture, for example one that has been focused by a very fast lens, will have components that approach the fiber at very off-axis angles.  These components are not within the angle that will be totally internally reflected within the fiber -- essentially, they are not part of the fiber mode.  This way of picturing the process is intuitive -- the real math is in the reference above.  Sometimes (often) it is good to carry a working picture in your head that allows your intuition to get the answer quickly -- and then work out the details rigorously.  Anyway, just wanted to share this picture of mode coupling.  The opposite case, a very low numerical aperture mode coupled to a high numerical aperture fiber (also, low coupling) is interesting to think about using the intuitive model.

Some results: I just placed the condenser with the convex side facing the laser and got a beam radius of 42 um, instead of 500. This is still not small enough as I would only have about 2 % of the transmitted power.

I found a different lens, with estimated focus of about 5 mm, and got 38 % transmission through the fiber. But I need a better setup to calculate the waist radius.

If none of my lenses are good enough, PI refuses to invest some money, and I need a really strong one to achieve a small spot, does that mean that I will be stuck with overfilled fiber aperture? My V number seems huge, so there will be some high modes in my beam which is not what I need.

Elena

38% coupling efficiency is quite respectable if you had to rummage through the spares box to find a suitable lens with less than 1/4 the optimum focal length.  The Gaussian approximation suggests only 19% coupling for a 5 mm lens, so perhaps you measured the back focal length (distance from focus to lens surface) rather than the effective focal length.  Could the EFL be closer to 8 mm?

If you have a weaker lens available, such as 80 - 200 mm focal length, you could use this in combination with the short focal length to reduce the spot size reaching the shorter focusing lens - equivalent to using a laser with a narrower beam.  At 70% of the distance between lens and focus, the beam width will have reduced from 5 mm to 1.5 mm diameter.  For a 100 mm focal length lens this would be 30 mm from focus.  Place an 8 mm focal length lens here, and in the absence of aberrations, your spot size will be around 5.7 microns.  Move the 8 mm lens a little further away and you reduce the beam width at its principal plane, and increase the diffraction-limited spot size at focus (which may give better results if the short focal length is actually closer to 5 mm).

Note the words "absence of aberrations".  If lens quality is inadequate, adding extra elements can make things worse - even if simple theory tells you otherwise.

Elena: "I need a really strong one to achieve a small spot, does that mean that I will be stuck with overfilled fiber aperture? My V number seems huge"

Can you explain what you mean by this?  What is your V-number?  From what you have told us, your 5 mm lens is too strong rather than too weak, with a predicted MFD of 1.4 microns at focus.  Your fibre with NA 0.14 and MFD 6 micron is comparable with Corning HI1060 fibre, which has cut-off wavelength around 920 nm.  Have you other data which suggests the fibre will support multiple modes?

Elena,

you have mentioned that the beam radius is still not small enough. Do you have the possibility to expand the laser beam diameter before focussing by the aspheric lens? The beam waist radius after focussing is directly related to the diameter before focussing: the higher the radius of the incoming laser beam, the lower its beam radius after focussing...

Christoph 

Dear Elena,

Your question is not clear enough to me . It needs more information about your problem. Anyway, converging lens with small focal length is proper lens  to your work.

Dear Elena,

Firstly I think you should check the M2 of the laser you're trying to couple. If it's not close enough to 1 then you may have problems to couple it efficiently anyway. I would measure it with a wide aperture lens that would focus it around 200-300 um diameter.

Also, you should make sure the input laser is collimated and the lens you're using has an aperture wide enough to accommodate the input spot-size without clipping or introducing aberrations. 

The way I normally couple in SM fibers, I initially use a cheap lens to initially couple any amount of light (even 10% would do at this stage).

After that you should use a test aspheric lens to re-collimate this little output light and measure the size of the collimated beam. Being the fiber SM any small amount of core light will be SM. Now you can either:

1 - reduce or increase the input size using a telescope to match the collimated output produced by your test aspheric lens AND use that test lens to couple it, or,

2 - try different aspheric lenses to re-collimate the fiber light and find the one that will match as close as possible the initial spot-size.

It is an empirical method that allows you to find the best matching input lens and I usually achieve easily 70-75% of coupling efficiencies on SM fibers (4-5 um MFD) with decent lasers (M2 < 1.5).

Hope it helps,

Kind regards

Alessio

Dear Elena,

• 1- First make sure that everything is clean, especially, the fiber input face, no lib, no particle as large as MFD of the fiber.
• 2-  Make sure that focused laser light is dropping on the fiber face.
• 3- Make sure you use an IR viewer to locate the whereabouts of the laser light.
• Make the laser beams parallel by a beam expander, then use convex lens in front of the fiber input face with higher focal length.. 
• Due to high wavelength of your laser, most of the light launched  into fiber propagate through fiber cladding which then leaks out of the fiber. 
• You should not expect coupled light more than few micro watts. Single-mode is hard to couple, specially higher wavelength, due to cladding propagation of Gaussian beam. 
• Good Luck

Hi Elena,

• I did couple 18-20% laser light into SMF twice last month. The focal length calculation is right. The crucial part is alignment. I use Nanomax 3 axis stage and associated components such as lens mount and fiber holder from Thorlabs for alignment. Here is the link.

            https://www.thorlabs.com/newgrouppage9.cfm?objectgroup_id=2386

• If your laser beam is in free space, use two mirrors before 3 axis stage to control the beam height and beam angle. I have shared the photo of the assembly with you. Make sure that beam is parallel to the top of the axis stage using a pinhole.
• Reversing the side of the lens helps sometimes.
• Make sure that your SMF is cleaved properly. Re-cleaving the fiber improves the performance a lot.

Good luck and have lot of patience.

So far, I ended up tapering a multimode fiber and using a 10x NA 0.22 objective. Could couple about 65 percent of light. Beam quality was, of course, compromised, and there is no way to polish it. The beam appears very speckly. I added the mirrors, too, but I don't touch them after I am happy with the direction of the beam.

Not very easy. You needd 2D translation stage, fiber holder and maybe objective(look at Thorlabs).  Easier way is using of a single mode colimator. Splice fiber of colimator to your fiber, then couple beam into it.

Dear Elena: The output with speckles is due to multimode fiber that you are using. If you coupled 65% is quite a success. Make sure your optical table is vibrationless. In addition, make sure all mounts and holders are tight enough to avoid displacement of components. Good luck!

The fiber is mounted on a XYZ micro stage. I tried an 11 mm Thorlabs collimator eith SMF before, but I ended up with no light output and line burns all around the connector hole. But I now have a better idea for alignment. I'll make sure to update if build a better setup.

Speckles specifically in my multimode fiber is a good point. It's AFS50/125y from thorlabs. I thought I had speckles because I couldn't polish the fiber tips properly, even though they looked decent under the microscope.

That is a good idea to re-setup your experiment. Good luck!

 To ANKITA FROM U DAYTON

Could you please specify your parameters?

Wavelength, laser BD, lens part # or parameters, waist in focus, fiber part # or parameters.

Thank you!

Problem solved. I used the procedure in this video. I could couple just above 25 % of light, which is comparable to Ankita's results. I used a collimator.

https://www.youtube.com/watch?v=kQvhbJbDG0M

Dear Elena,

That's a good news. Good luck !

dear Elena,

Good luck in your work.

Dear Elena,

In agreement with Alan I also recommend you to try the 10X. Additionally you may play with the beam diameter arriving to the 10X. In case this is not possible, move the laser-to-10X distance until matching the NAs of the 10X and the Optical fiber. Good luck and best regards.

Is it necessary to put two mirrors ?

or only mirrors are used to change the directions in a fiber lunch system !!

Inserting mirrors , we have some power losses due to reflection.

If I do not want to change the directions, can I couple the laser to the finer using two collimating lens like : Laser>>> lens to hv collimation>>>lens to focus>>>> fiber

Could any body suggest me ??

@Jitendra Behera

Mirrors are simply a convenient way to adjust the transverse alignment and direction of the beam as it enters the focussing lens.

What do you mean by "hv collimation"? With a linear system, you may be able to align the beam and the focussing lens with sufficient accuracy, then align the fibre at the focus with a 3-axis micro-positioner. An alternative is to provide transverse (x-y) or 3-axis (x,y,z) adjustments on the focussing lens. Final adjustment by translation stage has the advantage that the x, y (transverse) and z (focus) translations are orthogonal, but does require a stable, high quality positioning stage for single mode alignment.

The two mirror Z-beam scheme potentially offers good stability without the expense of a precision micro-positioner. Small adjustments to the beam angle entering the focussing lens correct for transverse misalignment of the fibre. The two mirrors need to be adjusted together to keep the beam centred on the lens, and avoid angular misalignment at the fibre facet.

Regarding reflection losses at the mirrors, these are minimised using narrow band dielectric coatings optimised for your chosen laser wavelength. In practice, it is probably more important to minimise aberrations in the optics, and match the spot size at focus to the fibre mode field.

The best is to use a commercial collimator connected to a fiber that matches the NA of your fiber. you can find this type of collimators in Thorlabs. After that splice the collimator's fiber to your fiber.

Good Luck

wrt Optimum fibre coupling lens:

The paper "MAXIMUM FIBER COUPLING EFFICIENCY IN THE PRESENCE

OF RANDOM ANGULAR JITTER FOR FREE-SPACE LASER

SYSTEMS AND ITS APPLICATIONS" by Toyoshima treats this matter. The following easy formula gives you the relationship between the coupling lens and the fibre:

w1/w0 = 1.709

in which w1 is the first zero of the Airy disc, i.e. w1=1.22 lambda/D *f (D being the lens diameter and f its focal length), and w0 is the mode radius of your fibre at 1/e. The latter can be calculated by Marcuses formula, see https://www.rp-photonics.com/mode_radius.html .

Good luck!

Link to paper

https://arc.aiaa.org/doi/abs/10.2514/6.IAC-04-M.6.03

Jitendra Behera Two mirrors helps in beam walking using which you can collimate any beam.

If you do not use two mirrors, it needs to be ensured that incoming beam is absolutely collimated which is not usually the case.

Two mirrors helps in course alignment of coupling beam into fiber. Later using three x, y and z translational stages fine alignment can be done to obtain desired coupling.

Elena Renzhiglova Dear elena you can achieve around 60 percent coupling efficiency in single mode fiber.

First you can couple the beam into a multimode fiber which is relatively easy. After achieving 80 percent efficiency for multimode, remove it and insert single mode then just by finely adjusting x and y translational stage where focusing lens is mounted, you can achieve around 60 percent efficiency.

I am curious to know if you consider the phase space of the light beam and the phase space which can be contained in the fiber to get an estimate of how much can be captured in principle. For a step index fiber it is roughly the core diameter times the NA. For the beam one can use Rayleigh criterion to estimate, and some better math about focusing to get an exact result. We used to work with these things, and there were cases where we could capture essentially all of the IR light in a MM fiber, but could show both theoretically and experimentally that only a small fraction could be captured in e.g. a 9 u core dia. SM fiber.

I am recently working on a visible system. The pumping source is 442 nm blue high-power semiconductor laser with beam size 4*4 mm. Also, I have an achromatic doublet lens with focal length of 7.5 mm. The single-mode fiber core diameter is 5.2 microns with NA 0.08. How can I correctly insert the laser beam into the fiber core? I set up the fiber at the closest distance to the lens. Is it okay or not?

Preangka Roy - for efficient laser coupling into single mode fibre:

• The laser output should correspond to a single transverse mode
• For a fibre with an approximately Gaussian mode field profile, the intensity profile of the laser spot should be similar with a beam quality parameter M2 close to 1. https://www.rp-photonics.com/m2_factor.html
• The laser output should be brought to a focus at the fibre core
• The field profile of the laser spot should match the near field of the optical fibre.

"How can I correctly insert the laser beam into the fiber core? I set up the fiber at the closest distance to the lens. Is it okay or not?"

Not OK. If the incident laser beam is collimated (not diverging or converging signicantly), then beam will be brought to a focus at the rear focal point of the lens. This is where you should position the fibre.

For a lens such as https://www.edmundoptics.co.uk/p/625mm-dia-x-75mm-fl-vis-0deg-coated-achromatic-lens/10440/ the back focal distance between focus and lens surface is 4.45 mm.

The back focus varies between lens designs. It is 5.01 mm for https://www.edmundoptics.co.uk/p/5mm-dia-x-75mm-fl-vis-0deg-coated-achromatic-lens/9610/

Mode field matching

For M2 ~ 1 the beam is close to diffraction-limited and close to Gaussian in profile. If the focused spot size is to match the fibre mode field size, it follows that the convergence angle of the laser must match the far field radiation angle of the fibre.

A lens with focal length 7.5 mm illuminated by a 4 mm diameter collimated laser beam will produce a focal spot with numerical aperture 2/7.5 = 0.267. This is 3.3 times greater than the numerical aperture of the fibre.

Maximum theoretical coupling efficiency with this mismatch is

η = 4 A12A22 / (A12 + A22)2 = 0.3

With a 4 mm diameter collimated beam, a 25 mm focal length will give a better match to the numerical aperture of the fibre.

An alternative is to use a longer focal length supplementary lens to converge the beam towards the 7.5 mm achromat, so that the radius of the beam at the lens is approximately 0.08 times the focal length, so 1.2 mm with a 7.5 mm focal length lens.

Note that there are multiple definitions of fibre numerical aperture in use, and the differences can be significant:

• Theoretical NA is defined in terms of core and cladding refractive indices
• NA = sqrt(n12 - n02)
• Sine of the far field radiation half-angle at 1% intensity. This is used in some fibre data sheets.
• Sine of the far field radiation half-angle at 1/e2 (13.5%) intensity. Free-space laser beams are frequently characterised in terms of 1/e2 values.

Numerical aperture derived from far-field radiation angles will vary with measurement wavelength.

Why did you align the lens flat side facing the laser? Was it recommended by the lens manufacturer? Usually, a lens is set facing a most curved surface to a laser. This will give you a smaller spot at the output. Alignment for an SM fiber is a very delicate procedure, especially in the IR.