There are certain conditions you need to satisfy in order to couple light into a fiber, involving shape and size of your light beam, angle of acceptance ecc..

https://www.newport.com/n/fiber-optic-coupling

Eventually light will make it through traveling from one input end to the output end...But you can't just shine light somewhere close to the core and expect light to propagate both ways, with no predetermined direction.

You can taper a fiber and do some interesting stuff with the evanescent filed stretching out of the fiber, but not coupling light into it from there I believe.

http://www.ijettjournal.org/volume-4/issue-3/IJETT-V4I3P240.pdf

I would be interested in knowing why you want to do this?

There are some Optical Fibers made that are called Scintillating Fibers. They take/absorb energy (UV / IR ) and converts the wavelengths to visible range and propagate to both ends. You will see both ends shining bright without any visible source attached. The trade name is OPTIBRIGHT. I am still looking for a commercial utilization of such fibers.

Dear Nirmal Pol ,

in the case of a straight optical fiber the principle of reversibility of light (paths) forbids the coupling of light across the cladding into the fiber core.

However for (very) stong bended fibers there might be a chance to couple light into the fiber.

At strong bended sections of the fiber the light from the core is partially coupled into the cladding and a part of that light is coupled out of the cladding.

So, when reversing the light path, one might be able to couple light into the fiber core via the cladding at strongly bended fiber sections. But I am sure that the coupling efficiency in very, very poor...

But I havn't done such experiments and I havn't found any descriptions of experiments in that direction.

The coupling of light via scintillation, as mentioned by Jay Madhav above, is an interesting alternative. Thanks Jay for that hint.

You are asking how to inject light into the middle of a fiber (not into the ends) without altering the fiber so that injected light propagates in both direction. No alterations means no evanescent coupling. So the good suggestions of P. Fonseca are disallowed. J. Madhav suggests generating the light from within the fiber -- fascinating! And G. Martens suggests bending the fiber so that the evanescent field extends to the outer surface of the cladding, which will work at the expense of high loss at the bend. If you are willing to consider Madhav's scintillating fiber idea and Marten's bent fiber idea, then maybe you would consider a fiber with a Bragg grating written within it that couples light injected into the cladding to the core. Focusing light tuned to the grating through the fiber cladding into the fiber core would couple to forward and backward propagating core modes of the fiber provided that the grating resonance condition were met. In this case, a cylindrical coupling lens would spread the injected laser beam over a length of core, but in any case, this type of coupling is likely to have relatively low efficiency. However, light injected into the cladding will propagate in both directions once it is diffracted by the Bragg grating at 90 degrees from the incoming direction.

Nirmal,

Apply a grating coupler on the clad layer of the fiber. Regards, Shigeo.

Check out my patents and papers on side illumination! I published two papers that review literature on this topic. They are all here in researchgate.net.

I have been injecting light into my fibers from the middle for the past 30 years and, recently, I patented a fiber coupler that uses side illumination as well.

There are several applications of side illuminated fibers: the most important one being an optical fiber sensor with a very high density of sensing points. For instance, with existing technology it is possible to produce a sensor with 1000 sensing points for every centimeter of fiber.

Dear Claudio Egalon ,

nice idea;

but I think, this principle only works if the fiber has

a) no cladding at its total length; being an overall bare fiber,

or

b) the refractive index of your sensing medium/liquid is very close to that of the fiber core (at least larger than that of the cladding).

Am I right?

Dear Gerhard Martens

It works in both cases:

In the first case I have been able to couple light into a side illuminated fiber with a cladding (this is discussed in one of my patents).

The second case is a common fallacy. Back in 1990 I discovered through computer simulations that, for the case of a fluorescent cladding fiber, the REVERSE is true. Ironically, even Dietrich Marcuse, who also worked modelling fluorescent cladding fibers, believed in this fallacy! Today, I clearly see this phenomenon in the lab with fibers without a cladding... But it took me 30 years to figure out all of that...

Thanks, Claudio Egalon ,

for feedback.

Surprising, that your experiments state that.

In theory we always assume ideal conditions.

I think, that the surface roughness of the boundaries, which in the ideal assumptions is neglected, may play an important role with respect to the efficiency of the optical coupling effect observed by you. Do you have analyzed that?

Dear Nirmal Pol ,

starting from the surface roughness argument mentioned above,

your idea on removed cladding/bare fiber core section, but with a roughened core surface will work. However, the efficiency will be quite poor. But this method still requires an altering/ modification of the fiber.

Dear Gerhard Martens

The reason for light coupling into a side illuminated fiber has to do with the following:

1. Multi mode fibers support low loss leaky modes that propagate for long fiber lengths. These are NOT bound modes but they do survive for long fiber lengths while undergoing partial (versus total) internal reflection at the core boundary and

2. Scattering centers in the fiber core that couple light into bound modes.

None at of those are factors in single mode ultra low loss fibers.

For clairity, Nirmal, do you require that your fiber be single mode or multimode -- or will either work for you? And, how stringent is your requirement to NOT alter the fiber that you side-couple to? Finally, are you interested in fiber-optic communication (suggesting long fibers) or sensing (long or short)? Your answers will determine how relevant the work of Egalon is to your solution. Interesting stuff, Claudio.

Dear Claudio Egalon ,

thanks for clearifications.

I wish you all the best for your research and stay safe and healthy;

best regards

G.M.

You could try the Keopsys technique - make a slit in the side of the fiber and launch light into it?

https://patents.google.com/patent/US6975792B1/en?oq=6975792