A question about collimated light and the collimator ?

[Accutronitis]

How does a collimator take the uncollimated light that is highly diverging and turn it into a narrow column of collimated light that is hardly diverging ?

Opps in the lower right hand corner is meant to say "collimated light" and of course the collimated light is diverging some....

 

[Radim]

Check this:
https://en.wikipedia.org/wiki/Collimator
https://en.wikipedia.org/wiki/Collimated_light

Kind of simple optics.

Edit: The point of collimator is basically to bring diverging light into parallel - positive lens might be used, which has focal point at certain place, placed a point light source at the focal point is imaged into parallel beams or from other side parallel beams are focussed to that point. Since various wavelengths refract under various angles, there are much more focal points (becomes issue in camera lens - further optics is needed), for lasers as being of monochromatic light, this does not prevent precise focus much, still there are further optical properties affecting it. However as the output from diode is not point source (diode has some shape of active region), so you cannot get perfectly parallel beam and often some other element (lens) is needed.

For gas and DPSS lasers (and some others) there is a bit different situation - since they are somewhat much more parallel (than diodes) on output due to different optical cavity design (parallel mirrors perpendicular to output beam), still those mirrors are often not flat (adjustment is too difficult) and somehow curved (these are easier to adjust), that increases divergence a bit (and makes it easier for mode hopping). (Also the OC design has some influence.) In these cases collimators acts as tiny beam expanders (bringing parallel beams with low diameter and high divergence into higher diameter with lower divergence - at least two lenses are usually used). When there are high intensities used, virtual focal points are used not to allow light to focus in real focal point, where the effects of high intensity radiation on air might damage beam.

Check here for beam expanders:
https://www.edmundoptics.com/resources/application-notes/lasers/beam-expanders/

 

[Accutronitis]

Then my question after this....

would be with diodes like the 44 and it's wider diverging fast axis, After the collimator the wider fast axis is still diverging when the slow axis is now focused very narrow and stays that way with little divergance, At that point and very close to where the beam exits the collimator why couldn't a single cylindrical convex lens be used to further collimate the fast axis so that it is diverging close to the same as the slow axis and in that way would produce a tight narrow beam on both axis's ?

 

[Alaskan]

In general, the smaller the emitter, the lower the divergence. Just a matter of the ratio of expansion relative to the active emitter diameter. If you had an infinitely small point source I believe the divergence would approach zero, after expansion and collimation. Then only the properties of the optics cause the limitation, or limits the ability to produce a perfectly collimated beam..

 

[Radim]

Cylindrical lens focusses only in one direction (vertical or horizontal), the rest remains unchanged (light just passes the same way). I guess you might use another cylindrical lens with different focal point (or some lens with different curvature in different directions).
Check here:

https://www.newport.com/n/beam-shaping-with-cylindrical-lenses

However I leave this question open for answers from the others as my practical experience with those available on market is very limited.

 

[Accutronitis]

In general, the smaller the emitter, the lower the divergence. Just a matter of the ratio of expansion relative to the active emitter diameter. If you had an infinitely small point source I believe the divergence would approach zero, after expansion and collimation. Then only the properties of the optics cause the limitation, or limits the ability to produce a perfectly collimated beam..

Alaskan is it just me or has what you have to say just got much more interesting for some unknown reason ???
:crackup::crackup::crackup: IE (avitar)

The thing is with the 44 the fast axis after expansion and collimation is still expanding while the slow axis is now collimated, So couldn't a additional single cylindrical lens be place on the still expanding fast axis to then finish collimating the fast axis so it would be similar to the slow axis ?

 

[Radim]

In general, the smaller the emitter, the lower the divergence. Just a matter of the ratio of expansion relative to the active emitter diameter. If you had an infinitely small point source I believe the divergence would approach zero, after expansion and collimation.

Exactly, however at infinitely small point source further effects of lens and scattering would occur increasing the divergence - if we take all these conditions as ideal (not real therefore), than the theory says it would be perfectly parallel (edit: in beam optics - so when wave is considered it is more complicated - as Steve001 mentioned below).

More here:
Spherical aberration (different refraction at the edges than in the centre):
https://en.wikipedia.org/wiki/Spherical_aberration

Here how point source is imaged under various spherical aberrations:

Chromatic aberration (different wavelength refracts differently - as different frequency corresponds to slightly different speed of light in that medium) is not that issue as I mentioned above. Whatever:
https://en.wikipedia.org/wiki/Chromatic_aberration

 

[Radim]

Alaskan is it just me or has what you have to say just got much more interesting for some unknown reason ???
:crackup::crackup::crackup: IE (avitar)

I think it is because he might miss women, when in service far far away from home. :crackup:

Alaskan, I think you need to get home very soon.

 

[steve001]

Alaskan is it just me or has what you have to say just got much more interesting for some unknown reason ???
:crackup::crackup::crackup: IE (avitar)

The thing is with the 44 the fast axis after expansion and collimation is still expanding while the slow axis is now collimated, So couldn't a additional single cylindrical lens be place on the still expanding fast axis to then finish collimating the fast axis so it would be similar to the slow axis ?

Ideally one would choose cylindrical lenses that compensate for expansion as quickly as each axis expands. Adding another lens would have to do that. But no matter what you do a beam of light still expands, that is the nature of light.

After some more thinking I think what I'm about to write is correct. To answer your first question. By using lenses to collimate a laser's light what is being done is extending the Rayleigh (RL) Length. Lower diverging beams have longer a RL.

 

[Radim]

Ideally one would choose cylindrical lenses that compensate for expansion as quickly as each axis expands. Adding another lens would have to do that. But no matter what you do a beam of light still expands, that is the nature of light.

Exactly, as light is wave and "waves go behind corners".

https://en.wikipedia.org/wiki/Diffraction

https://en.wikipedia.org/wiki/Diffraction-limited_system

https://en.wikipedia.org/wiki/Angular_resolution#Explanation

Great point - rep+.

 

[Alaskan]

Single mode laser diodes have very tiny emitters and all of them I own have the lowest divergence for a given size of collimation lens. Infinitely small point source not gonna happen for us, in theory, yes.... even if you could then the optical problems, can't win. Perhaps if you had a good enough parabolic mirror you could expand the beam and collimate it without as many problems, don't know, optics is what I'd like to know well but don't.

I'm just goofing off with the pix, not really hot to trot for any, except my Ukrainian wife.

 

[Julian95]

besides the spherical aberrations induced by spherical optics(which can be reduced greatly with aspherical lenses) the thing is that the output of the diodes are very spatially dirty with lots of stray rays, a collimator is a partial solution that have very low efficiency, blocks a lot of the incident light..a much better solution is a spatial filter, it's like a collimator with a second lens before the pinhole. (the pinhole must be small, 20-200um)
that REALLY improves the output characteristics of a laser without throwing away much light

 

[GSS]

Julian95, what have you been doing the last 5 years?

 

[paul1598419]

I have a lot of experience using spatial filters with gas laser for use in holography. Because the diode lasers we use have a large bandwidth, they are relatively pointless to use with these diodes. You want a laser with a long coherence length when using a spatial filter as they will "clean up" the beam as opposed to lasers with a short coherence length.

 

[Julian95]

i've been building a laser based optical lithography system for making photomasks for the last 6 months..it uses a 405nm laser, spatial filtering and high numerical aperture optics to make a focal point as small as possible, i'm currently at submicron focal points, it's very tricky to mantain focus, i have a depth of field of a couple of microns

 

[paul1598419]

These diodes have very good focus close to your objective, like a DVD in a Blue Ray. They keep the objective lens very close to the DVD for this purpose. Back in the 1980s, Philips had video disc players with He-Ne lasers in the large optical package that would move the whole sled over a huge disc to play videos. The sleds were metal and about 10 inches by 6 inches. All the optics were on the sled and the player was huge compared to what we have today.