Would a flattened (sheet glass) dispersion prism work?

[norlesh]

Could a thin sheet of glass with appropriate edge angles work as a dispersion prism? If the prism were laying flat my thoughts are the light should be contained via TIR in a plane that would remain perpendicular to the top and bottom while said plane dispersed horizontally according to wavelength - would polarization effect this? Does such a device have a name?

 

[blue_leaf77]

Could a thin sheet of glass with appropriate edge angles work as a dispersion prism?

Do you mean some kind of prism but with very small apex angle? Such device is usually called optical wedge, commonly used to tune the dispersion of a laser pulse.

 

[norlesh]

Do you mean some kind of prism but with very small apex angle?

No, the angles would be normal like say a 60:60:60 but the height would be much less than the widths of the three sides.

 

[blue_leaf77]

By height I think you meant thickness, am I wrong? If the thickness is too small so that the incoming beam diameter cannot be contained completely, I think such device will have no significant use. What do you want to achieve?

 

[norlesh]

By height I think you meant thickness, am I wrong? If the thickness is too small so that the incoming beam diameter cannot be contained completely, I think such device will have no significant use. What do you want to achieve?

Thickness is possibly a more accurate term, my use of height was in relation to the first post where I specified the geometry was laying flat... The thickness would of course be wide enough to accept the full beam (but no thicker) - the objective is to have a very large prism, the frequencies separated on one edge and a motorized sensor/collector that travels along the glass edge to where the appropriate frequencies are emerging.

 

[blue_leaf77]

Why is the prism height a restriction in your desired setup? In thorlabs website the thinnest prism they offer has 10 mm of thickness. And how large do you want your prism to be and why need that large?

 

[norlesh]

Depending on other parameters, the side the disbursed beam emerges from could be anywhere up to 1 meter (the larger the edge, the better the resolution) - it will be a polychromator for an experiment I'm planning.

 

[blue_leaf77]

You can use a very dispersive prism material to increase the resolution. Alternatively you can also disperse light using grating, in which case the overall size can be significantly smaller than when using prism.

 

[norlesh]

You can use a very dispersive prism material to increase the resolution. Alternatively you can also disperse light using grating, in which case the overall size can be significantly smaller than when using prism.

I will be using the highest refractive index I can find for the glass, and as for gratings I need to separate and collect the full beam while most of the beam in a grating ends up in the zero-order mode which is not separated.

Which leaves me with the original question of whether or not the flattened prism would function?

 

[blue_leaf77]

Well as long as the whole beam can go in, there is no difference between a thick and a thinner (but again, it can still accept all parts of the beam) prisms.

 

[norlesh]

Well as long as the whole beam can go in, there is no difference between a thick and a thinner (but again, it can still accept all parts of the beam) prisms.

It accepts the whole beam, but the beam will have diverged during its travel (unavoidable due to path length) inside the prism enough that it will interact with the top and bottom - so there is a difference.

 

[blue_leaf77]

I'm wondering just how large are the edges and how small is the height. For example a well collimated beam, e.g. 2 mrad of divergence angle, will have its diameter increased by 2 mm after traveling a distance of 50 cm. I think typical beam diameter in most applications is less or equal to 5 mm, this gives only 7 mm of beam diameter increase after traveling 50 cm. In that case, a thickness of 10 mm of the prism will be safe enough to suppress the reflection at the upper and bottom parts of the prism.

 

[norlesh]

10 mm is getting REALLY heavy if things go toward the 1 metre end of the scale. I appreciate your input and will definitely be taking note, unfortunately it still does not answer my original question.

 

[blue_leaf77]

I didn't think the size would reach that big, I think there should be available polychromator in the market that works just well, are they too expensive? This is my final suggestion though, if you are concerned with beam divergence inside the prism, you can consider focusing the beam inside the prism such that the whole beam path inside the prism still lies within the Rayleigh length of the focusing geometry, obviously it will require loose focusing. Consider beam waist at the focus of only 1 mm, refractive index 1.5, and wavelength 700 nm, if google had done no mistakes in calculating the Rayleigh length, it should yield 6.7 m of collimated region (this means 6.7 m to the left and to the right from the focal point inside the prism the beam radius only widens until 1.41 mm, in which case a prism thickness of about 3 mm should be enough. But a 1 m sides prism with thickness of 3 mm, can you imagine how fragile that thing would be?). You can probably use this site to estimate the required focal length for such arrangement: http://www.calctool.org/CALC/phys/optics/f_NA

By the way, I think you can also do something with that unusually big sides of your prism, for example you can apply some highly reflective coating on the exit surface of the prism, so that the dispersed rays will bounce back instead of going out and hence increasing the path length inside prism. I think this way you can reduce the size up to some extent.