Question about detector and lens

[phi_skr]

TL;DR Summary

Light from the sample is collimated light. For imaging, does the spectrometer requires a lens to focus the collimated light on the entrance slits of the detector.

Light from the sample is collimated light. For imaging, does the spectrometer requires a lens to focus the collimated light on the entrance slits of the detector.
Using the beam splitter, the collimated light from the sample is directed toward the spectrometer. Previously, we used a lens to focus the light on the spectrometer slits. However, the image quality is bigger and better without the lens. I am not sure which is the correct way.

 

[sophiecentaur]

A lens will produce a phase tilt across the field but a diffraction grating works best with a flat phase across it. The collimator does this by increasing the effective distance from the source a point source withought losing intensity.
I can't be too sure about this without a diagram but I think the collimator already does the right job. That would be without any additional lens.
If you google basic spectrometer optics images, the diagrams seem to use mirrors rather than lenses. That would produce less chromatic aberration.

 

[phi_skr]

A lens will produce a phase tilt across the field but a diffraction grating works best with a flat phase across it. The collimator does this by increasing the effective distance from the source a point source withought losing intensity.
I can't be too sure about this without a diagram but I think the collimator already does the right job. That would be without any additional lens.
If you google basic spectrometer optics images, the diagrams seem to use mirrors rather than lenses. That would produce less chromatic aberration.

Thanks a lot for your answer. It clears some of my doubts. But I am bit confused here. The collimating light in the picture is falling on imaging lens and focuses on the slit. Inside a typical spectrometer, there are already some collimating mirrors and gratings that help differentiate the light. I am curious if the light is really required to be focused on the slits or not before entering the spectrometer .

 

[Gleb1964]

I am curious if the light is really required to be focused on the slits or not before entering the spectrometer .

Yes, of cause, the feeding optics should fit the light within the acceptance etendue of spectrometer. That means you should aim 1) even spatial illumination within the entrance slit and 2) even angular illumination within the numerical aperture (the latest to obtain even illumination across dispersive grating).

 

[phi_skr]

Yes, of cause, the feeding optics should fit the light within the acceptance etendue of spectrometer. That means you should aim 1) even spatial illumination within the entrance slit and 2) even angular illumination within the numerical aperture (the latest to obtain even illumination across dispersive grating).

Thanks for your response. Actually, in our system, collimated light is incident on the sample and then, reflection from the sample is used for imaging. So, I was confused whether the reflection rays from sample is also collimated or not. If these reflected rays aren't collimated, then should we add additional collimation lenses or focusing lenses in front of the spectrometer's slit?

 

[Gleb1964]

Right, presumably you need a sort of relay optics to reimage illuminated sample spot onto spectrometer slit, aiming that the light cone would exceed the acceptance numerical aperture of the spectrometer.

 

[phi_skr]

Right, presumably you need a sort of relay optics to reimage illuminated sample spot onto spectrometer slit, aiming that the light cone would exceed the acceptance numerical aperture of the spectrometer.

Thanks a lot (^_^) If you don't mind, would you please recommend me a book to understand optics better for practical applications like making optical systems for experiments? I am having a hard time with it experimentally:(

 

[sophiecentaur]

I am curious if the light is really required to be focused on the slits or not before entering the spectrometer .

I am into aerials / RF rather than optics and gratings but the same principles apply for all waves. There are two requirements. You need as much light as possible through the system so the source needs to be squeezed into the slit. Then, the path length from all points on the collimator needs to be the equal to points on the grating. Having a curve allows this. You need a sharp image of the slit on the detector for best wavelength resolution. The grating will shift that image according to the wavelength.

I'd imagine that first off you want a sharp image in the detector and then you want it as bright as possible by getting the source focussed on the slit. That would be a way to approach it.

 

[phi_skr]

Thanks a lot. My understanding is better now. Actually, this is the system. From what I understand, the rays from the soure/sample should be collimated before it hits the slit and most light should pass through the slit. In that case, there should be collimation part between beam splitter and the slit ?

[Mentor Note: Apparatus image removed by the Mentors at OP request]

 

[Gleb1964]

.. Then, the path length from all points on the collimator needs to be the equal to points on the grating. Having a curve allows this. You need a sharp image of the slit on the detector for best wavelength resolution. .

That is not a necessary demand.
If resolution approaching diffraction limit, the slit width should be reduced accordingly and that would filter out phase aberrations.
If resolution defined by detector's pixel pitch, the slit width would be optimally twice the pixel pitch and again, the phase aberrations is not limited factor.
That means, it is enough to focus "somehow " illuminated sample to the entrance slit. The main subject is matching the illumination cone with the spectrometer internal acceptance cone and provide evenness of the light distribution within the cone.
Uneven light cone would result the an apparent wavelength shift compared to the wavelength calibration.

 

[sophiecentaur]

I hadn’t thought about this before but if the specimen has finite depth, then so will its image so wouldn’t you have to ‘put it at infinity’ to reduce this? Any form of imaging has to be a compromise.

 

[Gleb1964]

Yes, but use of slit relaxes demands for the imagine optics.
If a sample has some finite depth, the image would do the same, even if it would be moved to "infinity ".
Demands about imagine optics depens upon if you want to preserve and resolve its spatial structure or you prefer homogenise sample by blurring it.

 

[sophiecentaur]

imagine optics

I think you mean "imaging optics". Translation problem no doubt.