Normally you would want to focus and collimate using identical lenses. In principle, you could use the same rig on the other side. Precise alignment of the collimating lens (at least in the x-y) is a good thing, since it avoids introducing new aberrations by illuminating the lens off-axis.

That said, you could probably just use aspherics on both sides. Or microscope objectives on both sides.

Using microscope objectives with high-power ultrafast beams gives me pause - they have a lot of glass which can introduce dispersion into the beam, effectively stretching it. And optical cement has a much lower damage threshold than glass.

Depending on the size of your beam and how tightly you want to focus it, you could also use best-form lenses.

Dear Tomoyuki,     I visited NewPort to see if how they specify M-10x objective lens and 25 micron pinhole for spatial filtering.   The lens is specified as NA=0.25, 400-700nm, f=16.5mm while input beam diameter is 5.5mm.   But as my attachment shows, the input beam diameter 5.5mm seems to be too large for 25 micron pinhole.   Actually about 1 mm beam diameter fits the pin hole size, when Gaussian apodization is ignored.    You can use aspheric lens pair with NA=0.25 if take great care of input beam daimeter matced with pin hole diamter and ensure aspheric lens should be used in monochromatic system.   Regards, Shigeo.

Dear Tomoyuki,

I agree with the considerations made by Shigeo. However, when the objective is working at infinity, the maximum beam diamter is 5.5 mm. Your beam will have lower diameter, depending of what collimated source entering the objective you use (LED, laser?). It is also important to keep in mind the intensity distribution through the beam diameter in the calulations, i.e., uniform or Gaussian.

Now, you should consider the aperture of your collimating lens after the filter. If you use one of the same value of the microscope objective, you will get a collimated "clear" beam of roughly the same size. If you want a wider beam, you should use a collimating lens having a greater NA value, or one of the same NA followed by a beam expander (typically a Galilean telescope).

I often use an achromatic lens (doublet lens) after a spatial filter, they are quite well correted for spherical aberration, several apertures (focal to lens diameters) to choose, and cheap. These can even have aspherical surfaces. Again, depending of your specifications, a beam expander is a good solution to control your collimated, filtered beam, unless space is a limitation in your setup.

Hope this helps.

José Antonio

For a Class IV laser, you might need to verify the intensity of the laser at the focus and its potential to damage the material of aperture. This can place an upper limit on the NA, in which case a singlet or doublet might be diffraction limited. Of course, this means larger focal length and a larger physical footprint.

For a low intensity laser where you want a smaller footprint and thus higher NA, AR coated plastic asphere condensers may be the cheapest option. Plastics may be inappropriate for some laser types, I'm not sure exactly. Simple aspheres have higher chromatic aberration than microscope objectives or doublets, which is no concern if there is only one laser wavelength.

I also agree with the other comments.