Posted by
Mark Cannell on
URL: http://confocal-microscopy-list.275.s1.nabble.com/How-to-measure-objective-transmission-curves-tp590172p590209.html
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocalGuy Cox wrote:
> Search the CONFOCAL archive at
>
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal> People do seem to be making rather heavy weather of this,
> which would seem fairly simple to me. First off, there's
> no need to use oil - the acceptance angle of the lens is
> not affected in any way by the presence or absence of
> oil (resolution and aberration correction are a different
> matter!).
That's true but the reflective losses from the lens will depend on the
medium contacting the front lens. Of course this could be estimated from
the Fresnel equations and for a high NA oil immersion lens this will add
( I guess) about a 10%.
> The acceptance angle of an NA 1.4 lens is
> 67.5 deg. One could put an integrating sphere over the
> lens if one is really worried that transmission for paraxial
> and abaxial rays is different. One could also test this point
> by using a parallel beam and measuring first on the optic
> axis and then at an angle of (say) 60 deg.
Hi Guy,
I don't think that will work because of the apertures present. If you
shine a // beam into the front of a high NA lens the beam will be highly
focussed within the lens body and the rear aperture may block some of
the light coming out (it depends where the back focal plane of the
objective is -for highly complex high NA lenses I'm not sure that that
is at the rear aperture). In addition I think reflective losses within
the lens will change as the input is not a cone of light with the
correct NA (i.e. the path of most of the rays is now completely
different). Also, real integrating spheres are _not_ Lambertian and
comparing a highly divergent beam to a highly collimated beam will also
likely give rise to an error (this is a well known problem). If you have
any lens design software you can simulate the problem quite easily.
I'd say the problem is so 'not simple' and manufacturers avoided
publishing transmission curves beacuse they are very hard to measure
correctly. I'll guess that that the curves now published by
manufacturers are actually produced by the lens design programs they
use. Modern optical design programs can incorportae the anti-replection
coatings to be applied and so the need for measurement is
reduced/simplified.
However, if one just wants to compare two lenses with the same NA then
your method could work -but it will be necessary to make sure the input
beam is not clipped.
Cheers Mark
>
>
> Associate Professor Guy Cox, MA, DPhil(Oxon)
> Electron Microscope Unit, Madsen Building F09,
> University of Sydney, NSW 2006
> ______________________________________________
> Phone +61 2 9351 3176 Fax +61 2 9351 7682
> Mobile 0413 281 861
> ______________________________________________
>
http://www.guycox.net
> </exchweb/bin/redir.asp?URL=
https://www.mcws.usyd.edu.au/exchweb/bin/redir.asp?URL=http://www.guycox.net>
>
> ------------------------------------------------------------------------
> *From:* Confocal Microscopy List on behalf of Mark Cannell
> *Sent:* Wed 5/09/2007 9:15 AM
> *To:*
[hidden email]
> *Subject:* Re: How to measure objective transmission curves?
>
> Search the CONFOCAL archive at
>
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal>
> Hi Andrew
>
> All rays whose effective NA >1 get reflected back by the system you
> describe. There are more high NA rays than low NA rays (due to the area
> of an annulus increasing with it's radius : pi(2r dr). As I see it, one
> can't measure the true throughput with axial rays as the marginal rays
> encouner very different glass thicknesses and there may be internal
> field stops (or even phase rings!). The measurement of throughput is
> certainly not going to be easy. I don't see how this is related to the
> sine condition, which I believe simply defines a geometric constraint
> for an undistorted image viz:
>
>
http://en.wikipedia.org/wiki/Abbe_sine_condition>
> Cheers Mark
>
>
> Andrew Resnick wrote:
> > Search the CONFOCAL archive at
> >
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal> >
> > Mark,
> >
> > That's a good question. We attached a coverslip with a small drop of
> > oil to the objective as a way to deal with immersion objectives. I
> > don't know what it means that the "sample" is in air; if some light is
> > then totally internally reflected back into the coverslip, which could
> > be a spectral issue due to dispersion.
> >
> > Using two matched objectives is probably a better method, but I
> > imagine considerably more difficult due to alignment constraints and
> > obtaining matched objectives. The objectives we had obeyed the 'sine
> > condition', which I believe means that there are fewer high-NA rays
> > than low NA rays- this would help mitigate our experimental error as
> > well.
> >
> >
> > At 04:19 PM 9/2/2007, you wrote:
> >> Search the CONFOCAL archive at
> >>
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal> >>
> >> Hi Andrew
> >>
> >> While I can see that method would work for lenses designed to work in
> >> air, wouldn't total internal reflection be a problem for lenses with
> >> an NA over 1.0? The only way I can think of doing it is to use two
> >> objectives of the same type looking at each other at a common focal
> >> point with the right immersion medium between them. Then the overall
> >> transmission would be half that of a single lens...
> >>
> >> Cheers Mark
> >>
> >> Andrew Resnick wrote:
> >>> Search the CONFOCAL archive at
> >>>
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal> >>>
> >>> We've done similar measurements, it's not too difficult.
> >>>
> >>> The main trick is handling the large NA lenses. Our setup was light
> >>> source -> objective -> integrating sphere -> spectrometer. After
> >>> normalizing to the source, we obtained really good data. It's
> >>> pretty easy, actually.
> >>>
> >>> Andy
> >>>
> >>>
> >>> At 08:45 AM 8/31/2007, you wrote:
> >>>> Search the CONFOCAL archive at
> >>>>
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal> >>>>
> >>>> Dear all,
> >>>>
> >>>> one of the main differences of objectives is their transmission
> >>>> efficiency at certain wavelengths. One way to compare this is the
> >>>> trial-and-error method, however, this is not straight forward.
> >>>>
> >>>> My idea is to use a combination of spectrophotometer and a lamp
> >>>> with a more or less even spectra (i.e. Xenon) on some kind of
> >>>> optical bench. This would make the setup independent from the
> >>>> manufacturer. Beside the distance between the light source and the
> >>>> detector, there are obviously more things to consider: different
> >>>> diameter of the back focal plane, different focal lenghts...
> >>>>
> >>>> I would like to hear about your opinion about how to measure
> >>>> objective transmission. Have you ever done this in your lab? Did
> >>>> you find a setup that worked for you?
> >>>>
> >>>> cheers,
> >>>> Michael
> >>>
>
>
>