High NA, Low Mag
Low magnification, high numeric aperture (NA) microscope objectives are popular choices for multiphoton imaging. They offer excellent light collection properties, a wide field of view, and still permit high resolution imaging. However, they are challenging to design optically. Different manufactures have settled on different engineering compromises.
For example, Leica’s 20x/1.0NA objective is optically magnificent. I’ve been impressed with Leica’s coatings in the past, particularly for IR transmission, and they still outperform the competition. However, the thing is simply massive. It weighs 411 grams, and is 38 mm in diameter. It has a respectable 38 degree access angle, for getting pipettes underneath, but with only 2 mm of working distance, there isn’t much room for movement.
In contrast, Nikon’s 16x/0.8 NA objective is clearly inferior to the Leica optically, despite offering a wider field of view. However, it has some great ergonomics to offer. The working distance is 3 mm, the access angle is a comfortable 45 degrees, and it weighs a comparatively light 221 grams. Furthermore, the lower magnification of the Nikon partially compensates for the lower NA when it comes to image brightness.
For patching, the Nikon wins hands down. For pure imaging, the Leica wins hands down. But they’re not the only games in town. Olympus was the first manufacturer to really push this style of objective for multiphoton imaging, and they have a 25x/1.05 NA objective with a 2 mm working distance and 35 degree access angle. And Zeiss has a 20x/1.0 NA objective with a 38 degree access angle and a 2 mm working distance. This market has seen some really nice innovation in the past 5 years, and there are new designs coming out all the time. So be sure to check with several manufacturers before settling on a particular objective. They are very specialized, so the best objective for one experiment may be quite poor for another application.
[…] from quickly perusing the site: Printable bolt size charts, Tips on intrinsic optical imaging, Comparison of high NA, low mag objectives, and my favorite, Catalogs as textbooks. (I still remember a neuroscience faculty member here at […]
The downside of the 1.05 from Olympus is it is $20,000, IF you can buy. As of June 2011 they still only sell with their 2P microscope. That will change as Nikon offer a 1.1 NA without restriction, apart from the price!
Nice weblog
Has anyone directly compared the Olympus 25x and Nikon 25x? Are they pretty similar in terms of image quality, IR transmission, correction for aberrations, etc? It’s hard to get real data/specs on this!
Olympus has now changed their product policy and is offering the 25x to endusers directly. No need to buy a FV-system anymore. Also interesting is that they have a 4mm WD version optimized for Scale and are bringing out a 8mm WD Scale version soon. With a slight drop in PSF quality, these objectives should also be usable as water dipping … that might make pipette access sweet and easy. I couldn’t get any info on the NA of the Scale versions.
I would like to see some detailed measurements before I agree with your view of the Leica.
A few years back, we took the 16X/0.8 Nikon and compared it to the 1.05 NA/25X Zeiss and the 1.05 NA/25X Olympus. Turns out that while there was a marginal improvement in resolution with the higher NA objectives, they literally had 50% the collection efficiency.
Maybe the 20X/1.0 NA makes it better, but what I would REALLY love to see is a 10X, HUGE field of view ~1 NA objective.
PS, here is how I took that measurement: I wanted a uniform point source, so I dropped a single bead onto a petri dish (hardest part – basically a series of dilutions and a bit of luck ; it was glued into a fixed position). I then illuminated from below with a filtered light source. The emission side had a filter to block out the excitation light. I used two photon (no filters) to center the bead. Then I popped in the illumination & filters. The PMT reported ~50% of the light with the 25X objectives.
Thanks for the great comment, Chewbacca! I agree with you that a 10x/1.0 NA objective could be great, and that higher NA is not always better.
A number of things factor into these comparisons, including the anti-reflective coatings. For example, 10% worse IR transmission (2p excitation is proportional to the square of intensity) and 10% worse transmission in the visible range can lead to a very significant decrease in detected fluorescence.
The Leica had, at the time that I wrote that post in 2010, the highest transmission of both IR and visible light. Some of the objectives from competing manufacturers had particularly poor IR transmission. For example, an older design of the Olympus 20x/0.95 had under 60% transmission at 800nm. Their improved coating got closer to 75%, Nikon’s was 72% at 800nm, Zeiss was at about 80%, and the Leica was around 90% (these are numbers from the manufacturers themselves, and their coatings may have been improved in the past few years).
Of course, one can always just turn the laser up, but this isn’t always controlled for in casual tests. By contrast, it sounds like in your setup, Chewbacca, it would be easy to measure and both IR transmission and light collection for a fixed laser power, and I’m guessing this is what you did. What did you note about the IR transmission of different objectives?
It sounds like your preparation was relatively low scatter. In highly scattering tissue, there may be an advantage of moderate (as opposed to high) NAs.
http://labrigger.com/blog/2012/03/27/axial-resolution-vs-numeric-aperture/
A final note is that it’s important to control for the overfilling of the back aperture, and this is not always trivial to do.
http://labrigger.com/blog/2012/03/28/axial-resolution-and-numerical-aperture-part-ii/
Overall, I think empirical approaches like you used, despite their challenges, are very important. Like I said in the original article, “Be sure to check with several manufacturers before settling on a particular objective. They are very specialized, so the best objective for one experiment may be quite poor for another application.”
I’d like to add one important point to this discussion:
Magnification != Magnification.
Magnification as stated by objective manufacturers is relative to the tube lens focal length used in their scope ecosystem:
Magnification = Tube lens focal length / Focal Length Obj
This has huge consequences if you use off-brand objectives. On a Nikon/Leica like system (or anything with f_tube = 200mm, e.g. Thorlabs) the 25×1.05NA olympus is actually a 28x objective. In my hand this translates to a smaller FOV for the olympus and lets the Nikon25x1.1NA shine in performance/price:
At my lowest zoom settings (TL scopes) I get the following:
Nikon 25x
~700x700um
Olympus ’25’x
~621x621um
and for comparison the ‘workhorse’:
Nikon 16x
~1135 x1135um
If the same holds true for the Zeiss 20x1NA (I did not test this!) then this would be at a horrible 33x magnification (Zeiss f_tube = 165mm).
Along the same lines: Be aware that the 0.5NA 5x air monster of Nikons AZ100 is effectively a 10X objective at a f200 tube. It’s still good enough for 2p excitation through air (!) and a cool option for cleared brains (http://goo.gl/VRFDl).
At some point I’d really like to get my hands on the Leica Behemoth. 20x1NA sounds awesome. Did you test that, L.?
Cheers,
T
Tube length data:
http://www.microscopyu.com/articles/optics/cfintro.html
I am sorry: I screwed up the Zeiss calculation (damn it, Excel!):
165mm/20Mag = f8.25
on a f200mm tube system this means:
f200/f8.25 = 24.2Mag
So effectively the only objective that _really_ stands out here is the Leica.
T
[…] Scientifica have completely redesigned the optical path to overfill the large back apertures of low mag, high NA objectives. This new optical design is also available for the galvo scanhead. Alternatively, they also still […]
Does anyone have any experience with the Olympus 20x/1.0? how this objective is compared to the ones discussed above?
Thanks
EITAN
Hi Eitan,
We’ve had a couple of those Olympus 20x/1.0NA objectives in the lab for a few years now, one on a Scientifica 2-photon and the other on a BX-51 ephys rig.
One important thing noted above is the ergonomics – the Olympus 20X/1.0NA has the same 2mm working distance and 38deg cone angle as the Leica 20X/1.0NA, which is not too bad, but having a bit of extra room around the objective can make a huge difference if you need to cram in extra electrodes etc. I ended up having to spend a lot of time redesigning our custom imaging chambers in order to accommodate the new objectives. Weight can also be an issue if you’re considering doing fast z-stacks using a piezo stepper – fortunately the Olympus weighs about 230g (roughly half as much as that beast of a Leica!). Optically I have absolutely no complaints, and I think 20X is the sweet spot in terms of magnification for the sorts of imaging we’re currently doing.
I can’t really speak to the relative optical merits of the other Leica/Zeiss/Nikon options mentioned above, having never used them.
Alistair
Is there a good solution for a 5x high NA ? Air would be fine, WD = 2 or more.
I just purchased the new Nikon 25x 1.1NA that is corrected out to 1300nm. I have a project in the lab imaging Gcamp6 in dendrites in visual cortex.
Its UNBELIEVABLY better than the 16x! We are able to see much greater detail. It pains me to say it, but it is worth the $24K
Josh,
Do you have couple comparison images you can share? Is this keeping all other parameters roughly constant (pix/ um), intensity/excitation volume, etc? Also, do you feel it’s the increased z resolution or the increased collection efficiency that is making this objective qualitatively better (for this purpose)?
Thanks for sharing!
I fully agree. For spine / dendrite imaging the Nikon 25x 1.1NA is the way to go (preferable to the Olympus 25x for reasons stated above). My measured PSF of the 25x is ~twice as good in all dimensions in comparison to the 16x.
My only gripe is the correction collar – it’s tricky to use properly because of the huge z-shift. Also: it’s completely covered by the heavy duty piezo I use for multilevel dendrite imaging. A piece of a red bull can attached to it solved that mostly – but still it’s tricky.
Are you adjusting the collar for each dendrite or are you just going for one setting in the end?
…worth reading (Olympus centric):
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4541535/