Jon Daniels

Details here: bsky.app/profile/jsda...

Tagging @amsikking.bsky.social @andrewgyork.bsky.social @tanner-fadero.bsky.social @retof.bsky.social @loicaroyer.bsky.social and I'm sure I'm missing more but hopefully this will make the rounds...

High-res cellular imaging uses O1 > 1.0 with O2 being 40x/0.95 02. Short O2 WD means the grind has to sacrifice FOV in the depth axis, but FOV in the other direction (parallel the coverslip) is still 600um.

For larger samples e.g. zebrafish the grind position is wider to capture the entire FOV.

4. 55° grind so compatible with any-immersion concept.

5. Two different options for grind position: one for situations where O2 is NA ~0.95 and another for NA ~0.8. This one requires a bit more explanation.

A few nerdy details:

1. Generally better imaging quality especially for those pushing the FOV. I can send plots to those interested.

2. The image plane is a few microns off the glass tip to mitigate problem with dust.

3. Improved capture of high-angle rays despite slightly reduced NA of 0.99.

comparison table for v2 and v3

Production has started for a new #Snoutscope objective, AMS-AGY v3. It has significantly larger FOV and some other tweaks from v1 and v2. Currently accepting pre-orders. The price will need to increase modestly after deliveries start late October.

OK, makes sense. There is a loss in perfect 3D focusing away from the native focal plane described in the Botcherby paper and seemingly there in practice. Can that also be derived from this analytic approach?

Really cool! Could this be extended to focal planes other than the "native" one?

The equation from the paper is equivalent to 4.676 pixels per "Rayleigh diffraction limit square" (i.e. change that one number and my calculation lines up)

Even if you don't use all the "pixels" in such an objective, there is speed/brightness benefit to having such a large NA over such a large FOV.

Can you explain a bit more how you calculate it? I'm probably missing a factor.

FOV is π/4*(7200^2) in units of μm^2

Rayleigh diffraction limit is 0.61*λ/NA in units of μm

Can support 4 pixels per (diffraction limit)^2 (is this right?)

At 500nm: π/4*(7200^2)/(0.61*0.5/0.5)^2*4 = 438 MPix

I'm excited to be able to be in Zurich for the mesoSPIM symposium. This is a really fantastic project.

We're about to start sending quotes, should be delivered late summer / early fall. Happy to discuss via email.

v3 is coming which will functionally replace v2, and should be modestly better in a few respects. So no reason to make more v2, at least for now.

I'll have to think about that one

Correct. We have precision-thickness 1.5 coverglass AR-coated and that is what we cut to size. If the objective you are gluing to has a correction collar -- many do -- then it doesn't matter anyway.

We routinely see PSFs indistinguishable before/after gluing (flipping the coverslip that the beads are supported on).

We use silicone-based glue with working time of 10s of minutes but hardens overnight. It can be carefully peeled off the objective after hardening, but coverslip isn't recoverable.

(yes this is verbatim a Twitter post from 2022, but I've basically stopped using Twitter and wanted to be able to refer to it here)

Iterative leveling is guided by bead stacks: take stack of beads, inspect reliced images, poke again. It’s tedious but algorithmic once you determine which coma direction corresponds to which tilt direction. In our hands, once the coma is gone then other aberrations are gone too.

If you want to do this yourself the process we’ve come to is 1) cut AR-coated coverslip to size; 2) apply silicone glue around the edge of the objective tip; 3) place coverslip 4) level it by iteratively pushing the edge with a toothpick; 5) double-check with beads after curing.

We then analyze the stack with PSFj to make sure the performance lines up with theory. We found out the hard way that you can miss things if you don’t have sub-micron beads or rely solely on visual inspection.

Most high-NA air objective lenses are coverslip-corrected but sometimes you want to image in air, e.g. high-NA SOLS. We have glued coverslips onto objectives for this. To make sure it gets glued properly we take image stack of diffraction-limited beads supported on a coverslip. 🧵

I've thought a lot about pushing stage speed. I'll add

1. Hardware triggering is essential for top speed.

2. Settling time can be a big deal. There can be physical shaking long after the move "completes". For ASI stages set the landing tolerance to your needs (default is deep sub-diffraction).

As you move away from the optic axis in the sample plane it will eventually become impossible to fill the aperture stop from that position in the field. I think that is what you are proposing as a definition for FN. My sense is that other aberrations usually kick in sooner.

But what one user considers acceptable might be unacceptable to someone else. Especially if you are using the objective in an unconventional way, I think you just have to try it out yourself. Also nominally identical objectives can have performance differences that matter to the very picky.

FN is reported by the manufacturer based on some metric(s) that they choose. Usually the image size where wavefront error or some other aberration is deemed to exceed acceptable limits however they define acceptable. It is a perfectly fair question to ask the manufacturer what criteria they use.

E.g. if your camera is spatially undersampling then the usable FOV might be larger because you are less sensitive to degraded Strehl ratio at the edges. If you are insensitive to field curvature then your usable FOV might be larger than a colleague who really needs a flat field. Etc.

The FN is a spec of the objective. From it you can estimate the usable FOV. However there is not a uniform rigorous definition of FN or FOV. The lack of rigor is a defensible approach because different applications vary on what is "good enough" optical performance.

Excited to see this more accessible and publicized! When I first heard Andy describe the concept I was blown away by its elegance.

Good luck with whatever comes next! You have had a huge positive impact in the microscopy world.

DaXi—high-resolution, large imaging volume and multi-view single-objective light-sheet microscopy - Nature Methods The DaXi single-objective light-sheet microscope achieves fast, high-quality imaging of large volumes. DaXi’s design allows increased scanning range without sacrificing imaging speed or quality, multi...

DaXi is a SOLS design that simultaneously achieves sub-micron resolution (450nm laterally and 2um axially) and large FOV (volume of 3000 x  800 x 300 um with galvo, more with stage scanning). Light sheet gentleness and speed. Multi-view capability. www.nature.com/articles/s41...