MicrobeHunter.com Microscopy Forum

Does anyone here use this Polychromatic polarization technique?
. Im super interested in this and want to know if anyone has modded their scope to do this.

Had to look it up: https://www.nature.com/articles/srep17340

Do the two z-cut quartz flats have to be the same thickness? And what typical thickness is used?

See , so many unanswered questions

microb wrote: ↑

Thu Jun 10, 2021 11:36 pm

Do the two z-cut quartz flats have to be the same thickness? And what typical thickness is used?

Get two wedges and try?
oops no, they wrote z-cut...

I am assuming they are wave plates and need to induce a specific phase shift or at least a known phase shift

in act the article describes the "polarization state generator" as essentially a circular polarizer

BramHuntingNematodes wrote: ↑

Fri Jun 11, 2021 5:46 pm

I am assuming they are wave plates and need to induce a specific phase shift or at least a known phase shift

But do they need to be paired, as in quartz plates that were lapped together to be identical? Or can they be different? Could one just use a couple polymer waveplates instead of quartz?

I like the result, so I'd be curious to try it out.

I think it says that a 45 deg rotation at green frequency is introduced by the first plate and an equal but opposite rotation is resolved by the second by virtue of a 90 degree different orientation. There is only one thickness t described, so I assume it is shared by the two plates. Phase shift is introduced by the specimen it seems like.

microb wrote: ↑

Thu Jun 10, 2021 11:36 pm

Had to look it up: https://www.nature.com/articles/srep17340

Do the two z-cut quartz flats have to be the same thickness? And what typical thickness is used?

They have to be the same thickness and of opposite rotation.

If my calculations are correct to get a satisfactory effect between 400-700 nm something in the range of 9-10 mm thick Z-cut quartz windows should do the trick.

The effects are pretty cool, but I still haven't found any amateur reports on the technique.

That's an interesting and colourful technique. I'll have to see whether the polycarbonate prism diy dic (See the 'DIC question' thread) setup could potentially be adapted for the polychromatic technique.

Louise

Confused, I thought Z-cut means the cuts are perpendicular to the optic axis (so optic axis parallel to the optical axis of the microscope) but I think equation (1) in the report is for the case of propagation perpendicular to the optic axis (as if the cuts were parallel to the optic axis)...?

Just read some more
It is not using the birifrengence of quartz, but another bizarre property, the "optical activity": Along the Z-axis, the quartz crystal is arranged as a corkscrew, so it rotates the light passing through it. "without adding ellipticity" as written in the paper.
I fear that Wes is correct, we need high precision to get this polarization effect, as the birifrengence of cells and so on is very weak.
Using quartz in the normal way, like with the Wedge or quarter-wave compensator, adds "ellipticity", meaning that the light is not perfectly polarized. Same for Mica or plastic compensators. Also, with birefrengent compensator, I wouldn't know how to "rotate back"
Using the quartz in the Z direction, rotates the light leaving it perfectly polarized. Also, each color get rotated by heavily different amount. That is needed to separate afterwards the tiny birifrengence of the biological sample.

Glucose / sugar syrup should do the same rotation trick, less precise, but we all have some sugar at home. I try now to see if it is true.

Edit 1 hr after: yes, Glucose rotates polarized light! Pretty colorful too. Need a substantial thickness, like 10mm or more. I tried with some drops on a slide, but didn't work. A glass full of syrup, it works.
Next - how to put two precisely identical cups of D-glucose and L-glucose in the light path ok now getting quartz plates seems more reasonable. Or maybe, in reflected light, could we use the same rotator? Rotations always confused me.

How much sugar syrup rotates
https://file.scirp.org/Html/8-1000120_41497.htm
Quartz rotation /optical activity
http://quartzpage.de/gen_phys.html

Here what I've understood of the method.
Polarizer polarizes white light.
The first quartz plate (right-handed crystal) rotates colors polarization differently, clockwise, while keeping each single color perfectly polarized. Quartz screw magics. Sugar magic.
The sample (birefrengent) alters (rotate/ellipsize) a little bit the polarized light; different colors have different polarization directions, so they get altered differently.
The second quartz-plate (left-handed crystal) rotates back the colors in the original direction.
The final polarizer (analyzer) kills the main signal, only the modification induced by the sample passes. Different colors have been modified differently, so the result is colored.

Precision needed: the color must be put in the original polarization direction, with error smaller than the signal.
Crude estimate:
Signal from sample: ~~Birifrengence*sample thickness /wavelength
Rotation error quartz plates: ~~Tot rotation* (thicknessPlate1 - thicknessPlate2) / tot thickness = Rotatory power quartz * thickness difference

How thick need to be the quartz plates?
1 or bit more millimeters. Each 1mm rotates blue (400nm) by 49 degrees; red by 17. Rotatory power as from https://www.ias.ac.in/article/fulltext/ ... /0103-0113
That is already enough to get the effect. More mm will introduce more rotation, more distance between colors, thus possibly nicer images. 10mm seems the max (more rotations afterwards).
Right-handed and Left-handed quartz crystals surely have slightly different rotatory powers because God is notoriously right-handed; may need some "positive action" to compensate this unfair difference.

Below attempt to translate the equations in a drawing.
Possible presence of gross errors in the whole post.

Giorgio
Edit:
found that quartz plates, Z-cut, are normal items for sale. 2mm standard thickness, which should be enough for the job. Goodbye sugar syrup.
Right-handed, 50£
Left-handed, 250£ we live in a weird world
Not cheap but feasible. Spilling sugar syrup over the halogen lamp will be more expensive.

Makes sense. So If I understand correctly, ignoring the "polychromatic" aspect this is not expected to give significantly better sensitivity in terms of amplitude for given retardation than standard crossed polarizers? The advantages are that structures show up in any orientation and orientation can be determined at lower retardation than is possible with a lambda/sensitive-tint plate?

The color depends on the slow axis orientation cf the arachnodiscus photo

In the Arachnoidiscus comparison 2c appears significantly brighter than 2a, but in addition to the optical stuff...

The right picture (Fig. 2c) was taken with polychromatic polscope. It shows the diatom image after background subtraction.

...which I am guessing refers to some software processing? The methods section at the end doesn't elaborate. Maybe that one is using the arrangement mentioned near the beginning with analyzer parallel to polarizer and the gray background is being subtracted in software?

So I wonder how much the improved brightness/contrast (ignoring color) going from (a) to (c) is due to optics vs. software processing? Good chance I am misunderstanding something but I don't see how addition of the Z-cut plates would significantly increase brightness relative to just crossed polarizers.

hans wrote: ↑

Sun Jun 13, 2021 1:39 am

ignoring the "polychromatic" aspect this is not expected to give significantly better sensitivity in terms of amplitude for given retardation than standard crossed polarizers?

Probably yes, no true advantage. Which leads to the conclusion (happy or sad) that the same images can be obtained by stacking artificially colored photos taken with standard cross-polarizer, with the specimen rotated at different angles.
Instead of rotating and de-rotating the light with the quartz z-plate, we rotate the specimen with the rotatable stage of the microscope - as has been done since 200 yrs. Then we can de-rotate the images digitally, color and combine.
No more need for perfectly lapping together wig and tory quartz plates! No more sugar syrup spills!
If you're picky and have color filters, the photo for each angle can be taken with a different filter, so get natively colored. I have no color filters, so I've taken black and white photos, then colorized digitally.
Of course the specimen needs to be well dead and imbalsamated, shouldn't swim away or eat a neighbor in the middle of the stack. Miscollimations and objective distortion will pop up as mismatches in the stack (my case).
Probably this colored rotation stack trick has been done before and may have some name. Buried in the deep archives of German or British forums.

Below the illustration from this Sunday morning frenzy. Subject is a butterfly wing scale; not sure if it is truly birefrengent, grating, or just scattering, anyway, hope it makes the technique understandable, enjoy & make it better!

Something wrong? Too logorrheic, arrogant or naive posts? Did anybody tried the sugar syrup?
Anyway, here the last one

Saturday I was in full steam and wrote to Michael Shribak, author of the paper. He answered 18 minutes after!

He says that they are selling on order the setup with some success for pathology/medicine and fluid analysis. Price is high for amateurs, but not crazy (less than a confocal setup at least). Write him if you seriously want to buy the system.

The answer to the questions:

microb wrote: ↑

Thu Jun 10, 2021 11:36 pm

Do the two z-cut quartz flats have to be the same thickness? And what typical thickness is used?

is "confidential". Anyway he says that the quartz plates need to be precisely matched and adjusted.

Forward here the links he sent me. First is a nice video presentation of this technique.
4th is a pdf, with a photo of the system built into Olympus CX21
He recommends to read the "Supplementary" linked in the last paper, where they explain and quantify the color rotation key trick.

https://www.youtube.com/watch?v=K-If4NGdp80
https://figshare.com/articles/media/pla ... ds/9733829
https://www.mbl.edu/bell/current-faculty/shribak-lab/
https://www.mbl.edu/bell/files/2014/12/ ... hribak.pdf
https://www.nature.com/articles/s41598-021-85667-8

As from previous post, I've worked out that exactly the same images can be obtained with a normal polarization microscope, just by digital stacking. Shribak's setup is however still much cooler, and can be used for visual and live observation of hyperactive rotifers, as noted in the paper.

If somebody has 300£ to throw, can buy the two standard z-cut quartz plates and see if they're good enough. Likely outcome will be "it works, but wish they were more accurate". Then, the accurate quartz plates cost 20x more.
There is another zero cost, elbow grease workaround. Do you have a polarization scope? Then you're a mineralogist. With a hidden, huge stash of crystals. I know. Pick the largest quartz ones (left and right..), cut them, put on lapping machine. In a couple of weeks of hard work, may be feasible to get the two plates polished with sub-micron accuracy. I may try that when I'll be a pensionist (extremely far and improbable future).

Writing on this thread has been big fun for me, I had no idea of the existence of "optical activity" until three days ago; the z-cut quartz riddle got me thinking. Still no idea on exactly why and how quartz pulls this feat. And why, oh my, left-handed quartz is more expensive than the right-handed.

Disclaimers: One butterfly has been irreparably harmed during the course of the experiment. One polscope has been harmed by repeated sugar syrup spills, but thorough cleaning restored all its functionalities.

Optical activity is a fancy way of saying it is chiral and thus induces circular polarization.
The fact that they sell the systems at least explains why they used a processed image to demonstrate the effect while the other images didn't seem to be.
It does seem interesting though.

Scarodactyl wrote: ↑

Mon Jun 14, 2021 5:25 am

The fact that they sell the systems at least explains why they used a processed image to demonstrate the effect while the other images didn't seem to be.
It does seem interesting though.

Yes, in the image 2c of the paper; the background should have some color hue, due to slight error in the colors polarization realignment. That background has been removed digitally. The hue would have allowed us to estimate the thickness difference of the two quartz plates, or the severity of other problems in the light path.

patta wrote: ↑

Mon Jun 14, 2021 4:53 am

Something wrong? Too logorrheic, arrogant or naive posts?

I liked your posts. I also had not appreciated the distinction between linear vs. circular birefringence until thinking through this.

patta wrote: ↑

Mon Jun 14, 2021 4:53 am

Did anybody tried the sugar syrup?

No, but it reminded me of this video I saw a while back:
https://www.youtube.com/watch?v=XhU-nNiAgtI

Would require a similarly-impractical amount of olive oil, I think.

Scarodactyl wrote: ↑

Mon Jun 14, 2021 5:25 am

Optical activity is a fancy way of saying it is chiral and thus induces circular polarization.

I think a key point is that the propagation parallel to the optic axis in the Z-cut plate actually does not "induce" circular polarization in the usual way starting from linear (phase shift of ordinary relative to extraordinary linear components) but rather a pure rotation that preserves the initial "shape" of the polarization state, modeled as phase shift of right- relative to left-hand circular components.

After patta's post I played around plotting some LH vs. RH phase shifts to convince myself it worked, for example here two initial states (oriented at zero degrees on the graph, one linear the other slightly elliptical) rotated 45 degrees by shifting the RH and LH components. If I understand correctly it is not possible to get 45 degree rotation like this by phase shifting linear components as happens during propagation perpendicular to the optic axis in quartz, for example.

Patta,
I have enjoyed this very much. Fascinating stuff. The fact that they sell a custom-built product definately makes a lot of sense, and explains why the papers were rather vague with regards to some critical aspects. Thank you so much for clarifying much of it.

In one of the papers you linked to ( https://www.nature.com/articles/s41598-021-85667-8 ) the authors actually specify the thickness of the z-cut quartz plate: 8 mm.

Lots of great info here !!

last question. What would be the differnce in plate heights need to be if the first (SiO2) quartz z-cut is 10mm in thickness.

Or would these need to be an exact thickness.

Nice to hear you enjoyed that too! About last question; if you make one of the plates 10.000 mm, the other should be 10.000 mm with max error of 0.001? They should be of same thickness I think. Precision depends on how small critters you want to look at.
I tried to google to see if Left and Right handed quartz crystals actually have a noticeable difference in rotatory power, so they would need different thickness.
Couldn't get it. The whole internet is spammed with things like below. My opinion is that the actual magics of quartz, like our birifrengence and z-rotation, are much cooler.

Come to think of it, I remembwr my undergrad advisor used circular polarizers on one of his microscopes for getting better images of crystal growth experiments, sunce it gave great color on the crystals regardless of orientation. I will have to ask him about that some time.

I emailed the original author . Hopefully i can get a responce and not get brushed off.

I also emailed a producer of synthetic quartz lense to see if they are the same or better when compared to natural quartz in a right and left hand set....

Here is a good read to understand quartz crystals better


https://rudolphresearch.com/sample-calibration/

https://www.google.ca/url?sa=i&url=http ... AdAAAAABAD

Some time after, came up this ballpark estimate for the accuracy needed in the quartz plate
Top and bottom plate must rotate and de-rotate by the same amount, within objective resolution.

Let's say we are taking 4 MPx photos (2000x2000).
The color rotates around the center; at the side, we are 1000 pixels away from the center.
We want the de-rotation to match the initial image by less than one (1) pixel; hence the angle must be precise within 1/1000 (in radians; =0.06 degrees)
As above posts, quartz rotates blue light 49 degrees per mm.
So to have blue light back where it was, we need the quartz plates thickness (and flatness, and maybe parallelism) to be precise better than 0.06/49 = 0.0012 mm = 1.2 microns.

Pretty tight but not otherworldly. More precise, more MegaPixels.

Later edit: I think the above estimate is confused crap.

I've just discussed with others this "polychromatic stacking", as post #17, and in order to convince them that it's legit I had to make some more decent images:
small chunks of Mica
fibers of polyester
both at 4x NA 0.10
It also turned out that this method has been presented already in the 90's. Anyway.

In those images, the color is not as with normal polarization (note in the fiber, there is no Extinction direction)
Here color represent the direction of the axes; while the brightness represent the total retardance/thickness
Blue is when the axes are horizontal/vertical; red when they are at 45 degrees; yellow-green-purple, in between.

I've made a mistake before, the three photos to be stacked should be taken at 0, 22.5 and 45 degrees.
That is because I'm not using a lambda/4 plate, that allows to distinguish the slow and fast axis.