silica deposits

A forum to ask questions, post setups, and generally discuss anything having to do with photomacrography and photomicroscopy.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

iconoclastica
Posts: 486
Joined: Sat Jun 25, 2016 12:34 pm
Location: Wageningen, Gelderland

silica deposits

Post by iconoclastica »

For a long time I have been photographing silica deposits on the ridges of the stem of horsetails (plants: Equisetum). In fact, this was what guided me into macrophotography. The deposits are transparent, glass-like knobs and bars, some 50 to 200 micrometer wide, and taxonomically important. Here's one of the oldest photos I have made, about a dozen years ago. The knobs are the irregular pairs of features within the dark bands, not the wider space pairs at either side of the lighter bands, for those are the stomata. It's clear from this picture that they reflect the light like silver christmas balls. The first thing I understood once I learned anything about photography was that I had to find a way to control the specular highlights.

Image

Follows a number of years of experimentation, mainly trying all kinds of paints and coatings to get the glimmers off. Finally found a way by sooting the specimen in the flame of an oil lamp. Later I tried magnesium chloride, but that resulted in flakes that didn't resolve the fine detail I am after. Moreover, the sooting method I way easier to apply, albeit prone to burning the plant tissue. So I still use soot.

Image

The next step is using photomety to visualize the depth information (to see whether the bars above are equally high in every part of their length). The depth map produced by Zerene proved an excellent means to that goal:

Image

ImageJ, to be precise: the interactive surface plot plug-in, plots this depth map into a relief-map, onto which the picture output of Zerene can be applied as a skin:

Image

Image

This example I chose because here it all works out excellent. But in some cases the z-resolution is insufficient. I am not overly much concearned about the terraces that become visible on the slopes between the ridges and the valleys. But I do have an issue with the case where the top of the knobs are in focus simultaneously with their bottoms, i.e. the top of the ridge. For then I don't get z-separation in the most important part of the scene.

My setup consists of the 10x Mitutuyo, the Raynox 150 and a bellows, stepping typically 10 micrometers. I think I ought to reduce the DoF per slice, but how could I do that? Or are there other ways to get higher z-resolution?

Wim
--- felix filicis ---

Chris S.
Site Admin
Posts: 4042
Joined: Sun Apr 05, 2009 9:55 pm
Location: Ohio, USA

Re: silica deposits

Post by Chris S. »

Wim,

Your sooting technique is excellent! Thanks for sharing it.
iconoclastica wrote:My setup consists of the 10x Mitutoyo, the Raynox 150 and a bellows, stepping typically 10 micrometers. I think I ought to reduce the DoF per slice, but how could I do that? Or are there other ways to get higher z-resolution?
I think your sense is correct that you want to reduce your DOF in each slice. For that, the obvious attack would be using a lens with a higher numerical aperture. The Mitutoyo 20x/0.42 seems like an obvious candidate, and if you want a similar field of view to what you're getting now, you could use a shorter focal length converging lens.

If you don't mind a narrower FOV, the higher magnification, higher NA Mitutoyo objectives would also be candidates.

At considerably higher cost, Mitutoyo's HR series of objectives has even higher NAs. The Mitutoyo HR 50x/0.75 would probably be wonderful for what you're doing. However, list price on this lens is $5700 USD--ouch. But it's a bargain compared with the Mitutoyo HR 10x/0.42, at $8,100.

--Chris S.

Lou Jost
Posts: 5943
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Very neat trick with the soot. It would be great to find something similar that coats without heat....I have the same problem with my micro-orchids, but they do not resist any kind of heat.

Many of us use a smaller step size with the objective you mention. I use 5-8um.

The hi-res Mitu objectives may have smaller good-quality image circles (that was my experience with the hi-res Qioptic objectives), so I would go with the simpler solution, the normal Mitutoyo 20x with a lower-power tube lens. You did not mention your sensor size. If it is APS, I can highly recommend the 135mm Nikkor-Q (available for < $50 on eBay) as a tube lens. It works well on FF too except for the extreme corners. Also you can use good medium-format lenses, some of which are great even on FF sensors.

Chris S.
Site Admin
Posts: 4042
Joined: Sun Apr 05, 2009 9:55 pm
Location: Ohio, USA

Post by Chris S. »

Wim,

I'm tempted to try your sooting protocol on some of our local Equisetum species (commonly called horsetails, as you said), stack them with higher NA objectives I have on-hand, and send you the results for comparison with yours, in hopes the collaboration might provide a sense of what level of benefit is conferred by higher NA objectives.

On the downside, though we have several Equisetum species in my area, they are common enough that my brain hasn't recorded their locations, but not so common that I can find them without a lot of looking.

So I'd be interested in hearing your opinion of whether or not such an effort is worthwhile.

Also, would you mind describing your ingenious soot-coating protocol in sufficient detail that someone else could carry it out as you do? For example, what kind of oil do you use? What kind of oil lamp? And how do you adjust your oil lamp to limit oxygen enough to create an inefficient oil burn, in order to produce a high level of soot (aka "unburnt hydrocarbons)? Assuming you adjust the flame visually, what does this flame look like when you get it right? How high above this flame do you hold your specimen to obtain maximum soot deposition with minimal subject burning?

And lastly, are there any Equisetum species of particular interest to you? If so, please list them. If these species happen to occur in my area and I can find them, I'd start with these specimens.

Cheers,

--Chris S.

iconoclastica
Posts: 486
Joined: Sat Jun 25, 2016 12:34 pm
Location: Wageningen, Gelderland

Post by iconoclastica »

Chris,

The horsetails I am interested in (for this technique) are those of the subgenus Hippochaete, or the hard or wintergreen horsetails. I think the common ones in the central USA are E. hyemale, E. laevigatum, E. variegatum and the hybrid E. x ferissii. The skins of the softer and commoner horsetails such as the Field horsetail E. arvense don't have very pronounced silica deposits.

I don't think the lamp is very critical. Here's the one I am using. (If it looks a bit baroque: I was making memento mori-s at the time when I started this, so it was just the lamp that happened to be available :) ). It's filled with ordinary commercial lamp oil, which is a mixture of C5-C26 mineral oils.

Image

I light the lamp and then pull out the wick with a needle, just unto where the flame becomes unstable at its tip. There may be occasional whisps of smoke but it's not continuously smoking.
Then I move the object through the tip of the flame, thus causing further instability and with that incomplete burning of the oil. It's just moving the object repeatedly through the flame. Once we used to soot pieces of glass this way to watch the solar eclipse.

Success varies. Too little soot and then only the knobs are sooted, which is not too bad for my purpose; only the ridges become visibly darker than the valleys. Too much soot happens when the flame is too long. Burning occurs when I keep the object too long in the flame, especially when the flame is too short and produces too little soot. Horsetail stems are perfectly round in cross section, so you can roll them quickly between your fingers to get an even covering and heating.

Avoiding burning is a matter of agility. I can soot my own finger without feeling pain.

Wim
--- felix filicis ---

iconoclastica
Posts: 486
Joined: Sat Jun 25, 2016 12:34 pm
Location: Wageningen, Gelderland

Post by iconoclastica »

I am afraid another Mitty is beyond my budget right now. But I can start experimenting with the microscope objectives I do have at my disposal. I am especially thinking about the 20x Nikon M-Plan finite objective. I hesitate about the narrower field, though.
--- felix filicis ---

iconoclastica
Posts: 486
Joined: Sat Jun 25, 2016 12:34 pm
Location: Wageningen, Gelderland

Post by iconoclastica »

Lou Jost wrote:Very neat trick with the soot. It would be great to find something similar that coats without heat....I have the same problem with my micro-orchids, but they do not resist any kind of heat.

Many of us use a smaller step size with the objective you mention. I use 5-8um.
Interesting. Has this been established empirically or was it calcuated?

And perhaps smaller step sizes may help me. Even if I don't see differential sharpness, perhaps the software does. At least it would be the simplest solution.
Last edited by iconoclastica on Fri Sep 27, 2019 6:48 am, edited 1 time in total.
--- felix filicis ---

mawyatt
Posts: 2497
Joined: Thu Aug 22, 2013 6:54 pm
Location: Clearwater, Florida

Post by mawyatt »

Interesting somewhat common problem (specular highlights), and extremely clever solution!!

I recall that soot was partially composed of carbon Buckyballs (carbon molecules arranged in a soccer ball configuration) and why it feels "oily", yet no liquid. If so, and these are the cause of this effect, could the carbon Buckyballs be applied without the need for heat, maybe from some other source?

Thanks for posting.

Best,
Research is like a treasure hunt, you don't know where to look or what you'll find!
~Mike

iconoclastica
Posts: 486
Joined: Sat Jun 25, 2016 12:34 pm
Location: Wageningen, Gelderland

Post by iconoclastica »

I did a test varying the step size (10µm and 5µm). Without smoothing (averaging) I didn't see any significant difference. With low smoothing (2) there is a great difference with much more detail in the 5µm stepped one:

Image
left: 10µm; right: 5µm

Also the small step size increased the detail contrast and (but just a little) the detail sharpness.

Now this test was done with a badly prepared sample, with little soot and much incineration. I nevertheless used it for it esposed the silica-skeleton and increased the micro-relief. I'll try to get some critical material again to see if the increased vertical detail also results in a better suface plot, when realistically smoothed.
--- felix filicis ---

Lou Jost
Posts: 5943
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Several others on this forum had seen subtle improvements when going to 5um, though I don't recall any side-by-side tests like yours. I use 5um when stacking speed is not important (inaminate objects) and I think the results are slightly better than 8um.

Rik's table at http://zerenesystems.com/cms/stacker/do ... romicrodof calculates 9um for NA=0.25, and 6um for NA=30. Thus 7 or 8um would be its recommendation. Since rail accuracy is not always perfect, it seems like a good idea to use a slightly finer step size.

Pau
Site Admin
Posts: 6051
Joined: Wed Jan 20, 2010 8:57 am
Location: Valencia, Spain

Post by Pau »

Excellent results and a very interesting technical thread, congrats Wim!

I concur with other people: a higher NA objective and smaller steps would provide improved resolution.

Moving out of subject, the easiest and best approach would be SEM: all the work done in standard steps ad a much higher resolution image
:-k ...although you need a SEM and ion sputter BTW, maybe you could avoid the critical point dehydration, but most people do not have access to such equipment, take a look at:
https://www.researchgate.net/figure/SEM ... 4/download
Pau

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

Lou Jost wrote:Rik's table at http://zerenesystems.com/cms/stacker/do ... romicrodof calculates 9um for NA=0.25, and 6um for NA=30. Thus 7 or 8um would be its recommendation. Since rail accuracy is not always perfect, it seems like a good idea to use a slightly finer step size.
There are other reasons why finer step sizes can be helpful.

I computed the numbers in that table on the basis of 1/4-lambda wavefront error, which corresponds to a maximum contrast loss of 26%, for a sinusoidal grid of intensity variation, in a planar subject.

If the subject is planar, and diffraction is the only significant factor with all the rest of the process perfect, then the numbers in the table will result in a slight amount of focus banding, detectable by eye with study.

However, when the subject is not planar, especially when there are short-range changes in gradient, then other factors become significant. Most important, I think, are "utilized aperture" effects that cause surface features to apparently move laterally as focus is changed. In severe cases, with heavily sculptured surfaces, this can result in features appearing to "squirm around" as focus changes. An example is illustrated at http://www.photomacrography.net/forum/v ... 187#149187 . In cases like this, using a much smaller step size should give the software a better chance to get things right, though the result should always be considered with a certain amount of skepticism.

I'm guessing that Wim's sooted surfaces will suffer less from this effect than more shiny surfaces do. But it's an issue that should be checked for.

--Rik

mawyatt
Posts: 2497
Joined: Thu Aug 22, 2013 6:54 pm
Location: Clearwater, Florida

Post by mawyatt »

If the center point is in "perfect" focus, and the focus falls off away from this center in a bell shape, then the step replicates these bell shaped curves with a periodic repetition of the step size.

When stacked this produces a wavey (peak-valley-peak) focus line with focus ripples following the step trajectory as you step thru and stack the images. Using more steps produces a smoother focus line with less focus ripple.

Anyway, this is my simplistic way of understanding focus banding.

Awhile back I ran across another type of focus banding not directly attributed to step size, this was due to JPEG compression related to colors.

Best,
Research is like a treasure hunt, you don't know where to look or what you'll find!
~Mike

iconoclastica
Posts: 486
Joined: Sat Jun 25, 2016 12:34 pm
Location: Wageningen, Gelderland

Post by iconoclastica »

Lou Jost wrote:Rik's table at http://zerenesystems.com/cms/stacker/do ... romicrodof calculates 9um for NA=0.25, and 6um for NA=30.
That website always has something useful to offer that I hadn't found before. Thanks!

Pau wrote:Moving out of subject, the easiest and best approach would be SEM
True, and probably I even could do it here at close premises. Indeed, most publications on the subject have used SEM. There's three reasons I won't use it though, at least not on a regular base. One is that the SEM imagery defines very well how the structures are built, but I find it hard to translate them to optical images as seen by hand lens or stereo microscope. Second, I want to develop a technique that's available perhaps not to everyone, but at least to a dedicated photographer. Finally that I am on a no-budget project, and the university has become commercialized and won't help me out for free.
I concur with other people: a higher NA objective and smaller steps would provide improved resolution
I just ran a test with 20x/0.40 and it looks promising. I'll try more seriously in the weekend.

Wim
--- felix filicis ---

iconoclastica
Posts: 486
Joined: Sat Jun 25, 2016 12:34 pm
Location: Wageningen, Gelderland

Post by iconoclastica »

I tried a few more 20x stacks today. I must admit: it's a lot more difficult than 10x. First thing it's not a LWD-objective but has a WD of about 4-5mm. That means I have a hard time reflecting the flash light and the shadows tend to get very hard. A curved metal reflector works to some extent, but a second speedlight would be better. That, or a slow light. Also manipulating the object and finding a good spot is rather more than I can manage without an extra xy-stage. And finally the stacks get up to 4x longer, much more processing time but also the stems dry somewhat in that time and move sideways.

I suppose the best I could do rather than adapting the macro setup would be to finish the automated stacker project on my microscope. That will give me both the desired stage and axial lighting. And fixing the stem part to the microscope slide might also solve a problem or two. But for now I'll satisfy myself with 10x and smaller step size.

Yet, I can see where it might go and it looks promising. Thanks everyone!.

Image
--- felix filicis ---

Post Reply Previous topicNext topic