Best flash for studio macro?
Moderators: rjlittlefield, ChrisR, Chris S., Pau
Nathan, you've probably seen some posts on this, but just in case, be aware flash can induce violent subject motion with some subjects--a phenomenon that is more likely to be a problem at higher magnification than at low. I consider some subjects (low mass, especially dark colored/low mass) quite impossible to photograph with flash, no matter how brief the flash duration. Even at, say, 1/36,000 second, we are not able outrun quantum mechanics.
--Chris
--Chris
That is so cool that it makes me want to photograph it deliberately!!!
We have a high speed video camera at my company's lab that would likely work...
Anyway, it is good to keep in mind. The objects I am photographing (for now) are weighty enough that I want the short duration light to freeze camera and other motion and are not likely to get pushed around by the photons.
We have a high speed video camera at my company's lab that would likely work...
Anyway, it is good to keep in mind. The objects I am photographing (for now) are weighty enough that I want the short duration light to freeze camera and other motion and are not likely to get pushed around by the photons.
nathanm
Nathan,
If I had access to a sufficiently high-speed video camera, I'd definitely want to film and characterize this phenomenon. Working within the video camera's dynamic range would be a challenge, as one would need to image the subject before, during, and after the flash. A combination of strong continuous backlight with flashed frontlight comes to mind. A subject I'd consider is one of the jet-black scales of the sunset moth (Urania ripheus) (a species more known for its rainbow, mirror-like scales, but which also sports black scales). These scales have 1/2 micron detail that is readily recorded with both backlight and frontlight.
There is a question I'd like to answer: Can the effect be demonstrated in a vacuum? I've done enough testing to be pretty certain that flash-induced subject movement behaves as Sir William Crookes observed in 1873--visible light is absorbed by the subject's atoms, temporarily kicks these atoms' electrons into a higher orbital, and as these electrons instantaneously fall back to their normal lower orbitals, they emit photons down-spectrum from the visible light, in the infrared spectrum; these infrared photons instantaneoulsy warm the air next to the subject, causing the air molecules to move much more rapidly and give the subject a violent kick.
My question: These atoms radiate down-spectrum because, as their electrons return to their normal, lower orbitals, some energy is lost mechanically, as the atom physically "shimmies" in several dimensions. So I wonder--is this atomic shimmying alone able to induce movement in the subject these atoms comprise? Testing in a vacuum might answer this question. And doing so with high-speed video might permit a very cool demonstration.
Cheers,
--Chris
If I had access to a sufficiently high-speed video camera, I'd definitely want to film and characterize this phenomenon. Working within the video camera's dynamic range would be a challenge, as one would need to image the subject before, during, and after the flash. A combination of strong continuous backlight with flashed frontlight comes to mind. A subject I'd consider is one of the jet-black scales of the sunset moth (Urania ripheus) (a species more known for its rainbow, mirror-like scales, but which also sports black scales). These scales have 1/2 micron detail that is readily recorded with both backlight and frontlight.
There is a question I'd like to answer: Can the effect be demonstrated in a vacuum? I've done enough testing to be pretty certain that flash-induced subject movement behaves as Sir William Crookes observed in 1873--visible light is absorbed by the subject's atoms, temporarily kicks these atoms' electrons into a higher orbital, and as these electrons instantaneously fall back to their normal lower orbitals, they emit photons down-spectrum from the visible light, in the infrared spectrum; these infrared photons instantaneoulsy warm the air next to the subject, causing the air molecules to move much more rapidly and give the subject a violent kick.
My question: These atoms radiate down-spectrum because, as their electrons return to their normal, lower orbitals, some energy is lost mechanically, as the atom physically "shimmies" in several dimensions. So I wonder--is this atomic shimmying alone able to induce movement in the subject these atoms comprise? Testing in a vacuum might answer this question. And doing so with high-speed video might permit a very cool demonstration.
Cheers,
--Chris
Photon momentum transfer is a really tiny effect, and quite a bit more complicated than one might think. Here are the results of some measurements of photon-induced displacement in liquid:
http://phys.org/news/2015-06-physicists-pressure.html
Notice that photons can "pull" as well as "push"!
http://phys.org/news/2015-06-physicists-pressure.html
Notice that photons can "pull" as well as "push"!
Lou, I agree that photon momentum transfer is a very tiny effect, but that's not what we're talking about here. The applicable paradigm is quantum mechanics, not relativity.Lou Jost wrote:Photon momentum transfer is a really tiny effect, and quite a bit more complicated than one might think.
Consider a situation I encountered when photographing glass fibers, some of which were coated with black soot, at either 50x or 100x magnification (I don't recall which, offhand). The uncoated fibers reflected most of the light that hit them, appearing white; the coated fibers absorbed most of the light that hit them, appearing black. Both were easy to see under my (continuous) focusing light. But when I photographed a stack with flash, the soot-coated fibers appeared absent. The white fibers were nicely rendered, but the black fibers were simply missing. I found this confusing at first, assumed I'd missed focus, and tried again with the same result. After a period of frustration and head-scratching, I tried a stack with continuous light, and both white and black fibers came out very nicely.
What was happening is that the flash was causing the black fibers to jump around so wildly as to be blurred to photographic insignificance, in the regime of very narrow depth of field that one gets with a high magnification microscope objective.
Had the culprit been the impact of the photons themselves, the white fibers would have been effected just as much as the black fibers, because all fibers were being hit by the same number of photons. Instead, the white fibers--off of which the photons bounced--were uneffected, while the black fibers--which absorbed the photons--gyrated.
My observation is not particularly original: William Crookes demonstrated it in 1873, with his radiometer. (There is, unfortunately, quite a bit of misunderstanding about how a Crooke's radiometer works; for years, the ones sold as novelties were accompanied by a faulty explanation that cited photon momentum transfer, which still turns up on the Internet.)
With my white and black fibers, I tried turning my flash down to its minimum output, which gives about 1/42,000 second flash duration. The black fibers were still invisible due to wild gyration even at this brief output. In fact, I didn't see any notable difference in the amount of gyration regardless of flash setting. I suspect that if one tried this with a 1/1,000,000 second flash, the result would be the same--we can't outrun quantum mechanics.
When the black fibers absorb photons, electrons in some of the fiber's atoms shift to higher orbitals. This shift takes no time. One chemistry professor I know tells her students that electrons are like medieval angels--able to go from one place to another without traveling in between. My way of picturing it is that electrons aren't ever anywhere, really; they just pretend to be somewhere, and when the mood strikes them, they pretend to be somewhere else.
However, electrons are lazy, and want to get back to the lowest possible orbital as soon as they can. Like medieval angels, that means instantaneously. When they drop back, they emit a photon. That photon would be at the same wavelength as the absorbed photon, except that the atom wiggles a bit during this process, which costs a bit of energy, so the emitted photon is down-spectrum. In the case of my black fibers, the emitted photons were infrared--in other words, heat.
These infrared photons are instantly absorbed by the air around the fibers. This makes the atoms in the air vibrate faster, and makes the air expand in volume. This delivers a mechanical kick to the black fibers.
A Crooke's radiometer doesn't work in a vacuum--it needs the mechanical kick of heated air to turn. This is probably the case with my black fibers, as well. But I do wonder. With a sufficiently low-mass object, such as these fibers or a butterfly wing scale, would the atomic wiggle that occurs while the electron drops orbitals be enough to kick the subject visually? Regardless, neither of these motive forces is electron momentum transfer.
Cheers,
--Chris
I doubt that photon momentum transfer is as important as thermal expansion effects. During the short flash you will get asymmetric heating of the thin dark subject. Differential thermal expansion will cause a twitch.
Thermal expansion time scales are very fast for very thin objects.
There are mems devices that work on this principle. Indeed the old fashioned bi-metallic strip thermostats did too.
Here is a nasa paper that seems relevant http://ntrs.nasa.gov/archive/nasa/casi. ... 035876.pdf to the basic idea.
It is potentially complicated by the fact that many insect structures - especially butterfly and moth scales - are photonic crystals that have strongly anisotropic absorption and reflection (as well as wavelength dependence).
The radiometer effect is about differential heating of the boundary layer of (low pressure) air. So it is about thermal expansion of the air, while the effect above is about thermal expansion of the actual structure.
To some degree each of these physical effects could be present. It would require some careful calculation and then likely some experiments to determine which of several optical and thermal effects are most important because there is potentially a lot going on.
Thermal expansion time scales are very fast for very thin objects.
There are mems devices that work on this principle. Indeed the old fashioned bi-metallic strip thermostats did too.
Here is a nasa paper that seems relevant http://ntrs.nasa.gov/archive/nasa/casi. ... 035876.pdf to the basic idea.
It is potentially complicated by the fact that many insect structures - especially butterfly and moth scales - are photonic crystals that have strongly anisotropic absorption and reflection (as well as wavelength dependence).
The radiometer effect is about differential heating of the boundary layer of (low pressure) air. So it is about thermal expansion of the air, while the effect above is about thermal expansion of the actual structure.
To some degree each of these physical effects could be present. It would require some careful calculation and then likely some experiments to determine which of several optical and thermal effects are most important because there is potentially a lot going on.
nathanm
Chris, I tend to think of your second scenario (photon-induced movement in a vaccum) as transfer of photon momentum. If the atom moves in a vacuum, the change in momentum of the atom came from the photon. Momentum is conserved.
Anyway I think that movement would be very hard to detect, but not impossible. It was first detected more than a hundred years ago so it is macroscopically detectable. One device that detects it is called the Nichols radiometer (as opposed to Crookes' radiometer).
Anyway I think that movement would be very hard to detect, but not impossible. It was first detected more than a hundred years ago so it is macroscopically detectable. One device that detects it is called the Nichols radiometer (as opposed to Crookes' radiometer).
Awhile back I had discussed the subject with some colleagues of transient thermal photon induced heating regarding the same material, not different materials as in the bi-metal strips.nathanm wrote:I doubt that photon momentum transfer is as important as thermal expansion effects. During the short flash you will get asymmetric heating of the thin dark subject. Differential thermal expansion will cause a twitch.
Thermal expansion time scales are very fast for very thin objects.
There are mems devices that work on this principle. Indeed the old fashioned bi-metallic strip thermostats did too.
Here is a nasa paper that seems relevant http://ntrs.nasa.gov/archive/nasa/casi. ... 035876.pdf to the basic idea.
It is potentially complicated by the fact that many insect structures - especially butterfly and moth scales - are photonic crystals that have strongly anisotropic absorption and reflection (as well as wavelength dependence).
The radiometer effect is about differential heating of the boundary layer of (low pressure) air. So it is about thermal expansion of the air, while the effect above is about thermal expansion of the actual structure.
To some degree each of these physical effects could be present. It would require some careful calculation and then likely some experiments to determine which of several optical and thermal effects are most important because there is potentially a lot going on.
What happens during a flash exposure is the material surface heats up due to the flash exposure and the surface has a temperature rise. The heat will spread and be controlled by the effective time constant and thermal properties of the material. So you have a scenario where there is a brief period of a temperature gradient near the material surface induced by the flash exposure. This gradient will cause a brief deformation of the material at the surface and may induce a slight bend as the bi-metal strips.
Thanks for posting the NASA paper, I had not seen this.
Best,
Mike
A newsletter from Sky & Telescope desrcibes another case of photon induced vibrations caused by natural flashes:
A striking parallel to the trembling butterfly scales.
Troels
Link to: Sky & Telescope articleA New Take on the Audible Meteor Mystery
By: David Dickinson | March 7, 2017
The Sandia study proposes that strong millisecond-long flashes recorded in bright fireballs are intense enough to induce radiative heating in dielectric materials such as dry leaves, clothing, or even hair in the vicinity of the observer, via what's called the photoacoustic effect. The irradiated surfaces heat the air next to them, producing tiny pressure oscillations — in other words, sound. The study shows that a bolide around –12 in magnitude (about half as bright as a full Moon) can induce an audible sound in dielectric material of around 25 decibels, loud enough to be heard. For context, a whisper is 10 to 20 decibels, on the lower threshold of what is barely audible. The study even suggests frizzy hair (!) might be an even more effective transducer of the photoacoustic effect.
"It seems significant that people with frizzy hair are reported to be more likely to hear concurrent sound from meteors," the study notes. "Intuitively, frizzy hair should be a good transducer for two reasons. Hair near the ears will create localized sound pressure, so it is likely to be heard. Also, hair has a large surface-to-volume ratio, which maximizes sound creation.
A striking parallel to the trembling butterfly scales.
Troels
Troels Holm, biologist (retired), environmentalist, amateur photographer.
Visit my Flickr albums
Visit my Flickr albums
Back to the main topic of the my original question, here is what I have been using.
I have Profoto D2 monolights for large scale macro - I got them mainly because the offer high speed flash (1/63000 sec), but they work for fine for thing at 1:1 and lower. I have some Einstein flashes that also have a high speed mode but the Profoto is faster.
For high mag macro I use the Canon MT 24, even though I am not using Canon any more.
For my transmission light microscope (Nikon E800) I am using a Quantum Qflash with a set of condensers I built to give it Koehler illumination with flash.
Nathan
I have Profoto D2 monolights for large scale macro - I got them mainly because the offer high speed flash (1/63000 sec), but they work for fine for thing at 1:1 and lower. I have some Einstein flashes that also have a high speed mode but the Profoto is faster.
For high mag macro I use the Canon MT 24, even though I am not using Canon any more.
For my transmission light microscope (Nikon E800) I am using a Quantum Qflash with a set of condensers I built to give it Koehler illumination with flash.
Nathan
nathanm
Sorry for the delay - I had some deadlines to meet.
The issue with using flash with a transmitted light microscope is how you achieve Koehler illumination, which requires collimated light.
Most microscope illuminators use a tungsten source, and rely on the filament to be a quasi-point source.
Charlie Krebs solved the problem by disassembling a Vivitar 283 flash so he could take the flash tube exactly where the tungsten lamp was, and relied on the fact that the flash tube is pretty small.
http://www.krebsmicro.com/microsetup2/index.html
You still need continuous light for focusing and finding the right image. Charlie solved this by putting a white LED behind the flash tube. This possibly sacrifices some of the point-like light characteristics for continuous light, and puts flash first priority wise.
This web page (perhaps a forum member?) solves the problem differently
http://www.micrographia.com/articlz/art ... pc0100.htm
In this case the flash is run through a set of two identical aspheric condenser lenses. The first of these expands a point on the flash to a beam, the second refocuses it into a point. This is put behind the normal tungsten lamp and arranged so that the focal point of the 2nd condenser hits at or near the tungsten lamp filament.
The issue with this is that the flash has to try to focus through the lamp, which likely distorts things a bit. In a sense this is the opposite of Charlie's approach - Charlie gave the flash the best position, and hoped for the best with the LED shining through the flash tube. The Micrographica site leaves the tungsten lamp in the best position and hopes for the best shining a flash through the lamp.
My approach benefitted from both of these.
The Nikon E800 lamphouse has condensers in it and it outputs a collimated beam, so I found that if you just detach it and move it back 8 or 9 inches it still works just fine. Also, that means that instead of mimicking a point source with my flash I am mimicking an even, collimated beam.
I use three aspheric condensers from Thorlabs. Two are positioned curved-to-curved like Micrographica, but in my case there is a ground glass diffuser (also from thorlabs) in between them. This is a typical technique for evening out the light. It is possible that I didn't need to do this but I went ahead just in case.
The output of these two condensers focuses the flash to a point - as with Micrographica, but I don't want a point because I am trying to be like the entire Nikon lamphouse, not the lamp.
The third condenser expands that point into a collimated beam.
This only leaves the problem of how to switch between the tungsten lamp and the flash. I solved that with a diagonal beamsplitter (also from Thorlabs). This cuts my illumination by a factor of 2, but that has not been a problem - the flash has plenty of power, and so does the normal Nikon lamp.
What this means is that I don't have to decide whether flash has priority, or the tungsten lamp has priority - both are equally in the optical train. I originally wanted to have a mirror that would flip in or out of position, but the beamsplitter has worked so well that
For the flash itself, I originally took apart an older Canon 580 flash, but I eventually burned it out. I suspect that my stacks were a bit too close together
My new approach is to use a Quantum Qflash. It has a bare bulb which is easy to position in the right place, and it has an AC power adapter.
Here are some pictures - overall view
Detail of the optical train. You can see the flash bulb, the condensers and the beamsplitter at the bottom.
The optics are assembled with a Thorlabs 60mm cage system.
[/url]
The issue with using flash with a transmitted light microscope is how you achieve Koehler illumination, which requires collimated light.
Most microscope illuminators use a tungsten source, and rely on the filament to be a quasi-point source.
Charlie Krebs solved the problem by disassembling a Vivitar 283 flash so he could take the flash tube exactly where the tungsten lamp was, and relied on the fact that the flash tube is pretty small.
http://www.krebsmicro.com/microsetup2/index.html
You still need continuous light for focusing and finding the right image. Charlie solved this by putting a white LED behind the flash tube. This possibly sacrifices some of the point-like light characteristics for continuous light, and puts flash first priority wise.
This web page (perhaps a forum member?) solves the problem differently
http://www.micrographia.com/articlz/art ... pc0100.htm
In this case the flash is run through a set of two identical aspheric condenser lenses. The first of these expands a point on the flash to a beam, the second refocuses it into a point. This is put behind the normal tungsten lamp and arranged so that the focal point of the 2nd condenser hits at or near the tungsten lamp filament.
The issue with this is that the flash has to try to focus through the lamp, which likely distorts things a bit. In a sense this is the opposite of Charlie's approach - Charlie gave the flash the best position, and hoped for the best with the LED shining through the flash tube. The Micrographica site leaves the tungsten lamp in the best position and hopes for the best shining a flash through the lamp.
My approach benefitted from both of these.
The Nikon E800 lamphouse has condensers in it and it outputs a collimated beam, so I found that if you just detach it and move it back 8 or 9 inches it still works just fine. Also, that means that instead of mimicking a point source with my flash I am mimicking an even, collimated beam.
I use three aspheric condensers from Thorlabs. Two are positioned curved-to-curved like Micrographica, but in my case there is a ground glass diffuser (also from thorlabs) in between them. This is a typical technique for evening out the light. It is possible that I didn't need to do this but I went ahead just in case.
The output of these two condensers focuses the flash to a point - as with Micrographica, but I don't want a point because I am trying to be like the entire Nikon lamphouse, not the lamp.
The third condenser expands that point into a collimated beam.
This only leaves the problem of how to switch between the tungsten lamp and the flash. I solved that with a diagonal beamsplitter (also from Thorlabs). This cuts my illumination by a factor of 2, but that has not been a problem - the flash has plenty of power, and so does the normal Nikon lamp.
What this means is that I don't have to decide whether flash has priority, or the tungsten lamp has priority - both are equally in the optical train. I originally wanted to have a mirror that would flip in or out of position, but the beamsplitter has worked so well that
For the flash itself, I originally took apart an older Canon 580 flash, but I eventually burned it out. I suspect that my stacks were a bit too close together
My new approach is to use a Quantum Qflash. It has a bare bulb which is easy to position in the right place, and it has an AC power adapter.
Here are some pictures - overall view
Detail of the optical train. You can see the flash bulb, the condensers and the beamsplitter at the bottom.
The optics are assembled with a Thorlabs 60mm cage system.
[/url]
nathanm