Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Eye of Ellipses

  1. Apr 12, 2008 #1
    Photographing Magnetic Lines of Constant Scalar Potential​


    This post is intended as an introduction to my paper, Photonic Dipole Contours. I have been asked to address some deficiencies in my paper such as background information and prior work. Please excuse my informal first person writing style; it is my most practiced form of writing.

    In a nut shell, my work can be easily characterized as we took a bunch of lab photographs and found a pattern in every picture, and then found equations that produce the same pattern. I think this is the most pure form of physics, find something in nature, find the mathematics for it, and report it.

    Let’s start with the visualization instrument. All of our photographs use a Ferrofluid Hele-Shaw as the main subject. I think the first use of such an instrument was by D. E. Rosensweig in Ferrohydrodynamics (ISBN 0486678342). It is my understanding that Dr. Rosensweig used a Hele-Shaw cell to investigate the microscopic physical proprieties of the ferrofluids.

    Hele-Shaw cells themselves are mostly used in chemistry and fluid dynamics. Most readers probably don’t know much about them. My understanding is that their main use is to study fluid viscosity and density gradients. A Hele-Shaw cell consists of two flat plates that are parallel to each other and separated by a small distance. At least one of the plates is transparent.

    The instrument I used was built by PF member pinestone. He has a commercial business building these cells, he would glad to answer questions about the cells.

    I would point out that my photographs use a completely different mindset than Dr. Rosensweig. The Hele-Shaw cell and external magnetic fields were well known tools he used to study unknown ferrofluid properties. Years later, my approach is the reverse. The ferrofluid is a well known subject, but the light paths seen in cell have not been studied. My work brings the instrument and the external magnetic fields into question.

    Why does light curve when passing orthogonally through the cell? What mathematical pattern does it take? Those two questions are the foundation of my paper.

    Below is a picture of the cells used in the paper. These were made by hand and are quite large. The one on the left is 150mm diameter and the one on the right is 114mm diameter. The one on the right has a 38mm hole missing in the center. Notice the frames have radial holes for orthogonal light injection. In practice the magnets are placed right next to the glass for the large cell, and with the smaller cell, right in the middle of the donut. In the photos, there is a piece of black paper around the magnet in the center of the cell. The poles and light injection are in the xy plane, and the camera in the z plane.


    Below is a good picture showing the line pattern we solved. The magnetic poles are located at where the lines cross in the photographs, and we believe related to Rosensweig peaks seen in many ferrofluid experiments.

    In both the green photo above and the red photo below, the non-crossing lines is what we have studied. Notice each line starts at an edge radial light source and curves into the center of the cell where the magnet is located. The poles in this case are roughly at the 1’oclock and 7’oclock locations, nearly vertical. The lines we solved for are the curving horizontal ones.


    The curving lines are a bit problematic. Based on Snell’s law the speed of light must be changing in the medium related to the location of the external field. In other words the complex impedance of the ferrofluid medium must be changing proportional to the locations of the magnetic poles. This line of thinking is not that strange because we know that Helmholtz’s equations do have a magnetic component for plane wave propagation in a dielectric. In this case, the ferrofluid medium is a light mineral oil.

    But what are we seeing? It is clear that xy plane orthogonal light must be scattering off the particles in order to reach the camera on the z-axis. Maybe there are long chains of particles and the light is just showing us chains of particles.

    Well, in the photograph below, we have yellow and green light injection side by side. We were using fiber optics and our light source was misaligned. This gave us random red/yellow/green light sources at the edges. Notice in the photo below, around the 10 o’clock location there is a green line parallel to yellow line.

    I believe having two different color lines parallel to each other clearly proves the light is curving in the medium based on dipole field vectors, and not just scattering off particle chains.


    And finally, what pattern are the curve lines following? We found three different sets of equations that will plot the same lines within the margin of error of our experiments. The field dipole equation was chosen because we were dealing with magnetism.

    \text{V}_\text{m} = \frac{1}
    {{\sqrt {(x - x_{p1} )^2 + (y - y_{p1} )^2 + (z - z_{p1} )^2 } ^{} }} - \frac{1}
    {{\sqrt {(x - x_{p2} )^2 + (y - y_{p2} )^2 + (z - z_{p2} )^2 } ^{} }}\text{ }

    Were [tex]
    [/tex] is a constant solved for each of the contour lines.

    We also found the well known fluid dynamics doublet equation would give very similar contour lines.

    The first solution we found is not directly mentioned in the paper. If you connect the non-crossing curved lines into sets of elongated circles, you will find the same four foci points can be used to express all the lines into families of ellipses. This is where the working name came from, the round donut shape and field contour lines made the instrument look like an eye, and the lines can be completed into ellipses.

    Hence for two years, the paper’s working title was “Eye of Ellipses”.​


    For more information that you ever wanted to know about my work, you can download and view the pictures and movies at http://www.sendspace.com/folder/on9dhg
    Last edited by a moderator: Apr 12, 2008
  2. jcsd
  3. May 7, 2008 #2
    Particle focusing

    Look at your photos closely and you will see several lines crossing near the pole regions. The ellipse model does not explain this. I do think this will happen if you assume the particles form long chains close to each other. Diffraction of the light around a chain can lead to interference from light from nearby gaps. If the chains curve and follow the magnetic field, the diffraction pattern will also curve.

    It is straight forward to compute the light scattering from a chain of small particles. The magnetic field strength required to change light is very high, but the ability of a dielectric to change light direction is very easy. From a pure energy perspective, it makes more sense to view the microscopic ferro particles as dielectrics which cause the light to bend around them than to view the magnetic field as changing the light direction.

    What happens when you move the camera focal plane? Do the lines move, or stay the same?
    What happens when you change the angle of the incoming light? Is there a point where no light passes? Is it reflected at that point? How do the lines change as the magnetic field changes?

    A good model has to answer a lot of questions. Good luck, and enjoy the challenge!
  4. May 12, 2008 #3
    Dear Dr. Mike

    I grant that the crossing lines at the pole regions are a bit problematic. By definition traditional field vectors of any type are not allowed to cross. Yet, in all of our photographs they do. I have a example looped gif file of twelve photos, with the magnet rotated 30 degrees per photo.


    I would be willing to concede that the crossing lines are a diffraction affect, yet I can’t reconcile a diffraction affect with the seen magnetic voltage lines also in the photographs. My belief is that crossing lines are related to Rosensweig peaks, but this yet to be proven in our experiments.

    I have no filmed experiments of changing the focal plane, but know from experience when looking at the cell at different angles one sees different affects. All our pictures so far have been on axis.

    In fact, one of our upcoming experiments is to use two cameras for stereographic photographs and then try to use Matlab to produce a 3D vector model.

    When you look at your TV you see the same thing at different angles. These cells have a 2D visual interface like a TV, but one sees different things at different angles. By reconciling the mathematical differences in two concurrent photographs at different angles; I think I can produce 3D model from a 2D interface.

    I do have a movie (large file size) of the camera and cell in fixed axis locations and only moving the magnet. It shows the contour lines are consistent for any visual plane. Move it, rotate it, the contour lines move with it. :cool:


    The angle of incoming light has a very large effect. We only analyzed the orthogonal light sources in our paper. The non-orthogonal light sources produce some great photos but we have not been able to produce a model for the effects.

    Here are two photos with the light source 30 degree off axis to the right of the lens.




    By the way, I am reshooting the three-color experiment this week. By using three alternating colors, red, yellow, green I should be able to directly isolate the diffraction affects.

    Thanks for your post,

    Last edited: May 12, 2008
  5. May 12, 2008 #4
    I got the huge file onto my flash drive, my computer at work refuses to play it! At least I have a lot of bandwidth to get it with.

    It will be very interesting to see how the 2D -> 3D mapping works. Tomography works this way. Use lots of 2D pictures to figure out the 3D positioning. Since the magnetohydrofluid is thin, you should be able to make several approximations that will make your life easier.

    I don't know how easy this would be to do, but if you could "subtract off" the background, matching theory to the data would be a lot easier. So for example, in the above two pictures, take a photo with no light source and one with the light source, then subtract pixel by pixel the first photo from the second. That would eliminate all the background stuff, and it will be a lot easier to explain what is going on in the images. Alignment will be tricky, but it will help to see the effects more clearly.

    I'm definitely biased towards diffraction :) but I look forward to your further posts! I'll try to get to that movie this week - I'm sure my kids computer can play it.
  6. May 13, 2008 #5
    Very nice movie! I'm partial to the sound track - reminds me of "Space Odyssey"

    Note that because the ferrofluid can rapidly change, diffraction can easily explain what is observed. The light source and shadows play an important role too. But this is the point
    of scientific inquiry - to whittle down the variables and understand how they all play a role.

    To make your point strongly, I suggest adding an arrow for north-south orientation of the magnet along with subtraction of the background. You have to explain how this is "data clarification" not "data manipulation" very clearly, but I think the result will be a strong argument for what the facts are.
  7. May 17, 2008 #6

    Dear Dr. Mike,

    Well, I did the new experiment with the three color leds this week. I had believed I would see bands of different color lights curving through the lens, hence tending to disprove the diffraction theory of operation. :cool:

    The results do have paths of different color lights, but the problem is some of the colors do not match the leds! This lens had 45 blue leds around it in series, and had been in service for over a year. I removed the blue leds and replaced them with a red, yellow, and green leds. The colors just repeat 15 times around the perimeter of the lines. Turns out the red leds are the most efficient and flood the lens with red light but it isn’t too much of a problem.

    Also seen in the background, is a ferrofluid precipitate residue on the inside of the lens. This has happened before, and it is do to long magnetic field exposure but doesn’t seem to affect the normal cell operation. I was doing some time lapse videos of over three hours duration, earlier this week, and sometimes a residue will form. Given a few weeks it will dissipate back into the solution.

    In this case, it turns out the thin layer of residue is a much better scatterer of photons for the camera on the z axis. The normal fluid does this too, but the inside residue seems better at the task. Remember that all the light sources and are orthogonal to the camera.


    Notice in the first photograph that colors warp around the magnet, with its pole facing the camera. Lines of yellow and green are clearly visible originating from the edge of the cell. This photo would tend to strongly support the particle based bending of light around the magnet.


    Now in the second photo notice that the same magnet with the same pole facing the camera, now has symmetrical lines around it, but the lines do not match any of the colors of the cell. I have managed to take red, yellow, and green light sources and made lines of nearly white light.

    When I think of diffraction; I think of a diffraction grating breaking white light into different colors. In this case we have different colors at the edges forming into lines of white light in the center.

    This seems like the inverse case of diffraction. Refraction?

    And just to keep things interesting, you get complex cases like photograph three. Photo three is eight magnets with all the same poles facing the camera. As a bonus feature, the lines are dynamic. Even if the magnets and cell are stationary, in the movies the lines tend to drift a small amount among themselves.


    So we have a large amount of new data, but no better conclusions so far about the functionality of the cell.

    How does the photons get from the radial sources around the edges, and then show up in order to travel to the camera?

    I think it is variable refraction based on the alignment of the particles.

    I have two new movies. The first one is like photo #1; a basic one magnet version that shows good views from all six sides of a cube magnet.


    The second one is like photo #3; with eight magnets. It is mind boggling complex. :biggrin:


    What do you think?
  8. May 17, 2008 #7
    Diffraction means the light wave length is comprable to the size of the object it is bouncing off of. Refraction means the object is much much larger than a wave length, and ray tracing can be used. I think diffraction applies because the wave length of light is in the 500 to 700 nm range, but the particle size is in the 20 to 50 nm size. A whole lot of particles lined up can block the light in one direction, but let it pass through the holes between lines of particles. If the gap size is on the order of 200 nm, and the opacity of the ferro particles is high, then the gaps are the only way for the light to pass. The magnetic field strength in light is 1/c the electric field strength, so it would take a very large magnetic field to bend the light.

    The other option is reflection. The light is coming from the side, hits a line of particles and gets bounced off. As it passes through the ferro particles, it gets diffracted and certain directions have waves add, others have waves subtract. As the particles move, the pattern moves.

    A microscopic picture of the particles will be hard, but it would be interesting to compare the light patterns from different densities of particles in the fluid. So for the same light source, same magnet, and different fluid properties you might be able to see where the patterns stop appearing for both low and high density situations. If the light is being bent by field, high density should bend things more. If it is diffraction, high density should totally block any light getting through at all. Not sure how hard it would be to do though.
  9. May 17, 2008 #8
    Here's a microscopic image of a segment of the magnetite particle chains:
    http://www.nanomagnetics.us/Lens 1 Three minis 11-27-07 007.jpg

    The illumination source is a 20 Watt incandescent lamp below the microscope stage. Notice the "superprism effect" breaking the white light into different colors.

    Ref: http://www3.interscience.wiley.com/journal/116835748/abstract?CRETRY=1&SRETRY=0
  10. May 18, 2008 #9
    Which brings up a good point - the different colors makes sense from a diffraction point of view, but if it was magnetic field changing, the frequency of light would not matter. All wave lengths should be bent equally if the magnetic field strength was causing it.

    A fixed wave length laser might help too - then only a single wave length enters the problem.
  11. May 19, 2008 #10
    Actually, laser is bent equally. Look at these three images.
    The first shows the laser beam passing through the lens on to a screen without a magnetic field. The second image shows the screen with the application of a horizontally applied field and the third is with the field vertically polarized.

    1. http://www.nanomagnetics.us/laser-lens-no magnet 0002.JPG
    2. http://www.nanomagnetics.us/laser-magnet-lens horizontal 0001.JPG
    3. http://www.nanomagnetics.us/laser-magnet-lens vertical 0003.JPG

    What isn't apparent in these images is that the beam is deflected 90 degrees in opposite directions from the center. That's a 180 degree deflection of light from an applied magnetic field.
    Bear in mind, the beam is traveling through a .25 mm hole in an opaque mask covering the lens surface, and manages to get deflected completely perpendicular to the incoming plane.

    I see the link in my previous post referencing the superprism effect is no longer working.
    Here's another link: http://spie.org/x14414.xml?pf=true&highlight=x2416
    Last edited: May 19, 2008
  12. May 20, 2008 #11
    I don't know if all the wavelengths are bend evenly, but going by the photo #1 that was three posts back, they sure are bent!

    Here is a streaming video link for those who do not wish to download the 50 meg movie. :uhh:

    I love the pole pictures, like photo #2, there is a geometric shape if I have ever seen one.

    Anyone else have some questions or comments? I believe I heard the forum has over 80,000 members.

    Surely, someone has something to say. o:)
    Last edited by a moderator: Sep 25, 2014
  13. May 20, 2008 #12
    As of today @ 20:15 UT- Members: 98,116

    So there are 98,113 others available for comment :>
  14. May 27, 2008 #13
    I have two new pictures taken late last week and processed over the weekend. I think we can rule out normal diffraction or refraction, thought maybe it is a special case of them.

    The first picture is profile view of two magnets, with a north-south at the top, and a south-north at the bottom.

    To say the connecting lines twist would be an understatement. :cool:


    The second picture is the same profile view, but now it is north-north at the top, and south-south at the bottom. Notice how the parallel lines make it all the way across. You can even follow certain colors as they move across the cell.


    Because we can track paths of LED sourced colored light across the cell, I believe it can not be a diffraction affect.

    I think it is a magnetically aligned particle based nano-wave guide characteristic of the cell.
  15. Jun 18, 2008 #14


    User Avatar
    Science Advisor
    Homework Helper

    Hmm. 45 blue LEDs is going to require something like 90+ volts to light up, so I'm guessing the lens runs off of 120 VAC to a recitifier and then a current source?

    If this is not the case and the blue lens runs off of low voltage, then different LEDs require different voltages, with the higher energy (blue) light naturally requiring higher voltages. Red requires the least voltage. So when you put LEDs in parallel, the red will steal all the current and put out all the light. But I also seem to recall that they red tends to be the most efficient in terms of photons per amp.
  16. Jun 18, 2008 #15
    Hi CarlB,

    You nailed it. I use a series configuration to keep the current low as possible around the cell. A regulated current source of about 20ma puts the voltage around 100 volts. I used to use a full wave rectifier and large capacitor with current limiting resistors to supply the voltage, but a regulated lab power supply is so much better. I just dial in the milliamps and start shooting video now days.

    The red LEDS are more efficient, but since the camera is RGB based, I lower the red channel in post processing.

    Have you seen my latest movie? http://www.scivee.tv/node/6174

    I’m getting pretty good at processing the videos. Two years ago I remember spending weeks trying to get one good photo from a series of low light photos taken during an experiment. :rolleyes:

    Now my biggest problems are getting the right codecs to work. It takes a beefy computer to handle the video encoding; no one wants to download a 29 gig raw avi file <smile>

    I started with the Indeo 5.1 codec but some machines couldn’t read it, then I switched to the Microsoft Codec V3 which worked but the Mac users could not watch the movies. I finally switched to the H264 codec which is the newest standard (satellite TV, Blue Ray, etc) and figured people could download the ffdshow decoder.

    I am more than happy to answer questions about my research. o:)

  17. Jun 18, 2008 #16


    User Avatar
    Science Advisor
    Homework Helper

    A wonderful addition for me would be to label the N and S ends of the magnets. What is confusing is that there are two sets of lines, the field lines and the magnetic gradient or strength contour lines. To have both of them appear in the same image, so similarly, is disconcerting.

    Maybe another cool thing to do is to add a polarizing filter (or two) to all this, so you can split the LED light into horizontal and vertically polarized. Or maybe a circularly polarizing filter.
  18. Jun 18, 2008 #17
    In the three magnet video, the poles are marked with + signs. The sides are marked with single lines.

    Well, I could be wrong, but the images do not have field lines in them. I admit with more than two magnets it gets very confusing. Might want to watch the one and two magnet videos first.

    My rule of thumb is to pick any two voids (enclosed areas lacking photons) and track lines between them.

    If the lines are perpendicular between two voids then you have a NS or SN interface. If the lines twist and connect the two voids then it is a SS or NN interface.

    It even gets weirder when you realize we are watching dynamic photon paths. It is like having a school of fish and you can only see where some of them are, when they jump out of the water.

    Most of the photons are traveling unseen in the layer of ferrofluid. The only photons we are seeing, are the ones that leave the fluid and make it to the camera.

    The proof we have that these are the paths the different schools are taking is that we have difference colored fish, (different color leds), so wherever a certain color photon leaves the fluid, it must have followed a similar path as the other photons of the same color that left the fluid before them. :smile:

    BTW, I have tried polarizing filters without affect. There isn’t any coherent light to polarize coming from the leds. The filters just attenuate the light levels.
  19. Apr 12, 2009 #18
    Re: Eye of Ellipses - Yearly Update

    Well, a year of graduate school later, I thought I would post an update to my thread.

    I have had about thirty different professors examine my experiment. A rough consensus has been that the images are a result of molecular optics. The exact mechanism is still poorly understood.


    Paraphrasing their comments, ‘A simple way of explaining this work is to think of the old magnetic iron filling experiments; but now each iron filling is a molecule floating in a liquid. Each molecule has the freedom to act as an independent lens that can be aligned by an external magnetic field. Because each lens directs light, the external field can be analyzed by following the light paths captured in the photographs.’

    Also there has been strong evidence that the light separates into left-handed and right-handed pathways inside the cell, producing different paths; polarizing-helicities does seem to play a large role in the images.

    Notice that some of the light paths bend to the left around a pole, and other paths bend to the right around the same pole.

    I have updated my posted paper from last year.


    On the other hand, the research continues. Rocking On. :smile:
  20. Dec 24, 2009 #19
    Here's my yearly update of my research. o:)

    It's my second year in grad school and am getting some good datasets. :approve:


    Here's a link for a 7 meg jpg poster, it's good.


    Here's a link for a youtube slide show, it's not too bad.

    Here's a link for my paper that is a bit subpar; but is being rewritten...



    Y'all Have a Merry Christmas :smile:
    Last edited by a moderator: Sep 25, 2014
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?

Similar Discussions: Eye of Ellipses
  1. Polarization Ellipse (Replies: 1)

  2. Eyes FPS (Replies: 6)

  3. Eye optics (Replies: 6)

  4. Bionic eye (Replies: 1)