Philips Develops Fluid Lenses

Lars T. writes "Digital Photography Review has a short report indicating: 'Philips Research at the CeBIT exhibition is demonstrating a unique variable-focus lens system that has no mechanical moving parts. Suited to a wide range of optical imaging applications, including digital cameras.' Here is Philips' press release and the Heise News article (in German) where I first heard about it. The latter also mentions that Philips has recently used the same electrowetting effect in an 'ePaper' display prototype."

Comments

Applications to Eyewear

[zalas]

Would it be possible to adapt this type of lens to eyewear by enlarging the size? Instead of using bifocals or trifocals, you might be able to have just one lens that changes shape according to a microcontroller, which is then hooked to either a button, or perhaps tapped into a nerve, which can then be trained to send the appropriate signals.

Re:Applications to Eyewear

[spectrokid]

I guess that if you fill a plastic bag with water you automatically get a lens which has a () shape, as opposed to the )) shape you normally get in glasses. It might work, but you not gone look like the coolest kid on the bloc wearing two fish-bowls on your nose.

Re:Applications to Eyewear

[stuffman64]

My guess is that it could be possible to cancel the effects of gravity (or other acceleration) on the fluid using the same technology. One set of electrodes would shape the lens, the other would 'hold' the fluid in place.

In my opinion, the biggest thing preventing this from being used in eyeglasses is the fact that the lens must always be as thick as the greatest magnification 'setting' on the lens. Also, since there is at least 4 different indexes of refraction (air, glass, fluid 1, fluid 2), there is a much greater likelyhood of chromatic abberation and other artifacts. But who knows what another century or so of research would do for this technology.

Re:Applications to Eyewear

[mikerich]

I guess that if you fill a plastic bag with water you automatically get a lens which has a () shape, as opposed to the )) shape you normally get in glasses. It might work, but you not gone look like the coolest kid on the bloc wearing two fish-bowls on your nose.

About ten years ago a UK charity demonstrated a pair of spectacles for the developing World that used just this principle. The lens were made of two plastic films separated by a small gap. Syringes filled with water (?) could be attached to the arms of the spectacles. The person needing the glasses would put them on, then the syringes would be depressed and water pushed into the gap between the films. When the wearer saw a sharp image, the syringe could be disconnected.

No need for precision lens grinding technology, no need for a trained optician and most of all cheap to make and replace.

Never saw it again, I guess the curse of 'Tomorrow's World' struck this one down.

Best wishes,
Mike.

Re:Applications to Eyewear

[SB9876]

It's not going to be possible to scale these things up much beyond what they've already demonstrated. The reason is that the curvature is solely determined by the interaction of the liquids with the electrodes on the edges. If the lens gets much larger, most of the liquid will be minimally affected. Look at a glass of water, for example: the water has a meniscus edge where it curves upwards to meet the glass. It looks as if most of the water is flat and the edges are curved but this is not true. In actuality, the entire surface is curved but only the water right by the edge of the glass has enough curvature for it to be visible.

If you want the lens to have a uniform curvature so that it is useful as a lens, the overall diameter of the lens must be less that a certain critical length scale which is the length at which the surface tension effects of the edge are still dropping off in a relatively linear fashion. This is determined by the relative surface tensions of the liquids and the electrodes around the edge. While you can tweak those parameters a bit there will be a practical limit to the lens size and I'm guessing that this model is probably pushing that limit already.

Furthermore, the lens is gonna be kinda crappy - it's actually analagous to the lenses in an electron microscope - magnetic lenses that bend electron paths. You can change the current level to change the focus but the lens itself is not shaped properly for good optics - as a result, electron microscopes don't have nearly the resolution they could have on paper. (by about a factor of 100 or so). This new lens has its shape determined by the interplay of surface tension and it's fairly certain that this shape won't make an ideal lens. The result is that you'll have awful sperical aberration. (probably nasty chromatic aberration as well) The result is that the image quality will be poor.

However, that said, these lenses are pretty cool. For point 'n shoot cameras, cellphone cameras and webcameras, these things would rock. Also, security cameras would probably benefit from being able to focus in a reduced total size as well. Any application where fantastic picture quality isn't necessary would benefit from this.

Another potential application is machine vision. There, the machine can easily be programmed to compensate for the lens distortions. Also, if you replaced the ring electrode with a ring of small electrodes around the circumference of the lens, you could set up an uneven charge distribution. This would allow a lens that has a focal point than can be moved off the optic axis. Not only can you focus, you can also bend the light coming through the lens - an electronic equivalent to a gimbal mount. A robot using these eyes can focus and 'look around' with 10 ms speed. This is a potential killer app for these things.

A Glimpse

[novalisg]

Just like the oil lenses in Dune

The lens diagrams are wrong.

[numist]

The lenses in the diagrams at this link do not work: link

The first lens has a positive magnification, and the second a negative magnification. The light rays drawn on the diagrams are reversed.

Others should be able to confirm this.

cool but i wonder

[lingqi]

ever mix a cup of oil and water? now - there is always a surface between the oil and water since they don't mix, but now, shake the darn thing up and a lot of "oil bubbles" will appear in the watery side, and vice versa. I am sure it will not be good for the optical qualities!

I also noticed that their prototype is extremely small - wouldn't a bigger one be subject to gravitational pull / buoyancy (in respect to eachother) of the liquids depending on lens orientation - and therefore causing a distortion to the optical surface?

Artificial Eyes?

[chendo]

Could this be a step for manufacturing artificial eyes? Being able to actually zoom in with my eyes would be cool, and if it has NekkidVision(TM), it would be even better :)

Good for photography...maybe

[mikeophile]

Modern photographic lenses have special coatings to reduce reflections from the air gap between two lenses.

For single lens cameras, no coatings are not that big a problem.

For multiple lens cameras, it can lead to a lot of chromatic aberations.

If these oil lenses can accept liquid optical layers, look out Karl Zeiss.

Stop saying "no moving parts", please

[SuperBanana]

Make a lense without moving parts

Everyone keeps saying this. I looked at the diagram, and at least one part of the lens moves. That's a moving part, folks. Stop saying it "has no moving parts".

Now, here are some predictions:

• They'll still break. Electrodes will corrode. Membranes will rupture. Say they discover after extended operation that the first units put a little too much voltage through the lens or something. You get the idea.
• Materials used, such as the membranes, will age. Either becoming stiff, brittle, or simply change properties enough that the lens doesn't focus the way it was supposed to
• the curve won't be as perfect as everyone is hyping and initial cameras will have excessively blurry images, or images that are blurry in parts but not others due to inconsistencies in high-volume manufacturing of the membranes(think LCD screen "acceptable bad pixel count")
• light loss will be significant. Whereas in the glass optics field we have multicoated lenses that are incredibly efficient, none of those coatings could be applied to the materials on this lens, and furthermore, you've got(for each element) 4 surfaces, not two, for light to pass through.
• Color balance will be odd despite calibration efforts, and will change as the fluid/membranes age(probably from UV exposure).
• It will be useless on anything other than consumer point&shoots. The sensor on a Canon 10D DSLR for example is almost twice the width of that prototype they showed, and uses lenses 2-3x larger still.

Re:Stop saying "no moving parts", please

[Zocalo]

It will be useless on anything other than consumer point&shoots. The sensor on a Canon 10D DSLR for example is almost twice the width of that prototype they showed, and uses lenses 2-3x larger still.

Actually, it's too small for consumer P&S cameras too if the picture is anything to go by, but it might be workable for disposables and video phones though. I'd hate to think what the accuracy of the lens will be in practice though - it's a fluid, so must have some vicosity, which means it's going to move about, which I would assume would impinge on image quality. Of course, there are other uses for lenses other than in cameras and spectacles where this type of lens might do very well.

BTW, your scaling for the EOS 10D is *way* off - the Philips lens is 3mm across according to the press release, most P&S camera sensors are around 8x6mm, although some are as low as 4.5x3.5mm - the 10D's is more like 23x15mm. As to the glass, for a 35mm camera format the glass would typically be in the region of 40mm-100mm in diameter depending on the type of lens and the location of the individual piece of glass in it.

Old news

[sakusha]

Fluidic adaptive lenses have been around for several years. I recall many years ago I read about a tech Academy Award for someone developing a fluidic lens for 70mm movie cameras, it was rather primitive, just a blob of transparent gel sandwiched between two plates of optical glass that could be moved by motors, but he got there first. I can't find a citation since the AMPAS database doesn't search on tech awards.

Why not a spinning liquid lens?

[zakezuke]

A neat and easy way to form a parabola is taking a liquid and spinning it. I've formed parabolic mirrors on my turn table just using ordinary epoxy and spray on silver paint. Not to say this isn't cool, but there seems to be an easier means to achieve a variable focus lens via spinning a clear liquid such as water, or perhaps even a reflective liquid like mercury.

Hermetically closed underwater camera!!!

[pesc]

I am a scuba diver and I always wanted to have a small camera when i dive. The ones you can buy now are quite expensive and fragile. Most models use a underwater housing for a standard digital camera that is quite fragile. There is a rubber O-ring around the enclosure to keep the thing water tight. But sometimes you get some dirt on the rubber and the camera leaks when you are 30m below the surface, spoiling your camera.

So I have always dreamt of a hermetically closed camera. You could fill it with a liquid (oil?) to reduce the pressure stress on the enclosing. (This is what current scuba computers do.) By using a digital camera, you don't have to open the camera to access the film. The problem so far has been how to construct a zoom lens since these vary in volume. This kind of lens seems to fix that problem!

A brave new world of imaging!

[Genda]

This is way cool...

This makes a lot of things possible that would have been prohibitively expensive, mechanically improbable, or optically restrictive. A small lense with fast focusing, which is high quality, shock resistent (this would depend on oil viscosity and lenses size), and remarkably cheap to manufacture in large numbers would revolutionize;

* Robotic vision,
* Consumer electronics,
* Security and Research imaging,
* Medical Imaging, and Lense Replacement.

You could cover a robot with cheap eagle-eye imaging devices, create a central imaging system that sews all the images together to produce an ultra-highres 360 degree whole world views. This machine would literally have eyes in the back of it's head. Give the critter broad spectrum vision, and spectrospopic analysis, and this robot could be used for anything from public safety, to mineral evaluation for mining. If you're going to buy a robot, make sure it has "Phillip's whole world vision(tm)".

This makes disposable highres digital cameras and camcorders totally practical. It makes low end devices possible, products for tens of dollars or less, that have the optical features you would expect to find in products that now cost hundreds of dollars. This is especially true if you combine glass element(s) to the lens. You get the power and optical benefit of a glass front lense, a large optical aperture for light gathering, with simple focus and zooming capabilities provided by liquid lenses. A superior lense with a huge list of advantages. Sign me up!

Now that you have a high quality cameras selling for $10.00, you can put them anywhere and everywhere. Imaging for a whole host of purposes becomes ubiquitous (orders of magnitude more prevalent than today.)

Beside giving medical devices better vision, replacing the lense in the human eye, with one that is for all intents and purposes perfect, would be a godsend to millions of people with cataracts, degenerative lense desease, and missing or injured lenses. In the end, this might become so common place, that when you get to that age where folks noadays begin buying multifocal glasses, our descendents will simply get a super lense implant, and have bionic visual abilities that we can only imagine. Would you trade your eyes in for one's that gave you superwide angle and telescopic capabilities? Oh, and for those folks with astigmatic trouble, one could circle the inside of the lense barrel with panels, and apply differing voltages to the panels so as to create a lense shape consistent with any corneal asymetries. This would be the hot new product among the rich and graying!

Genda

Aberations

[]ix[]

Hmmm. I wonder what kind of aberations one could expect from these? Normal lenses are spherical and expensive special-purpouse lenses are hyperbolical, This lens seems more likely to have a Bessel-shape. It will probably work better than the pinhole cameras in modern cellphones but I wouldnt expect them in any high performance imaging equipment anytime soon.

Aberations aside, its always cool to se new technology emerge in the field of visual optics. The field of optical science is realy realy old and still there are many more things to be discovered.

Re:News?

[JohnPM]

I think the size of the prototype is just right. The real breakthrough application area right now is camera phones. I have a camera on the back of my Sony Ericsson T610 that I've never used. Mobile phone cameras are held back by one massive obstacle: They're currently fixed zoom and fixed focus. Solve that cheaply and a truly massive market is ready made to adopt your solution.

Another fascinating application mentioned in NewScientist's coverage of this stuff is variable zoom security cameras. A security camera could zoom and focus selected portions of the field of view without needing to tilt or swivel the housing. Imagine a kind of moving fisheye effect within the rectangular frame of view. You bolt one of these very cheap cameras to the wall with a very wide field of view and then your operator/software invisibly controls the lens to follow objects closely. Awesome.

Suitability for Precision Use

[perfessor multigeek]

I take this to mean that it is not ready for precision applications and that it may not be.
Well, duh. Think about the vibration concerns. Any movement not only jiggles the thing but it also takes time for it to settle back down.

Back when I designed a version of this w a a a a y back in the early eighties I was quite paranoid about the issue of how do ya keep the thing from accumulating stuff near the resonant frequencies. I'm not seeing anything in the brief English-language piece about this at all. My puppy allowed for the option of changing focal length by changing ring diameter which, oh btw, made things potentially even worse on that front. On the other hand, IIRC, I made a point of the importance of being willing to switch ring materials to optimize for stuff like ability to dampen vibration.
I wonder if they've figured out yet that when you've got a liquid lens that changes properties by changing electrical charge, you can add impurities to the liquid such that charging the liquid, the liquid will change color. Very precise, very neat, and entirely reversable, at least for as many cycles as they would need for a consumer product.

As I mentioned below, I really *am* gonna have to dig up my old drawings and writeup.

*sigh*
This because I have nothing else to do with my time.

Yeah, right.

Rustin

Military Applications

[HullBreach]

There are some uses for these that aren't necessarily obvious. For a long time the US Military has been researching ways to protect troops eyes from blinding lasers (Lasers that scan for relflective surfaces like eye-balls which they then pulse with enough energy to damage). One Idea that has repeatedly come up was to give the troop a sort of "blind cockpit" to operate from. This would be acheived by making a pair of goggles with high res screens on the inside and an exterior studded with an array of small, inexpensive cameras. By feeding the data from the cameras through a processor then to the troops veiw, you seperate his/her eyes from the danger. This would also allow the troop to have a sort of HUD overlay as well as easily integrated nightvision (IR receptive CCD's scattered amoung the visible light CCD's).

Re:One idea for using those

[SB9876]

Yes, the lens can refocus in basically no time but that's not the primary limitation. What presently limits the speed of taking picutres is the time it takes for the processor to determine whether the image is in focus. Although your lens might be able to change focus in 10 ms, it'll still take a second or two of focuss fiddling for the camera to lock down the focus range for either of those focus ranges. If high speed image processing comes along with a similar time scale, yes, this would be practical.

Actually, an idea just occured to me - you could do all the autofocussing first, establishing the various focal lengths and then hammer out the different shots as fast as the CCD can pull down the data. If you did things that way, your idea might actually work. However, its tempered by the fact that these lenses are very low quality, optics-wise. (they may be flawless but they don't have the optimal curvature for a lens and they're pretty small to boot) Any scheme would have to take into account that these lenses are going to be strictly disposable camera and cell-phone camera material.

For the other respondants:

A tripod would be useful for taking two shots with different focus but relys upon the photographer to take both shots which probably involves a significant delay between those shots. (hence the tripod, I assume) However, it is quite likely that objects in the shot will move in that time period, making it impossible to composite the shots. If the camera can be used to grab two sequential shots within a fraction of a second of each other, the need for a tripod is greatly reduced.

Yes, a small aperture will increase depth of field but at the expense of incoming light. A pinhole camera has everything in focus but requires a very brightly lit scene or long exposure times. IIRC, Ansel Adams used this techniques to get his photos and they required him to use obscenely long exposure times.

Where I want such lenses: car headlights

[timothy]

Headlight beams are entirely too coarse; I'd love to be able to dial in a perfect throw, depending on road and conditions. Lenses like this would be one good component of the perfect headlight system. Other parts would be intelligent swiveling mechanisms (left and right as well as up and down) and colored gels (or a chemical layer with a variable color) to best match the day and the driver's vision ... but I digress :)

timothy