Birefringence and Automotive Window Tinting

[RedneckPhysics]

Hi all, I'm new to the forums (in posting at least!). I recently graduated with a B.S. in Applied Physics, and have viewed topics here on and off throughout my educational experience. I definitely love this forum!

I've been fortunate enough to have the opportunity to buy a new car, and recently had the windows tinted. The tint job itself is another story altogether... not the best quality tinting material and workmanship in the world. Now, I know some of you have probably seen this effect in person before: when wearing polarized glasses, the tinted windows seem to impart coloration to scenery viewed through the window.

I'm a bit weak in the optics discipline (I focused on E&M, Engineering Physics and Laboratory Electronics and Automation), but I have had some radiometry, and have extensively studied the causes of induced birefringence via the Faraday Effect. I was hoping to run the issue by everyone here on the forums, to see if anyone knew more about what exactly is happening with the effect.

So, first, some background on the problem:

When I first got the car, I immediately started using a pair of dark amber (almost a bronze) tinted polarized glasses for daytime driving. I could immediately tell that the front side windows were nicely tempered, as they distinctly showed a pattern of darkened areas. The pattern resembled a matrix of 2cm to 3cm circles, with no obvious hue at all (just a black "coloration", causing a darkening of the color transmitting through the window, through the glasses, and to my eyes). The windows are PGW (formerly PPG) Solargreen glass. I found a super handy transmittance and reflectance chart on their website:

Like many manufacturers, they didn't provide hard data points, but we shouldn't need them to solve this issue. Look at that sharp UV cutoff! Also, the curve peaks nicely around the 500's, which I assume is great for driver visibility from the photometry/safety standpoint. Still, I'm not at all sure how the glass achieves these optical characteristics. Perhaps a patent search is in store?

After tinting, the film still had a good amount of water (or perhaps water and soap... whatever they use these days) trapped between the film-window interface. I put the glasses on for a drive home, and lo' and behold, it looked like I was swept back to the 60's. Pinching myself a few times, and making sure the new car scent wasn't actually something else, I noticed that the film was definitely the culprit of a newfound "rainbow road" effect. It really looked like I was driving a car protected by soap bubbles.

After some time in the sun, the tints apparently absorb enough radiation to self heat and smooth out... a process many refer to as "curing". So, the rainbow gradually went away day by day... but plateaued at the level it is now after about 4 sunny days outside. The color shift effect is there, but it only gets noticeable when viewing a reflection off of a smooth surface (more on that in a minute).

One thing that I did notice is that the effect initially displayed the full spectrum of colors, but now the "rainbow tint" seems to be dichromatic, with only varying intensities of magenta and cyan being noticeable. Here are two pictures, taken through a photographic quality linear polarizer, rotated 90 degrees after the first shot:

My sunglasses also showed a very similar colored effect as the polarizing filter, so I won't show a comparison shot with the glasses (although, as I should know, our eyes are very selective instruments, and the digital camera may in fact reveal that there is actually more than meets the eye (pun totally intended))

Naturally, being the curious kid that I am, I ended up wanting to get to the cause behind the effect itself. Like many topics in physics, the more I thought about it and the more I researched, the more uncertain I became about what's really happening with the windows. Clearly, this phenomenon involves polarization at several points in the optical chain. The real question is, what exactly is occurring at each point along the way?

Given that the formulation from first principles was going to take a while, and the fact that I go back to have the windows re-tinted in a few areas on Monday, I wanted to get to the bottom of this a bit quicker. It is still at the point of being annoying when driving around, so it will either be new tints, or new glasses if it continues. So, I decided to do a few "cheap and dirty" experiments (I mean after all, I did just call myself "redneckphysics", right?) in order to deductively eliminate certain theories.

First, I repeated the above two shots with a "party-cloudy sky" background, for comparison:

This first shot shows a similar pattern that I saw with the untinted window darkening, of "holes" and/or dots, with the now magenta coloration.

I don't have the pictures for some reason (will try to re-create tomorrow), but I then took pictures near the top edge of the window. This revealed a more consistent pattern (perhaps due to the stress, or perhaps due to the tints, it's unclear which), and the colors flip-flopped, with magenta and cyan hues switching places upon a 90 degree rotation.

Then, I decided to make a simple polarizer network with the photographic filter and my sunglasses. As I'm sure may of you have seen demonstrated before, linear polarizers arranged with their direction of polarization will cause an almost complete extinction of transmittance, due to the selective sequential filtration process. I laid the filter on top of an led flashlight, then took pictures through the sunglasses positioned at around 0 or 180 degrees separation between polarizer directions (the two are basically indistinguishable), and at 90 or 270 degrees separation. The glasses were actually nicely polarized, as you can see here in the following shots:

(And yes, that did cover up the writing underneath!)

Now we can analyze the window, soapy film and tint for their effects on the overall polarization of transmitted light in the system. Photographing the light passing through the photographic polarizer first, then window soapy film and tint, and then finally the glasses, gave some interesting results!

First, I rotated the photographic filter to what appeared to be the extinction point with respect to the sunglasses, and placed everything into position, so that half of the photo filter transmits light above the window pane, and half of the photo filter is aligned to transmit through the tinted glass. Note that there is a very small strip of untinted glass still visible as well:

First off, there are a lot more colors present than simply magenta and cyan! Note that the un tinted strip does appear to be monochromatic, but also shows transmitted light that was not blocked by the sunglasses. Does this indicate that the glass does indeed cause optical rotation?

Then, I tried to rotate the filter as close to 90 degrees away from the initial point as possible:

Note how the coloration again seems to shift in a complementary fashion! Interesting effect!

Just for good measure, have another set of photos:

And the second (albeit a bit under exposed!):

My initial theory is that the glass itself is causing varying degrees of optical activity (some polarization and possibly optical rotation of polarization through stress-induced birefringence). I do recall that rotating the sunglasses did cause a change in the observed pattern on the window, so I can say that the glass itself is somehow selectively polarizing. It's perplexing, because as we know, optical rotation in many optically active materials shows a spectral dependence (or rather, the angle of rotation increases as wavelength decreases, due to the nature of refraction, including the speed of light in various media, etc.).

Based on the very first two shots I showed you all, I would say that perhaps the tint is causing an additive color bias by shifting the overall transmitted hue to blue (or reducing the visible red and green). Then I think the soap film or the tint is somehow dichroic, causing red and green wavelengths to be absorbed at different angles of orientation. Perhaps those different angles of rotation just so happen to be perpendicular to one another? Thus, you would get the magenta and cyan switching places. However, this sounds like an incomplete explanation of the phenomenon. The tint film is likely of low quality (but still pro grade), and thus uses a dye (or several) to filter light. I do find it interesting that polarized rays cause the effect to increase (that light reflecting off of the satin concrete finish and car body is likely linearly polarized due to the reflection).

Something that keeps bugging me is the change in effect intensity and diminishing colors with the evaporation of the water/soap mixture over time. Perhaps there is some thin film interference involved as well? If the film left behind is indeed causing some constructive and deconstructive interference, how would this appear? I wonder if some of the coloration apparent when viewing the glass edge through the polarizer network is caused by the variation of thickness in the underlying soap film, vs. stresses in the window or tint alone? Is birefringence the only cause, or is it even a factor at all?

So, I would love to hear your thoughts on this! Does the initial theory sound correct? Again, it seems like I'm not tying everything together, but that's likely due to my limited knowledge of optics. Anything else that you can think of to test things further? I imagine a filter network test repositioned amongst the matrix of dots is in order... to perhaps see if the window's polarization properties influence the color variation, vs. the tint or soapy film. Without a control window to work with, it is difficult to deduce things experimentally!

Thanks for your help on this!

Mike

 

[Drakkith]

Wow, that's an interesting effect. I'll be eager to see if anyone has any thoughts on this!

 

[Andy Resnick]

I agree- nice job 'documenting' the effect, as well. I've often seen the pattern you show in image #3 and wondered what the cause is.

Some initial questions:
1) What is used to tint the windows? (thin plastic film, etc.?), and do you have any specs/product info?
2) I don't understand what you mean by 'soapy film'. Is that something you applied?
3) is the photographic filter used in images 4-9 a linear or circular polarizer?

 

[A.T.]

3) is the photographic filter used in images 4-9 a linear or circular polarizer?

Since the sunglasses are most likely linear, the photographic filter is linear as well, given what happens with only the two at different angles. The carglass seems to depolarize the light, so when you put it between the two linear filters which are at 90°, the light gets through.

 

[RedneckPhysics]

Thank you Andy! It's a puzzling effect is it not?

To help answer your questions:

The tint film itself is a mystery tint. I wasn't able to speak with the tinting professional (although I asked) before he applied the film. Moral of the story, if you ever buy a new car and want your windows tinted, go to an independent shop!
Perhaps I can find out what it was that they applied. Even if I can figure it out, manufacturers often don't share the radiometrics of their tint films, outside of overall visible light transmission, which would be a photometric quantity I suppose.

I'm not entirely certain, but I believe a dilute solution of soap and water are used in the tint application process.i believe it serves the purposes of removing particles of dirt/debris from the window, making the adhesive less tacky for application (thus aiding in air pocket removal) and reducing the coefficients of friction when squeegeed. It is sprayed on both sides of the tint, and a squeegee is used to remove some of the solution trapped behind the film by running it across the exterior side of the film. Some solution is squeezed out, the film is de-wrinkled and left semi-wet to dry in direct sunlight over the next few days. Of course, when the water evaporates, some soap would be left behind, creating a very thin film between the tint and window. I would imagine that such a thin film of soap would be on the order of a few microns, much like that of a soap bubble. Perhaps this could cause a bit of thin film interference? Much like we would see in an interference type optical filter.

I wish I had some fresh tint to apply using nothing, or water alone. Then, perhaps we could rule out the thin film's contribution.

I believe the photographic polarizer was linear in nature, but the actual specs weren't listed on the filter as they often are in laboratory grade equipment. I wish I could say "ThorLabs to the rescue!" Here. It seemed to behave like it was linear though. Also, the sunglasses should have used linear polarization as well. I think that sunglasses employ a linear polarization, oriented vertically with respect to the horizon, in order to block reflected rays from horizontal surfaces, as the reflected rays would pick up a good bit of linear polarization through reflection. Using the test-bed (of the led flashlight and paper) I observed that the photo filter and sunglasses combination approximately followed Malus's law upon a 360° rotation. I am not too familiar with circular polarizers, but I would imagine that the sunglasses would block more light at the theta = 0° angle (theta being the difference between the orientation of polarization of each filter) if the photo filter was not a linear polarizer.

Just to clarify, both the 2nd and 3rd pictures used the photographic polarizer as the only filter. I held it close to my cellphone camera's lens, so it wasn't apparent. Sorry if I confused you!

I do have a circular polarizer, which I actually tried the day I took those photos. It had little to no noticeable effect however. Again, it's photographic, so no specs available on that filter either.

To give you a qualitative description of everything: the linear photo filter is a 67mm, with a slight tint (~OD 0.5), the sunglasses should be linear, with a bronze tint of ~OD 1, and the circular polarizer (that as I mentioned, I didn't use in any photo) was darkest, with an OD of ~2 to 3.

I hope that helps, and thanks for the reply!

 

[RedneckPhysics]

Well said AT! It seems as though the glass and/or tint is causing a change in polarization angle.
If the glass only acted as another polarizing filter, and not as an optical rotator, then we would expect to see a dark area there as well, as the initially polarized rays would be blocked by the glass/tint (and any remaining original polarized light blocked again by the sunglasses), assuming the photo filter is an effective (>90% or so) polarizer.

Now if we can only explain why the colors switch back and forth!
My hunch is that the window tint has varying degrees of optical rotation (maybe due to the stresses of tempering) with respect to the spatial plane of the window itself. When linear polarized light is shone through the glass and reaches the tint, it would have varying angles of polarization. Therefore, rotating one of the linear polarizers when repeating the experiment on an un-tinted window would show various regions of "light and dark". Perhaps the tinting somehow adds a spectral dependence to this polarization. The tint already gives a blue hue to transmitted light to give it a "black" look. One of the resulting angles of polarization blocks more green light than red (or transmits more red light than green) resulting in magenta. The complementary angle of polarization (perpendicular to the first) would block more red light than green (or transmit more green light than red) leading to a cyan hue.
This all seems to make sense (in my head at least) as the effect is strongest when we use a single linear polarizer (eg, the sunglasses), and view light that has been reflected at a shallow angle off of a semi-gloss or glossy surface. Such a reflection would impart a certain amount of linear polarization to the light, and thus, act similarly to the first filter in the optical chain in the above experiments, yet still allowing some un-polarized light to make it through un-effected. That un-polarized light would also have a slightly blue hue. This could explain why we don't see the full "rainbow" spectrum of colors as we do with the experiment.

For some good reading, there's a series of articles on Wikipedia about stress-induced birefringence and utilizing polarizing filters to view that phenomenon. Cool stuff! May be part of what's going on here.

 

[A.T.]

Now if we can only explain why the colors switch back and forth!

I have seen similar effects when playing around with circular polarizes (3D cinema glasses). If you have those newer models lying around, play around with them and the car-glass. It might be that the car-glass act as a waveplate, converting linear to circular polarization:
http://en.wikipedia.org/wiki/Waveplate#Quarter-wave_plate

Different wavelengths get different amounts of circularity, depending on how much λ/4 deviates from the waveplate delay. The more elliptical the polarization is, the more sensitive it is to the orientation of a linear filter. And depending on whether the λ/4 < delay or λ/4 > delay, the orientation of the ellipse goes one way, or the other. So the final linear filter blocks different wavelengths, depending on how you orient it.

 

[Andy Resnick]

Since the sunglasses are most likely linear, the photographic filter is linear as well, given what happens with only the two at different angles. The carglass seems to depolarize the light, so when you put it between the two linear filters which are at 90°, the light gets through.

Not necessarily- my photography polarizers (and many other standard ones) are circular polarizers, for obscure (to me) rationale based on autofocus technologies.

 

[A.T.]

Not necessarily- my photography polarizers (and many other standard ones) are circular polarizers,

But this photography polarizer is obviously linear, given how it works together with the sunglasses.

 

[Andy Resnick]

Thank you Andy! It's a puzzling effect is it not?

To help answer your questions:

The tint film itself is a mystery tint. <snip>

This problem is fascinating...

Ok, after a cursory search I found three manufacturers: 3M, SolarGard and Gila, and each company has a wide variety of films. All the films seem to be applied as you say- apply a cleaning solution, then the film, and let it dry. I couldn't find a proper spec sheet on any of the products, but all of the films claim to:

1) block UV transmission
2) block IR transmission
3) "block glare"

The first two simply mean the plastic absorbs or reflects UV and IR light, most likely with a dye(s), although many plastics absorb UV on their own. The third claim implies that the film has polarization properties.

This site has a schematic of the tinting film layers, but I don't know how applicable it is:

http://www.tintmyridenewcastle.co.uk/car-window-tints-available/

As you can see, there are (possibly) multiple layers of plastic, each being birefringent.

In any case, after the applied film is dry and ready, if there was any contaminants left between the film and window, these would probably appear like 'bull-eye" patterns due to thin-film interference.

The window glass itself is likely 'safety glass'- two layers of glass with a thin plastic film in between. Again, this plastic film is birefringent and results in that 'grid' pattern like this:

http://3.bp.blogspot.com/_Y1xTrEn3pwg/S9Jg2wUbY2I/AAAAAAAACJY/1LzjYcaHrq8/s1600/NJ041002c4.jpg

What is interesting (to me) is that polarization control seems to be achieved without an explicit polarizing layer- at least it's not called out on the URL above. I suppose this makes sense in terms of application; if the orientation of the film from manufacturing to installation can't be controlled, then use of linear polarizing films would be suboptimal.

Polyester films, possibly used in the tinting films, appear to have some controllable polarization properties:

http://www.nps.gov/museum/coldstorage/pdf/2.3.1b.pdf
http://www.pstc.org/files/public/smith-06.pdf

I should get a roll of the stuff for lab...