Chapter 28. Bacterial and biofilm iridescence part 2. Refraction and diffraction

The rainbow iridescence of the aquarium biofilm and the colonies of rainbow coloured bacteria I observed within it had got me thinking about rainbows, probably because they are so familiar. I imagined that the coloured tetrads might be acting like drops of water in the air, refracting the light due to their having a different refractive index from water they're floating on, and so causing the dispersion of the visible wavelengths of light. But after some reading I discovered two problems. Firstly, the difference in refractive index between a bacterial cell (1.38) and water (1.33) is much less than between the atmosphere (1.0) and a drop of water (1.33). Would that be enough to cause dispersion? Even allowing for the EPS which is reported here to have an RI of 1.4, it seems unlikely. Second, a rainbow can only be seen in the sky opposite the sun and from an angle of 42 degrees from the direction opposite the sun. In other words, rainbows are only visible under epi-illumination at a fixed angle. When looking down a microscope I'm staring directly at the sun and the biofilm dispersion at the aquarium surface was visible under trans-illumination and a range of angles of epi-illumination so rainbows don't make sense as a model. I decided to stop thinking about them, what I'm trying to understand is obviously very different.

Further reading of the wikipedia page on iridescence led me to other ways light waves can be interfered with in order to split them into different visible colours. These effects are often caused by multiple reflections from two or more semi-transparent surfaces. They're responsible for many examples of iridescence in biology. One example is thin film interference (TFI), and I think because the description includes the word 'film' I was convinced I had found the answer. In TFI some wavelengths of light are cancelled out and some accentuated when reflected from the surface and lower boundary of a thin film. The colours affected dependent on the thickness of the film, which must be in the range of the wavelength of visible light so ~0.4-0.75 microns. It's the reason for the iridescence of oil on wet tarmac and it was while I was reading about it that I took the following photo.

Iridescent aquarium biofilm formed on the top of a powerhead.

To my mind this looked oily, which fitted with TFI being responsible, but the more I read the less likely that seemed. The first problem is the thickness of the biofilm. Even if it is one cell thick that's still almost 1 micron, and when I looked at a sample of the above biofilm under the microscope it looked like it had topography.

Aquarium iridescent biofilm x100. I didn't see a single coloured tetrad in this sample.

The second problem is refractive index again but in a slightly different sense. When light waves meet a boundary between to mediums with the same refractive index they pass straight through. So I'm not sure there would be enough reflection from the lower boundary of the biofilm to cause TFI. The third problem is that TFI seems to be all about reflected light so I'm not sure it would be visible by trans-illumination.

It turned out the answer had been staring me in the face all along, if only I had read the caption to the photo of an iridescent aquarium biofilm in the wikipedia page on iridescence. The phenomenon responsible is diffraction not refraction/reflection. In the context of biofilms it seems the idea is that bacteria can grow in ordered patterns and form a diffraction grating. The spaces between the bacteria form slits, which have to be small enough to be within the wavelength of visible light, so as to split the light into it's constituent colours. Importantly "Diffraction will produce the entire spectrum of colors as the viewing angle changes, whereas thin-film interference usually produces a much narrower range". That fits with the full spectrum effects I observed in chapter 27 but not so well with the photo above which is dominated by red blue and green. I occurred to me that diffraction might also be responsible for the apparent iridescence of the coloured terads. In that case they would have to have some kind of repetitive structure in their cell walls, perhaps a protein, in order to diffract the light. Towards the end of this article there's a very interesting discussion of the possibility.

I couldn't see any order or pattern to the growth of the bacteria in the biofilms I examined at x100. It looked totally random to me no matter how much I digitally zoomed in. When I put a coverslip on this sample and used a higher power objective, I did perhaps see some evidence of structure. The edges of the image are out of focus but in the areas where the bacteria seem to be forming slits, the slits do seem thinner that the width of a bacteria, which puts them in the range of the wavelength of visible light.

Aquarium iridescent biofilm x1000. Could the gaps between the bacteria act as a diffraction grating?

The problem was, when I put a coverslip on an iridescent biofilm sample, I lost the iridescence, so I'm not sure if this image is an accurate representation of the biofilm structure. I needed a way to examine iridescent samples at high magnification but without a coverslip and completely by accident I found one.

Aquarium biofilm iridescence. Fresh sample trans (top left) and epi (top right) illumination. Dried sample trans (bottom left) and epi (botton right) illumination.

Just as you would expect, the fresh biofilm sample is iridescent under trans (top left panel) a epi (top right panel) illumination. What I noticed after not cleaning my slides one evening, is that the same is true after the sample has dried onto the slide over 24 hours. This is interesting because it suggests that the biofilm structure, in terms of it's ability to diffract light, is the same after drying. The only other possibility is that the iridescence of the fresh sample is due to diffraction and the iridescence of the dry sample is due to TFI. This is more plausible than for a fresh sample because the dry biofilm might be thin enough and have a sufficiently different refractive index to glass (~1.5) to cause it. But it doesn't explain the iridescence under trans-illumination (I think) and I can see a full spectrum of colours in the dry sample iridescence which is more in keeping with diffraction. Because the sample had dried I could examine it with higher power objectives without a coverslip.

Iridescent aquarium biofilm dried onto microscope slide x400 (digitally zoomed x3). The maze-like patterns formed by the bacteria apparently act as a diffraction grid. The circular objects range in size from 2.5 to 10 micrometres and are probably dried amoebae.

A strong candidate for a diffraction grating was revealed I think. The lawn of bacilli, which appeared to be unstructured at x100 magnification, actually formed complex maze-like patterns at x400. It is possible to view the dried film x1000 but it is not possible to use lens oil! As a result x1000 images should suffer a loss of resolution and be effected by spherical aberration.

Bacterial diffraction grating x1000 (x2 digital zoom)

I think there is a loss of reolution around the edge of the image but I get that with any of my objectives. The way the bacteria might be forming slots through which light waves would be diffracted is clear. Remarkably, the film appears to be one cell thick which I would never have guessed from looking at the biofilm under epi-illumination.

Aquarium surface film x100 epi-illumination.

I think both images also reveal that any small gaps in the biofilm are populated by other rod shaped bacteria which don't like to be crowded. Perhaps they and the tetrads produce an antibiotic which is why they're always in clearings in the film.

However, in terms of whether the coloured tetrads were iridescent, the x400 image was deeply troubling. True, the coloured tetrads were still coloured (which would be expected if some kind of nanostructure in their cell wall was diffracting light) but then so were many of the objects in the image. The bacteria in the gaps in the diffraction grating appeared purple and the amoebae green. As far as I know amoebae aren't green and I can't start claiming to have discovered iridescent bacteria willy nilly. More concerning still, was that neither object was coloured at x1000 demonstrating that my optics/imaging are capable of giving a misleading representation of colour in certain circumstances. All this doesn't look good for the 'iridescent' tetrads.

I needed to pause, do some more reading about chromatic aberration and also examine how my microscope in particular is affected by chromatic aberration.

Chapter 27. Bacterial and biofilm iridescence.

Why were some of the bacteria I observed in my last post coloured? This video is digitally zoomed and as you can see the bacteria that form tetrads are not only coloured but are changing colour. My first thought was that it was some kind of digital imaging artefact. That's human nature I think. If you see something you can't explain you rub your eyes and throw the vodka bottle away. And there's reason to be concerned. The digital camera attached to my microscope has a CMOS detector which are apparently less sensitive and more susceptible to noise than alternatives. Also, entry level digital cameras take a probabilistic approach to estimating the colour of any given pixel and, when you watch the linked videos, there is a strange static oily rainbow pattern covering the field of view. Having said all that, if it is an imaging artefact, then it's an artefact that only effects a sub-set of the bacteria in the film, which is interesting anyway.

Firstly I needed to know if the observation could be repeated so I continued to allow the biofilm to form by detaching the venturi valve from my powerhead and angling its exhaust away from the tank surface. As before I sampled by dipping a microscope slide into the aquarium and scooping up a fragment of the film in situ. I then viewed at x100 magnification without a coverslip. This image was taken around two weeks after first contact.

Aquarium surface biofilm in situ x100 magnification. The almost continuous sheet of bacilli is punctuated by the occasional island of coloured bacteria.

The oily rainbow effect is much reduced in this image although there is still variation in background colour. I think the effect is a microscopy artefact. I can recreate and accentuate it by setting the condenser to about half way between the light and the stage of the microscope.

The coloured bacteria were much less common than in the previous sample but were still easy to find. And as before they change colour over time. To produce the video I used VirtualDub to crop from 1280x1024 to 484x448 (19%) but any media player I use expands it to full screen when playing and must do some interpolation or the video would be far more pixelated. Below is a still from the cropped video expanded to 1106x1024 using GIMP to do the interpolation.

Six day old aquarium biofilm x100 zoomed x5,  Iridescent bacteria?

I can't find any images like these on the internet. The bacteria seem to be all the colours of the spectrum, twinkling away. There is a genus of cyanobacteria that form similar grid like rafts. Merismopedia cells divide on two planes and so produce tetrads. They are reported to contain reddish, blue/green, violet or pink content but they seem to form much more orderly structures than these and there's no mention of cells changing colour. Importantly, the bacteria described here are not coloured when viewed under a cover-slip (Chapter 26 and subsequently confirmed) which suggests that pigmentation is not important. I might be the first person to ever observe this species so I thought I would have first go at describing them.

Stubby rods 1 to 1.5 micrometres wide and 1 to 3 micrometres long. Division by binary fission on two planes to form tetrads. Forms colonies in a freshwater/air interface aquarium biofilm. Colonies range from two to several hundred cells. Cells exhibit varied dynamic colouration under direct trans-illumination. 

Aquarium biofilm with a colony of coloured bacteria. Four frame .GIF. The frames were taken 1-2 seconds apart. 

This GIF (made with GIMP) shows the changing colours nicely. They look washed out to me when I upload my videos to YouTube, I guess YouTube has a different interpolator (which does make the point that interpolation may be accentuating the effect). The obvious question I had was; what is known about coloured bacteria? I couldn't find much information but I did find an article about bacterial iridescence from the blog "small things considered". In it the author explains that iridescence has been observed for bacterial colonies but is poorly understood and has only been studied in detail since 2012. Very helpfully she links the original research papers and they concern the iridescence of quite large colonies of bacteria cultured on growth media. There is no mention of individual cells being iridescent although there is some microscopy of the colony fringes. However the research is still interesting because it describes a category of iridescent bacteria colonies called 'rainbow diffuse' that

"exhibited all spectral colors ranging from red to blue only under the condition of trans-illumination"

I think they're talking about oblique trans-illumination from reading about their experimental set up, but the above description broadly fits my observations albeit at a much larger scale. So, what is iridescence, what causes it and could it explain what I have described here?

It seems iridescence is an optical phenomenon of surfaces in which hue changes with the angle of observation and the angle of illumination. Given that the angle of illumination and observation are fixed when looking down a microscope, if iridescence is responsible for the changing colour of these bacteria then it must be because the biofilm fragment I'm observing is floating across the surface of the drop of water on the microscope slide. You can see this movement in the videos I have linked (movement is not obvious in the above GIF because, for artistic reasons, I nudged the layers to align better). What struck me about the wikipedia page on iridescence was that one of the the examples used was an aquarium biofilm but the iridescent pattern didn't look anything like the oily effect I remembered from my aquarium or oily effects in general. The photo on wikipedia shows a perfect rainbow with all the colours in the expected order and that's not what comes to mind when looking at oil on wet tarmac. It's usually a bad idea to rely on memory so I thought I had better start photographing the biofilm as well as examining it under the microscope.

Aquarium biofilm iridescence under epi-illumination (left panel) and trans-illumination (right panel).

The photo on the left was taken the day before the microscopy I've described above and the biofilm has produced a series of rainbows, one for each of the five LED bulbs in my strip-light. The effect was angle dependent. In the left panel the rainbow effect was still visible under trans-illumination. So it seemed that either my memory was at fault or I was observing a different effect. Either way, I couldn't help wondering if the rainbow coloured bacteria I saw under the microscope had anything to do with the rainbow effect on the aquarium surface.

If they did then I would have expected them to be iridescent under epi-illumination just like the aquarium surface.

Aquarium biofilm x100 under epi-illumination. Two frame GIF digitally zoomed x4.

I captured these images by turning off the microscope lamp and illuminating the sample from the side using an LED torch. The cells in the colony of coloured bacteria in the centre seem to be different colours from the surrounding bacilli but there is no great range of colour or change in colour. I tried a variety of angles of illumination and couldn't see any evidence of iridescence in any of the tetrad colonies I examined. I wondered if converting a microscope to epi-illumination in the simplistic way I had could produce meaningful images. From what I read here there seem to be two concerns. First, my achromatic objectives don't have the anti-reflective coatings found on epi-illumination objectives. The linked video certainly has some strange arcs of light but seems reasonably well resolved. Second, all my achromats have a lens corrected to allow for a 0.17mm coverslip.

This made me think because all the images I've uploaded were captured without coverslips. What effect would not using a coverslip have under trans-illumination when using such an objective and could it be that the phenomenon I have described is a microscopy artefact after all? Achromatic lenses are designed to prevent chromatic aberration  (and spherical aberration) and they do this by focusing the different wavelengths of light on a single focal point under the expectation that the light entering the objective has been split into different wavelengths by passing through a coverslip. If you don't use a coverslip, I guess the light enters the objective in phase and is then split. As I mentioned earlier, when I put a coverslip on biofilm samples I loose the coloured effect from the tetrads so CA does make sense. However, and I feel a future retraction coming as I write this but, I don't think chromatic aberration could explain my observations because high numerical aperture objectives like the x100 I used are much less susceptible to CA than high power objectives and CA only seems to effect red and blue light.

So, if the coloured tetrads were not iridescent under epi-illumination then that suggested that they didn't contribute to the biofilm iridescence. I wanted to understand what did cause the biofilm iridescence in the hope that it would help me understand why the tetrads were coloured.

Coming soon: Imaging aquarium biofilms in situ at x400 and x1000 magnification and the mystery of the oily effect is solved.