Illumination for Imaging moving erythrocytes

[flying fish]

Hello all,

This I suppose this is probably 80% of an optics/ microscopy question, maybe only 20% of a biology question, but I thought that people in this section probably have the most experience with this sort of thing.

I am using a setup than involves an objective lens suspended over a microfluidic chip. A camera is positioned behind the the lens (well more precisely, the image actually reflects off a beam splitter into a camera, the beam splitter for the purpose of allowing optical tweezers to be used on the sample while simultaneously imaging on the camera).

The student who worked on this before me was able to see the erythrocytes just sitting there on a slide (and trap them with the laser), but in the chip they were nowhere to be found! I attempted the same and so far have also failed. What I probably want to do is maximize scattering/ minimize light hitting the camera that is from anything other than the cells (I suppose they call this dark field imaging?). I have not been successful in seeing anything after several days of messing around with the lighting (the light source is free to move basically anywhere). Some of my co-workers suggested lighting from above, or from the side, but still the only time I can even see the channel walls decently is when lighting from below ("below" meaning pointing at the front of the objective, through the chip). I have "frosted" the surface of the transparent fixture plastic that the chip sits on, again to maximize scattering. This improved the contrast of the chip walls a little bit, but I still see no erythrocytes. I have been able to see 6 micron plastic beads (which were supposed to represent the erythrocytes) but when I put the real thing in I see nothing. So now that I've ranted sufficiently, does anyone have any tips for lighting techniques for this sort of application?

 

[Monique]

The best thing to do is contact the student who has experience with the system. Can you get an image of the erythrocytes on a slide? Some other questions: what kind of magnification are you using, how do you know you are looking at the right focal plane, have you adjusted the condensor to the optimal position.

 

[Andy Resnick]

I can help you (and we can continue the discussion offline if you prefer), but some parts of your description are not clear:

1) are you using an actual microscope, or just holding a microscope objective over the sample? What are the specifications (NA, working distance, correction) of the objective lens?

2) What light source are you using? Where is the light source currently? How is the sample illuminated?

2.5) Camera specifications: are there eyepieces, or can you only image via the camera? What is the make/model of camera?

3) Are you sure there are erythrocytes present?

4) What is the dimensions of the microfluid channel? How fast are the erythrocytes moving?

That's enough to get a good start on troubleshooting. As a simple thing to try, often low contrast objects are easier to see when the NA of the illumination is low- the resolution goes down, but the contrast goes up (to a point).

 

[flying fish]

Thanks for the responses. I guess I left out a lot of details. Here is a rough schematic:

The light source is a high intensity fiber illuminator (adjustable).

I'm not sure about the camera, it is just a cheap CCD camera. Through the camera is the only way of viewing the sample.

I can see erythrocytes in the sample solution when looking under a "real" microscope.

The channel is 100x100 um. I don't know the speed as I haven't actually seen them, but a slow flow rate of maybe ~100 um/s or so is what I am aiming for.

I did manage to make some improvements today in the contrast by blocking off (with tape) all areas on the sample fixture where light could get to the camera without first going through the chip (basically just trying to reduce noise as much as possible). Whether this will help me image the erythrocytes I have yet to find out (maybe tomorrow!)

 

[Andy Resnick]

Oh my... I'm amazed you can anything to work at all with this.

First- the objective NA is really low for tweezing- are you trying to get a 3-D trap, or just hold the cells against the bottom surface?

Second- the path length you gave is really suggestive of a 160 mm finite-conjugate objective lens, as opposed to an infinity corrected objective. Either way, are there lenses on the camera? If you replace the camera with a sheet of paper, what do you see- an image plane?

Third- the main point- you have no illumination control at all. This is going to lead to all kinds of problems (flare is an obvious one).

I would start by looking at a target- get rid of the fluid chamber and put something down, anything, that has a known shape- print out (tiny) concentric circles on a piece of paper, or draw a small picture on a slide, if you have resolution targets sitting around (you can get some cheap reticles from Edmund Optics that can work). Looking at a known object will help you figure out what is wrong.

The flow speed is going to cause problems also- blurring. 100 um/s, even at 30 frames/s, will produce ~3.5 microns displacement during an exposure, which at 20X leads to about 6 or more pixels (depending on the camera).

The first step should be trying to image something simple, to get the optical performance optimized. Then start imaging your sample.

 

[flying fish]

Thanks for the detailed advice! As you can probably see I don't have much experience with this sort of work, but I'm glad to be learning now.

I'm not sure (at the moment) whether the camera has a lens on it. Now I'm interested in finding out though. The fiber illuminator I believe does give me some control over lighting...since I have all the standard optical bench hardware, and so I have some easy options for mounting the fiber light source. I was even thinking of putting it on a stage for fine tuning.

I don't think it is a 3D trap (not knowing enough about optical tweezers to know what that means!). It is a "conveyor" for bulk particle sorting, done by shining a "bar" of laser light across the sample flow. Only radial trapping about the "bar" is desired. Sort of an "optical water slide" to divert the particles to one arm of the microfluidic chip.

I was finally able to see the erythrocytes today. After staying in the lab until 8:00 PM messing around with the lighting! I had the fiber illuminator pointing at the sample from below at roughly a 45 degree angle, with some random object to diffuse the light in between. It took a lot of fine tuning to hit a "contrast sweet spot", but I finally got it. The resolution was actually better than the plastic beads, perhaps because of the flatter geometry I'm guessing. And then of course I ran out of HD space to store the video!

One more "little" detail I forgot to mention - I also have to see platelets! No details are required; they can just look like little dots floating by. I'm guessing I need fresh blood to see platelets, right? Or do they still float around in there regardless of whether or not they are dead?

 

[flying fish]

So to bring this thread up again...I'm still working on the same project. I managed to get GREAT images using a very obvious technique that we should have been doing all along - adding a condenser lens!

There is still a problem though. This is part of a platelet aggregometry project, and I need to see platelets as well as red cells. The contrast of platelets is just no good. So my first question, is what kind of (relatively simple) microscopy techniques are best for platelets?

On a hunch (let me know if you think I'm on the right track) I started to turn the setup into a "darkfield" microscope, but I'm not having much luck seeing anything in the dark field. Do I need a ridiculously bring illumination source? I'm already using an extremely intense fiber illuminator. It will start to melt any black plastic parts you put within a couple mm of the fiber tip! The darkfield schematic I found on wikipedia ( http://en.wikipedia.org/wiki/Image:Dark_Field_Microscope.png ) shows that you should clip off the "halo" before the objective lens. But can you do it after, as long as it is before it hits the camera?

 

[Monique]

So to bring this thread up again...I'm still working on the same project. I managed to get GREAT images using a very obvious technique that we should have been doing all along - adding a condenser lens!

A condenser is essential, do you kohler it? I would suggest a nomarski prism to get optimal contrast.

 

[flying fish]

A condenser is essential, do you kohler it? I would suggest a nomarski prism to get optimal contrast.

My condenser is just a DIN 10x that I stuck in front of the fiber illuminator. I haven't really optimized it or anything.

How difficult is it to set up a normarski prism?

As for the dark field thing I was experimenting with, I was able to see red cells as white dots on a black background, but the image was blurry and I could only see it with the camera filter off (and it has to be on while trapping). I figure in order to get enough light with the filter on, I'd need an airplane light coupled to the illumination fiber!

Another thing I was thinking - perhaps I can just use a sample that is highly concentrated with platelets (presumably by centrifuging and extracting platelet rich plasma off the top), so that I fill the whole screen with platelets. Then they should be noticeable no matter how poor the contrast is!

 

[Monique]

My condenser is just a DIN 10x that I stuck in front of the fiber illuminator. I haven't really optimized it or anything.

How do you know that your condenser is in the right focus plane? Look up kohler illumination.

How difficult is it to set up a normarski prism?

I only work on high-quality microscopes that have all the components installed. The prism I use is part of the objective lens. You really need differential interference contrast (DIC) to get optimal images.

 

[flying fish]

I actually did suggest to my supervisor today..."Why don't we just shoot the laser through the eyepiece of a REAL microscope." Seems like the only real challenge would be how to mount the laser, but I think even that would be relatively easy; I could mount the laser so that it can rotate parallel to the vertical optical breadboard. Then just set the microscope on the table in the appropriate orientation so that the laser can be shot into the eyepiece. My supervisor however didn't like the idea too much, especially since it involves buying a new microscope!

Another thing that I tried today was simply putting one linear polarizer under the condenser, and one in front of the camera, 90 degrees to one another...And with the light source turned all the way up, I could see the particles glowing a bit over a dark background. It was very interesting, but the image really was not any better than the brightfield image. And as before, the light source wasn't strong enough to see anything through the red absorbing filter.

I will look more into Kohler illumination. I would think proper setup of the condenser should yield more light getting through. Or perhaps use a camera flash to get a short pulse of high intensity light (since all I need is stills anyway) If all else fails, I wonder if all I need for DIC as you mentioned is a specialized lens, or if there are more components that have to be added to the system? Anyway, thanks again for your help.

 

[Ygggdrasil]

Although I don't have enough practical experience with microscopy to help you, I can point you to a useful site with lots of good tutorials on microscopy:

http://www.microscopyu.com/

The microscopy tutorial contains information on DIC and phase contrast which may be useful techniques to consider and also info on darkfield imaging in the stereomicroscopy section.

 

[Monique]

I actually did suggest to my supervisor today..."Why don't we just shoot the laser through the eyepiece of a REAL microscope."

I really don't understand why you are developing all this by yourself. Go to a department that has all this set up and use their equipment, or at least learn how it is best done. I know too little about the technical side of laser trapping, but wouldn't you be able to use a confocal microscope?

 

[flying fish]

This is basically unfunded research in a small start up company. We are currently trying to get a grant for researching and developing a platelet aggregometer. Before we apply for the grant, we want to have quality images of on-chip laser trapping separation of blood components. The laser trapping part actually does work quite well, and we have videos of it working under the makeshift bright field setup that I worked on before. However, now we want quality still images that clearly show platelet separation, but under the current setup the contrast is really too pathetic to produce such an image.

We do work with universities that do have the proper microscopes, but we would need to take our laser off-site to work on it - which actually might be fine for a temporary solution, but for continued studies we really need something on site. And my boss, being a physicist with great theoretical knowledge but no practical microscopy knowledge, has more faith in playing with polarizers than actually obtaining the proper equipment or using a proven technique...

So yeah, that's the basic story there. Thanks again. Also, thanks for the link Ygggdrasil, lots of good reading material.

 

[Monique]

This is basically unfunded research in a small start up company. We are currently trying to get a grant for researching and developing a platelet aggregometer. Before we apply for the grant, we want to have quality images of on-chip laser trapping separation of blood components. The laser trapping part actually does work quite well, and we have videos of it working under the makeshift bright field setup that I worked on before. However, now we want quality still images that clearly show platelet separation, but under the current setup the contrast is really too pathetic to produce such an image.

We do work with universities that do have the proper microscopes, but we would need to take our laser off-site to work on it - which actually might be fine for a temporary solution, but for continued studies we really need something on site. And my boss, being a physicist with great theoretical knowledge but no practical microscopy knowledge, has more faith in playing with polarizers than actually obtaining the proper equipment or using a proven technique...

So yeah, that's the basic story there. Thanks again. Also, thanks for the link Ygggdrasil, lots of good reading material.

I understand, it is never easy to get new things set up. Good luck, I hope you'll find an easy fix

 

[Moonbear]

I'm still a bit confused why you don't do your work at a university with proper equipment. It would be cheaper to pay user fees for their equipment than to spend all this time creating it from scratch. Once you can show proof of concept on someone else's equipment, you can then write into your own grant the cost of purchasing the equipment for yourself if it's then integral to the project.

One thing you haven't mentioned, and I'm not sure I understand your set-up enough to figure it out from what you've already told us. Do you need live cells when you're doing the microscopy, or can you stain them first to make them easier to visualize?

 

[flying fish]

Now that I think about it...I don't think they need to be live for the laser trap sorting, but they do need to be fluid-suspended, and they do need to retain their physical, refractive, and absorptive properties (at least, at the wavelength we are using, which is 670nm). I never thought of staining, but it sounds like a good idea.

The system that I'm working with basically creates an entire line with the appropriate beam profile for trapping transparent particles. Particles will ride along the line like an "optical conveyor" allowing you to steer red cells and other particles of similar size and transparency. Platelets, being smaller, slip right through thus sorting by size can be accomplished.

I think one of the major issues with taking down the system and setting it up at a university is just the fact that it works (except for the platelet imaging part), and my supervisor is worried that if we take it down and set it up again...it might not work. He is only willing to take that risk AFTER we have the images we want. I personally do not think the system is that sensitive. But I'm not the one in charge...

I should also say that I'm not sure how easy it will be to convert our current system to run on an enclosed microscope. In theory it should work, but we still need to add in a heavy-duty filter to cut out the red laser light before it gets to the camera, and we still need a way to align the laser and launch it into the microscope. Sounds easy, but you never know what could go wrong. There's also the issue that this thing would be hard to set up without optical breadboard, and the live cell microscopy equipment and optical breadboard are from two different departments in the university. I do agree though, it is definitely worth a try.