Has anyone explored using flow-through cells with the OFM?
Back in 2023, I chatted with @r.w.bowman about real-time monitoring of Spirulina cultures using image recognition. He mentioned that some folks were looking into flow-through cells. I didn’t follow up then, as our focus shifted to hydrogen-oxidising bacteria.
Recently, another bioreactor manufacturer expressed interest in real-time culture monitoring. Unfortunately, I can’t find my notes from the conversation with Richard…
Is anyone here working on or interested in collaborating on a flow-through cell design for the OpenFlexure Microscope?
I cannot be much help with flow cell design. On the optics / mechanics side, you will probably not need x-y motion as things are moving past in x and confined to a channel in y. The separate z-actuator can either have the square top to screw through (as on the upright microscope) or it can have the triangluar top to screw into.
Stability will be important, I expect autofocus will not work if things move, even quite slowly. If the objects are small and sparse then autofocus will struggle even on stationary samples. With careful design you might get away with no fine motion at all, setting it up once and for all at the start. It depends on the magnification that you want. Shimming will not allow you to position within the focal depth of a lens, but will allow you to position the focus to be inside a 100um channel. Plenty of light will be needed for fast shutter speed for pictures of flowing things.
You could also consider multiple optics systems along a channel, if you wanted an additional magnification or additional imaging mode. Using different sides of the channel would allow the viwing area to be close together along the channel.
I recently met Dr. Lillehoj from Rice University. Just yesterday, his paper was published, and I thought it would be a good fit to integrate it with a modified version of the OFM. Please let me know if this is what you’re looking for, and I can connect you both.
Thanks @dgrosen – the micrographs look brilliant. If we can build a large enough team interested in doing this with OFM, it would indeed be fantastic to involve Dr Lillehoj.
However, at the last OFM-OSS meeting, @gerrit mentioned customers have been requesting fluorescence capabilities. I imagine that might be a higher priority for the experts we’d need to engage…?
I would imagine that fluorescence is an Engineering challenge and also expensive. What if you could achieve similar results without using fluorescence?
Fluorescence is something that we have managed to do on OpenFlexure Microscope, but is not turnkey yet. There is some tightening up of the physical hardware that is needed, but also some further discussions about sourcing the correct filters and LEDs, this is something that is going to change depending on the specific application.
I think the engineering work is achievable but not trivial. I would think a specific funded research project, with a specific application in mind is probably the best way to push this forward. As a self contained package of work it may be possible to fund within a university?
We should probably start a new thread on Fluorescence. Does the same comment apply to Flow-Through Cell development:
The reason I mentioned Fluorescence was to gauge whether it’s further up in the priority queue than Flow Through Cell development. If so (unless we have parallel groups) engaging Dr Lillehoj may be premature.
There are a number of florescence threads we can resurrect. I think flow-through seems like something that woulld be best developed separately to OpenFlexure and compatible with anything that can hold a microscope slide?
I know @biodotpe has been working on fluid transport combined with the flat top microscope. I am not sure if this sort of setup might work?
Thanks @biodotpe and @j.stirling - yes I had seen your microfluidics posts before with great interest @biodotpe and yes, I think strobe-enhanced microscopy is likely critical.
I think the PlanktoScope could be very easily modified to work for bioreactor work large enough organisms. They claim a 0.75 um pixel size, but using reversed M12 lenses makes me think the images won’t compare favourably with an OFM and comparably priced 40x objective.
The OFM itself is at the boundaries of what we’d like to do. It should be perfect for many fungal bioreactors, and pretty good with yeast. But for work with bacteria (AMYBO’s current focus) I’m thinking we’d need staining and 100x oil immersion objectives to get useful information - our main candidates are around 0.3-1.0 x 0.6-6.0 µm.
Getting useable results with bacteria may be a pipe dream, but I believe a flow-through cell OFM could be ideal for low cost realtime monitoring of the likes of Spirulina production.
Combining a liquid holder and a 100x objective will require some very thin material on the imaging face. A normal coverslip is 0.17mm, the working distance is about .36mm. Assuming you want any depth of imaging into the cell you probably want something not much more than .25mm.
Sounds really interesting. I guess it depends what (if any) bioreactor you’ll be using. We have been using Pioreactors and my (possibly naïve) assumption was that their standard peristaltics would suffice, given that flow should be stationary when they are not powered.
The pump selection will depend on the precision, stability, and flow rates users need for their experiments. Sometimes, a cheap peristaltic pump is enough to infuse liquid into a chip. Additionally, the liquid volume is not a limitation, such as in the syringe pumps.
In this image from an ElveFlow video, you can see the performance of the different pumps:
I made plenty of differnet flow cells for single-molecule fluorescence. After 2 decades of trying every double sided tape on the market, silicone, PDMS, parafilm etc as spacers (which are still used by most people in the field) came to a design based on a 3mm-thick borosilicate slide with milled channels (top side), a coverglass (bottom side) and a 10-micron thick layer of epoxy. to hold them together. Milling was all at the lab on a sherline endmill, with diamond bits. The components were recyclable with soaking in an organic solvent for 3 days to dissolve the epoxy. The flow was created using a peristaltic pump (“suck-in”). Stability was on the order of a few nanometers per hr, and I could store them at -80 w/o cracking (made a few in one shot). The chamber had 8 channels, each channel was about 2.5 ul. Temperature control was to within 0.1C with a resistive heater [2 omega kapton strips+temp probe] glued to the top with same epoxy (also removable in the same solvent to recycle). So, not a proper “microfluidics” setup, but allowed me to save on solvents when cleaning and reusing (as opposed to “standard” size slides). Been thinking of writing the protocol up for a few years, but had to retire from my university position due to a post-covid sickness, so I doubt I will ever bother to write this up.
Edit: found a screenshot from a file I sent to a japanese company for milling of the slides. They charged about $100 per slide. Each slide is 9x40x3mm. To be honest, my milling was higher quality than theirs – their CNC mill generated microcracks that caused the slides to break after 2-3 recycling treatments. Mine survived 20-30 cycles (organic solvent, acidic piranha, KOH).
– their CNC mill generated microcracks that caused the slides to break after 2-3 recycling treatments. Mine survived 20-30 cycles (organic solvent, acidic piranha, KOH).