Fluorosphere separations with SiN-NP SEPCONS

Karl and I attempted once again to filter particles with the SiN-NP 5 slot SEPCON chips.  (wafer # 1019, 27% porosity, avg. pore dia. – 54 nm in the center, 30%, 61 nm at the perimeter, histograms below)

 

We started trying to first pass 20 nm red fluorospheres, then block 100 nm yellow fluorospheres, and finally pass 20 nm red fluorospheres.  Both particles came at 2 mg/mL.  We diluted them to 100:1 volume ratio, then attempted to pellet out the particles by spinning for 15 min at 25 krpm in the OPTIMA Ultracentrifuge in the TLA 55 holder.  The 100 nm fluorospheres pelleted while the 20 nm did not.  We removed the supernatant and resuspended the 100 nm particles and left the 20 nm particle solution as is.  We then further diluted the solutions to 1000:1 volume ratio in water.

Using 300 µL in the top of the SEPCON and wetting the bottom with 12 µL, we initially tried to centrifuge at 2 krpm but had very little pass (~12 µL?) even after 10 minutes.  We increased to 3 krpm and passed ~100 µL after 15 min.

We used the TECAN and the fluorescent microscope to try to detect fluorospheres in the filtrate.  I ran initial baseline scans for both “red” and “yellow” wavelengths but didn’t actually subtract them below because the signal was so low and consistent across wells.

The following plot shows a TECAN scan exciting at 450 nm.  It shows a relatively strong fluorescence for the stock solution of 100 nm particles, but no fluorescence from the filtrate.  So 100 nm fluorospheres are not passing through the membrane.  Alas….

In the next two plots we show the results of the TECAN scans for the 20 nm red fluorospheres excited with 540 nm.  The strong peak is the stock solution and the rest of the lines are from the 1st and 2nd filtrate, before and after the 100 nm’s were attempted.  So according to this, neither 20 nm , nor 100 nm fluorospheres are passing.  (I show a closeup on the low signal lines to show that there is a slight increase for one of the filtrates, but it is so low in comparison to the stock that I effectively consider it no signal)

As mentioned, we tried verifying the TECAN measurements with fluorescent microscope images.  I washed slides and coverslips with ethanol and used 1 µL of solution.  We used the “FITC” filter to look at the 100 nm fluorospheres (Exposure-100 ms, Gain-10) and the “Rhodamine” filter to look at the 20 nm fluorospheres(Exposure-100 ms, Gain-2).

 

 

While the 100 nm fluorospheres were fairly consistent in that I could find several of them per viewing field, I saw very little evidence of the 20 nm fluorospheres in the field.  The following images show the fluid front and the quick drop off in particles as you move towards the center.

 

 

I would be tempted to say that not finding 20 nm fluorospheres in the field with the microscope indicates that there is something wrong with our 20 nm stock, but I’m afraid it is more likely that something with my technique is flawed.  To gain more confidence in the 20 nm fluorospheres, we also did a particle size analysis using the malvern.  Here’s a table showing results for 20 nm fluorospheres and 10 nm Au NPs.  Honestly, I’m not even sure how to interpret the values provided, but in general is looks like the fluorospheres are at least smaller than 40 nm.  (I included all of the data in the summary because I don’t know if the PDL or the background ratio gives us a degree of confidence.  I need to read up on all of this…)

Not passing 20 nm fluorospheres through a membrane with an average pore size of 54 nm was disconcerting at the time, so we decided to try smaller particles.  We don’t have smaller fluorospheres, but we do have Au NPs.  We diluted 10 nm Au NPs by adding 4:1 volume water:Au.  (I couldn’t find an initial concentration labeled on the bottle, but Karl said this worked well for him in the past so I ran with it)  The plot below shows the absorbance results using the cuvette in the TECAN.  Gold passed through the exact same membranes that had blocked the fluorospheres in the tests above.

I discussed the above results with Jim and Chris and they both pointed out that diluting the fluorospheres in water would lead to an effective radius due to charge much greater than that of the physical fluorosphere.  In lieu to reading up on that phenomenon and truly understanding it, I just diluted our initial 100:1 solutions of 20 and 100 nm fluorospheres with 1X PBS in a 10:1 ratio.  (So I mixed 1000 µL of 1X PBS with 100 µL of 100:1 aqueous fluorosphere solution)  I assume that’s sufficient?

Well I won’t keep you in suspense, I did the same experiment as above with new membranes and neither the 20 nm nor the 100 nm fluorospheres passed, even in PBS.

Just for the heck of it, I did the same dilution detailed above but replaced 1X PBS with 10X PBS.  Still no 20 nm fluorospheres passed.

So what now you ask?  Great question!

The good news is that I am much more confident using the current wavelength and gain (150) settings on the TECAN and using the 1000:1 dilutions of fluorospheres.  I’m also getting a better feel of the flow rates expected out of the SiN-NP SEPCONs.  We have not had a problem getting the filters to pass fluid. (possibly due to the high humidity?)

So I think we just need to find a way to pass 20 nm fluorospheres, or come up with a good explanation as to why they shouldn’t be passing.      Hopefully someone sees a flaw in my work up above that can be easily remedied.  Below is a schematic showing the direction of flow for the fluorospheres “through” the filters.  We’ve discussed the possible implications of have the dual-layer of pnc-Si on SiN.  Perhaps that is adding to our woes?  I could try flipping the chips to flow past the pnc-Si first, but I hesitate to do that until I have a pretty good reason to do it.  JP has also stripped the pnc-Si from some chips and I could try separations with them to see if there is any difference.