Transport through 'non-porous' membranes

There’s been lots of discussion about “non-porous” membranes and the possibility that they are actually porous.  By “non-porous”, I mean that a TEM of that membrane doesn’t have obvious pores by pore processing/our eyes.  This could mean a couple things: either the pores are so small that they are hard to resolve by TEM or there are bigger pores that are out of the TEM regions of interest.  A couple of months ago, I showed that H2O2 (34 Da) passed SC348, which pore processing identified as nonporous (here).  This is particularly interesting with respect to my cell culture studies because one of our hypotheses is that permeability drives vacuole formation in endothelial cells on pnc-Si transwells.  In this post, I showed vacuole formation on SC348, which surprised me because (at the time) I assumed SC348 was non-porous.  Since then, I’ve wanted to show that SC348 was actually porous, or at least porous enough for the cells to recognize and form vacuoles.  The post mentioned above was the first experiment to support this point.

I’ve had one of our high school students, Nikita, working on salt diffusion with pnc-Si transwells.  He has collected some data to support the H2O2 results mentioned above.  The graph below summarizes his data to date.  For these experiments, we set up a beaker filled with 200mL DIH2O and then floated a Sepcon with 600uL 2M NaCl in the apical well.  The DIH2O was stirred and conductivity measurements of this ‘filtrate’ were taken for about a day.  I labeled the last data point with the actual conductivity for clarity.

The blue diamonds and maroon squares represent conductivity of pnc-Si chips without membranes.  After a day, conductivity measured 1.7 uS/cm.  The purple X’s show NaCl transport through SC501 – a low porosity pnc-Si membrane – which showed a final conductivity of 559 uS/cm.  SC348, the ‘non-porous’ membrane, was in between these two samples, with a final conductivity of ~ 47 uS/cm.  However, these data are further evidence that ‘nonporous’ SC348 passes salts and is actually porous.