Diffusing Membranes

The idea is the development of an expendable "concentrator" of molecules. We want to provide an easy way to deposit molecules from a solution only to specific areas of a sample. This approach could be interesting for drug delivery, cell sorting and analysis.

The basic idea is to put a silicon thin membrane on a gold chip.

The membrane should be FIB milled to create some holes in it. At this point we drop water with a certain quantity of phluorophores on it. The solution will dry, passing through the holes and depositing there the molecules. Then the membrane will be removed, washed and be ready for another deposition, while the gold chip will be analyzed.

It is important that the drop doesn’t leak beneath the membrane. So we use two additional steps to avoid this phenomenon. The first one is to provide a layer of Teflon on both sides of the membrane, creating an hydrophobic barrier that pushes away the drop. Another help will come from a good adhesion between the membrane and the chip, achieved with the help of a vacuum chamber. It is enough to put the membrane on the chip and then insert them in the sputter machine. If the membrane doesn’t jump, it will adhere correctly.

The holes to be milled will have different diameters, from 3 to 15 µm, in order to understand which will be the minimum size to "bend" the water inside the wells. The gap between the holes should be around 5 µm. We performed some tests in order to select the best dimension and gap. A lower gap will provide an area with a bigger "equivalent hole" respect to a bigger gap.

Putting the chip on a hot plate at around 40-50 °C will increase the drying without affecting the "bleeding" through the holes.

For the first experiments we used two different types of beads:

1. Latex beads, carboxylate-modified polystyrene fluorescent yellow-green.
   λex ∼470 nm; λem ∼505 nm
   2,5 % solution, re-suspended 1:10 water. 30 nm size medium

2. Latex beads polystyrene
   10% Solution, re-suspended 1:200 water, 100 nm size medium

The first type of beads are fluorescent, so that we can have a quick analisys of the concentration under the UV-microscope. I deposited around 10 μl of solution on each membrane. Then I observed the first membranes at the fluorescence microscope. As expected, the first membrane showed a residual of particles on the borders, while the second had a residual exactly in the center, over the milled holes. So, looking the substrate underneath, I could see this result:

The leakage is evident but it can be improved. The second type of beads, instead, have been observed under the SEM, since the particles are not fluorescent. We changed the geometry of the hole to search for the optimum.

As proof of rpinciple, the amount of particles is huge, so ohter experiments were performed with picomolar solution.

Another approach is to use molecules instead of particles, marking the substrate with a fluorophore. We selected Rhodamine, Fluorescein and Cresyl Violet. Their concentration can be tracked with Raman imaging, so that any trace will appear. A good idea should be not to use gold on substrate for these tests, in fact, there could be quenching of the plasmons, the fluorescence could be adsorbed by the gold layer and lose visibility. However the gold provide a better marker to track the right spot in which there was deposition. Below we see an example of umolar deposition of cresyl violet.

These concentrator can be improved with better shapes, better adhesion or also with different approaches. Currently we are thinking about using PCL masks (polycaprolactone), that is a soft, reusable layer of polymer that we can adapt to our substrate, functionalize and then peel off. The process is the following:

OH^- radical. We have to take a piece of silicon or silica well cleaned and create OH radicals on it to enhance the adhesion of the following layer of silane. So we have to clean the sample with the usual Acetone, Isopropyl alcohol and deionized water, then a Piranha solution 1:3 fore 15 minutes, again a rinsing in water (two times). When the sample is well cleaned, we expose it to an O₂ plasma at 100 W for 120 seconds. Now the surface of the samples will have OH^- radicals.

Silanization. The goal of silanization is to create a hydrophobic surface on the samples, so that it will be easy to remove the PLC mask. So I drop 10-20 µl (depending on the sample dimensions) of Trichloro (1H,1H,2H,2H-perfluorooctyl)silane (448931-10G Sigma Aldrich) on the samples, I put them in a glass flask covering the top and I insert them in a oven at 100 °C for 30 minutes. Then I put the samples on a external hot plate at 100°C for 30 minutes. In the end I will have the sample silanized.

PLC spinning. I spin a layer of PLC on the sample. The PLC is put in Tetrahydrofuran (THF) in different concentration, from 5 to 30%; the resulting layer thickness will increase for bigger concentration, from 1 to 30 μm. It is important to know the thickness, since we have to know the etching time in the end of the processing. To obtain a final thickness around 30 μm, we spin a 17% solution at 750 rpm. After pre-baking the layer at 60° C for 3 minutes, we deposit 100 nm of Chromium by sputtering.

Chromium mask. We have to create a proper mask for the chromium layer, in order to etch small holes in the PLC. So we spin on the sample a layer of SPR 220 4.5 at 5000 rpm and we pre-bake it on a hot plate at 115°C for 90 seconds. We will obtain a thickness of around 4 µm. Then we put a pillar mask and we expose the sample with soft-contact for 15 seconds. We develop the SPR in MF-319 for 60 seconds and we obtain holes exposing the chromium underneath. Putting the sample in the Chrome etch for 1 minute will remove the metal from the holes. Then, to remove the unexposed SPR, we do an oxygen plasma at 100 W for 1 minute and we repeat 3 times.

ICP-RIE. We do a dry etch with SF6 and Ar to create a rough surface.

At this point the PLC layer is ready and can be used as a sort of reusable inkjet mask for the concentration of analytes.