Team:Paris Bettencourt/Notebook/Assay
Week 4th - 10th July
Wax impression on cellulose paper
We used candle wax (vanilla scented) from Franprix (Supermarket in France). Wax was melted at 95 degrees. The 96 well plate template was used to create wax impressions on the cotton. Since wax is hydrophobic and the cotton, we used, has hydrophobic coating, it spread quickly, the diffusion was fast which makes this cotton futile to carry out experiments or assay development. Using hydrophobic cotton in the development of assay is merely not pragmatic for a reliable assay.
Cellulose chromatography paper was used to carry out further trials for assay development. The template dipped in melted wax is pressed against the paper for sometime (1-5mins). We created successful impressions of wax on the paper where the wax penetrates vertically into the paper and creating specific boundaries. At times, we had to reheat the template pressed against the paper to further improve the penetration (but no longer than 10 secs at 100degrees). We could successfully recreate the assay on the paper. The wells are not uniform given the impression and dipping the template were done manually.
Anthocyanin of volume 20μl, 10μl, 2μl and 1μl were used to test the wells. The Anthocyanin did not spread and they were contained in the wells. We think 10μl can be used to perform experiments on cellulose paper. For experiments on assay, we need cotton which is untreated and unbleached in nature.
Week 18th - 24th July
Wax impression on cotton
We are still counting on wax, we have sent emails to the Musée Grévin, wax specialist, and iMakr that seemed to have printed candles in the shape of celebrities. We await their response.
Scandles, 3D printed candles
In the interim, we used the CNC (Computer Numeric Control) on a block of wax we've found in the Openlab of CRI-Paris, to make 96 wells. The aim is to heat it and press it against cotton fabric to create wax impression of wells in which we could pour liquid without having crosstalk between wells. We made the model in 3D to be fed into CNC and it really worked well, wax is soft so it was easy to make the wells using a drill bit but still quite long, approximately 3 hours. The diffusion of wax on chemically untreated cotton is very fast even though cotton is hydrophilic this time and wax is hydrophobic.
Two component mold
We made up our mind to make a mold in order to pour wax so that it would create circles of wax on cotton when the cotton is placed beneath the mold. We designed the three dimensional mold in Solidworks to 3D print it. We use Makerbot Replicator 2x 3D printer in Openlab of CRI-Paris to print our molds. We first did it in one part and the 3D print was not possible because some parts didn't have support and hangings are not typically supported by 3D printer. Therefore, we then apportioned the design into two parts which we 3D printed individually. The way 3D printing works is lateral deposition. It does not have shear strength. Shear failures and delamination failures are typically observed. The pillars in the model were very tall, and consequently very fragile, two of them broke easily as one of our teammate played with it. The base of the pillars against which we press the fabric are not uniform. Therefore, we cannot get uniform circles and avoid crosstalk when we pour wax through the mold.
Week 25th -31th July
This week we had a good talk with our advisor, Edwin, and realised together it would be much more interesting if we created an actual 96-well plate, that would be cheap and standardised to test products directly on fabric. So we are going to change our way of thinking from 2D to 3D.
Using 3D printed wax on fabric
First we still are looking options with wax as we saw wax can be 3D printed, indeed we've found a wax filament. To know more about it we've planned a skype call with Arthur Dalaise, one of our teammate's cousin that is funding a 3D printer service startup. The good news is that he agreed to try out printing with this wax filament for free and gave us really good advice !
Our only concern with 3D printing is that we don't know if wax will stick to the fabric, not burn it and make real wells that don't allow liquid to spread.
Using PDMS to create our own 96-well plate
Teja had a revelation and thought of using PDMS, a silicone-based polymer, to create a 96-well plate where we would sandwich cotton fabric between two layers of PDMS with wells. The bottom side of PDMS is then bonded to glass using a plasma cleaner so that the fabric can be scanned using a flat bed scanner.
Our first try shown in the above figure looked promising. The top layer of PDMS is prepared with 10:1 ratio of silicone polymer to curing agent. The bottom layer of PDMS is prepared with 5:1 ratio of silicone polymer to curing agent. Once the mixture is prepared, it is degassed twice to remove bubbles and then heated at 75 degrees for two hours in a convective heater for curing. PDMS is punched to create the wells using a 6mm diameter puncher. The punching went well but not the spacing between the wells. Manual punching resulted in non uniform spacing. The bottom layer of PDMS is first bonded to the glass using a plasma cleaner. A small layer of 1:5 mixture that was first prepared is applied to the bottom layer of PDMS. A The cotton fabric is then sandwiched between two layers of PDMS with wells. The alignment of wells is not precise with the fabric sandwiched between PDMS layers. Then the whole sandwich is heated at 75 degrees for one hour in a convective heater for curing. Although there is a strong bond between cotton and PDMS, the PDMS got diffused into the wells which hinders the use of fabric in the microplate. We are in pursuit of finding new ways to have a precise result.
A 3D printed mould for PDMS
We decided to 3D print a mould with pillars to pour PDMS on it and have a layer of PDMS with wells in it. The pillars are 1cm high but half of it is slightly conic so that we can remove the PDMS easier. The temperature of the platform of 3D printer is 120 degrees and the temperature of the extruder is 220. We used ABL filament for 3D printing.
So this mould didn't work as well as planned ... The PDMS layer is really hard to remove from the mould and we broke both the PDMS and the mould while removing it. But the wells created on the parts that were not broken are really nice so we will do another try with a different mould with pillars that won't be as high as this one.
Week 1st - 7th August
A new try for PDMS' 3D printed mould
On this picture you can see our new 3D model with smaller pillars the total height is around 6mm. Each of pillar comprises of two sections one bottom cylindrical section and one top conical section. The perfect cylindrical bottom section is 6mm in diameter and 2mm in height. The height of the top conical section is 4mm and the top diameter is 4mm. We are very confident with this model.
And this is the result of the 3D printing that went really well this time, we did a 80% infill so that the pillars don't break this time. The result of the PDMS is on the left, it is really what we expected and we are really happy with this. We created a 96 well PDMS!!
Lasercutting Cotton Fabric
We also decided to laser-cut cotton, our research on internet have shown it is already done and seems very efficient. We used the laser cutting machine in the openlab, CRI-Paris. The minimum width of the cotton needs to be 2mm for precise cut. We want to have circles of cotton that are linked so we don't have to use a whole piece of cotton fabric on PDMS. Besides that, we want to minimize diffusion through cotton as much as possible. It worked really well and we managed to lasercut multiple layers at the same time (around 8).
The borders of the lasercut cotton is then a bit burnt, which was totally expected but we fear it may interfere with the biological experiments. So we tried soaking it in water and it disappeared instantly as water dissolves the burnt part but the fabric shreds slightly after being wet.
Combining the different layers
The cotton seems to fit really well, but we tried to stick the PDMS, cotton and glass layers on a small scale and it was not very good yet, the liquid spreads because the cotton is creating a small gap between the glass and the PDMS. The gap (thickness of the cotton) is sufficient enough for crosstalk.
3D printed wax on cotton
Thanks to Arthur Dalaise who helped us we have a functional assay using 3D printed wax filament on cotton. The first layers were printed with heated platform (100°C) with a 190°C temperature for the extruder and then the sheet of fabric was laid on, and finally the rest of the layers were printed on top of this with a unheated platform and a very hot extruder (240°C). The first step lasts for 30 mins and the second one for 6 hours. So the result is exactly what we want but the time it takes is really too long, so we will still continue the PDMS experiments.
Week 8st - 14th August
Crossroads design
Lasercut cotton when inserted between flat PDMS surface (with 96 wells) and the glass creates a gap between the surfaces which results in the diffusion of fluids from one well to the other. This is due to the high diffusion properties of cotton and the thickness of cotton. Adding to this, there is a possible misalignment of wells and the cotton fabric which might result in measurement problems. To avoid this, we designed 3D printed mold in such a way to easily insert the lasercut cotton directly without any misalignment and diffusion. The mold has extrusions which matches the dimensions of laser cut cotton with good tolerance. When PDMS is taken out of the mold after curing, PDMS will have the grooves (negative of extrusions) to allow us to keep the cotton safely inside without creating a problem of gap or misalignment.
There is a potential problem of diffusion through the grooves which can be avoided, if we can fill the connection between the wells using pdms or hydrophobic material. Having said that, filling the connections with pdms or hydrophobic material may eventually result in the bumps which does not allow us to bind PDMS to glass using Plasma cleaning.
Week 15st - 21th August
PDMS Microplate
Since cotton is highly diffusive, we came to a conclusion that we need to cut circles of cotton for each well. Each cotton circle should be kept in a PDMS well firmly. PDMS should then be bound to glass using plasma cleaning. Cotton should be scanned using the glass face. Cotton should not come out of the microplate when flipped. The whole process of producing this microplate should be automated.
The bottle neck of whole process is how can we cut 96 cotton circles and place them in PDMS wells. If performed manually, this process cannot be automated. We need a cheap, reliable solution which can be automated. We tried laser cutting cotton circles of diameter 6mm, 6.5mm, 7mm on a glass plate. Of course, they are light and they did not stay still. By sheer serendipity, we got an idea of wetting the cotton and repeat the process. It was splendid. When we lasercut wet cotton placed on a glass plate, if you remove the fabric after the process, the circles would remain in their respective positions. They adhere to the glass plate possibly due to the surface tension of water and cotton absorbs water which is more than twenty times of its own mass. But once you let the cotton dry, there is a high possibility of disorderedness of cotton circles. So, the cotton circles on the glass must be immediately used. This process can be easily automated to create the circles of cotton in 96 well plate format.
The 3D printed mold has 96 pillars in the following design. Each Pillar has two sections, the bottom section is a cylinder of diameter of 7mm and height of 1mm. The top section is a cylinder of height 5mm and diameter of 5mm. All the sharp edges of the pillars are filleted and chamfered with 0,5mm. The PDMS microplate made from this mold has 96 wells with two sections which are negatives of the sections mentioned above. The idea is to bind the glass with 96 pieces of cotton to the PDMS microwells using plasma cleaning. We were afraid, if presence of cotton makes plasma cleaning problematic. But instead, the cotton stayed at their respective positions during the process of plasma cleaning. When we tried to bind them, we came across a strange but fundamental problem of binding two surfaces using plasma cleaning. The surfaces we bind should be flat for efficient binding which is not the case as the PDMS is taken out of 3D printed mold. 3D printing does not produce perfectly flat surfaces needed for plasma binding. So, we could not bind the surfaces using plasma cleaning. We thought it might be because of the presence of cotton. We repeated the same trial without using cotton, the result was same as before. We could not bind the PDMS and the glass using plasma cleaning as the surface of the PDMS is not flat which makes the process inefficient.
Week 22st - 28th August
PDMS Microplate 2
In the previous week, we tried to use plasma cleaning to bind the surfaces. We got an idea of binding them using a fluid glue which solidifies quickly after bringing the surfaces to be bound. We successfully could bind the PDMS surface with glass using the glue without any diffusion from one well to the other. In this experiment, we tried to bind PDMS surface and glass without any cotton to check the proof of concept. There was leak at the borders due to manual gluing.
Week 29th August-4th September
PDMS Microplate 3
We implemented gluing PDMS 96 well plate to plastic/glass transparent bottom using a glue with cotton circles of 6.5mm in the PDMS wells. Once the cotton is trapped in the PDMS wells, cotton cannot escape the well because of the two different sections. Manual gluing did not give desired results. One must not forget that the PDMS is not perfectly flat. it might also be reason for diffusion. There were leaks almost everywhere.
Plastic Microplate
We found a plastic analogue of the PDMS design we made. Plastic has perfectly flat and smooth surface therefore, we expunged the influence of diffusion due to roughness of PDMS. We attached the plastic analog to transparent bottom plate using an adhesive. The plastic analog once attached with the bottom plate can trap the cotton circles efficiently. We could see clear diffusion from well to well another. In the process of solidification of the glue, it left some gaps which are clearly responsible for diffusion.
Week 5th September-11th September
Automated 96 well plate of fabric
We got an idea of using fluorescent 96 well plate, placing cotton circles inside the wells and gluing the tip holder(which is yellow in the images). This way there is no diffusion, cotton is trapped inside the wells, as the well's inner diameter is bigger than cotton circle diameter of 6.5mm and the circles of tip holder are less than 6.5mm. This is an arrant analog of plastic microplate we tried earlier and the pdms microplate.
Dilution test
We did a a dilution test to see if we can clearly witness the change in the intensity of the wine of different dilutions in the Microplate. The results are terrific. We could find the difference in intensity of wells from each other and between before dilution test and after dilution test.
+33 1 44 41 25 22/25
igem2016parisbettencourt@gmail.com 