Archive for the ‘Fabrication’ Category

Parallel fabrication of an array of holes in PDMS

D. Qi, C. K. Chan, and A. C. Rowat*
Department of Integrative Biology and Physiology, University of California, Los Angeles, 610 Charles E. Young Drive East, Los Angeles, CA 90095, USA
* e-mail: rowat[at]ucla.edu

Why is this useful?


A major challenge in microfluidics is interfacing micron-scale flow channels with fluid samples and peripheral equipment of the macroscopic world, such as plate readers and liquid-handling robots. To exchange fluids in and out of the device, inlet and outlet holes are essential. In polydimethylsiloxane (PDMS) microfluidic devices, small <1 mm inlet- and outlet-holes are typically used to interface fluidic samples and micron-scale channels. These holes are typically fabricated by manually excising PDMS slabs to produce holes, for example using a needle or razor-sharp biopsy punch. However, to achieve scale-up of microfluidic devices can require ~10^2 inlets and outlets required for parallelization of multiple samples.

Manually punching holes is tedious due to the following factors: (1) it is often difficult to determine where to punch holes due to the micron-scale dimensions of the channels and transparency of PDMS; (2) the resulting holes do not always have smooth edges, and resultant chunks of PDMS can clog micron-scale channels; (3) biopsy punches used to excise holes become dull after multiple uses and thus frequently need to be replaced; (4) it is difficult to manually punch holes that are vertically straight and have reproducible spatial position; this makes it challenging to achieve the precision required for fabricating devices that robustly interface with existing high throughput equipment, such as liquid-handling robots. Here we present a method to easily and reproducibly fabricate smooth, vertically straight arrays of interfacing holes in PDMS using standard arrays of 96 or 384 pipette tips.

What do I need?


  • Common items to prepare PDMS devices: petri dish, master wafer or device mold, PDMS silicone elastomer (base and curing agent), vacuum desiccator, and oven.
  • Micropipette tip racks. Depending on microfluidic device designs, users can build their own tip racks; for example, we machined a 96-tip holding rack out of Perspex (polymethylmethacrylate).
  • Micropipette tips, e.g., 10-, 20- or 200- μl volume.
  • Four 3-inch steel bolts with nuts.

What do I do?


  1. Assemble tip racks using four steel bolts with nuts (see Figure 1.); the bottom rack is fixed and the top one is adjustable vertically along the bolts.
  2. Put the petri dish (with the master wafer inside) on the bottom rack.
  3. Position pipette tips in the top rack.
  4. Align tips to the inlet- and outlet-hole positions of the desired microfluidic channels; move the top rack down until the tips physically touch the master wafer.
  5. After the tip alignment, put a flat light weight object, e.g., a Perspex block ~155g, on top of the tips to hold them in position, while also making sure that tips and the master wafer are in physical contact.
  6. Prepare PDMS silicone elastomer (base-to-curing agent ratio-10:1) and mix well by vigorous stirring.
  7. Carefully pour the silicone elastomer mixture into the petri dish.
  8. Put the entire setup in the vacuum desiccator and degas for 20-30 minutes.
  9. Move the setup into an oven and thermally cure silicone elastomer at 65 degree Celsius for >1 hour.
  10. After curing, remove the setup from the oven, and let cool to room temperature.
  11. Using tweezers, gently pull out tips from the thermally cured PDMS.
  12. Using a razor, cut out the PDMS device with molded channels and holes from the petri dish.
  13. Bond PDMS to glass coverslips or other substrates, as required for your application.

Figure 1
Figure 1. Images of hole-molding setup with (a) standard pipette tip rack, and (b) custom fabricated Perspex holder. Images of the molded holes in PDMS are shown in (c) and (d); inset in (d) shows one micropipette tip with tiny PDMS plug at its tip.

Figure 2
Figure 2. Typical inlet/outlet holes in a PDMS slab (a) molded by our molding method using a 200 μl micropipette tip; and (b) punched by a 0.75mm Harris Uni-Core (TED PELLA) punch. The PMDS slab is patterned with 50 μm cylindered posts.

What else should I know?


  • The arrangement of tips can be tuned for specific applications by using a custom-fabricated pipette tip rack to accommodate unique inlet- and outlet-hole arrangements; such as the Perspex rack as shown in Figure 1. (b).
  • During molding, a small amount of silicone elastomer is taken up into the micropipette tip ends during degassing in the vacuum chamber (Fig. 1d, inset); this occurs due to the imperfect sealing between the micropipette tips and the master wafer. This tiny plug of PDMS is easily removed from the device together with the pipette tip, resulting in holes that are free of any obstructions.
  • When using tubing to interface microfluidic devices with macroscopic equipment, the outside diameter (OD) of the tubing should be larger than the diameter of such mold holes in PDMS to avoid leakage. We have successfully used both 20- and 200- μl micropipette tips to mold holes, which are ~0.7 mm in diameter at the end; 10 µL pipette tips have similar diameter. Both 1⁄32″ and 1⁄16″ OD tubing (VICI Valco Instruments) interfaces well with holes of these dimensions.
  • By our proposed molding method, the size of the holes closely follows the outer diameter of the pipette tips; the holes do not contract after the tips are removed since they are formed during PDMS thermal curing. As shown in inset of Figure 1. (d), the PDMS plug in the micropipette tip is tapered, which helps to produce clean holes when the micropipette tip is pulled out of PDMS.
Digg This
Reddit This
Stumble Now!
Share on Facebook
Bookmark this on Delicious
Share on LinkedIn
Bookmark this on Technorati
Post on Twitter
Google Buzz (aka. Google Reader)

Facile and plasma-free bonding of PDMS

J. Waynelovicha, Da’Kandryia Petersb, P. Salamona,  A.M. Segallb, Nicholas Sam-Soonc*
aDepartment of Mathematics & Statistics, San Diego State University
San Diego, CA 92182
bDepartment of Biology, San Diego State University
San Diego, CA 92182
cDepartment of Engineering, San Diego State University
San Diego, CA 92182
Email: nsamsoon[at]gmail.com

Why is this useful?


Microfluidic-based approaches have become widely adopted for analysis of cells, molecules, reactions, and processes. While plasma bonding has become the de facto way of bonding PDMS to other substrates, this method is expensive and, does not in fact work (without further additional steps) on plastics [1]. Here we show a simple way of using cheap and readily available materials to irreversibly reseal leaking plastic connections and bond PDMS to itself and other plastics.

What do I need?


  • PDMS (Polydimethylsiloxane) arts / microdevices
  • Loctite Plastic Bonding System (two part cyanoacrylate mix)
  • Plastic / silicon rubber based tubing

What do I do?


The procedure can be broken down into two steps; the priming of the substrates, and the application of the glue.

1) Priming
The end of the tube which contacts the leaking part is taken out and roughened up with sand paper. The activator is then applied on the around the surface as well as on the top of the device and left to dry for 30 seconds. (Figures 2A & B).

2) Glue application
The tube is reinserted into the hole and glue is then quickly smeared around the tubing and hole. (Figure 3).  Excess glue can be removed with an exacto knife. Allow at least 10 minutes to dry.

Figure 1

Figure 1, A) Loctite Plastic Bonding System B) Droplet formation due to poor seal at tube.

Figure 2

Figure 2. A) Applying the activator to the end of the tubing and B) to the top of the device.

Figure 3

Figure 3. Application of glue and bonding of the tubing to the part.

Figure 4

Figure 4. Bonding PDMS to A) PDMS B) Glass

What else should I know?


Loctite Plastic Bonding System can also be used to bond PDMS to PDMS (Figure 4A). This allows the creation of thick mounting blocks on thin devices providing a more durable friction fit.
If the device is to be used in a biological capacity, adequate time should be allowed to ensure complete polymerization of the glue. This along with flushing the device with buffer prior to using should minimize the possibility of any toxicity from unpolymerized material.
This system can also be used to bond PDMS to glass, PET and possibly other plastics. (Figure 4B)

References


[1]   Lee, Kevin S, and Rajeev J Ram., Plastic-PDMS bonding for high pressure hydrolytically stable active microfluidics. Lab Chip, 2009, 9, 1618-1624.

Digg This
Reddit This
Stumble Now!
Share on Facebook
Bookmark this on Delicious
Share on LinkedIn
Bookmark this on Technorati
Post on Twitter
Google Buzz (aka. Google Reader)

A stacked microfluidic device for improving experiment throughput

Jiandong Wuab, Xun Wubc and Francis Linabcd*
a Department of Biosystems Engineering, University of Manitoba, Canada
b Department of Physics and Astronomy, University of Manitoba, Canada
c Department of Immunology, University of Manitoba, Canada
d Department of Biological Science, University of Manitoba, Canada
Email: flin[at]physics.manitoba.ca

Why is this useful?


Owing to the advantages in miniaturization and cellular microenvironmental control, microfluidic devices have been increasingly applied to cell biology research [1]. Particularly, microfluidic devices can precisely configure chemical concentration gradients and flexibly manipulate the gradient conditions in space and in time [2, 3]. Various microfluidic gradient-generating devices have been used for studying cell migration and chemotaxis [2, 3]. These studies rely on live cell microscopy and usually only one experiment can be performed at a time. Previously, a double gradient device was demonstrated for parallel cell migration experiment with a motorized stage to image cells in different gradient channels [4]. However, the full XYZ motorized stage is expensive and thus often times limits the practical use of high-throughput microfluidic devices.

To overcome this limitation, here we report a stacked microfluidic device that allows parallel live cell imaging experiments on a single chip with only a Z motorized stage. This device is fabricated with multiple stacked layers of PDMS devices, and the cell imaging channels in each layer are aligned so they all fit into a single microscope viewing field. Thus, by only adjusting the vertical focus using a Z motorized stage, multiple cell channels can be imaged repeatedly over time. If a full XYZ motorized stage is available, the throughput of the stacked device can be further increased along horizontal dimensions. Making a stacked device is straightforward and this strategy can be useful for improving experiment throughput, especially in a limited microscopy facility.

What do I need?


  • Two identical PDMS chips and a glass coverslide
  • An oxygen plasma cleaner
  • A fluorescent microscope equipped with a programmable motorized Z stage or a full XYZ motorized stage, and a CCD camera

What do I do?


  1. Fabricate your SU-8 mask using standard photolithography. We used a simple ‘Y’ shape design with a 350µm wide main channel.
  2. Make two PDMS replicas of the same design from your SU-8 masks using a standard soft-lithography method.
  3. Bond the two PDMS chips using O2 plasma treatment. The channels should all face down and the main channels in the two chips should be vertically aligned. To avoid overlap of the inlets, we align the two layers so that their inlets are separated by some distance, while the main channels of the two layers are still within the same viewing field in the microscope. We used a 10X objective in our experiment, and this strategy works well with a 350µm channel. However, this method may not work if higher magnification is required, and it may also cause shifted gradients in the two layers. As an alternative strategy, we can align the two layers perfectly along the vertical direction, but punch inlet holes for the two layers with different distance relative to the ‘Y’ junction so the inlets of the two layers can be separated (this will require punching inlet holes for the top layer before plasma bonding). Two masters with different inlet designs can also be used to separate inlets of the stacked device.
  4. Punch the holes for making fluidic inlets and outlets.
  5. Bond the double-layer PDMS chips to a glass coverslide using O2 plasma or air plasma treatment to complete the simple stacked device (Figs. 1A & 1B).
  6. Image the channels and cells using a microscope equipped with a Z motorized stage.
  7. We used food coloring to show the channels in the two layers of the stacked device (Fig. 1B). Furthermore, we show that we can generate chemical gradients in each layer of the stacked device by mixing buffer and FITC-Dextran as well as imaging cells loaded to the main channel of each layer only with a Z motorized stage (Fig. 2).
  8. Finally, using a different design that consists of two gradient channels in each layer and a full XYZ motorized stage, we demonstrate that four individual channels can be imaged in the stacked double-layer device. Again, food coloring is used to show the channels in the upper layer and the bottom layer (Fig. 1C).

Figure 1
Fig. 1. Illustration of the stacked microfluidic device. (A) The schematic drawing of the stacked double-layer device that consists of 2 identical ‘Y’ shape channel; (B) A real picture of the stacked device of the same design. Food coloring is used to show the channels in the 2 layers. (C) A real picture of the stacked device of the design that consists of 2 gradient-generating channels in each layer. Again, food coloring is used to show the channels in the 2 layers.

Figure 2
Fig. 2. Gradient generation and cell images in the double-layer stacked device. (A) Gradient of FITC-Dextran 10kDa generated in the bottom layer channel using the ‘Y’ shape design, and images of Jurkat cells in the same channel. (B) Gradient of FITC-Dextran 10kDa generated in the upper layer channel using the ‘Y’ shape design, and images of Jurkat cells in the same channel. Only a Z motorized stage is used for the imaging.

What else should I know?


Here we demonstrate the simple double-layer stacked device. More layers of PDMS can be stacked to further increase the throughput. In addition, in the current demonstration, the channel in the bottom layer is formed between PDMS and a coverslide while the channel in the top layer is formed between PDMS. If needed, the double-layer chip can be first bonded to another piece of PDMS before bonding to the glass substrate. This way, both layers of channels are in PDMS for consistency.

References


[1] G. B. Salieb-Beugelaar et al., Latest developments in microfluidic cell biology and analysis systems. Anal. Chem., 2010, 82, 4848-4864.
[2] S. Kim, H. J. Kim, and N. L. Jeon, Biological applications of microfluidic gradient devices. Integr. Biol., 2010, 2, 584-603.
[3] J. Li and F. Lin, Microfluidic devices for studying chemotaxis and electrotaxis. Trends Cell Biol., 2011, 21, 489-497.
[4] W. Saadi et al., A parallel-gradient microfluidic chamber for quantitative analysis of breast cancer cell chemotaxis. Biomed. Microdevices, 2006, 8, 109-118.

Digg This
Reddit This
Stumble Now!
Share on Facebook
Bookmark this on Delicious
Share on LinkedIn
Bookmark this on Technorati
Post on Twitter
Google Buzz (aka. Google Reader)

A refinement of a method to prevent sagging during the bonding or lamination of chips with high aspect ratio chambers

Brian Miller, Stewart Smith and Helen Bridle*
School of Engineering, University of Edinburgh, 3.17 William Rankine Building, Kings Buildings, Edinburgh, EH9 3JL, UK
Email: h.bridle[at]ed.ac.uk

Why is this useful?


Jie Xu and Daniel Attinger previously described a method to prevent the sagging of high aspect ratio channels during bonding [1]. The method involves careful placement of salt crystals into the channel prior to bonding to create a supporting structure. A limitation of the technique is the depth of channels this method can be used on, which must be within the size range of the salt crystals (~100 μm).

Described here is a method that builds on this technique to allow salt structures to be created with much smaller surface profiles, down to between 25-35 μm maximum heights of profile. This will allow the technique to be applied to shallower channels for devices in which performance is sensitive to channel height, for example, inertial focusing devices (Fig.1) [2,3].

This refined technique also helps to simplify the handling of the devices after preparation; reducing the risk of contamination of equipment as, once they are applied, the salt crystals are adhered to the surface of the device.

Inertial focusing devices
Figure 1. Inertial focusing device A; with solution applied to high aspect-ratio sections as indicated in pinched-flow segment B and output leg C. Device depth is 30μm (max. aspect ratio 60:1 in PDMS)

What do I need?


  • High purity KCl (potassium chloride salt)
  • De-ionised (DI) or reverse osmosis (RO) water
  • Weighing balances/scales
  • Decon 90 surfactant
  • Hypodermic needle (thin gauge)
  • Magnetic stirrer
  • Beaker and syringe

What do I do?


  1. Prepare a solution of the salt and surfactant by measuring 100ml of DI or RO water into the beaker. Add 1.5g of KCl to the water and stir gently for two minutes. Use the syringe to add 5ml of Decon 90 to the solution and stir very gently for a further two minutes, taking care not to foam the solution. Do not allow the solution to rest for more than a few minutes after stirring.
  2. Bending the sharp point of the hypodermic needle on your workbench or other hard, clean surface can help to ‘grab’ sufficiently small quantities of solution from the beaker. Use the needle or a similar applicator to carefully apply a small quantity of solution to the high aspect ratio section of your channel (Fig. 2). Only a very small volume is required and it is important to not allow the solution to overflow the channel. Dabbing the needle on a dust/lint-free wipe can help to regulate the amount of solution delivered to the surface of your device. If you accidentally over-apply the solution, clean the device with DI/RO water and IPA, and re-attempt application once it has dried.
  3. Allow the solution to evaporate at room temperature without assistance of any kind (no added air-flow or heat). An area of crystal formation should form where the solution was applied. The edges of this area tend to grow larger due to the ‘Marangoni effect’. The edge will typically be a maximum of 30-35μm in height, with the centre of the area typically yielding crystal formations between 10 and 25μm tall (as measured on a surface profiler over 4 repetitions), which should suffice to prevent the unintended bonding of the ceiling to the floor of the channel.
  4. Bond your device to your substrate layer following your normal bonding procedure (we used oxygen plasma bonding of PDMS to a glass microscope slide in this example).
  5. Before using the device, run water or buffer through to dissolve away the crystal formations (Fig. 3). The surfactant helps to quickly remove the salt structures and leave the device clean of any remnants.

Application of solution into channels
Figure 2. Careful application of very small quantities of solution into channels, using a hypodermic needle

Support structures before and after rinsing
Figure 3. Top: water salt formations in the support structure before rinsing with DI/RO water. Bottom: the same area after rinsing through, illustrating that practically no residue is left in the device

Notes


  • Failure to use the surfactant will result in much larger crystal formations, as the crystal structures nucleate in very few locations and form much deeper structures. It is conjectured that the surfactant suspends the salt ions with a larger interspaced distribution throughout the solution, causing a much more distributed nucleation of the crystals and yielding the lower profile crystal formations.

References


[1]  Jie Xu and Daniel Attinger, How to prevent sagging during the bonding or lamination of chips with large aspect ratio chambers, Chips & Tips (Lab on a Chip), 24 July 2009.
[2] D. Di Carlo, D. Irimia, R. G. Tompkins and M. Toner, Continuous inertial focusing, ordering, and separation of particles in microchannels. Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 18892-18897.
[2] D. Di Carlo, J. F. Edd, D. Irimia, R. G. Tompkins and M. Toner, Equilibrium separation and filtration of particles using differential inertial focusing. Anal. Chem., 2008, 80, 2204-2211.

Digg This
Reddit This
Stumble Now!
Share on Facebook
Bookmark this on Delicious
Share on LinkedIn
Bookmark this on Technorati
Post on Twitter
Google Buzz (aka. Google Reader)

Easy and inexpensive fabrication of PDMS films of different thicknesses

Rodrigo Martinez-Duarte
Microsystems Laboratory (LMIS4), École Polytechnique Fédérale de Lausanne, Station 17, CH-1015 Lausanne, Switzerland

Why is this useful?


The following tip describes an easy and inexpensive way to fabricate PDMS films of different thicknesses. The main advantages are that no infrastructure (e.g. a spin coater) is needed for fabrication and that the materials needed are readily available. The idea here is to use a film of a specific thickness as spacer between two plates. PDMS is first deposited on one plate and squeezed in between plates to get a PDMS film with thickness similar to that of the film used as spacer. A similar methodology could be used to make films of other materials as well.

This methodology allows for very quick fabrication of a wide range of PDMS films using off the shelf components which are common in a traditional laboratory and office or can be easily purchased. The final goal is to fabricate films which feature different thicknesses, according to the original film used as spacer, and different hardnesses, by changing the ratio of PDMS to cross linker. The film can be stored and used as needed to cut off parts such as gaskets, spacers, etc. Here in the lab we use them to make microfluidic chambers of specific thickness.

What do I need?


  • 2 rigid plates with at least one flat surface per plate. They can be glass, PMMA (>3-4 mm thickness) or other material as long as they are rigid, preferably a fair thermal conductor and do not soften at temperatures up to 100°C. The use of transparent plates is not necessary but recommended to monitor the squeezing process of the PDMS and minimize the presence of bubbles in the final PDMS film
  • Pieces of a film, which can be tape, pieces of a plastic bag or any film that does not compress or  absorbs PDMS, with similar thickness to that being targeted for the PDMS film
  • Paper clamps, binder clips
  • General purpose soap or Mylar® film, the Mylar® film can be replaced by any film to which PDMS won’t adhere
  • And of course, PDMS and the equipment recommended to process it: balance, degasser and oven. Although this equipment is recommended, it is not necessary, as already suggested by Aung K. Soe and Saeid Nahavandi in a previous Tip [1]

Where do I get it and how much will it cost me?


In principle all the materials required are already available in the lab. The two plates can be two pieces of glass, as simple as two glass slides or two old wafers, or pieces of plastic, polycarbonate or PMMA for example. I have used PMMA plates and old glass wafers. The cost of the plates can be minimal and should not be more than $10. For example, you could go and buy a couple of very cheap picture frames with a glass piece of the size you want and use those. The pieces of film you use depend on the final thickness of the PDMS film and they can be obtained from a variety of sources, use your imagination! The paper clamps or binder clips are usually available in the office, if not you can buy them for several cents apiece.

What do I do?


  1. Wear gloves before starting. PDMS can be messy! The materials needed are detailed above and shown in Fig. 1.
  2. A. Dip your rigid plates in soapy water for a couple of minutes to deposit a layer that prevents PDMS from adhering to the plates. Remove the plates and blow them dry or let them dry naturally, do not wipe dry!
    Or alternatively:
    B. You may want to use a Mylar® film in between the rigid plate and PDMS. Take care not to scratch the Mylar® film. The rationale behind using this Mylar® film is to eliminate the need for surface treatment of your plates to prevent adhesion of the PDMS to them. PDMS does not stick to Mylar® and therefore you can easily just peel the Mylar® film off after you fabricate your PDMS film. A further advantage is that you don’t need to worry about scratching your plates or about cleaning the plates each time you use them. Storing your PDMS film between Mylar® films can also help protect it against scratches.
  3. Prepare your PDMS. Different ratios between the polymer and the cross-linker give different hardnesses. A 20:1 mix results in a gluey, soft film that easily sticks to a variety of materials. A traditional 10:1 results in a harder film. 5 grams of mix should be plenty for a couple of films featuring an area of 8 by 8 cm and thickness of up to 200 µm.
  4. A. If you treated the surface of the plates with soap: after the plates are dry (but still with the dried soap layer on them), you can position your spacer film close to the edges of the plate as shown in figure 2. Do this on only two edges to allow plenty of space for the PDMS to flow out from between the plates during step 7.
    B. If you are using the Mylar® films: cut the film into a couple of pieces similar to your plates. You can then position the spacer film close to the edges of one of the Mylar® films. Leave some space in between the spacer films to allow PDMS to flow out while pressing in step 7.
  5. After the PDMS is well mixed, manually deposit the PDMS mix on one plate treated with soap, or on the Mylar® film positioned on top of the plate, as shown in figure 3.
  6. Degas the previously deposited PDMS until the mix looks homogeneous (around 40 minutes in a general purpose degasser).
  7. Remove the arrangement from the degasser, place it on a flat surface and use the other plate to squeeze the PDMS in between the plates if you are using the soap-treating approach. If you are using the Mylar® films, then first lay the second film on top of the PDMS and then squeeze with the other plate. In both approaches, it is recommended to start applying pressure in one side and work your way towards the other side to avoid introducing bubbles in the PDMS.
  8. Use the paper clamps to clamp A) the plate-PDMS-plate or B) plate-Mylar®-PDMS- Mylar®-plate sandwich together at the location of the spacers as shown in figure 4.
  9. Bake for 1 hour at 80°C.
  10. Remove from oven and let cool for 5 min.
  11. A. If using soap: use a knife to remove the PDMS accumulated on the edges of the plates. This is important because it will facilitate the release of the PDMS film in the next step.
    B. If using Mylar® films: unclamp your sandwich and retrieve your PDMS film in between the Mylar® films. You are done!
  12. If using soap-treated plates: slowly but firmly separate the two plates, the PDMS film is likely to remain on one plate from where you can peel it off as shown in figure 5. Make sure you do it slowly to avoid rupturing the film. You may use fine tweezers to aid you during the release.
  13. You are done! You can store your film and later use a hole puncher, knife, etc. to cut off any feature you want.
  14. Clean the PDMS from your plates and store them clean for the next time. Avoid scratching the surface of your plates!

Notes


  • Tape (175, 140, 70 and 50 µm thick) and plastified paper (50 µm thick) have been used as spacers.
  • This methodology has been used to fabricate films as thin as 50-70 µm. The exact thickness of the PDMS film is difficult to measure due to the soft nature of the film.
  • The use of soap may interfere with further applications of the PDMS film. However, here in the lab we haven’t seen any adverse effects so far.

Materials needed
Fig.1 Materials needed: paper clamps, rigid plates and pieces of film (red arrows)

Film on edges
Figure 2. Pieces of film (red arrows) positioned close to the edges of the rigid plate shown on the left.

PDMS deposition
Figure 3. Manually deposited PDMS (blue arrow) on one of the plates

Squeezed PDMS
Figure 4. Squeezed PDMS between rigid plates using paper clamps

Film release
Figure 5. Release of the PDMS film from one of the plates after it has been cross-linked by heating

References


[1]  Aung K. Soe and Saeid Nahavandi, Degassing a PDMS mixture without a vacuum desiccator or a laboratory centrifuge and curing the PDMS chip in an ordinary kitchen oven, Chips & Tips (Lab on a Chip), 26 May 2011.

Digg This
Reddit This
Stumble Now!
Share on Facebook
Bookmark this on Delicious
Share on LinkedIn
Bookmark this on Technorati
Post on Twitter
Google Buzz (aka. Google Reader)

Fabricating microporous PDMS using a water-in-PDMS emulsion

Juyue Chena, Rui Zhangb and Wei Wang*a
a National Key Laboratory of Science and Technology on Micro/Nano Fabrication, Institute of Microelectronics, Peking University, Beijing 10871, P. R. China
b School of Pharmaceutical Sciences, Peking University, Beijing 10871, P. R. China

Why is this useful?


Microporous PDMS has been proposed as a functional PDMS material for cell culture related microfluidics applications where high gas perfusion is required to improve cell survival and functions. The phase separation micro-molding (PSμM) technique, which is widely used in microporous polymer preparation1, creates difficulties in fabricating microporous PDMS as there are many restrictions on the solvent – including high boiling point, low volatility, stability and appropriate compatibility with nonsolvent. Yuen reported that microporous PDMS can be simply prepared by curing PDMS pre-polymer with a porogen, such as salt or sugar particles, and then dissolving and washing away the porogen.2 However, it is difficult to obtain a microporous PDMS with pore size of micrometer scale by using this solid porogen. The small porogen makes the soaking and washing time cumbersome for most applications. As mentioned by Yuen, the porogen should be dissolved and washed away by soaking the PDMS and washing it in ethanol solution in an ultrasonic cleaner for at least 3 hours (longer may be required for smaller particles of porogen).

Fabrication of microporous PDMS

Figure 1. Fabrication of microporous PDMS by water-in-PDMS emulsion

Here, we propose a simple way to fabricate microporous PDMS using an emulsion of PDMS and water, as illustrated in  Figure 1. After manually blending the PDMS pre-polymer and water (1% SDS inside), water droplets were dispersed inside the PDMS pre-polymer. When the mixture is heated at a relatively low temperature (80oC) and in a relatively highly humid environment, the pre-polymer partially cures with the water droplets, keeping their original state. Through further curing at a higher temperature (120oC), the water trapped in the PDMS evaporates and leaves pores inside the PDMS matrix. Pore density is determined by the ratio of the water volume to the PDMS pre-polymer volume.

What do I need?


  • PDMS (Sylgard 184, Dow Corning Co.)
  • SDS (dodecyl sulfate sodium salt)
  • Deionised water
  • Petri dish
  • High temperature durable container
  • Oven

What do I do?


  1. Mix the PDMS according to the manufacturer’s instructions, with a mass ratio of base to curing agent of 10:1.
  2. Make the SDS solution, with a mass ratio of SDS to DI water of 1:100.
  3. Pour the PDMS pre-polymer and water (1% SDS inside) into a Petri dish with a given volume ratio, and manually blend them until a uniform emulsion (milky and opaque) is achieved. The water can be added step by step to facilitate the blending. Pore density is determined by the ratio of water volume to PDMS pre-polymer volume, as shown in Figure 2.

    SEM photos of microporous PDMS

    Figure 2. SEM photos of the prepared microporous PDMS. R = Vwater:VPDMS. (a) R = 0.01; (b) R = 0.05; (c) R = 0.3; (d) R = 0.7. All the scale bars represent 20μm

  4. Add some DI water in a high temperature durable container, and put the Petri dish with the water-in-PDMS emulsion inside on the water, then cover with the container lid. Put the container into the oven for about 2 hours at 80oC.
  5. Once the PDMS has been partially cured, remove the Petri dish from the container and finish curing it at a relatively high temperature in the oven for about 1 hour. After all the trapped water droplets have evaporated, the finished result is microporous PDMS.

References


[1] L. Vogelaar, R. G. H. Lammertink, J. N. Barsema, W. Nijdam, L. A. M. Bolhuis-Versteeg, C. J. M. van Rijn and M. Wessling, Phase separation micromolding: a new generic approach for microstructuring various materials, Small, 2005, 1, 645–655.
[2] P. K. Yuen, H. Su, V. N. Goral and K. A. Fink, Three-dimensional interconnected microporous poly(dimethylsiloxane) microfluidic devices, Lab Chip, 2011, 11, 1541-1544.

Digg This
Reddit This
Stumble Now!
Share on Facebook
Bookmark this on Delicious
Share on LinkedIn
Bookmark this on Technorati
Post on Twitter
Google Buzz (aka. Google Reader)

Simple and rapid fabrication of paper microfluidic devices utilizing Parafilm®

E. M. Dunfield, Y. Y. Wu, T. P. Remcho, M. T. Koesdjojo and V. T. Remcho
Department of Chemistry, Oregon State University, Corvallis, OR 97331, USA

Why is this useful?


Paper-based microfluidics offers several advantages over conventional microfluidics, and has great potential to generate inexpensive, easy-to-use, rapid and disposable diagnostic devices. Unlike traditional microfluidics, which often requires pumps to move fluid through the microfluidic channels, paper microfluidics can be performed without such instrumentation due to the flow of fluid being driven by capillary action through the paper. Hence, paper-microfluidics is well-suited for use in point-of-care diagnostics and in developing countries where expensive instrumentation is not available. There have been several advancements in the fabrication of paper microfluidic chips. Fabrication of paper-based chips can be done by photolithography [1], wax printing [2, 3], plasma etching [4], inkjet etching [5], and use of a cutting plotter [6]. However, such techniques may require the use of organic solvents during the fabrication process, they can be labor-intensive and expensive, and may impose limits on the types of materials that may be used in the chip. A novel technique is presented that utilizes a hydrophobic film material, Parafilm® “M”, which is heated to above its melting temperature of 60οC, and pressed into a piece of paper. The channel mask is cut from polycarbonate (PC) film, and sandwiched between the paper and the Parafilm®. The PC film mask prevents the melted Parafilm® from penetrating into the paper in the channel region, and therefore defines the hydrophobic boundaries of the paper channel. This technique has been tested on a wide variety of paper materials including Whatman Grade 1 filter paper, VWR light-duty tissue wipes and Kimtec Kimwipes, among other papers. While in some cases the paper can be cut directly into the desired channels, thin paper can easily tear and complex patterns proved to be challenging when cut by a cutting plotter. In contrast, the more rigid PC film can easily be cut into simple or complex patterns with a cutting plotter. Furthermore, the process is inexpensive, rapid, and does not require the use of organic solvents during fabrication.

What do I need?


  • Paper, such as Whatman filter paper, Kimtech Kimwipe or VWR light-duty tissue paper
  • Polycarbonate film, thickness approximately 100 µm
  • Parafilm® “M”
  • Aluminium foil
  • Scissors or x-y cutting plotter (for more complex channel patterns)
  • Hot press

What do I do?


  1. Cut out the desired channel patterns in polycarbonate film. For better cutting precision and accuracy, an x-y cutting plotter can be used.
  2. Create an assembly consisting of paper, polycarbonate (PC) cutout, and Parafilm® “M” as shown in figure 1. The paper can be Whatman filter paper, Kimtech Kimwipes, VWR light-duty tissue wipes or other paper of desired properties.

    Paper, polycarbonate and Parafilm® stack

    Figure 1. Paper, polycarbonate and Parafilm® stack set-up before heating and pressing.

  3. Cover both sides of the paper, PC, Parafilm® stack with aluminium foil to prevent sticking of the Parafilm® to the hot press plates, and place the whole assembly into the hot press.
  4. Heat the hot press to above 60οC, and apply ~200 psi of pressure for 1 minute. Note: 200 psi is necessary when using Whatman Grade 1 filter paper. Applied pressure varies depending on the thickness and porosity of the paper used.
  5. Allow the aluminium packet to cool, and then remove the foil from the paper microfluidic chip.

    Completed paper microfluidic chips made using Parafilm®

    Figure 2. Completed paper microfluidic chips made using Parafilm®. (a.) Spiral design has a channel width of 1 mm. (b.) Design has circles of 4mm diameter and 8 straight channels of 2mm width and 10mm length. (c.) Blue dye added to the paper microfluidics chip shown in b.

What else should I know?


For heavier weight paper such as Whatman Grade 1 filter paper, it is necessary to apply higher pressure (~200 psi) such as in a hot press to produce the microchips. However, for lighter weight paper such as Kimtech Kimwipes and VWR light-duty tissue wipes, microchips can simply be made by heating the plates of a heating element such as a hair straightener, and then applying gentle pressure to the paper, PC, Parafilm® stack to produce the paper-based chips.

References


[1]  A. W. Martinez, S. T. Phillips, B. J. Wiley, M. Gupta, and G. M. Whitesides, Lab Chip, 2008, 8, 2146-2150.
[2] Y. Lu, W. Shi, L. Jiang, J. Qin, B. Lin, Electrophoresis, 2009, 30, 1497-1500.
[3] E. Carrilho, A. W. Martinez, G. M. Whitesides, Anal. Chem., 2009, 81, 7091-7095.
[4] X. Li, J. Tian, T. Nguyen, W. Shen, Anal. Chem., 2008, 80, 9131-9134.
[5] K. Abe, K. Suzuki, D. Citterio, Anal. Chem., 2008, 80, 6928-6934.
[6] E. M. Fenton, M. R. Mascareñas, G. P. Lopez, S. S. Sibbett, ACS Appl. Mater. Interfaces, 2009, 1, 124-129.

Digg This
Reddit This
Stumble Now!
Share on Facebook
Bookmark this on Delicious
Share on LinkedIn
Bookmark this on Technorati
Post on Twitter
Google Buzz (aka. Google Reader)

Design of an inexpensive spin coater (with a touch-screen interface)

Gurucharan V. Karnad*a,b, R. N. Ninad*b and V. Venkataraman a
a Dept. of Physics, Indian Institute  of Science, Bangalore, India
b Dept. of Electronics and Communication Engineering, Amrita School of Engineering, Bangalore, India
* Corresponding authors: Gurucharan V. Karnad, R. N. Ninad
Email: gvkarnad[at]physics.iisc.ernet.in, rnninad[at]gmail.com

Why is this useful?


Spin coating is one  of  the  coating  techniques  used  to  apply  thin uniform  films  to  flat  substrates. A Spin Coater is a machine used for spin coating, and is one of the most ubiquitous instruments in any laboratory dealing with microfluidic devices.

Most commercial spin coaters are extremely expensive (≈ > US $1,995 [1]) and come with features and specifications not necessarily needed for fabricating and experimenting with polymer based microfluidic devices.

The cost of this instrument should not act as a deterrent for groups intending to venture into fabricating and experimenting with simple devices, hence the effort in this direction.

We have designed an inexpensive spin coater (with a touch-screen interface) costing less than US $350. The user input is through the touch-screen interface, where parameters such as spin duration and speed can be entered. Real time speed is also displayed alongside. A microcontroller forms the intelligence of the system and manages the inputs, display, and speed and duration control. The real time speed is sensed by the microcontroller using an optical encoder, and a control loop keeps it within acceptable error limits.

The substrate is mounted either using a double-side tape or a set of clamps. Improvements with regard to vacuum chuck, computer interfacing etc., can be done as per necessity.

What do I need?


  • PIC Microcontroller Programmer (with support for 18F4550)
  • PIC 18F4550 (≈ US $04.47 [2])
  • Graphic LCD 128×64( JHD 12864E or equivalent) (≈ US $16.80 [3])
  • TouchPanel (≈ US $07.00 [4])
  • TouchPanel Connector Board (≈ US $04.00 [5])
  • Infrared Distance Sensor (equivalent – Cytron IR01A Medium Range Infrared Sensor) – Optical Encoder-Sensor (≈ US $08.10 [6])
  • Crystal oscillator circuit (11 Mhz Crystal Oscillator, 22pf Capactitor (2) – connection as given in PIC 18F4550 Datasheet) (≈ US $01.00 [7])
  • IR 2110 (≈ US $07.25 [8])
  • IRFP 150N (≈ US $02.77 [9])
  • 1N 4744A (≈ US $00.21 [10])
  • BYQ 28E200 (≈ US $00.98 [11])
  • Brushed DC Motor (Como Drills 719RE850 or equivalent) (≈ UK £21.69 [12] ≈ US $35)
  • Any PC CPU Case (for Power Supply-SMPS and as instrument control box) (≈ US $50.00 [13])
  • PC (to program the microcontroller initially)
  • Software – MPLAB IDE (Free), HI TECH C Compiler for PIC 18 MCU (Free Version)
  • Machining, raw material, workshop access ( for chuck, optical encoder mount and motor cabinet) (≈ US $200.00)

What do I do?


  1. Interface and connect the PIC Microcontroller Programmer to the PC as per the instructions given in its manual
  2. Burn the Program01.hex file into the microcontroller with the help of the programmer
  3. Connect the microcontroller and other components as given in the circuit diagram (Figure 1). The circuit and display can be appropriately mounted in a PC CPU case
  4. Power up the circuit
  5. The Graphic LCD should display data similar to Figure 2. Figure 3 and Figure 4 will appear on the display if appropriate icons on the Menu “Selected Values” screen are touched
  6. There can be dissimilarities with each touch panel and hence it may not respond to input due to change in touch co-ordinates. Hence, there would be a need to modify the Program01.c. If there are no problems, jump to instruction 11
  7. Install the software mentioned above
  8. Modify the program, by finding out the new co-ordinates of the icons by following the instructions given in readme.txt
  9. Compile the modified Program01.c and program the microcontroller as per the instructions given in the programmer manual
  10. Repeat instruction 2 to 5
  11. Make an appropriate box to mount the motor
  12. An aluminium ( light and easy to clean) chuck (Figure 6, Figure 7) to fit the motor shaft  has to be machined
  13. A simple optical encoder set up  which includes mounting of a slim white acrylic piece on the bottom of the motor shaft and the IR optical sensor has to be made ( Figure  5)
  14. Put together all the components appropriately. The spin coater (Figure 8 ) should now be ready ( appropriate modification of  the chuck may result in a centrifuge too)
  15. Mount the wafer samples for spin coating either with a double-sided tape or a set of adjustable clamps

What else should I know?


  • Basic knowledge of C is essential to modify the code to suit individual requirements or specification
  • The spin coater has been set to receive input of speeds from 1000-7000 RPM ( limited only due the motor used, can be modified easily)
  • The maximum spin time is 999 seconds (increments of 1)
  • Due to nonlinear response of the brushed DC motors to voltage ( and hence varying PWM values),  the speed response of the motor to variation in PWM values has to be plotted and an appropriate equation has to be estimated. The equation in Program01.C has to be modified (see the readme.txt).  Note: This needs to be done only if a Brushed DC motor other than Como Drills 719RE850 is used, or if further fine tuning of response is required.
Spinner control schematic
Fig 1. Spinner control schematic
Menu display screen
Fig 2. Menu display screen

Runtime settings – spin duration (in seconds) screen
Fig. 3 Runtime settings – spin duration (in seconds) screen
RPM settings screen
Fig. 4 RPM settings screen
Optical encoder set-up
Fig. 5 Optical encoder set-up
Chuck with adjustable clamps
Fig.6 Chuck with adjustable clamps
Chuck without adjustable clamps – samples mounted using a double-sided tape
Fig.7 Chuck without adjustable clamps – samples mounted using a double-sided tape

Spin coater
Fig 8. Spin coater

References


[1] http://www.mtixtl.com/DeskTopSpinCoater0-5100rpmwithCompleteAccessoriesVTC-100P.aspx

[2] http://us.element14.com/microchip/pic18f4550-i-p/ic-8bit-mcu-pic18f-48mhz-dip-40/dp/74K8623?in_merch=Popular%20Products

[3] http://www.ebay.com/itm/Graphic-Matrix-LCD-Module-LCM-JHD-12864-E-128X64-    /200392901987?pt=LH_DefaultDomain_0&hash=item2ea8590563#ht_4043wt_906

[4] http://www.mikroe.com/eng/products/view/277/various-components/

[5] http://www.mikroe.com/eng/products/view/207/touchpanel-connector-board/

[6] http://www.robotshop.com/2cm-10cm-infrared-distance-sensor-2.html

[7] http://us.element14.com/

[8] http://us.element14.com/international-rectifier/ir2110-1pbf/ic-mosfet-driver-high-low-side/dp/19K8414?in_merch=Popular%20Products

[9] http://us.element14.com/international-rectifier/irfp150npbf/n-ch-mosfet-100v-42a-to-247ac/dp/63J6848?Ntt=irfp+150n

[10] http://us.element14.com/multicomp/1n4744a/zener-diode-1w-15v-do-41/dp/90R9431?in_merch=Popular%20Products

[11] http://in.mouser.com/Search/ProductDetail.aspx?qs=kop3EGWU%252bIt67Q6mQhs7oA%3D%3D

[12] http://uk.rs-online.com/web/p/dc-motors/3213192/

[13] http://www.amazon.com/Computer-Cases-Add-Ons-Computers/b?ie=UTF8&node=572238#/ref=sr_nr_p_n_feature_keywords_0?rh=n%3A172282%2Cn%3A%21493964%2Cn%3A541966%2Cn%3A193870011%2Cn%3A572238%2Cp_n_feature_keywords_browse-bin%3A3606145011&bbn=572238&sort=price&ie=

Digg This
Reddit This
Stumble Now!
Share on Facebook
Bookmark this on Delicious
Share on LinkedIn
Bookmark this on Technorati
Post on Twitter
Google Buzz (aka. Google Reader)

Easy and robust interconnection methods for PDMS-based microfluidics

Shuo Wang, Huaiqiang Yu, Wei Wang and Zhihong Li
National Key Laboratory of Science and Technology on Micro/Nano Fabrication, Institute of Microelectronics, Peking University, China

Why is this useful?


PDMS (polydimethylsiloxane) is one of the most important materials in microfluidics and is widely used because of its optical transparency, ease-in-fabrication, low cost and air permeability. A widely used interconnection approach for PDMS chips is the “press-fit” method [1]. However, the seal is only achieved by the compression of PDMS. An unexpected disturbance to the needle may damage the PDMS around it and produce small cracks leading to leakage around the needle. The mechanism of disturbance-caused leakage is shown in Fig. 1.

Fig. 1 Mechanism of disturbance-caused leakage

Here, we report two easy methods of fixing needles by a secondary PDMS fabrication. In these “cure and fix” methods, uncured PDMS is poured and cured to fix needles. Sectional views in Fig. 2 show the two schematic fabrication processes respectively. Cover packaging methods can be applied to produce covers in large number used as standard components. After PDMS chips are made, simply bond them with these standard covers to seal reservoirs. We can also employ whole packaging method to fabricate one specific device with lots of reservoirs.

Fig. 2 “Cure and fix” method

What do I need?


  • PDMS chip peeled off from Si mould
  • Uncured PDMS (base:cure = 10:1)
  • Silicon or glass substrate
  • Unmodified needles
  • Scalpel and tweezers
  • Hole puncher
  • Oxygen plasma etching machine or corona charging device

How do I do it?


Whole packaging method

  1. Punch holes for reservoirs on a PDMS chip and bond the chip with silicon or glass substrates using oxygen plasma or corona treatment [2].
  2. Plunge unmodified needles into reservoirs laterally using the “press-fit” method. Clear away PDMS scraps with a pair of tweezers.
  3. Seal all reservoirs by bonding PDMS blocks. Cast uncured PDMS onto the chip until lower half of the needle is submerged.
  4. After curing PDMS at 70°C for 1 hour, cut the chip into proper size.

Cover packaging method

  1. Punch a hole for the reservoir on a flat PDMS block and bond it with another PDMS block
  2. Plunge an unmodified needle into the reservoir laterally using the “press-fit” method. Clear away PDMS scraps with a pair of tweezers.
  3. Put the PDMS cover on a flat culture dish and cast uncured PDMS.
  4. After curing PDMS at 70°C for 1 hour, cut the cover into proper size.
  5. Bond the cover with a PDMS chip to seal reservoirs.

What else should I know?


In order to plunge the unmodified needle into reservoirs successfully, the PDMS cannot be too thin. The thickness should be larger than 3 mm.
Be careful in step 4 of whole packaging method because silicon and glass are brittle.

Fig. 3 A) Vertical view and B) side view of device using whole packaging method C) vertical view and D) side view of device using cover packaging method

References


[1] A. M. Christensen, D. A. Chang-Yen and B. K. Gale, Characterization of interconnects used in PDMS microfluidic systems. J. Micromech. Microeng., 2005, 15, 928-934.
[2] K. Haubert, T. Drier and D. Beebe, PDMS bonding by means of a portable, low-cost corona system, Lab Chip, 2006, 6, 1548-1549.

Digg This
Reddit This
Stumble Now!
Share on Facebook
Bookmark this on Delicious
Share on LinkedIn
Bookmark this on Technorati
Post on Twitter
Google Buzz (aka. Google Reader)

Adding colour to PMDS chips for enhanced contrast

Marco A. Cartas-Ayala and Suman Bose
Department of Mechanical Engineering, Massachusetts Institute of Technology, USA.

Why is this useful?


Most materials used to fabricate microfluidic devices are transparent to facilitate sample visualization (e.g. PDMS), but this property has several drawbacks too. Alignment and visualization of the channels is difficult when the channels are completely transparent, making bonding of polymer devices difficult. Additionally, when multilayer polymer devices are manufactured, sometimes it is necessary to distinguish between different layers to easily evaluate functionality. Finally, having a way to add permanently colour to any kind of transparent channel can become really handy when creating permanent exhibitions displaying the devices created in the lab.

What do I need?


  1. PDMS (Sylgard 184)
  2. SILC PIG. blue silicone pigment, from Smooth-On, Inc
  3. 3 mL Syringe
  4. Blunt pieces of stainless tube (1/2 inch long, diameter smaller than PDMS holes, from New England Small Tube)
  5. Tygon tubing that fits the blunt needles and the stainless pieces of tubing
  6. Blunt needles for 1 mL syringe (diameter selected accordingly to tygon tubing diameter)

What do I do?


  1. Mix PDMS (Sylgard 184) in the recommended 10:1 ratio
  2. Add to the mix 5% w/w of the rubber paint and mix completely. If the mixture is not mixed thoroughly, pockets of paint can be formed in the final mixture, if you have problems with the mix, reduce the paint ratio
  3. Degas the mixture for 30 minutes
  4. Load 0.1 mL of the sample into the syringe with the blunt needle and tubing
  5. Inject into the channels to visualize. Be careful to not introduce bubbles, while air in PDMS leaks out when enough pressure is applied, air has to be flown out from glass devices
  6. Cure PDMS at 70 C for 1 hour
  7. Devices are ready for display. Notice the enhanced contrast of the colour filled channels vs the empty channels for the same device in Figure 2. While channels are visible only from some directions when they reflect light, colour-PDMS devices can be observed from every direction. Additionally, different device layers or areas can be specified by colour. In the figure control layers are blue and flow layers are red

Fig. 1 Injection of PDMS through the channels. Air trapped inside the syringe provides a way to regulate the pressure applied to the device to minimize de-bonding. Compressing the air to 1/3 of original volume should provide enough pressure to drive the PDMS through.


Fig. 2 Enhanced channel contrast after injection, devices on the left side have empty channels and devices on the right have color PDMS inside.

Fig. 3 Different device zones can be identified by color. Here control layer is blue and flow layer is red. Secondary regulation channels are practically invisible when not filled.

References


[1] http://www.upchurch.com
[2] http://www.smooth-on.com

Digg This
Reddit This
Stumble Now!
Share on Facebook
Bookmark this on Delicious
Share on LinkedIn
Bookmark this on Technorati
Post on Twitter
Google Buzz (aka. Google Reader)