Chips and Tips

Jonathan Siegrist, Mary Amasia, and Marc Madou
Departments of Biomedical, Chemical, and Mechanical Engineering, University of California, Irvine, CA, 92697, USA

Why is this useful?

The fabrication and bonding of rapid-prototyped polymer-based microfluidic devices is of great interest. The emphasis is not only on the ability to produce large numbers of devices rapidly, but to perform fast design and test iterations. Here, we present a method for rapidly cutting thin plastic films into complex shapes, such that sophisticated 2.5-D microfluidic chips can be created. The plastic layers are bonded together using traditional, low-pressure thermal bonding to minimize channel deformation while achieving high bond strengths capable of withstanding high-pressure & temperature operations, such as polymerase chain reaction (PCR). The authors have found this method to be particularly amenable to fast design iterations, as one can easily go from a design on the computer screen to testing of a real part within 3 hours.

The method presented here of cutting plastic films using a commercially-available knife-based cutter/plotter avoids the use of traditional CNC milling, which can leave behind burrs and rough edges. While hot-embossing and laser machining can be superior alternatives to CNC milling, they are considerably more complex and expensive than the method presented here. Also, by using a computer-controlled cutter, obvious advantages are gained in terms of possible design complexity and reproducibility as compared to previous Tips [1, 2]. Finally, this cutting method can be used for many types of plastic films (< 400 um thick, depending on the machine being used) without the need for extensive optimization of cutting speeds or feed rates.

The use of thermal bonding allows for much higher bond strengths as compared to the use of pressure-sensitive adhesives or double-sided tapes, and also provides a large dynamic range in terms of the types of plastics that can be used. For example, thin films of either high glass-transition temperature (Tg) materials such as polycarbonate or low-Tg cyclic olefin copolymers can be used with this method, as compared to lamination methods that are limited to low Tg materials only [3].

What do I need?

• Hydraulic Thermal Press, ideally with a max. temperature of at least 200ºC and a max. pressure of at least 1 MPa (the device used here was an MTP-8 from Tetrahedron Associates, Inc., CA, USA)
• Computer-Controlled Cutter/Plotter (the device used here was a CE-2000 from Graphtec America, Inc., CA, USA)
• CAD software
• Bare Si Wafers, at least single-side polish
• Plastic Films (the films used here were polycarbonate from McMaster-Carr, CA, USA)
• Double-sided Tape (3M Scotch brand was used here)
• Aluminium Foil
• Tweezers
• Alignment Pins (optional)
• Oxygen Plasma System (optional)

What do I do?

1. Design your microfluidic device using common CAD software (Fig. 1). Ensure the drawing for each layer is in a “polyline” format (i.e., each continuous line is defined as a single object) to facilitate smooth cutting. Import the CAD file into the software that controls the cutter/plotter.

2. Attach your plastic film (using double-sided tape) to a plastic sheet that will act as your support/sacrificial layer (e.g., 0.75 mm thick). Place this structure in the cutter/plotter, and cut the film (Fig. 2). Cutting force and number of cuts may need to be optimized depending on the film thickness. We found that 2x low-force cuts worked well for 127 µm-thick polycarbonate films, and 4x medium-force cuts worked well for 254 µm-thick films.

3. After all cuts are complete, carefully remove the part from the sacrificial layer and use tweezers to remove and discard the cut-out plastic features (Fig. 3). Repeat this for all layers, and then clean the plastic layers using isopropanol and water.

4. Align all of the layers, and hold together using a clip. You may include alignment pin features on your layers, and use pins to assist in alignment. For example, you could use 1 mm-diameter pins (McMaster-Carr, CA, USA – #91585A001).
5. Place the assembled microfluidic device layers on the mirror-finished side of a Si wafer, and remove the clip(s) (Fig. 4). Then place the second wafer mirror-finished side down on top of the microfluidic device, being careful not to disturb the alignment.

6.  Place the assembly in the thermal press sandwiched between two pieces of foil, and thermally bond. Approximate parameters we found to work well for our proof-of-concept, multi-layer, polycarbonate device (1-top layer with loading/venting holes: 254 µm thick, 2-film with serpentine channel: 127 µm thick, 3-film with access holes and reservoirs: 127 µm thick, 4-film with linear channel: 127 µm thick, and 5-bottom layer: 254 µm thick) are shown in Fig. 5.

7.  Remove the microfluidic device from the thermal press, and test (Fig. 6).

What else should I know?

Thermal bonding parameters, such as bonding temperature, pressure, dwell times, and ramping rates, will need to be optimized to ensure complete bonding of the plastic layers. The main disadvantage of thermal bonding is possible microchannel deformation, which can affect fluidic function. Thus, the Tg of your material, the number of layers, and the required bond strength will all need to be considered. The previous Tip on prevention of chamber-sagging has obvious uses here [4].

The limitations on feature size using this method should also be noted. The thickness (z-axis) of your channels/chambers is determined by the thickness of the films. We found the feature size (x-y plane) limitations when cutting films to be ~ 200 µm, as limited by the cutter/plotter being used. Newer machines list resolutions on the order of 10s of µm.

Finally, while these plastic parts are inherently hydrophobic, they can be made hydrophilic via oxygen-plasma treatment. Common plasma-treatment parameters for our devices are 200 W for 2 mins at 200 mTorr O2-pressure to provide hydrophilicity for many weeks. This can facilitate liquid loading and also serve as a sterilization step.

Acknowledgements

We would like to thank the DARPA-MF3 center for funding, and Dr. Albert Yee of UC Irvine for generously allowing us to use his thermal press.

References

[1] R. J. Holmes and N. J, Goddard, Rapid prototyping of microfluidics, Chips & Tips, (Lab on a Chip), 15 February 2007.
[2] R, Kumar, R. L. Smith, and M. G. Pappas, A method for rapid fabrication of microfluidic devices, Chips & Tips, (Lab on a Chip), 30 June 2009.
[3] D. Olivero and Z. Fan, Lamination of plastic microfluidic devices, Chips & Tips, (Lab on a Chip), 30 July 2008.
[4] J. Xu and D. 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.

S. Ghorbanian, M. A. Qasaimeh and D. Juncker
Biomedical Engineering Department, McGill University and Genome Quebec Innovation Centre, McGill University, 740 Dr. Penfield Avenue, Montreal, QC, H3A 1A4 Canada

Why is this useful?

Polydimethylsiloxane (PDMS) is widely used for the fabrication of microfluidic systems because it can readily be molded into the desired shape, is easy to seal onto substrates, and is transparent thus permitting visualization of the sample [1]. However, the fabrication of PDMS microfluidic devices depends on a microfabricated mould that needs to be made in a clean room using photolithography and microfabrication methods, all of which are costly and time consuming and beyond the reach of many researchers. Rapid prototyping techniques that circumvent the requirement for a clean room have been proposed, such as the use of double sided scotch tapes, but lack precision and control [2,3].

Here we present a method for rapid prototyping of branched microfluidics in PDMS with control over the architecture, channel width and depth. We propose using capillaries as the mould.  They are cut to size, then arranged on a flat PDMS according to the desired architecture and covered with PDMS which is then cured. The size of each channel can be adjusted by selecting a capillary with the desired diameter, and different branch architectures can readily be produced. The capillaries are then simply removed from the PDMS replica and leave behind a network of channels.

A key challenge is the gaps formed at the intersections between abutting capillaries. We found that a small flake of paraffin can be used which is then melted to fill up this gap. The use of capillaries with a square cross-section further facilitates the moulding, and allows for making complex networks with ease and good yield. This protocol only requires materials which are commercially available and comparatively inexpensive and takes less than one hour of hands-on time, followed by three hours of curing.

What do I need?

1. A layer of flat cured PDMS, Figure 1d
2. Uncured PDMS
3. Petri dishes
4. Square capillaries (preferably) or circular capillaries [4], Figure 1a
5. Ceramic cutting stone [4], Figure 1b
6. Paraffin [5], Figure 1c
7. Oven or hot plate
8. Sharp tweezers and razor blade, Figure 1(e,f)
9. Double-sided tape [6]

Figure 1. Required materials: (a) Glass capillaries, (a1) Cross section of the square glass capillary and (a2) the round glass capillary [4]. (b) Capillary cutting stone [4]. (c) Paraffin flakes [5]. (d) Flat cured PDMS piece in a Petri dish. (e) Razor blade. (f) Tweezers

What do I do?

1. Cure a layer of PDMS in a Petri dish. Remove the PDMS, and cut it to the desired size. Flip this piece to obtain the flat surface on top and place it inside another Petri dish as shown in Figure 2.

2. Cut two pieces of capillary, Figure 1a (or as many needed to make the branched channels) using a ceramic cutting stone. The capillaries should be cut to a length that facilitates handling, typically three or more times the length of the final channels of interest and shorter than the diameter of the Petri dish. Also see “What else should I know I”. After cutting the capillaries, dip their tip at each side in melted paraffin or liquid glue and let it solidify. This helps prevent air trapped in the capillaries from exiting while degassing the PDMS, which can lead to bubbles and displacement of the capillaries.

3. Under the stereomicroscope grind the tip of the capillaries (which will be making the connections between capillaries) to remove the bumps seen in Figure 3a using a polishing stone, such as the side surface of a cutting stone, with horizontal movements, Figure 3b and flatten the tip completely, Figure 3c, which can be done at angles other than 90 degress as well, Figure 3d.

4. Stick double-sided tape at the outer extremities (preferably not at the connection sites) where the capillaries will be placed on the flat PDMS in the Petri dish, Figure 4a. Under a stereomicroscope place the capillaries according to the desired architecture. Figure 4b illustrates the placement of capillaries for fabricating T-shaped microchannels.

5. Examine the gaps and connections between the capillaries to ensure good contact. A T-shaped connection and 45 degree connections are shown in Figure 5 below.

6. Carefully place a small piece of paraffin using sharp tweezers on each of the capillary connections, Figure 6a. Heat the tip of the tweezers on a hot plate for a minute, and then carefully approach the tip to melt the paraffin, which fills the gaps between the capillaries due to capillary effects and joins them to one another, Figure 6b. The excess melted paraffin can be removed carefully wiht the tip of a razor or sharp tweezers.

7. Pour a layer of uncured PDMS over the connected capillaries as shown in Figure 7 and degas the PDMS by placing the Petri dish in a vacuum desiccator to remove all air bubbles, for alternative methods see “What else should I know II”. Then, place the Petri dish in the oven at 65°C for 3 hours or more. For shorter curing time see “What else should I know III”.

8. Remove the whole piece of cured PDMS from the Petri dish. Cut the sides of the cured PDMS using a razor blade, leaving a significant amount of the capillary exposed outside as shown below. Then carefully pull out the capillaries from the sides of the PDMS using pliers as shown in Figure 8. To make this process easier, the network can be immersed into or washed with acetone which will swell the PDMS and expand the channels prior to pulling out the capillaries.

If residues of paraffin are left inside the microchannels these can be dissolved and washed by flushing the microchannels with acetone.

9. Trim the microfluidic device to the desired shape using a razor blade or a cutter, Figure 9.

What else should I know?

Several alternative fabrication tips are listed below:

I) To fabricate an open channel microfluidic network replace the flat cured PDMS layer (in step 1) with a clean glass slide. Once the PDMS is cured it can be separated from the glass slide. In this process smaller capillaries can be used to make the channels which can be removed using tweezers after detaching the PDMS from the substrate. This open network can also be bonded to another PDMS layer after the capillaries are removed.
II) If no vacuum desiccator is available to degas PDMS, the sample can be left to be degassed and cured at room temperature overnight followed by post-curing in an oven at 65°.
III) To increase the speed of PDMS curing in step 8, the Petri dish may be replaced by an aluminum foil or a glass plate and allow to use much higher temperatures inside an oven or on a hot plate to cure the PDMS within a few minutes.
IV) It is preferable to use square capillaries because there are no gaps formed at the connection sites due to the square shapes of the capillaries as opposed to the round capillaries which will have a small gap formed at the connection sites due to the rounded shape of the capillary walls, Figure 1(a1). These gaps can, however, get filled with melted paraffin.

References

[1] D. Duffy, J. McDonald, O. Schueller, G. Whitesides, Rapid prototyping of microfluidic systems in poly (dimethylsiloxane), Anal. Chem., 1998, 70, 4974-4984.
[2] R. J. Holmes and N. J, Goddard, Rapid prototyping of microfluidics, Chips & Tips, (Lab on a Chip), 15 February 2007.
[3] R, Kumar, R. L. Smith, and M. G. Pappas, A method for rapid fabrication of microfluidic devices, Chips & Tips, (Lab on a Chip), 30 June 2009.
[4] Polymicro technologies
[5] Fisher Scientific
[6] Scotch Tape