Here we present a protocol to design and fabricate custom microfluidic devices with minimal financial and time investment. The aim is to facilitate the adoption of microfluidic technologies in biomedical research laboratories and educational settings.
Microfluidic devices allow for the manipulation of fluids, particles, cells, micro-sized organs or organisms in channels ranging from the nano to submillimeter scales. A rapid increase in the use of this technology in the biological sciences has prompted a need for methods that are accessible to a wide range of research groups. Current fabrication standards, such as PDMS bonding, require expensive and time consuming lithographic and bonding techniques. A viable alternative is the use of equipment and materials that are easily affordable, require minimal expertise and allow for the rapid iteration of designs. In this work we describe a protocol for designing and producing PET-laminates (PETLs), microfluidic devices that are inexpensive, easy to fabricate, and consume significantly less time to generate than other approaches to microfluidics technology. They consist of thermally bonded film sheets, in which channels and other features are defined using a craft cutter. PETLs solve field-specific technical challenges while dramatically reducing obstacles to adoption. This approach facilitates the accessibility of microfluidics devices in both research and educational settings, providing a reliable platform for new methods of inquiry.
Microfluidics enables fluid control at small scales, with volumes ranging from microliters (1 x 10-6 L) to picoliters (1 x 10-12 L). This control has been made possible in part due to the application of microfabrication techniques borrowed from the microprocessor industry1. The use of micro-sized networks of channels and chambers allows the user to take advantage of the distinct physical phenomena characteristic of small dimensions. For example, at the micrometer scale, fluids can be manipulated using laminar flow, where viscous forces dominate inertial forces. As a result, diffusive transport becomes the prominent feature of microfluidics, and can be studied quantitatively and experimentally. These systems can be properly understood using Fick’s laws, Brownian motion theory, the heat equation, and/or the Navier-Stokes equations, which are important derivations in the fields of fluid mechanics and transport phenomena2.
Because many groups in the biological sciences study complex systems at the microscopic level, it was originally thought that microfluidic devices would have an immediate and significant impact on research applications in biology2,3. This is due to diffusion being dominant in the transport of small molecules across membranes or within a cell, and the dimensions of cells and microorganisms are an ideal match for sub-millimeter systems and devices. Therefore, there was significant potential for enhancing the way in which cellular and molecular experimentation is conducted. However, wide adoption of microfluidic technologies by biologists has lagged behind expectations4. A simple reason for the lack of technology transfer may be the disciplinary boundaries separating engineers and biologists. Custom device design and fabrication have remained just outside of the capabilities of most biological research groups, making them dependent on external expertise and facilities. Lack of familiarity with potential applications, cost, and the time required for design-iteration are also significant barriers for new adopters. It is likely that these barriers have had the effect of disrupting innovation and preventing the widespread application of microfluidics to address challenges in the biological sciences.
A case in point: Since the late 1990’s soft-photolithography has been the method of choice for the fabrication of microfluidic devices. PDMS (polydimethylsiloxane, a silicone-based organic polymer) is a widely used material because of its physical properties, such as transparency, deformability, and biocompatibility5. The technique has enjoyed great success, with lab-on-a-chip and organ-on-a-chip devices continually being developed on this platform6. Most of the groups working on these technologies, however, are found in engineering departments or have strong ties to them4. Lithography usually requires clean-rooms for the fabrication of molds and specialized bonding equipment. For many groups, this makes standard PDMS devices less than ideal due their capital costs and lead-time, particularly when there is a need to make repeated design modifications. Furthermore, the technology is mostly inaccessible to the average biologist and to students without access to specialized engineering laboratories. It has been proposed that for microfluidic devices to be widely adopted, they must mimic some of the qualities of materials commonly used by biologists. For example, polystyrene used for cell culture and bioassays is inexpensive, disposable, and amenable to mass production. In contrast, industrial manufacturing of PDMS-based microfluidics has never been realized because of its mechanical softness, surface treatment instability, and gas permeability5. Because of these limitations, and with the goal of solving technical challenges using customized devices built “in-house”, we describe an alternative method that utilizes xurography7,8,9 protocols and thermal lamination. This method can be adopted with little capital and time investment.
PETLs are fabricated using polyethylene terephthalate (PET) film, coated with the thermoadhesive ethylene-vinyl acetate (EVA). Both materials are widely used in consumer products, are biocompatible and are readily available at minimal cost10. PET/EVA film can be obtained in the form of laminating pouches or rolls. Using a computer-controlled craft cutter commonly found in hobbyist or craft stores, channels are cut out of a single film sheet to define the device’s architecture11. The channels are then sealed by applying additional film (or glass) layers that are bonded using an (office) thermal laminator (Figure 1A). Perforated, self-adhesive vinyl bumpers are added to facilitate access to the channels. Fabrication times range from 5 to 15 min, which allows rapid design iteration. All the equipment and materials used to make PETLs are commercially accessible and affordable (<350 USD starting cost, compared to thousands of USDs for lithography). Therefore, PETLs provide a novel solution to two main problems posed by conventional microfluidics: affordability and time effectiveness (See PDMS/PETL comparison in Supplementary Tables 1, 2).
In addition to providing researchers with the opportunity to design and fabricate their own devices, PETLs can be easily adopted in the classroom because they are simple and intuitive to use. PETLs can be included in high school and college curricula8, where they are used to help students better understand physical, chemical and biological concepts, like diffusion, laminar flow, micromixing, nanoparticle synthesis, gradient formation and chemotaxis.
In this work we illustrate the overall workflow for the fabrication of model PETLs chips with different levels of complexity. The first device is used to facilitate imaging of cells and micro-organs in a small chamber. The second, more complex device consists of several layers and materials, and is used for research in mechanobiology9. Lastly, we built a device that displays several fluid dynamics concepts (hydrodynamic focusing, laminar flow, diffusive transport and micromixing) for educational purposes. The workflow and device designs presented here can be easily tailored for a large range of purposes in both the research and classroom settings.
1. Design
2. Cutting
3. Alignment
4. Lamination
5. Inlet/outlet ports
6. Testing
Figure 1: Fabrication. (A) An office laminator and a craft cutter are the only two pieces of equipment required for fabrication. Both are available online or at crafts/office supplies stores. Other required tools include scissors and tweezers. (B) Channel and chamber architectures can be composed digitally using any software program that includes drawing tools (vector graphics may be preferred by some users, but are not required). Lines and shapes are drawn in black with a white background. The file or a screen capture of the design can be imported into the craft cutter software by dragging and dropping. (C) Craft cutter software is available free to download and is required to control the cutter. The software acquires the design and allows for modifications, like sizing. It also provides drawing tools. (D) The cutting mat carries the film for cutting. It is slightly adhesive, allowing for immobilization of the materials to be cut. The figure shows four different materials ready for loading: 3 mil-thick PET/EVA film (top), 5 mil-thick PET/EVA film (middle), 6 mil-thick EVA/PET/EVA (bottom left) and PVC film (bottom right). (E) Cutter is open to display blade (in black) unit and loaded mat. (F) After cutting, individual layers are lifted using tweezers. Cut outs of channels and chambers remain attached to the mat and are later removed and discarded. (G) Individual layers are aligned and superimposed for lamination. Small pieces of double-sided tape (arrow) are often used to aid in aligning and prevent layer shifting during lamination. (H, I) The device is fed at the top of the laminator and recovered through the slot. Lamination provides a robust seal, leaving channel paths open. (J, K) In order to access the channels, it is necessary to add perforated, self-adhesive vinyl bumpers. Image in (J) displays the “reverse” approach for alignment, in which the bumper is placed from the back, allowing visual alignment of the inlet/outlet with the bumper perforation. Please click here to view a larger version of this figure.
In addition to low cost and rapid iteration, PETL technology can be easily customized to solve specific challenges. First, we describe a simple device consisting of a glass coverslip, a chamber layer, a channel layer, and an inlet/outlet layer (Figure 2). This device was designed to facilitate the imaging of cells and micro-organs under constant flow. Culture medium is replenished at low flow rates to encourage nutrient and gas exchange. The round chamber features a glass bottom, which allows for imaging using an inverted microscope. There are at least two reasons for the use of glass in this device. The first one is optics. PET and EVA are thermoplastics used for their optical transparency and flexibility, and can be used as an interface for imaging (particularly at low magnifications9. Light transmittance of PET in the visible spectrum ranges from 87 to 90%12. Glass, however, has better optical properties and is the standard used in biological imaging. The second reason to use glass is that the cells so far tested (mammalian cell lines), tend to attach more readily to it than to (untreated) PET/EVA.
Figure 2: Simple chamber for inverted microscopy. (A) The device consists of one glass layer and three PET/EVA layers (3 mil thick). A glass coverslip (24 mm x 60 mm) is the bottom layer. The next layer features the bottom of the imaging chamber. The next layer features the top-half of the chamber and connects it to the in/outlet channel. Thus, the height of the channel is only 75 µm, while the height of the chamber is 150 µm. The width of the channel is determined by the user (500 µm is shown here). The top layer seals the chamber/channel path and provides access to the inlet/outlets. The superimposed layers are shown to the right. (B) The finished device is shown infused with red dye for visualization. Loading can be achieved using a micropipette and tip, laboratory or medical/surgical tubing outfitted with a blunt needle, or PEEK tubing, as shown. (C) The channel/chamber design can be iterated in one single device (e.g., to facilitate observation of several individual specimens). Please click here to view a larger version of this figure.
The dimensions of the channels and chamber in this device are worth describing. Height in PETLs is always a function of film or layer thickness. Commercially available PET/EVA has a thickness measured in thousandths of an inch (1 mil ≈ 25 µm). Therefore, channel and chamber heights are usually multiples of 25 µm. Standard PETLs are built using 3 or 5 mil PET/EVA film, which results in features having a height of 75 or 125 µm. The device shown in Figure 2 has channels with a height of 75 µm, and a chamber defined by two layers, with a total height of 150 µm. It should be noted, however, that layers can be composed of different materials (e.g., glass, foil, PVC, paper) and can present varying thickness, usually ranging from 25 to 250 µm.
Figure 3: Cell imaging. (A) The simple chamber PETL can be used for short term culture of adherent cells. Cells adhere to the glass exposed in the chamber and can be observed using an inverted microscope. (B) Rat basophils were stained with Hoechst (blue) and plasma membrane (red) fluorescent dyes for visualization on an inverted confocal microscope. (C) Bright field image of cells in a simple chamber device. (D) Phase contrast image. White scale bar is 200 µm. Please click here to view a larger version of this figure.
The nature of PETL fabrication allows for significant complexity in fluid path design. The simple chamber device consists of four layers containing features in two levels of the z-axis (channel and top of the chamber in one level, bottom of the chamber in the second level). An advantage provided by PETLs is the ease with which 3-dimensional channel/chamber architectures can be built. Addition of features like cooling or heating channels, dialysis membranes, electrical circuits or pressure lines (see Figure 5) is achieved by connecting multiple layers in three dimensions. A caveat so far encountered is the limit on the number of layers that can be laminated. Heat transfer necessary for EVA curation has been found insufficient in devices with an overall thickness above 800 µm. This limitation can be addressed in some devices. In many instances, it is possible to laminate every time a new layer is added. This is not possible when a new layer requires the thermoadhesive (EVA) to face the outside of the device.
Figure 4: Micro-organ imaging. (A) The simple chamber PETL is used to image a wing disc of the embryo of Drosophila melanogaster (2x magnification). The wing disc dimensions are approximately 90 µm x 250 µm x 500 µm. One or several organs can be imaged through the coverslip window. Increased magnifications of another wing disc are shown in (B) 20x/0.75-air, (C) 40x/1.30-oil, and (D)100x/1.49-oil objectives using a spinning disc confocal microscopy. Please click here to view a larger version of this figure.
The study of cells in culture benefits from tools that provide dynamic state conditions like constant flow or mechanical stimuli. Figure 3 provides an example in which a mammalian cell line is cultured and imaged in a simple chamber device. Medium can be constantly exchanged during imaging, allowing not only for maintaining ideal growth conditions, but also for the controlled introduction of chemical stimuli while imaging in real time. This is also true for the imaging of ex-vivo micro-organs, as shown in Figure 4. Channel and chamber structures may be designed with specific dimensions to fit different biological specimens, from organs or tissues to whole organisms (e.g., Drosophila embryos and imaginal discs or C. elegans).
Figure 5: Mechano-PETL. The simple chamber PETL is modified by adding a compression chamber. (A) The device consists of seven layers featuring four different materials: a glass bottom layer (coverslip, not shown), four PET/EVA 3 mil layers (channel/chamber layers, spacer layer and inlet/outlet layer), an EVA/PET/EVA 6 mil layer (channel/chamber sealing and PVC adhesion) and a deformable PVC layer (for compression). (B, C) The specimen channel/chamber path is visualized using red dye. The compression channel/chamber path contains only air. (D) Air pressure is manually (or mechanically) applied to the compression path, resulting in the expansion of the PVC film at the top of the chamber. The expansion displaces the dye in the chamber. Please click here to view a larger version of this figure.
Mechanical perturbation of biological specimens improves our understanding of cellular physiology and sheds light on processes such as embryonic development and differentiation. Figure 5 describes a PETL device consisting of a simple channel/chamber array and a compression chamber. It consists (in its simplest form) of six layers, one of which is a PVC film. The PVC film deflects when air pressure is applied, causing it to compress specimens within the chamber. This device is an example of the use of materials other than PET/EVA, and it has been successfully employed9 in replacing PDMS/glass devices utilized to study mechanical load on Drosophila micro-organs13 (as shown on Figure 6). PETL devices are reusable. However, because of the low cost of fabrication, reduced footprint, and the potential for delamination after continuous manipulation or washing, we recommend the use of new devices at the start of every procedure.
Figure 6: Mechanobiology imaging. (A) A DE-cadherin::GFP expressing Drosophila wing disc is imaged within a mechano-PETL using a spinning disc confocal microscope at 20x magnification. (B) Pressure through the membrane above the chamber can be applied by actuating an air-filled syringe manually or using a syringe pump. The ideal gas law is used to estimate the amount of force applied to the membrane9. Disc pouch area (white dotted line) increased by approximately 30% (red dotted line) with the application of ~4 psi. Please click here to view a larger version of this figure.
Because of the ease of fabrication of PETL devices, we have explored their use in educational settings like chemistry, biology and engineering classrooms and teaching laboratories. An example of an educational PETL is shown in Figure 7. The device is designed to display some of the basic features of fluid flow at the micro scale (e.g., laminar flow). It consists of four layers of 5 mil PET/EVA film (Figure 7A) and a channel architecture that includes three converging input channels and a serpentine structure. Circular “depressions” or “down” steps have been added to the path to promote micromixing14. Using a syringe pump, phenol red solution is infused through the outer ports while pH 9 solution is infused through the center port. Hydrodynamic focusing15 is visualized with the outer fluid flow forcing the inner flow into a smaller stream (Figure 7C). Laminar flow in the device prevents convective mixing, and gradual diffusive mixing is shown along the length of the channel (arrows). Devices like the one shown can be used to teach concepts (e.g., diffusion, laminar flow) in fluid dynamics and biotransport. Alternatively, students may be invited to design and fabricate their own devices, a project that can be undertaken in a regular laboratory session lasting two to three hours8.
Figure 7: PETLs in the classroom. (A) The device is fabricated using four layers of 5 mil PET/EVA film. The second layer (from right to left) features circular chambers that will be positioned under the channel path. (B) The finished device has been loaded with the pH indicator phenol red (2 mM, yellow) and a transparent pH 9 solution (center channel). Phenol red turns magenta when in contact with basic solutions. The boxes indicate areas shown in (C) through (F). (C) Hydrodynamic focusing. (D, E) laminar flow and diffusion. (E, F) micromixing. White scale bar is 2 mm in all panels. Please click here to view a larger version of this figure.
Supplementary Figure 1: Bumper/port perforation. (A) A drill press setup holding a rotary tool facilitates bumper perforation. Drill bits of sizes 1/32” and 3/64” are used. (B) The process is efficient, and large numbers of bumpers can be processed in a short amount of time. (C) Biopsy punch perforation is an alternative to drilling. (D) A revolving punch plier is used for larger perforations. These perforations can be used to load large specimens (by liquid withdrawal instead of infusion) or as media reservoirs. Please click here to download this figure.
Supplementary Figure 2: Tubing. (A) Laboratory or medical/surgical tubing (1/32” ID, 3/32” OD) is the simpler option. It is flexible and easy to cut. It requires the use of (18 G) blunt needles. (B) One of the needles is attached to the syringe using the (pink) Luer adaptor, which is removed from a second needle so it can be fitted to the tubing. (C) Researchers already familiar with PEEK tubing (0.010” ID, 1/32” OD) may use it with PETLs. (D) PEEK fittings. (E) Syringe pump set up is the same for both types of tubing. (F) Laboratory or medical/surgical tubing set up will require perforations with a 3/64” drill bit, while PEEK tubing will need 1/32” perforations. Perforations made with a 1 mm biopsy punch can accommodate both sets of tubing. Please click here to download this figure.
Supplementary Figure 3: Alignment using a wire jig. The design of the device can include perforations that may serve as guides for the alignment of several layers. Wire jigs are commercially available for around 20 USD. Please click here to download this figure.
Supplementary Figure 4: Size limitations. Although craft cutters are able to cut straight channels that are ~100 µm in width (A), the accuracy of the cut patterns is greatly diminished for features that measure 150 µm or less (B). The dimensions next to the shapes indicate channel width. Please click here to download this figure.
Supplementary Table 1: Time and cost for fabrication of microfluidic chip in PDMS. *Manufacture time when wafer/mold is readily available and PDMS can be cured using an oven. Any design modification represents a delay of several days. Please click here to download this table.
Supplementary Table 2: Time and cost for fabrication of PETL microfluidic chip. Please click here to download this table.
While microfluidics are increasingly present in the toolbox of laboratories around the world, the pace of adoption has been disappointing, given the potential for its positive impact16. Low cost and high efficiency of microfluidic device fabrication are essential to accelerate adoption of this technology in the average research laboratory. The method described here uses multiple film layers to create two and three-dimensional devices at a fraction of the time and cost required by lithographic methods. Standard lithography costs thousands of dollars (USD) to start-up, and requires days to fabricate, PETL fabrication startup cost is less than 350 USD and devices can be fabricated in minutes. This facilitates their adoption not only in the research laboratory but also in settings where rapid iteration is advantageous (e.g., prototyping for standard PDMS devices), or where the industrial production of inexpensive, disposable devices are required. For example, PETLs can be fabricated using biodegradable materials, and may be adapted for their use in the healthcare field, making them ideal as diagnostics tools. They can be used in the classroom, either as pre-fabricated learning materials or as a creative challenge, in which students design, fabricate and test their own devices.
PETL fabrication is meant to be uncomplicated. It is, however, helpful to identify critical steps and current limitations of this technique. Some users will find that gas exchange in PETL devices is reduced in comparison to PDMS devices, which is compensated for by having continuous flow of media during experimentation. Another limitation is sizing. Channels and other features smaller than 150 µm are below the resolution limit of the cutter (Supplementary Figure 4). We recommend working with channels with a width in the range of 200–900 µm. These limits are flexible, and tend to vary particularly at the upper threshold. For example, channels with a height of 75 µm will collapse when the width of the channel is 950 µm or more, but remain open if the height increases. Although the architecture of the device will vary according to the application, we routinely use channels with a height of 75 or 125 µm, and a width of 400–600 µm.
Attention to detail when aligning layers and bumpers is important. Most of the few complications arising from PETL fabrication are a result of alignment issues. Exposed EVA at the time of laminating can adhere to the internal rollers and make them unusable. Fluid infusion may be blocked by an ill-positioned bumper. Fortunately, PETLs are not only cheap but are also rapidly built, therefore faulty devices can be easily replaced or modified.
PETLs can withstand infusion flow rates similar to those used in other microfluidic devices. Although 0.01 to 100 µL/min is the range used by our group, flow rates of up to 500 µL/min (and perhaps higher when using hand-actuated micropipettes) can be used. We have found that PETLs can withstand pressures in the range of 30 to 57 psi8. Syringe pumps are recommended for most experimental settings, although they are not an absolute requirement. In the classroom, burettes were used to test student’s devices15. Peristaltic pumps are useful in certain settings like cell culture, particularly since gas exchange is limited in PETLs. PDMS may be more advantageous in this regard, although leaching may be a concern5. We have attempted to produce hybrid PET/EVA-PDMS, yet EVA will not directly adhere to PDMS; it is possible that surface modification of the latter (e.g., plasma treatment or surfactant treatments) may solve this issue. Another approach that can be compared to PETLs is the micro machining of channels using CO2 laser ablation17,18 of PMMA. We have found that laser cutting is incompatible with PET/EVA film, since the heat produced tends to cure EVA and to produce uneven channel edges. The use of adequate laser equipment might also significantly increase fabrication costs.
In summary, PETLs offer multiple advantages over current technologies: (i) Costs are significantly lower than traditional methods due to the use of consumer-grade materials and equipment, making it easily accessible to both researchers and students. (ii) Devices can be designed, cut, and assembled within minutes, which allows for rapid prototype iteration and facilitates time-effective experimentation. (iii) Multiple devices can be fabricated simultaneously, allowing for high throughput production. (iv) A variety of materials can be incorporated, adding versatility and allowing extensive customization. Current and future development of new functionalities using this technology rests in the creativity and requirements of new users. Wide adoption of PETL microfluidic devices will likely result in the inclusion of new materials and configurations, as users develop novel designs and approaches for their particular needs.
The authors have nothing to disclose.
The work in this manuscript was supported in part by the National Science Foundation (NSF) (Grant No. CBET-1553826) (and associated ROA supplement) and the National Institutes of Health (NIH) (Grant No. R35GM124935) to J.Z., and the Notre Dame Melchor Visiting Faculty fund to F.O. We would like to thank Jenna Sjoerdsma and Basar Bilgiçer for providing mammalian cells and culture protocols and Fabio Sacco for assistance with supplementary figures.
Biopsy punch (1mm) | Miltex | 33-31AA | Optional, replaces rotary tool set up |
Blunt needles | Janel, Inc. | JEN JG18-0.5X-90 | Remove plastic and attach to Tygon tubing |
Coverslips | Any | 24 x 60 mm are preferred | |
Cutting Mat and blades | Silhouette America or Nicapa | www.silhouetteamerica.com/shop/blades-and-mats | Re-use/Disposables |
Double-sided tape | Scotch/3M | 667 | Small amounts, any width or brand |
PEEK tubing | IDEX/any | 1581L | Different configurations available. Consider using Tygon tubing intead, if not already using PEEK |
PET/EVA thermal laminate film | Scotch/3M & Transcendia | TP3854-200,TP5854-100 & transcendia.com/products/trans-kote-pet | 3 – 6 mil (mil = 1/1000 inch) laminating pouches or rolls. |
PVC film – Cling Wrap | Glad / Any | Food wrapping | |
Rotary tool-drill | Dremel/Any | 200-121 or other | 1/32 and 3/64" drill bits from Dremel recommended |
Rubber Roller | Speedball | 4126 | To facilitate adhesion, any brand will work |
Scissors & tweezers | Any | Fiskars-Inch-Titanium-Softgrip-Scissors |Cole-Parmer –# UX-07387-12 | Quality brands are recommended |
Silhouette CAMEO Craft cutter | Silhouette America | www.silhouetteamerica.com/shop/cameo/SILHOUETTE-CAMEO-3-4T | Preferred craft cutter |
Silhouette Studio software | Silhouette America | www.silhouetteamerica.com/software | Controls the craft cutter and provides drawing tools (free download MAC and PC) |
Syringe Pump | Harvard Apparatus or New Era | 70-4504 or NE-300 | Pumps are ideal, pipettes or burettes can be used. |
Syringes | Any | 1-3mL | |
Thermal laminator | Scotch/3M | TL906 | Standard home/office model |
Tygon tubing (E-3603) | Cole-Parmer | EW-06407-70 | Use with blunt needle tips |
Vinyl furniture bumpers | DerBlue/3M/ Everbilt | Clear, self-adhesive (6 x 2 mm and 8 x 3 mm) | Round bumpers are recommended |