We detail a method to fabricate three-dimensional paper-based microfluidic devices for use in the development of immunoassays. Our approach to device assembly is a type of multilayer, additive manufacturing. We demonstrate a sandwich immunoassay to provide representative results for these types of paper-based devices.
Paper wicks fluids autonomously due to capillary action. By patterning paper with hydrophobic barriers, the transport of fluids can be controlled and directed within a layer of paper. Moreover, stacking multiple layers of patterned paper creates sophisticated three-dimensional microfluidic networks that can support the development of analytical and bioanalytical assays. Paper-based microfluidic devices are inexpensive, portable, easy to use, and require no external equipment to operate. As a result, they hold great promise as a platform for point-of-care diagnostics. In order to properly evaluate the utility and analytical performance of paper-based devices, suitable methods must be developed to ensure their manufacture is reproducible and at a scale that is appropriate for laboratory settings. In this manuscript, a method to fabricate a general device architecture that can be used for paper-based immunoassays is described. We use a form of additive manufacturing (multi-layer lamination) to prepare devices that comprise multiple layers of patterned paper and patterned adhesive. In addition to demonstrating the proper use of these three-dimensional paper-based microfluidic devices with an immunoassay for human chorionic gonadotropin (hCG), errors in the manufacturing process that may result in device failures are discussed. We expect this approach to manufacturing paper-based devices will find broad utility in the development of analytical applications designed specifically for limited-resource settings.
Paper is widely available in a range of formulations or grades, can be functionalized to tune its properties, and can transport fluids autonomously by capillary action or wicking. If paper is patterned with a hydrophobic substance (e.g., photoresist1 or wax2), the wicking of fluids can be controlled spatially within a layer of paper. For example, an applied aqueous sample can be directed into a number of different zones to react with chemical and biochemical reagents stored within the paper. These paper-based microfluidic devices have been demonstrated to be a useful platform for the development of portable and inexpensive analytical assays3,4,5,6,7. Applications of paper-based microfluidic devices include point-of-care diagnostics8, monitoring of environmental contaminants9, detection of counterfeit pharmaceuticals10, and delocalized healthcare (or "telemedicine") in limited-resource settings11.
Multiple layers of patterned paper can be assembled into an integrated device where hydrophilic zones from neighboring layers (i.e., above or below) connect to form continuous fluidic networks whose inlets and outlets may be coupled or left independent.12 Each layer can comprise a unique pattern, which enables the spatial separation of reagents and multiple assays to be performed on a single device. The resulting three-dimensional microfluidic device is not only capable of wicking fluids to enable analytical assays (e.g., liver function tests13 and electrochemical detection of small molecules14), but it can also support a number of sophisticated functions (e.g., valves15 and simple machines16) common to traditional microfluidic approaches. Importantly, because paper wicks fluids by capillary action, these devices can be operated with minimal effort from the user.
Since reagents can be stored within the three-dimensional architecture of a paper-based device, complex protocols can be reduced to a single addition of aqueous sample to a device. Recently, we introduced a general three-dimensional device architecture that can be used for the development of paper-based immunoassays using the wax-printing technique to create patterned layers.17,18 These studies focused on how aspects related to the design of the device-number of stacked layers used, composition of the layers, and the pattern of the three-dimensional microfluidic network-controlled the overall performance of the immunoassay. Ultimately, we were able to use these design rules to facilitate the rapid development of a multiplexed immunoassay19. In this manuscript, a previously developed immunoassay for human chorionic gonadotropin (hCG; pregnancy hormone)17 is used as an example to illustrate the strategies that we have developed for the assembly and manufacture of three-dimensional paper-based immunoassays. Accordingly, we focus on the assembly and operation of a device rather than the development of an assay.
In a sandwich immunoassay, which is the format used to detect hCG, a capture antibody specific to one subunit of the hormone is coated onto a solid substrate, which is then blocked to limit the non-specific adsorption of a sample or any subsequent reagent. This substrate is most often a polystyrene microwell plate (e.g., for an enzyme-linked immunosorbent assay or ELISA). The sample is then added to a well and allowed to incubate for a period of time. After rigorous washing, an antibody specific to the other subunit of hCG is added and allowed to incubate. This detection antibody may be conjugated to a colloidal particle, enzyme, or fluorophore in order to produce a measurable signal. The well is again washed prior to interpreting the results of an assay (e.g., using a plate reader). While commercial kits rely on this time-consuming multistep process, all of these steps can be performed rapidly in paper-based microfluidic devices with minimal intervention to the user.
The device used for the hCG immunoassay comprises six active layers, which are, from top to bottom, used for sample addition, conjugate storage, incubation, capture, wash, and blot (Figure 1). The sample addition layer is made from qualitative filter paper. It facilitates the introduction of a liquid sample and protects the reagents in the conjugate layer from contamination from the environment or accidental contact by the user. The conjugate layer (qualitative filter paper) holds the color-producing reagent (e.g., colloidal gold-labeled antibody) for the immunoassay. The incubation layer (qualitative filter paper) allows the sample to travel laterally within the plane of the paper to promote binding of the analyte with reagents before reaching the next layer, the capture layer. The capture layer (nylon membrane) contains ligands specific for the analyte adsorbed to the material. After the assay is completed, this layer is revealed to enable visualization of the completed immunocomplex. The wash layer (qualitative filter paper) draws excess fluids including free conjugate reagents away from the face of the capture layer into the blot layer (thick chromatography paper). The six-layer device is held together by five layers of patterned, double-sided adhesive: four layers of permanent adhesive maintain the integrity of the assembled device and one layer of removable adhesive facilitates peeling of the device to inspect the results of the immunoassay on the capture layer.
For the purpose of this manuscript, we use only negative and positive control samples of hCG (0 mIU/mL and 81 mIU/mL, respectively) to provide representative results of a paper-based immunoassay, which permits a dedicated discussion of the relationship between fabrication methods and the performance of a device. In addition to demonstrating how to manufacture devices successfully, we highlight several manufacturing errors that could lead to the failure of a device or irreproducible assay results. The protocol and discussion detailed in this manuscript will provide researchers with valuable insight into how paper-based immunoassays are designed and fabricated. While we focus our demonstration on immunoassays, we anticipate that the guidelines presented herein will be broadly useful for the manufacture of three-dimensional paper-based microfluidic devices.
1. Preparation of Paper-based Microfluidic Device Layers
2. Preparation of Paper Layers: Sample Addition, Conjugate Storage, Incubation, and Wash Layers
3. Preparation of Nylon Membrane Layer: Capture Layer
4. Creating Hydrophobic Barriers in the Printed Layer
5. Preparation of Adhesive Layers
6. Backing of Device Layers with Adhesive
7. Treatment of Conjugate Layer with Reagents for Immunoassays Prior to Device Assembly
8. Treatment of Lateral Channel with Reagent for Immunoassays Prior to Device Assembly
9. Treatment of Capture Layer with Reagents for Immunoassays Prior to Device Assembly
10. Assembly of Three-dimensional Paper-based Microfluidic Devices
11. Performing a Paper-based Immunoassay
Obtaining reproducible assay performances in three-dimensional paper-based microfluidic devices relies on a fabrication method that ensures consistency among devices. Towards this goal, we have identified a number of manufacturing processes and material considerations, and discuss them here in the context of demonstrating a paper-based immunoassay. We use a wax printing method to form hydrophobic barriers within paper-based microfluidic devices (Figure 2A).2 This method is ideal because it relies only on widely available office equipment, requires minimal procedural steps to complete, and does not require the use of chemicals (e.g., photoresists) that might interfere with protein adsorption or alter the wettability of paper fibers. Further, wax printing produces fluidic pathways with reproducible dimensions, which is critical for assays with repeatable performances and duration times. After the hydrophobic barriers are formed, adhesive sheets are applied to layers to facilitate assembly of three-dimensional devices (Figure 2B). Any reagents required for the immunoassay can be applied after the adhesive film is added to the back of a layer (Figure 2C). This procedure is useful for fabrication processes in an academic laboratory because many devices can be prepared in parallel. The assembly process for an immunoassay device is completed after all layers of the device are stacked and laminated together (Figure 2D). We add sample to begin the assay. In this example, we use a urine control set for pregnancy tests, which contains negative and positive samples of hCG in buffer, as samples to demonstrate the operation of our devices and the reproducibility of assays performed using them. Two aliquots of wash buffer are then added sequentially. Once the final aliquot of wash buffer has completely entered the device, the assay is considered complete. The top three layers are then peeled away to reveal the capture layer (Figure 3A). This step irreversibly damages the device ensuring that it cannot be used again. The completion of a paper-based immunoassay results in a qualitative color readout that can indicate a negative or positive output upon visual inspection. The objectivity of these results is apparent in uncorrected images acquired using a flatbed scanner (Figure 3B).
Failed experiments can often highlight certain procedural steps whose importance may be otherwise imperceptible when the analysis of an experiment is focused on successful results. We demonstrate three errors in the manufacture and assembly of three-dimensional paper-based microfluidic devices that result in failures of the immunoassay: (i) Occasionally, device failures are not apparent until after an assay is completed. For example, a misalignment between layers comprising the incubation channel and capture zone can cause the development of an irregular pattern in the positive signal, which may result in the misinterpretation of the qualitative signal by a user (Figure 4A). (ii) If the wax is not printed in a sufficient amount or not allowed to melt completely through the full thickness of the paper, then the integrity of the resulting hydrophobic barriers may be compromised. Incomplete formation of these barriers will cause a loss of control over wicking and lead to leaks within the device. For example, instead of confining flow to a channel within a layer, a semi-permeable wax barrier will allow fluid to wick elsewhere in the plane of the paper. Without defined channels, the sample is unlikely to reach the capture or wash layers. The user will perceive this kind of error as a greatly shortened assay duration time. We demonstrate this device failure by applying a solution of red food coloring to a layer whose wax pattern was not allowed to melt for the full 30 sec (Figure 4B). An immunoassay using such a layer was "completed" in 6 min, which is clearly different than the expected duration of 15 min. (iii) Assays that take longer than expected to complete may indicate a malfunction in the fabrication of a device. For example, improperly cut adhesive or occluded pores due to the application of an excessive amount of reagents (e.g., blocking agents or colloidal gold) could prohibit a sample or wash buffer from entering the device (Figure 4C).
Overall, our manufacturing protocol is useful to fabricate numerous paper-based microfluidic devices in parallel on a scale that is useful for an academic laboratory. We demonstrate the performance of the hCG paper-based immunoassay prepared using this method by performing 70 assays in parallel: 35 negative replicates and 35 positive replicates. For the purposes of this demonstration, we prepared a set of layers with the designs of our device, affixed the layers of paper with adhesive, and then cut the sheets into rows of devices. Each sheet was cut into 7 rows, which contained ten devices. This facilitated the arrangement of the layers onto the smaller acrylic frames where the layers are taped and then treated with reagents needed to perform the assays. This method of device preparation is suggested in a note in the protocol. Following the treatment of layers, the devices were assembled in strips of ten and then laminated. After the final device fabrication steps were completed, the devices remained in the strips of ten and sample was added to each device. We observed a 0% failure rate for devices fabricated using our protocol. We used an open-source image processing software20 to quantify the results of these assays. While a number of methods are available to analyze the intensity distribution in circular spots (e.g., radial or linear distributions)21 we measure the mean intensity from the green channel of an RGB image of the device using the entire detection spot as a region of interest.17,18,19 We then normalize the measurements of both positive and negative assays by subtracting the raw negative data (Figure 3B). We determined the coefficient of variation for each data set to be 1% for assays performed using negative samples and 3% for assays performed using positive samples.
Figure 1: Schematic of three-dimensional paper-based device. This illustration shows the hydrophobic and hydrophilic regions that define the fluidic pathway within the device, as well as the patterned layers of permanent and removable adhesive that hold layers together. Each layer is labeled by the function it performs in the assay. The red, blue, or green outline on each layer indicates the material used to fabricate that specific layer (red: chromatography paper, blue: nylon membrane, green: thick chromatography paper). Dimensions are given for each zone within the device in mm. Please click here to view a larger version of this figure.
Figure 2: Procedure used to fabricate immunoassays from three-dimensional paper-based microfluidic devices. (A) Images of the front and back of a sheet of chromatography paper patterned using wax printing before and after heating. (B) A sheet of chromatography paper backed with a film of patterned adhesive. (C) Treatments applied to the hydrophilic zones of a strip of patterned nylon membrane. (D) Assembly of strips of a multilayer device using a light box and alignment holes as a guide. Please click here to view a larger version of this figure.
Figure 3: Interpreting results of a paper-based immunoassay. (A) The top three layers of the paper-based device are peeled back to expose the capture layer and interpret the results of the assay. (B) Graphical representation of the performance of a paper-based immunoassay for hCG. The results depicted are the averages of 70 replicates performed simultaneously where 35 replicates each are used for positive and negative samples of hCG. Error bars represent the standard deviation of the data set. Uncorrected, representative images depicting positive (red color) and negative (white color) results from an hCG immunoassay are shown above their respective data. Please click here to view a larger version of this figure.
Figure 4: Examples of manufacturing errors. (A) Due to misalignment of the lateral channel above the capture layer, the positive signal is concentrated in a small area of the readout zone. A "wet" circular region (dashed outline) can be observed to the right of the readout zone resulting from contact between the misaligned lateral channel with the capture layer (left). Image of a positive readout on the capture layer of a properly aligned device (right). (B) Incomplete melting of wax throughout the thickness of a layer can lead to leaks within the device. Food coloring has been added to the solution to assist with visualization of sample in layers with incomplete or fully-formed hydrophobic barriers. (C) Improperly cut adhesive can block the fluidic network between layers of paper, which stops the flow of a sample. Please click here to view a larger version of this figure.
Figure 5: Manufacturing three-dimensional paper-based microfluidic devices. The schematic depicts the assembly and lamination of multiple layers of patterned paper into completed three-dimensional devices. In this example, 70 devices can be made simultaneously. The layers of adhesive and alignment holes have been removed from the schematic for simplification purposes. After assembly, individual devices can be removed and used in assays. The red, blue, and green outlines on each layer indicate the material used to fabricate that specific layer (red: chromatography paper, blue: nylon membrane, green: thick chromatography paper). The Scale bar = 25 mm except for the separate device (right), which has dimensions of 12 x 28 mm2. Please click here to view a larger version of this figure.
Identifying a reproducible manufacturing strategy is an essential component of assay development.22 We use a sequential, layer-by-layer approach to manufacture three-dimensional paper-based microfluidic devices. In contrast to those methods that apply folding or origami techniques to produce multilayer devices from a single sheet of paper23,24 additive manufacturing offers a number of advantages: (i) Multiple materials can be incorporated into a single device architecture without modification to methods for the printing, alignment, or assembly of layers. (ii) Patterned adhesive films can be integrated into the assembly process. These films affix adjoining layers, and, based on the strength of the adhesive, can be reversible to enable peeling and evaluation of internal layers. Moreover, adhesives provide structural integrity to the three-dimensional device, which precludes the need for binder25 clips or machined enclosures.23 (iii) Individual sheets of US Letter chromatography paper can accommodate an array of replicates, which can greatly improve the throughput of laboratory-scale manufacturing (Figure 5). This is particularly beneficial when evaluating numerous experimental conditions that require technical replicates. By this approach, 70 three-dimensional paper-based devices can be prepared simultaneously. (iv) Similar multilayer lamination approaches are used for the high-volume manufacture of numerous commercial products in healthcare (e.g., wound care dressings and transdermal patches), which consequently lowers the production barrier to translating these three-dimensional paper-based microfluidic devices.
In addition to facilitating peeling and assembly, the choice of adhesive is also critical to the design of the three-dimensional fluidic network. An adhesive film can serve as an additional barrier between layers of paper, which can enable masking of hydrophilic zones on adjacent layers. In practice, the use of thin layers of adhesive is desirable. If the adhesive is too thick (e.g., many double-sided tapes), then the gap formed between layers of paper will be too large to facilitate wicking and must be filled with a hydrophilic substance (e.g., cellulose powder) to regain function.12 While this additional step adds complexity to manufacture and the substance used may interfere with some assays, these gaps become a useful feature for the production of controllable, fluidic push-down valves.15 Other forms of adhesive have been used in the manufacture of three-dimensional paper-based microfluidic devices. Adhesive sprays offer a simple method to affix layers to each other.26 Using this method, the adhesive material is applied uniformly onto both the hydrophobic and hydrophilic area of the paper. An advantage to this method is that additional equipment (e.g., knife plotter or laser cutter) is not needed to design the pattern for the adhesive layer. However, the conditions for the uniform application of the adhesive spray must be determined experimentally for each type of material used. The topography of the material may affect the adhesive-material interface and longer spray times may be needed for rougher surfaces. In addition, spraying adhesive onto the hydrophilic zones of the fluidic pathway may result in impaired wicking by altering the wettability of the paper. Alternatively, the use of stencils27 or screen printing8 may be used to pattern adhesive directly onto layers of patterned paper.
Two major considerations for the development of three-dimensional paper-based microfluidic devices are the choice of materials and the design of the fluidic network. (i) We select materials and combinations of materials based on wicking rate, wet strength, thickness, and protein-binding capacity. Wicking rate can influence the duration of an assay and the amount of time reagents have to react or bind within a layer. Different grades of paper are characterized by wicking rates based on, for example, the treatment of the paper, its porosity, and its thickness. It is possible to use multiple layers of paper to increase the effective wicking rate of a device.28 A good wet strength is desirable for applications that require handling (e.g., peeling an immunoassay) after the device has been saturated with a sample. Materials that are too thick or that cannot be passed through commercial printer due to fragility will require an alternative method to produce patterned channels (e.g., photolithography). However, in contrast, thicker materials are ideal for blot layers (or sinks) to draw fluids through the device. Many grades of nylon membranes are available commercially, which may differ in their ability to bind proteins irreversibly to the capture zone. Material substitutions (e.g., nitrocellulose instead of a nylon membrane) can also influence binding capacity, which may affect the sensitivity of the assay. (ii) The use of symmetry in the design of fluidic networks ensures that the unique channels patterned into three-dimensional devices behave identically (e.g., filled simultaneously), which is critical for multiplexed assays.19 Symmetry further simplifies layer design, assists with layer alignment when assembling full sheets of devices, and can minimize waste. Modifications to the device design can influence the performance of the assay. For example, increasing the length of the lateral channel in the incubation layer will affect the duration of the assay, because the fluid will wick a proportionately longer distance before reaching the outlet.17 In applications that rely on the binding of a target biomolecule, a longer assay time may be advantageous because it can increase the fraction of bound, labeled species prior to immobilization on the capture layer.
In conclusion, we have presented a method to fabricate three-dimensional paper-based microfluidic devices that support the development of immunoassays. This method, which uses a type of additive manufacturing to produce multilayer devices, facilitates the production of devices at a scale that is suitable for laboratory research. The protocol described in this manuscript is specific for paper-based immunoassay devices; however, we expect the procedures related to the assembly of these immunoassays-wax printing, patterning adhesive, aligning layers, and lamination-will be readily extendable to numerous three-dimensional paper-based microfluidic device architectures. An understanding of fabrication methodology can lead to the development of new point-of-care assays with a broad range of applications in healthcare, environment, and agriculture.
The authors have nothing to disclose.
This work was supported by Tufts University and by a generous gift from Dr. James Kanagy. This material is based upon work supported by the National Science Foundation Graduate Research Fellowship Program under Grant No. (DGE-1325256) that was awarded to S.C.F. D.J.W. was supported by a U.S. Department of Education GAANN fellowship. We thank Dr. Jeremy Schonhorn (JanaCare), Dr. Jason Rolland (Carbon3D), and Rachel Deraney (Brown University) for helping develop the design of the three-dimensional paper-based microfluidic device and immunoassay.
Illustrator CC | Adobe | to design patterns for layers of paper and adhesive | |
Xerox ColorQube 8580 printer | Amazon | B00R92C9DI | to print wax patterns onto layers of paper and Nylon |
Isotemp General Purpose Heating and Drying Oven | Fisher Scientific | 15-103-0509 | to melt wax into paper |
Artograph LightTracer | Amazon | B000KNHRH6 | to assist with alignment of layers |
Apache AL13P laminator | Amazon | B00AXHSZU2 | to laminate layers together |
Graphtec CE6000 Cutting Plotter | Graphtec America | CE6000-40 | to pattern adhesive films |
Swingline paper cutter | Amazon | B0006VNY4C | to cut paper or devices |
Epson Perfection V500 photo scanner | Amazon | B000VG4AY0 | to scan images of readout layer |
economy plier-action hole punch | McMaster-Carr | 3488A9 | to remove alignment holes |
Whatman chromatogrpahy paper, Grade 4 | Sigma Aldrich | WHA1004917 | |
Fisherbrand chromatography paper (thick) | Fisher Scientific | 05-714-4 | to function as blot layer |
Immunodyne ABC (0.45 µm pore size ) | Pall Corporation | NBCHI3R | to function as material for capture layer |
removable/permanent adhesive-double faced liner | FLEXcon | DF021621 | to facilitate peeling |
permanent adhesive-double faced liner | FLEXcon | DF051521 | |
wax liner | FLEXcon | FLEXMARK 80 D/F PFW LINER | to assist with patterning adhesive |
acrylic sheet | McMaster-Carr | 8560K266 | to fabricate frame |
self-adhesive sheets | Fellowes | CRC52215 | to use as protective slip |
absolute ethanol | VWR | 89125-172 | to sanitize work area |
bovine serum albumin | AMRESCO | 0332 | |
Sekisui Diagnostics OSOM hCG Urine Controls | Fisher Scientific | 22-071-066 | to use as positive and negative samples |
anti-β-hCG monoclonal antibody colloidal gold conjugate (clone 1) | Arista Biologicals | CGBCG-0701 | to treat conjugate layer |
goat anti-α-hCG antibody | Arista Biologicals | ABACG-0500 | to treat capture layer |
10X phosphate buffered saline | Fisher Scientific | BP3991 | |
Oxoid skim milk powder | Thermo Scientific | OXLP0031B | |
Tween 20 | AMRESCO | M147 |