Metering the Capillary-Driven Flow of Fluids in Paper-Based

Apr 22, 2010 - Abstract. This article describes an exceedingly simple and low-cost method for metering the capillary-driven flow rate of fluids within...
0 downloads 9 Views 5MB Size
Download PDF
Anal. Chem. 2010, 82, 4181–4187

Metering the Capillary-Driven Flow of Fluids in Paper-Based Microfluidic Devices Hyeran Noh and Scott T. Phillips* Department of Chemistry, The Pennsylvania State University, University Park, Pennsylvania 16802 This article describes an exceedingly simple and low-cost method for metering the capillary-driven flow rate of fluids within three-dimensional (3D) microfluidic, paper-based analytical devices (µPADs). Initial prototypes of 3D µPADs control the spatial distribution of fluids within a device, but they provide little control over how quickly (or slowly) fluids move within the device. The methods described in this article provide control over when and how quickly a fluid is distributed into detection zones. These methods are inexpensive (the metering regions are composed of paraffin wax), the devices are easy to fabricate, and they are capable of controlling the flow of fluids to detection zones with precise time delays (e.g., (6% of the total wicking time). We anticipate that this type of precise control over fluid distribution rates will be useful particularly for point-of-care assays that require multiple steps (where each step requires that the reagents interact for a defined period of time) or for simultaneously displaying the results of multiple different assays on a single device. Paper-based microfluidic devices1-7 and other paper-based detection platforms8-19 are emerging as convenient and low-cost * Corresponding author. E-mail: [email protected]. (1) Martinez, A. W.; Phillips, S. T.; Butte, M. J.; Whitesides, G. M. Angew. Chem., Int. Ed. 2007, 46, 1318–1320. (2) Martinez, A. W.; Phillips, S. T.; Carrilho, E.; Thomas, S. W.; Sindi, H.; Whitesides, G. M. Anal. Chem. 2008, 80, 3699–3707. (3) Martinez, A. W.; Phillips, S. T.; Wiley, B. J.; Gupta, M.; Whitesides, G. M. Lab Chip 2008, 8, 2146–2150. (4) Martinez, A. W.; Phillips, S. T.; Whitesides, G. M. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 19606–19611. (5) Ellerbee, A. K.; Phillips, S. T.; Siegel, A. C.; Mirica, K. A.; Martinez, A. W.; Striehl, P.; Jain, N.; Prentiss, M.; Whitesides, G. M. Anal. Chem. 2009, 81, 8447–8452. (6) Carrilho, E.; Phillips, S. T.; Vella, S. J.; Martinez, A. W.; Whitesides, G. M. Anal. Chem. 2009, 81, 5990–5998. (7) (a) Nie, Z.; Nijhuis, C. A.; Gong, J.; Chen, X.; Kumachev, A.; Martinez, A. W.; Narovlysansky, M.; Whitesides, G. M. Lab Chip 2010, 10, 477– 483. (b) Dungchai, W.; Chailapakul, O.; Henry, C. S. Anal. Chem. 2009, 81, 5821–5826. (8) Pelton, R. Trends Anal. Chem. 2009, 28, 925–942, and references cited therein. (9) Westman, E.; Ek, M.; Wagberg, L. Holzforschung 2009, 63, 33. (10) Abe, K.; Suzuki, K.; Citterio, D. Anal. Chem. 2008, 80, 6928–6934. (11) Hossain, S. M. Z.; Luckman, R. E.; Smith, A. M.; Lebert, J. M.; Davies, L. M.; Pelton, H.; Filipe, C. D. M.; Brennan, J. D. Anal. Chem. 2009, 81, 5474–5483. (12) Li, X.; Tian, J.; Nguyen, T.; Shen, W. Anal. Chem. 2008, 80, 9131–9134. (13) Zhao, W.; Ali, M. M.; Aguirre, S. D.; Brook, M. A.; Li, Y. Anal. Chem. 2008, 80, 8431–8437. (14) Su, S.; Ali, M. M.; Filipe, C. D. M.; Li, Y.; Pelton, R. Biomacromolecules 2008, 9, 935–941. 10.1021/ac100431y  2010 American Chemical Society Published on Web 04/22/2010

platforms for running assays with microliter volumes of fluids.2 3D paper-based microfluidic devices (µPADs)4 are particularly useful because they permit fluid movement in the x-, y-, and z-directions, and therefore, they can accommodate more assays on a smaller footprint than typical 2D, lateral-flow devices.1 A 3D µPAD is capable of distributing a sample from a single entry point into hundreds of test regions.4 In addition, 3D µPADs are (i) exceedingly inexpensive, (ii) easily fabricated for rapid prototyping of new designs, (iii) made from abundant raw materials, (iv) conveniently incinerated for rapid disposal of hazardous waste, and (v) stand-alone devices that do not require external pumps or other complicated equipment to move fluids. These attributes make µPADs, and 3D µPADs in particular, an exciting platform for building a new generation of low-cost, high-performance pointof-care diagnostic devices with analytical characteristics that eventually may surpass polymer-based microfluidic devices.20,21 3D µPADs are a nascent technology, however, and substantial development is needed before their full capabilities can be realized. Certain useful features, like the ability to control flow rate, interaction times between sample and reagents, and mixing of fluids, are well developed for polymer- and glass-based microfluidic devices, but similar technologies are unavailable for µPADs. We reason that control over flow rate (by means of metering) will enable better control over mixing and interaction times of reagents and, therefore, will constitute the first step toward developing these other useful features as well. We also anticipate that the development of passive flow control in µPADs will broaden the potential applications for µPADs. Many colorimetric assays that are used in the clinic, for example, require sequential mixing of reagents with the analyte and require defined incubation times at each step. Flow control within µPADs should help enable the translation of these types of multistep assays onto paper without increasing the cost of the devices. Likewise, flow control within 3D µPADs should enable time-controlled signal amplification reactions within a device, such as enzyme-linked antibody-based assays commonly associated with enzyme-linked immunosorbent assays (ELISA). (15) Ali, M. M.; Aguirre, S. D.; Xu, Y.; Filipe, C. D. M.; Pelton, R.; Li, Y. Chem. Commun. 2009, 6640–6642. (16) Su, S.; Nutiu, R.; Filipe, C. D.; Li, Y.; Pelton, R. Langmuir 2007, 23, 1300– 1302. (17) Wong, R.; Tse, H. Lateral Flow Immunoassay; Humana Press: New York, 2009. (18) Edwards, R. Immunodiagnostics; Oxford University Press: Oxford, UK, 1999. (19) von Lode, P. Clin. Biochem. 2005, 38, 591–606. (20) Martinez, A. W.; Phillips, S. T.; Carrilho, E.; Whitesides, G. M. Anal. Chem. 2010, 82, 3–10. (21) Whitesides, G. M. Nature 2006, 442, 368–373.

Analytical Chemistry, Vol. 82, No. 10, May 15, 2010

4181

We anticipate that an early application of metering in 3D µPADs will be in the context of devices that run several different types of assays to detect multiple biomarkers simultaneously. Quantitative versions of colorimetric assays are time-dependent, but the ideal diagnostic device would provide the results of all assays simultaneously. To achieve this goal, the device must be capable of synchronizing the assays, and the most straightforward method for achieving synchronization is to control when each assay begins. We reason that methods for metering flow rate will provide this capability by precisely controlling the time required for the sample to wick into each assay conduit. The goal of this article is to describe in detail a method for metering the flow rate within paper-based devices. The end-goal of this effort is to enable cost-minded medical practitioners in the developing world to accurately and rapidly test for specific biochemical markers of various diseases.22-26 Fully developed µPADs also should enable low-cost assays for a broad spectrum of applications in the developed world as well, such as routine home-based and personalized medicine, rapid tests in emergency rooms, measurements of the levels of contaminants in food and water, and diagnostic applications in the military.27,28 EXPERIMENTAL DESIGN Initial Prototypes of 3D µPADs versus 3D µPADs with Flow Control. Initial prototypes of 3D µPADs4 consist of stacked, alternating layers of patterned paper and micropatterned, doublesided adhesive tape (Figure 1a). Microfluidic channels are formed in these devices when the hydrophilic paper is patterned with wax, which creates hydrophobic walls through the entire thickness of the paper. These devices are capable of distributing a sample from the top of a device into multiple detection zones on the bottom of the device. In the device shown in parts a and c of Figure 1, for example, a 10 µL sample of water traveled from the 2.4 mm diameter circular entry zone patterned into Whatman chromatography paper no. 1 (on the top layer of the device) down to the fifth layer of the device, where the dye Yellow 5 had been predeposited and dried into the paper prior to assembling the device. The resulting yellow aqueous solution then wicked into four circular regions of hydrophilic paper on the bottom of the device; these regions mimic detection zones for colorimetric assays. This entire wicking and distribution process requires ∼1 min and is consistent across the four detection zones (i.e., 53 ± 6 s). The rate of wicking within this (or any) µPAD depends on the characteristics of the paper, the dimensions of the channels, the viscosity of the fluid, and the temperature and humidity of the environment. Controlling precisely when a fluid reaches a region within a typical 3D µPAD is difficult, but this type of precision is critical for next-generation µPADs, particularly in the (22) Chin, C. D.; Linder, V.; Sia, S. K. Lab Chip 2007, 7, 41–57. (23) Sia, S. K.; Linder, V.; Parviz, B. A.; Siegel, A.; Whitesides, G. M. Angew. Chem., Int. Ed. 2004, 43, 498–502. (24) Daar, A. S.; Thorsteinsdo´ttir, H.; Martin, D. K.; Smith, A. C.; Nast, S.; Singer, P. A. A. Nat. Genet. 2002, 32, 229–232. (25) Yager, P.; Edwards, T.; Fu, E.; Helton, K.; Nelson, K.; Tam, M. R.; Weigl, B. H. Nature 2006, 442, 412–418. (26) Mabey, D.; Peeling, R. W.; Ustianowski, A.; Perkins, M. D. Nat. Rev. Microbiol. 2004, 2, 231–240. (27) Nathan, B. Nat. Methods 2009, 6, 683–685. (28) Zhao, W.; van der Berg, A. Lab Chip 2008, 8, 1988–1991.

4182

Analytical Chemistry, Vol. 82, No. 10, May 15, 2010

Figure 1. Comparison of an initial prototype of a 3D µPAD4 (a) with a 3D µPAD that contains meters (b). Each paper-based device consists of fused layers of patterned paper and micropatterned, double-sided adhesive tape. The images in parts a and b show an expanded view of the devices to highlight the features in each layer. (c) Photograph of an initial prototype of a 3D µPAD that distributes 10 µL of water into four hydrophilic regions of paper. The time required for water to pass from entrance to exit was 53 ( 6 s (N ) 7). The dotted lines around the device mark the edges of the device. (d) Timelapsed photographs of 3D µPADs with meters. The dyes used in the photographs are (listed counterclockwise starting from the upper right): Yellow 5, green dye, Red 40, and Blue 1. The wax concentrations are (listed counterclockwise starting from the upper right): 0, 15.7, 31.4, and 47.1 µg µm-3. All measurements were taken at 25 °C and 40% humidity.

context of time-based assays that require the sample to reach the reagents at well-defined points in time. One approach for controlling the fluid distribution rate in µPADs is to fabricate new devices with distribution channels of different lengths and/or widths. Longer and wider channels, however, increase the volume of sample required for an assay and increase the footprint and cost of a device. There is an upper limit to the size of a paper-based device, especially when only microliter volumes of sample are available (e.g., tears, blood from a finger prick, saliva, sweat, etc.), and therefore, adjusting channel size is not an effective approach for tuning distribution rates. In addition, this approach (i.e., tuning channel size) is not practical when distribution rates must differ by several minutes or more on the same device; in this case, the sample volume and size requirements for the device are prohibitively large. An alternative and unexplored method for controlling flow rate in paper-based microfluidic devices is to modulate the wetting properties of the paper. Parts b and d of Figure 1 illustrate our design, which involves predepositing microgram quantities of paraffin wax per cubic micrometer of paper into specific locations within the µPAD. We refer to these paraffin wax-infused regions as “meters” because they affect the wetting properties of the paper and, hence, the rate at which a sample permeates through layers

in the z-direction of a 3D µPAD. These wax-based meters have little to no effect on the quantity of sample needed for an assay: the 3D µPAD in Figure 1b requires 7 µL of fluid to fill completely, as does the same device without added paraffin wax. Although, these meters cannot be adjusted while a sample is wicking through a device (in contrast to traditional, electronically controlled valves that are used in polymer- and glass-based microfluidic devices), this apparent weakness is offset by the ease with which wax-based meters are constructed and by the low cost and scalability of the meters. Choice of Paper. Whatman chromatography paper no. 1 (200 mm × 200 mm) was used without further adjustment of size. We chose this type of Whatman paper because of its uniform composition (relative to other types of paper) and because it lacks additives that affect flow rate. Choice of Material for Meters. We chose paraffin wax for constructing meters in paper because (i) it is composed of hydrocarbons (CnH2n+2) and, therefore, lacks functionality (e.g., amines or carboxylic acids) that could interact with and potentially sequester polar analytes in a sample; and (ii) it is insoluble in water, so the composition of the metering component will not change over time as the sample wicks through the µPAD. EXPERIMENTAL SECTION Aqueous Dyes for Revealing the Flow of Fluids within µPADs. Synthetic food dyes (Assorted Food Colors & Egg Dye; Wal-Mart brand) were used to give colorimetric responses and to track the distribution of fluids within a device. The synthetic food dyes contain the following components: RED 40 (disodium salt of 6-hydroxy-5-[(2-methoxy-5-methyl-4-sulfophenyl)azo]-2naphthalenesulfonic acid), BLUE 1 (disodium salt of ethyl[4-[p[ethyl(m-sulfobenzyl)amino]-R-(o-sulfophenyl)benzylidene]-2,5cyclohexadien-1-ylidene](m-sulfobenzyl)ammonium hydroxide inner salt plus p-sulfobenzyl and o-sulfobenzyl salts), YELLOW 5 (trisodium salt of 4,5-dihydro-5-oxo-1-(4-sulfophenyl)-4-[4-sulfophenylazo]-1H-pyrazole-3-carboxylic acid), and green dye (which is a 1:1 mixture of YELLOW 5 and BLUE 1). The dyes were used as 1:5 mixtures of dye-distilled water. Designing and Printing Microfluidic Channels in Paper. CleWin (PhoeniX Software, The Netherlands) was used for designing patterns in paper and adhesive tape. Designs were saved as postscript files, which were converted into PDF files for printing. A Xerox Phaser 8560N color printer was used for depositing solid wax (Genuine Xerox Solid Ink Black) onto paper in defined patterns according to the procedures reported by Carrilho et al.29 Printing quality was set at the highest resolution for photoquality printing. Printed papers were placed on a hot plate set at 150 °C for 2 min. During this time, the wax ink penetrated through the paper in the z-direction to create hydrophobic barriers within the paper. Solid inks are composed of a mixture of hydrophobic carbamates, hydrocarbons, and dyes; when combined, these ingredients melt at 120 °C. The patterned paper was cooled to room temperature and was ready for further processing after 10 s. Patterning Tape. An Epilog Laser (Epilog Mini, 45 W) CO2 laser cutter was used to cut holes in double-sided adhesive tape (29) Carrilho, E.; Martinez, A. W.; Whitesides, G. M. Anal. Chem. 2009, 81, 7091–7095.

(ACE plastic carpet tape 50106). The patterns for these holes were designed in CleWin, as described previously.4 Fabricating Meters in 3D µPADs. 3D µPADs with meters are prepared using a slightly modified version of the methods reported previously for patterning paper29 and tape.4 Meters are prepared in these devices by depositing 0.4 µL of a solution of paraffin wax (Aldrich; mp 58-62 °C) in hexanes (Aldrich) (with concentrations ranging from 1-50 mg mL-1) into a 2.4 mm diameter × 0.18 mm thick hydrophilic region of patterned paper. The hexanes evaporated within ∼30 s of depositing the wax solution. The layer of paper was flipped over, and another 0.4 µL of the same wax solution was applied to the back of the same region of paper. Since this method uses only microgram quantities of wax, eventually it may be compatible with waxbased printers for rapid and simultaneous printing of microfluidic channels and valves. Once the hexanes had evaporated, the paper was incorporated as a layer in the 3D µPAD; the method for assembling a 3D µPAD is described in ref 4. Measuring the Time Required for Green Dye to Pass from the Entrance to the End Point of a 3D µPAD. The entrance of each 3D µPAD was dipped into 10 µL of green dye solution. The time required to completely fill a device was determined when the circular end point at the bottom of the device was filled completely with green fluid. Each measurement was repeated seven times using seven separate 3D µPADs that were fabricated independently from one another. RESULTS AND DISCUSSION Fluid Dynamics within 3D µPADs that Contain Meters. The 3D µPAD shown in parts b and d of Figure 1 contains four meters (one in each of the four wicking paths) and the entire device is composed of nine layers of paper and tape. The first four layers distribute the fluid from one entry point to four exit regions. The fifth layer contains different concentrations of paraffin wax in each circular region of paper (the wax concentrations starting from the upper right and moving counterclockwise are 0, 15.7, 31.4, and 47.1 µg µm-3). The seventh layer contains different colored dyes (deposited into the paper as 1:5 dilutions of food coloring-distilled water); the dyes starting from the upper right and moving counterclockwise are Yellow 5, green dye, Red 40, and Blue 1. The ninth layer contains four hydrophilic regions of paper that serve as the end points for the fluid in the device. Layers six and eight are patterned adhesive tape. Addition of 10 µL of water to the circular entry zone (2.4 mm in diameter) of this device resulted in rapid distribution of the water (24 µg µm-3), the time required to completely fill a device increases dramatically (Figure 3b). In other words, wax-based meters can delay the filling time by up to 2 h or by as little as 30 s within a single device. Multilayer Meters. The relationship between total quantity of paraffin wax (i.e., the total wax from all meters in a conduit) and the time required for a sample to fill a device is nonlinear. For example, a given quantity of wax has very different effects depending on whether the wax is in a single meter or distributed across multiple meters in series. Increasing or decreasing the quantity of wax per meter provides coarse-level control over the wicking rate (Figure 3), but increasing the number of meters per conduit provides fine-level control (Figure 4). Figure 4a shows a 3D µPAD with multiple meters in series. Each meter contains a low concentration of paraffin wax (i.e., 1.6 or 3.2 µg µm-3, depending on the meter) distributed evenly through a 2.4 mm diameter × 0.18 mm thick region of paper. As the number of meters increases, the total quantity of paraffin wax in a conduit increases as well. Likewise, as the number of meters increases, the time required for fluid to wick from the entrance to end point increases. With multiple meters in series, the flow of a fluid within a particular conduit can be controlled with high precision using minimal quantities of wax. This latter point is important, particularly if the fluid contains hydrophobic components that could nonspecifically adsorb to the paraffin wax (e.g., hydrophobic small molecules or proteins); smaller quantities of paraffin wax should minimize this type of nonspecific adsorption. For a given quantity of wax, the time required for the fluid to reach the exit depends on the number of meters through which the wax is distributed. For example, a conduit containing 32 µg of paraffin wax divided between four meters (8 µg of paraffin wax per meter) retains fluid 3.7 min longer than a conduit that contains only two meters (16 µg of paraffin wax per meter). Likewise, a device with 48 µg of paraffin wax spread over six meters in series (8 µg per meter) retains the fluid 12.3 min longer than a device with the same amount of wax spread over three meters (16 µg per meter). Because of this nonlinear relationship between the number of meters and the fluid retention time, we can use a constant quantity of wax and alter the flow rate dramatically simply by changing the number of meters in the device. Effect of Wax on Paper. The presence of wax changes the wetting properties of paper and, consequently, the rate at which (30) This change in the quantity of wax requires fabrication of a new 3D µPAD; the fabrication process, however, requires only a few minutes when the patterned paper and tape are pre-prepared. (31) Mukhopadhyay, R. Anal. Chem. 2008, 80, 3949.

Figure 4. Meters in series: (a) expanded view of a device that contains meters in series; (b) number of meters versus the time required for a sample to wick from the top of the 3D µPAD to the end point. The circles represent meters that contain 1.6 µg of paraffin wax per µm-3 of paper, and the open squares represent meters that contain 3.2 µg of paraffin wax per µm-3 of paper. The data points and error bars represent the averages and standard deviations of seven measurements. (c) Total quantity of paraffin wax (per device) versus retention time. Each meter consists of a 2.4 mm diameter × 0.18 mm thick (5.1 µm3) region of paper. The meaning of the circles and open squares is the same as in part b. For example, the square data point at 82 µg of paraffin wax per device was obtained from a device that contained five meters in series, each containing 3.2 µg of paraffin wax per µm-3 of paper.

water wicks into the paper in the z-direction. The paraffin wax increases both the contact angle of paper and the time required for a sample to absorb into the paper (Figure 5a). Ultimately, there is a maximum quantity of wax that can be applied to paper if absorptive capabilities are to be maintained (see, for example, the intersection point in Figure 5a). In other words, beyond 78 µg of wax per µm-3 of paper (the intersection point), additional wax will not significantly alter the contact angle but the wicking rate will be much slower. Paper infused with 78 µg of paraffin wax per µm-3 has a contact angle of 136°, whereas paraffin wax itself has a contact Analytical Chemistry, Vol. 82, No. 10, May 15, 2010

4185

Figure 6. Effect of sample composition on the total fill time for a 3D µPAD. The design of the 3D µPADs is shown in Figure 3a. We tested two meters (one with 5 µg of paraffin wax per µm3 of paper and the other with 31 µg of paraffin wax per µm3 of paper), and we normalized this data to wicking rates obtained in the absence of a wax meter. Each experiment was repeated seven times.

Figure 5. Effect of wax on the wetting properties of paper. (a) Graph of the total quantity of wax (in 2.4 mm diameter × 0.18 mm thick pieces of Whatman chromatography paper no. 1) versus the contact angle of the paper when exposed to a drop of distilled water (circular data points). This graph is superimposed on a graph of the quantity of wax versus the time required for a 10 µL drop of distilled water to penetrate fully into the same area of paper (square data points). (b) Scanning electron microscope images of pure Whatman chromatography paper no. 1 and Whatman chromatography paper no. 1 infused with 78.4 µg µm-3 paraffin wax.

angle of 107°. The larger contact angle of wax-infused paper relative to paraffin wax is due to the combination of increased hydrophobicity of paper (compared with unmodified paper) and the rough surface topography of paper.32 These features likely decrease the surface energy of the cellulose and thus decrease the rate of absorption of water into the paper. The rate of absorption of water into wax-treated paper also may be slowed by decreased pore sizes within the paper. Scanning electron microscope images of unmodified paper versus paper infused with 78 µg of paraffin wax per µm-3 appear identical: no macroscopic differences in surface topology are apparent (Figure 5b). Gurley porosity measurements (which measure the resistance of the paper to passage of air), however, reveal a decrease in porosity as the quantity of paraffin wax increases: for example, a sample with 0 µg of paraffin wax per µm-3 has a Gurley porosity of 2.5 ± 0.1 (s), while samples with 47 and 94 µg of paraffin wax per µm-3 give Gurley porosity values of 43.9 ¨ stmark, E.; Hult, A.; Malmstro (32) (a) Nystro ¨m, D.; Lindqvist, J.; O ¨m, E. Chem. Commun. 2006, 3594–3596, and references cited therein. (b) Quan, C.; Werner, O.; Wågberg, L.; Turner, C. J. Supercrit. Fluids 2009, 49, 117– 124.

4186

Analytical Chemistry, Vol. 82, No. 10, May 15, 2010

± 5.7 and 66.8 ± 18.1 (s), respectively. The latter values indicate nonporous surfaces according to the Gurley TAPPI T460 protocol.33 Effect of Sample Composition on the Performance of the Meter. We chose paraffin wax as the metering component for first generation meters because the hydrocarbons will not participate in noncovalent interactions with charged and polar solutes. The corollary to this choice is that paraffin wax likely will interact with (and sequester) solutes that readily adsorb to hydrophobic surfaces. Molecules that adsorb on top of the wax will alter the contact angle of the meter and affect the breakthrough time of the sample. Figure 6 depicts the behavior of four types of fluids tested in the context of two types of meters: one with only 5 µg of paraffin wax per µm3 of paper and a second with 31 µg of paraffin wax per µm3 of paper. Both meters were constructed using the configuration of 3D µPAD shown in Figure 3a. We normalized the time required for a sample to fill a device by subtracting Twax from Tpaper (the time required for a sample to fill an identical device that lacked a wax meter in the third layer). This normalization procedure accounts for differences in wicking rate caused by differences in the viscosity of each solution. Samples (10 µL) containing 10 mM glucose or 0.1 M sodium phosphate buffer (pH 7) wick at essentially equal rates as ddH2O (Figure 6). In other words, the compositions of these fluids have little effect on the metering capacity of the paraffin wax. Wicking rates for samples that contain 200 µM bovine serum albumin (BSA), however, were affected by the wax meter, resulting in a 49% increase in fluid retention time when 200 µM BSA is present in the sample compared to when it is not. Exactly how BSA leads to an increase in fluid retention time is unclear at this point, but we are currently studying this process with the intention of developing a quantitative model for understanding fluid flow in these wax-based meters. (33) Knauf, G. H.; Doshi, M. R. Proc. TAPPI 1986 International Process and Materials Quality Evaluation Conference; TAPPI Press: Atlanta, GA, 1986; p 33.

The precise relationship between metering capacity and sample composition will have to be determined empirically for now. It is worth noting, however, that 200 µM BSA (i.e., 13 mg/mL of protein) is 98-fold more concentrated than the quantity of protein found in the urine of healthy patients (0.13 mg/mL)34 and is only 4-fold less concentrated than the quantity of protein found in the serum of healthy adults. CONCLUSIONS The addition of microgram quantities of paraffin wax to defined regions in hydrophilic paper provides a low-cost and easily implemented method for controlling the flow rate of fluid within a 3D µPAD. These meters also offer the convenience of controlling flow rate within each conduit in a 3D µPAD (i.e., each conduit can have a different flow rate). Until now, the only method for controlling flow rate within µPADs has been to lengthen or widen a microfluidic channel. These changes, however, have little impact on flow rate (i.e., delay times of seconds to a few minutes) in a compact analytical device (i.e, devices with footprints of