Universal Microfluidic Automaton for Autonomous Sample Processing

The Mars Organic Analyzer (MOA) is a highly sensitive capillary zone electrophoresis .... Automated microfluidic program for analysis of a mixture of ...
8 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Universal Microfluidic Automaton for Autonomous Sample Processing: Application to the Mars Organic Analyzer Jungkyu Kim,† Erik C. Jensen,‡ Amanda M. Stockton,§ and Richard A. Mathies* Department of Chemistry, University of California, Berkeley, Berkeley, California 94720, United States S Supporting Information *

ABSTRACT: A fully integrated multilayer microfluidic chemical analyzer for automated sample processing and labeling, as well as analysis using capillary zone electrophoresis is developed and characterized. Using lifting gate microfluidic control valve technology, a microfluidic automaton consisting of a two-dimensional microvalve cellular array is fabricated with soft lithography in a format that enables facile integration with a microfluidic capillary electrophoresis device. The programmable sample processor performs precise mixing, metering, and routing operations that can be combined to achieve automation of complex and diverse assay protocols. Sample labeling protocols for amino acid, aldehyde/ketone and carboxylic acid analysis are performed automatically followed by automated transfer and analysis by the integrated microfluidic capillary electrophoresis chip. Equivalent performance to off-chip sample processing is demonstrated for each compound class; the automated analysis resulted in a limit of detection of ∼16 nM for amino acids. Our microfluidic automaton provides a fully automated, portable microfluidic analysis system capable of autonomous analysis of diverse compound classes in challenging environments.

C

and mixing operations that have the potential to achieve total automation of diverse assay procedures.29,30 The Mars Organic Analyzer (MOA) is a highly sensitive capillary zone electrophoresis analysis system designed for the detection of a wide range of biomarkers and chemical compounds, including amino acids,5,31,32 aldehyde/ketones,33 carboxylic acid,34 and thiols.35 By manually connecting a programmable microfluidic platform with a CZE microchip via PEEK tubing, on-chip carboxylic acid labeling was previously demonstrated.36 However, the lack of robust integration and large dead-volumes associated with this tubing interface result in lengthy sample processing times, which can negatively impact the quality of results. An integrated platform for sample labeling and CZE analysis of amino acids has also been demonstrated.37,38 However, this platform requires a complicated fabrication process, including the etching of features in three separate glass layers and the microfluidic design is specific to a single and fixed assay protocol. To perform a variety of distinct assays with a single instrument, a universal microfluidic sample processor integrated with a chemical analyzer is needed. In this paper, we introduce a new type of wafer-to-wafer integration, which exploits the unique properties of lifting gate microfluidic control technologies23 to create a fully integrated sample processor and analyzer. The lifting gate microfluidic sample processor is integrated with a glass CZE microchip, providing extremely low dead volumes between components

apillary zone electrophoresis (CZE) is a powerful chemical analysis method that has been widely used for environmental monitoring,1−4 astrobiology,5−7 and biosensing.8−11 CZE assays often require complex, manual sample processing procedures which involve sample metering, transporting, mixing, and storing actions that can be timeconsuming, labor-intensive, and sensitive to contamination. Commercial platforms for automated capillary zone electrophoresis have become available for high-sensitivity chemical and biological analysis.12−15 However, these platforms use conventional robotic systems to dispense or withdraw reagents, mix solutions, and transport samples between locations that are large, expensive and not field deployable in challenging environments. Microfluidic systems offer the prospect of miniaturization, automation, and reduction in sample volume requirements for chemical and biochemical sensing. Various microfluidic mechanisms for fluidic control have been demonstrated, including electro-wetting,16,17 acoustic,18,19 electrokinetic,20,21 and pneumatic microvalve actuation.22−24 Among these mechanisms, pneumatically actuated microvalves provide the unique combination of precise metering, high density integration, and the ability to scale to large arrays of fluidic control structures. Microfluidic control systems comprised of microvalves and pumps have enabled the automation of a wide range of sample processing procedures for biochemical analyses.25−29 Using monolithic membrane microvalve technology, our group developed a programmable microfluidic processor based on two-dimensional (2D) microvalve array technology. Using this programmable microfluidic “automaton”, we have demonstrated metering, transporting, routing, © 2013 American Chemical Society

Received: December 26, 2012 Accepted: May 15, 2013 Published: May 15, 2013 7682

dx.doi.org/10.1021/ac303767m | Anal. Chem. 2013, 85, 7682−7688

Analytical Chemistry

Article

four CZE wells with the appropriate fluidic channels as shown in Figure 1A. The integrated microfluidic device was then

because complex via-holes and PEEK tubing interfaces are eliminated. A fully automated microchip is developed and used to analyze diverse compound classes, including amino acids, aldehydes/ketones, and carboxylic acids in the same device and in the same run. The demonstrated ability to perform diverse sample processing operations on a common microchip format is a critical step toward multipurpose, high-sensitivity, portable chemical analysis systems for applications, including exobiology and terrestrial environmental monitoring.

■

MATERIALS AND METHODS Sample and Reagent Preparation. Amino acid, aldehyde, ketone, and carboxylic acid samples were purchased from Sigma-Aldrich and prepared as 20 mM stock solutions in Millipore-filtered water. Each standard solution was prepared by diluting these stock solutions with the appropriate buffer. Sodium tetraborate and boric acid were obtained from SigmaAldrich and used to prepare 400 mM aqueous stock without pH adjustment. The 30 mM borate buffer solutions at pHs 9.2, 5, and 3 were prepared by dilution and pH adjustment. Pacific Blue succinimidyl ester (Invitrogen, Carlsbad) was dissolved in dimethyl formamide (DMF, Sigma-Aldrich) to 20 mM. Cascade Blue hydrazide (Invitrogen, Carlsbad) was dissolved in Millipore-filtered water to 10 mM. EDC [1-ethyl-3-(3dimethylaminopropyl carbodiimide)] was dissolved to 20 mM in distilled acetonitrile (Sigma-Aldrich), divided into 100 μL aliquots, dried down, and stored at −20 °C until rehydration with pH 3 borate buffer. Fabrication of Microfluidic Capillary Zone Electrophoresis Chip. Conventional photolithography and wet chemical etching were used to prepare channel features in 1.1 mm thick borofloat glass wafers.33 The glass wafer was coated with 2 nm of amorphous polysilicon using LPCVD (low pressure chemical vapor deposition) and then spin-coated with 2 μm S1818 photoresist (Shipley Company). After patterning and developing the photoresist, regions of exposed polysilicon were removed using a sulfur hexafluoride (SF6) plasma. The exposed glass was then etched with 49% hydrofluoric acid (HF) to a depth of 35 μm. Photoresist and polysilicon were stripped off with acetone and SF6, respectively. The four inlets to the CZE channel for the target sample, cathode, anode, and waste were drilled with 2.5 mm diameter diamond-tipped bits, using a CNC mill. The etched glass wafer was then thermally bonded with a 650 μm thick backing glass wafer (Alphadyne) using an oven (Vulcan furnace, Neytech, USA). Fabrication of Programmable Microfluidic System. With the use of standard soft-lithography, SU-8 molds for fluidic and pneumatic layers were fabricated to obtain a 45 and 100 μm feature height, respectively. After silanizing the SU-8 molds, using CVD (chemical vapor deposition), the mold for the fluidic layer was spin-coated with a 10 (base material):1(curing agent) ratio of PDMS (Dow Corning) at 250 rpm and thermally cured at 65 °C to fabricate a ∼250 μm thick PDMS membrane structure. For the pneumatic layer, PDMS was poured on the mold and cured on a 65 °C hot plate for one hour. Holes were punched in the pneumatic layer replica for pneumatic connections and then were aligned and permanently bonded to the PDMS fluidic layer using oxygen plasma treatment (PETS Inc.) to form the automaton. Integration and Operation. To integrate the programmable automaton with the CZE chip, the contacting surfaces of both the chips were first activated with UV ozone. Then, the automaton was bonded to the CZE glass chip by aligning the

Figure 1. (A) Schematic of the microfluidic automaton integrated with the Mars Organic Analyzer (MOA) Capillary Zone Electrophoresis (CZE) analysis chip. The top PDMS lifting gate microvalve array, including a pneumatic layer (red) and a fluidic layer (blue), is integrated with the bottom layer, including the CZE glass chip (black). (B) Detailed view of cellular microvalve array for universal sample processing. (C) Cross-sectional view of the CZE sample well interconnection showing lifting gate microfluidic features. (D) Cross-sectional view of anode, waste, and cathode CZE well interconnections and lifting gate microfluidic structures.

assembled with the pneumatic control system and used to perform programmable sample processing. By controlling vacuum and pressure to inlets on the pneumatic layer, each valve is actuated by a series of off-chip solenoid controllers, using a vacuum pump (−87 kPa) and positive closing pressures (40 kPa). The basic operation for the fluidic transfer in a microvalve array is initiated by opening a single microvalve, drawing fluid from an inlet well (see Figure 1B). An adjacent microvalve is then opened, pulling fluid from the first valve. By closing the first valve with an applied pneumatic pressure, the remaining fluid in the first valve moves into the second valve. 7683

dx.doi.org/10.1021/ac303767m | Anal. Chem. 2013, 85, 7682−7688

Analytical Chemistry

Article

borate buffer pH 9 at a 1:1 ratio for analysis in the CZE platform. The aldehyde/ketones were labeled with Cascade Blue hydrazide (CB), using analogous automated sample processing. In Figure 3 (panels A and B), one part of the

By programming the route of actuation, all the samples and reagents can be transferred from any input well to any output well. In addition, by performing repeated cycles, precisely metered total volumes can be transferred to a selected output. Laser-Induced Fluorescence Detection System for CZE Analysis. Injection of sample from the CZE inlet, separation, and laser-induced fluorescence detection were performed using the Mars Organic Analyzer (MOA).5 The MOA consists of high-voltage power supplies to generate the electrophoretic force for separation and an optical system for laser-induced fluorescence detection. In brief, a 405 nm diode laser is reflected off a dichroic and focused through an objective to a ∼20 μm spot within the separation channel. The fluorescence signals, collected by the optical system, are measured with photomultiplier (PMT) and processed with Labview.5 Automated Sample Processing. After the addition of the samples, reagents, and buffer in the appropriate input wells of the microfluidic processor, preprogrammed pumping and valving sequences were used to execute all labeling steps, including metering, mixing, transferring, and incubating processes. Figure 2 (panels A and B) illustrate the automated

Figure 3. Microfluidic program for aldehyde/ketone analysis. (A) One part aldehyde/ketone standard and one part 1 mM Cascade Blue are combined, and 27 μL are delivered to a reaction well by 50 pumping cycles. (B) After a 15 min incubation, the reaction mixture is combined with one part 30 mM borate buffer, pH 9, and pumped to the sample well of the MOA CZE chip for analysis. (C) Electropherograms of the aldehyde/ketone standard resulting from automated processing is compared with manual processing. Each of the three aldehydes (propionaldehyde, acetaldehyde, and formaldehyde) and one ketone (methyl ethyl ketone) are clearly identifiable and have similar migration times and efficiencies in both electropherograms.

aldehyde/ketone sample and one part of 1 mM Cascade Blue (CB) in 30 mM borate buffer, pH 5, were combined and a total volume of 27 μL was delivered to a reaction well by 50 pumping cycles. After a 15 min incubation at room temperature, the labeled aldehyde/ketone samples were then mixed with 30 mM borate buffer, pH 9, at a 1:1 ratio or higher for analysis. The labeling process for the carboxylic acid samples is illustrated in Figure 4 (panels A and B). First, one part of the standard in dH2O, one part of 1 mM CB in 30 mM borate buffer, pH 3, and one part of 10 mM EDC in 30 mM borate buffer, pH 3, were combined and a total volume of 40 μL was delivered to a reaction well by 50 pumping cycles. After reacting for 15 min at room temperature, the mixture was diluted with a 30 mM borate buffer, pH 9, at a ratio of at least 1:3, to adjust the pH, and analyzed on the CZE platform. The automated labeling steps for the amino acids, aldehyde, ketone, and carboxylic acid were compared with the results of the manual labeling process. Sample carryover was reduced below the detectable threshold by flushing the valving structure with buffer at least 3 times as previously reported.37,38 For analysis, target samples first were electrophoresed through the cross-injector by applying −1500, −1500, −1500, and 0 V to anode, sample, sample waste, and cathode,

Figure 2. Microfluidic processor program for amino acid analysis. (A) One part amino acid and one part 40 μM Pacific Blue are combined and a total volume of 27 μL is delivered to a reaction well by 50 pumping cycles. (B) After a 15 min incubation, the reaction product is combined with 1 part 30 mM borate buffer, pH 9, and pumped to the CZE sample well for analysis. (C) Electropherograms of an amino acid standard resulting from automated sample processing is compared with manual processing. Each of the 7 amino acids (citrulline, valine, serine, alanine, glycine, glutamic acid, and aspartic acid) are clearly identifiable and have similar migration times and efficiencies in both electropherograms.

labeling process for amino acids. First, one part of amino acid standard, including seven amino acids in dH2O, and 1 part of 40 μM Pacific Blue (PB), in 30 mM borate buffer pH 5, were combined and a total volume of 27 μL was delivered to a reaction well by 50 pumping cycles. After incubating at room temperature for 15 min, the mixture was diluted with a 30 mM 7684

dx.doi.org/10.1021/ac303767m | Anal. Chem. 2013, 85, 7682−7688

Analytical Chemistry

Article

Figure 4. Microfluidic program for carboxylic acid analysis. (A) One part carboxylic acid standard, one part 1 mM Cascade Blue, and one part 10 mM EDC are combined and a total volume of 40 μL is delivered to a reaction well by 50 pumping cycles. (B) After a 15 min incubation, one part of the reaction product and three parts of 30 mM borate buffer, pH 9, are pumped to the CZE sample well for analysis. (C) The electropherogram of the carboxylic acid standard resulting from automated sample processing is compared with manual processing. Each of the five carboxylic acids (formic acid, acetic acid, propionic acid, butyric acid, and valeric acid) are clearly identifiable and have similar migration times and efficiencies in both electropherograms, demonstrating the utility of the automated processing. Figure 5. Automated microfluidic program for analysis of a mixture of amino acids, aldehydes, ketones, and carboxylic acids in a single efficient process on the Mars Organic Analyzer. (A−C) Separate aliquots of the mixture, which includes four amino acids (citrate, valine, serine, and glycine), two carboxylic acids (acetic acid and formic acid), one aldehyde (formaldehyde), and one ketone (acetone) are sequentially labeled with Pacific Blue (PB) and Cascade Blue (CB) in the microfluidic processor and transferred to separate reaction reservoirs. (D) After a 15 min incubation, all the labeled reaction products are combined together with 30 mM borate buffer, pH 9, and transferred to the MOA sample well for analysis. (E) Each of the species present in the standard is successfully identified in the combined program for amino acids, aldehydes, ketones, and carboxylic acids.

respectively, for 30 s. By changing the voltages to 0, 0, 0, and 15000, separation of the amino acids was performed and fluorescence signals were detected using the MOA platform. Simultaneous Analysis of Spiked Mixture Sample. We developed a program to automatically label aliquots of a mixed sample with three different labeling procedures (amino acid, aldehyde/ketone, and carboxylic acid), combine the products, and electrophorese the combined sample for simultaneous analysis. Running buffer, washing buffer, labeling dyes, and a standard consisting of four amino acids, one aldehyde, one ketone, and two carboxylic acids were loaded into inlets of the lifting gate processor. First, the anode and waste wells of the CZE chip were automatically filled with a 30 mM borate buffer, pH 9. After the electrophoresis channel is filled by capillary action, the cathode well was then automatically filled with the same buffer. Figure SI1.A of the Supporting Information shows the procedure for filling the electrophoretic channel. Three different aliquots of the mixed standard were then sequentially labeled for amino acids, aldehydes/ketones, and carboxylic acids, and transferred to three different output reservoirs for a 15 min incubation, as shown in Figure 5. Between each labeling program, a 10 s washing procedure was performed to remove residual reagents from the microvalves of the 2D array (Figure SI1.B of the Supporting Information). After a 15 min incubation, one part of each of the labeled samples was combined, along with one part running buffer, and the mixture was pumped to the CZE sample well for 35 cycles. Results from

this mixed sample analysis were compared with the individual labeling procedures described previously.

■

RESULTS AND DISCUSSION Lifting Gate Programmable Microfluidic Processor. Figure 1 illustrates the microfluidic processor with 16 perimeter microvalves for selecting reagents and 16 internal microvalves for routing and mixing reagents that was fabricated with soft lithography, using our lifting gate microfluidic technology.23 The microvalve array enables programmable microfluidic sample processing including metering, transferring, routing, mixing, and storing of liquid samples (see the video of the Supporting Information). In this microfluidic processor or automaton, the maximum volume pumped per cycle for the internal microvalves was calculated as described previously.23 7685

dx.doi.org/10.1021/ac303767m | Anal. Chem. 2013, 85, 7682−7688

Analytical Chemistry

Article

aldehydes/ketone, and carboxylic acids in a fully automated integrated process. Processing speed is critical for labeling reactions, dilution, and transfer of labeled samples to the CZE inlet. For instance, the imide bonds formed between aldehydes/ketones and CB are unstable at pH 9, which is the optimum pH for CZE. In addition, slow loading of the CZE inlet can result in diffusion of the sample into the cross injector, resulting in baseline drift. The 16 μL/min pumping rate of the 2D microvalve array enables complete filling of the ∼27 μL CZE reservoirs within 2 min, thereby mitigating both of these potential problems. Fully Automated Mars Organic Analyzer Operation. A complex sample, including four amino acids, one aldehyde, one ketone and two carboxylic acids, was loaded into a single fluidic reservoir, and aliquots were automatically processed using three different labeling programs. The products of each labeling program were then transferred to the sample well of the CZE microchip for simultaneous electrophoresis and analyses. Figure 5A shows the labeling process for the amino acid only. After the washing step shown in Figure SI1.B of the Figure SI1.B, aliquots of the complex sample were mixed with CB to specifically label aldehyde/ketones and transferred to a second reaction well for 15 min (Figure 5B). After another washing step, aliquots of the complex sample were combined with EDC and CB to specifically label carboxylic acids and transferred to a third reaction well for a 15 min incubation (Figure 5C). The three labeled samples were then analyzed individually and also simultaneously. For individual analysis, the same ratio of labeled sample and running buffer described previously was combined and transferred to the CZE sample well. Figure 5D presents the procedure for the combination of all three reaction products with running buffer and transferring the mixture to the CZE well for simultaneous analysis. Figure 5E shows the separation results for the mixed compound class program and the individual compound class programs. By a comparison of the individual separation results with the mixture results, all the amino acids, aldehyde/ketone, and carboxylic acids were identified; all the individual compounds showed almost identical peak efficiency, and using a single electrophoresis run to analyze three different chemical compound classes enables rapid screening of complex target samples without any reduction in sensitivity (Table SI2 of the Supporting Information). The ability to analyze multiple compound classes simultaneously significantly reduces the time to result, enabling rapid, portable, and multiplexed screening of samples for the presence of any analytes. This approach is particularly desirable for planetary applications where biomarkers are expected to be rare and at low concentration. When greater sample complexity is encountered, it is easily resolved, as shown in Figure 5E by performing individual analyses. The coupling of the lifting gate microfluidic processor or automaton with the MOA produces a simple modular assembly, integrating the sample processing functions and the downstream electrophoresis and analysis. The fully integrated microdevice is formed by simply aligning and bonding the PDMS lifting gate microfluidic processor with the glass MOA chip. This same approach could easily be used to integrate sample processing units with a variety of microchip or other wafer-based analysis systems. Since the top of the integrated glass device forms one wall of the microfluidic channel, there is no dead volume between the modules. Thus, this integrated platform enables fast transfer of the labeled samples to the CZE

Given the volume of the internal microvalves and the 86% pumping efficiency, the processing programs transferred 270 nL per pumping cycle. The 32-microvalve array has 232 unique operational states, which enables the serial implementation of a wide range of sample processing programs with nanoliter to microliter sample volume scales. The automated processor design was selected to address two basic problems: the introduction of bubbles in the CZE microchip wells and leftover buffer in the microfluidic processor, which can result in electrical short circuits when CZE is performed. To minimize bubbles, the anode, sample, and cathode wells were designed to include an additional channel for connection to the sample processor (Figure 1, panels C and D). A drilled hole at the other end of this channel connects to output channels of the microfluidic processor. With this design, the buffer is pumped into the bottom of these wells, which effectively prevents bubble entrapment in the wells during loading. To inhibit the short circuit resulting from leftover buffer in the channels of the microvalve array, the previously used microvalves were actuated while all inlet valves were closed. Due to the gas permeability of PDMS, air bubbles were formed in the microvalves, thereby isolating the anode, sample, and waste of the CZE channel. Automated Labeling of Amino Acids, Aldehydes, Ketones, and Carboxylic Acids. To perform the amino acid analysis, an amino acid standard, including citrulline, valine, serine, alanine, glycine, glutamic acid, and aspartic acid was labeled with Pacific Blue (PB), using both manual and automated labeling processes. Figure 2C presents the electropherogram of the autonomously labeled amino acid standard, as well as the off-chip label. The 7 different peaks from this standard were readily identified in both off-chip and on-chip labeled samples. The automated analysis demonstrates very similar migration times and peak efficiencies to those resulting from the manual process (Table SI1−1 of the Supporting Information). The limit of detection (LOD) of 4 different amino acids (citrulline, valine, serine, and alanine) was measured using the automated sample processing program based on a signal-to-noise ratio of 3. From the log−linear relationship between concentration and peak area, as little as 16 nM of these amino acids can be detected from spiked samples (Figure SI2 of the Supporting Information). Aldehyde and ketone analysis was performed on a standard, including three different aldehydes and one ketone. Figure 3C presents the results of the automated microfluidic labeling process and CZE analysis for this standard. The four targets, methylethyleketone, propionaldehyde, acetaldehyde, and formaldehyde, are all observed and well-resolved. Comparison of the on-chip labeling results with the off-chip control reveals peak efficiencies and peak intensities that are identical (Table SI1−2 of the Supporting Information). Carboxylic acids were also analyzed using the fully integrated microfluidic organic analyzer. Under pH 3 buffer conditions, and using EDC as an activator, a standard comprised of five different carboxylic acids was labeled on-chip with CB using the automated processor. Figure 4C shows electropherograms of results from the on-chip and the off-chip labeling processes. For the automated results, the five different carboxylic acids in the standard are identified with labeling and peak efficiencies (Table SI1−3 of the Supporting Information). From these experiments, we confirm that the programmable microfluidic sample processor can effectively label and analyze amino acids, 7686

dx.doi.org/10.1021/ac303767m | Anal. Chem. 2013, 85, 7682−7688

Analytical Chemistry

Article

(2) Marle, L.; Greenway, G. M. TrAC, Trends Anal. Chem. 2005, 24, 795−802. (3) Ginterova, P.; Marak, J.; Stanova, A.; Maier, V.; Sevcik, J.; Kaniansky, D. J. Chromatogr., B 2012, 904, 135−139. (4) Marak, J.; Stanova, A.; Vavakova, V.; Hrenakova, M.; Kaniansky, D. J. Chromatogr., A 2012, 1267, 252−258. (5) Skelley, A. M.; Scherer, J. R.; Aubrey, A. D.; Grover, W. H.; Ivester, R. H.; Ehrenfreund, P.; Grunthaner, F. J.; Bada, J. L.; Mathies, R. A. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 1041−1046. (6) Skelley, A. M.; Cleaves, H. J.; Jayarajah, C. N.; Bada, J. L.; Mathies, R. A. Astrobiology 2006, 6, 824−837. (7) Aubrey, A. D.; Chalmers, J. H.; Bada, J. L.; Grunthaner, F. J.; Amashukeli, X.; Willis, P.; Skelley, A. M.; Mathies, R. A.; Quinn, R. C.; Zent, A. P.; Ehrenfreund, P.; Amundson, R.; Glavin, D. P.; Botta, O.; Barron, L.; Blaney, D. L.; Clark, B. C.; Coleman, M.; Hofmann, B. A.; Josset, J. L.; Rettberg, P.; Ride, S.; Robert, F.; Sephton, M. A.; Yen, A. Astrobiology 2008, 8, 583−595. (8) Zeisbergerova, M.; Reminek, R.; Madr, A.; Glatz, Z.; Hoogmartens, J.; Van Schepdael, A. Electrophoresis 2010, 31, 3256− 3262. (9) Despeyroux, D.; Walker, N.; Pearce, M.; Fisher, M.; McDonnell, M.; Bailey, S. C.; Griffiths, G. D.; Watts, P. Anal. Biochem. 2000, 279, 23−36. (10) Bossi, A.; Piletsky, S. A.; Righetti, P. G.; Turner, A. P. J. Chromatogr., A 2000, 892, 143−153. (11) Gulbis, B.; Fontaine, B.; Vertongen, F.; Cotton, F. Ann. Clin. Biochem. 2003, 40, 659−662. (12) Solangi, A. R.; Memon, S. Q.; Mallah, A.; Khuhawar, M. Y.; Bhanger, M. I. Biomed. Chromatogr. 2009, 23, 1007−1013. (13) Lanz, C.; Kuhn, M.; Deiss, V.; Thormann, W. Electrophoresis 2004, 25, 2309−2318. (14) Bossuyt, X.; Marien, G. Clin. Chem. (Washington, DC, USA) 2007, 53, 152−153. (15) Bossuyt, X.; Lissoir, B.; Marien, G.; Maisin, D.; Vunckx, J.; Blanckaert, N.; Wallemacq, P. Clin. Chem. Lab. Med. 2003, 41, 704− 710. (16) Abdelgawad, M.; Wheeler, A. R. Adv. Mater. 2009, 21, 920−925. (17) Fair, R. B. Microfluid. Nanofluid. 2007, 3, 245−281. (18) Rife, J. C.; Bell, M. I.; Horwitz, J. S.; Kabler, M. N.; Auyeung, R. C. Y.; Kim, W. J. Sens. Actuators, A 2000, 86, 135−140. (19) Neild, A.; Oberti, S.; Dual, J. Sens. Actuators, B 2007, 121, 452− 461. (20) Jacobson, S. C.; McKnight, T. E.; Ramsey, J. M. Anal. Chem. 1999, 71, 4455−4459. (21) Piruska, A.; Branagan, S.; Cropek, D. M.; Sweedler, J. V.; Bohn, P. W. Lab Chip 2008, 8, 1625−1631. (22) Grover, W. H.; Skelley, A. M.; Liu, C. N.; Lagally, E. T.; Mathies, R. A. Sens. Actuators, B 2003, 89, 315−323. (23) Kim, J.; Kang, M.; Jensen, E. C.; Mathies, R. A. Anal. Chem. 2012, 84, 2067−2071. (24) Thorsen, T.; Maerkl, S. J.; Quake, S. R. Science 2002, 298, 580− 584. (25) Easley, C. J.; Karlinsey, J. M.; Bienvenue, J. M.; Legendre, L. A.; Roper, M. G.; Feldman, S. H.; Hughes, M. A.; Hewlett, E. L.; Merkel, T. J.; Ferrance, J. P.; Landers, J. P. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 19272−19277. (26) Fan, H. C.; Wang, J.; Potanina, A.; Quake, S. R. Nat. Biotechnol. 2011, 29, 51−57. (27) Gomez-Sjoberg, R.; Leyrat, A. A.; Pirone, D. M.; Chen, C. S.; Quake, S. R. Anal. Chem. 2007, 79, 8557−8563. (28) Hong, J. W.; Studer, V.; Hang, G.; Anderson, W. F.; Quake, S. R. Nat. Biotechnol. 2004, 22, 435−439. (29) Jensen, E. C.; Zeng, Y.; Kim, J.; Mathies, R. A. J. Lab. Autom. 2010, 15, 455−463. (30) Jensen, E. C.; Bhat, B. P.; Mathies, R. A. Lab Chip 2010, 10, 685−691. (31) Chiesl, T. N.; Chu, W. K.; Stockton, A. M.; Amashukeli, X.; Grunthaner, F.; Mathies, R. A. Anal. Chem. 2009, 81, 2537−2544.

well, thereby reducing the diffusion of the samples in the injector. In addition, the two devices are independently fabricated and then assembled to achieve the fully integrated system, in contrast to more complex, error intolerant doublesided glass fabrication approaches.38 The modular assembly also facilitates replacement of individual components, thereby simplifying and improving fabrication yields and reliability and reducing the cost of the final, integrated device. Additionally, while the entire device is reusable for multiple (100+) assays, the more expensive, fabrication-intensive, all-glass microcapillary electrophoresis device can easily be removed from the integrated system for reuse in other systems or for chemically harsh cleaning, if system fouling occurs and the lifting-gate processing structure can be cheaply refabricated and reattached once the molds have been made.

■

CONCLUSION AND PROSPECTS A fully automated microfluidic organic analyzer has been developed, exploiting the facile integration of lifting gate microfluidic processor technology with microchip capillary zone electrophoresis. All the procedures, including buffer filling, labeling, and dilution, were automated with the automaton, and the processed samples were analyzed to identify biomarkers, such as amino acids, and oxidized biomarker compounds, such as aldehydes/ketones and carboxylic acids. All of the target molecules were analyzed within 30 min, using totally automated processing. Integration of the automaton-MOA device with any of a number of small laser-induced fluorescence detection systems will now enable fully autonomous analytical systems for portable environmental monitoring, as well as planetary exploration. Furthermore, this platform can be applied to almost any type of aqueous chemical and biochemical analysis by simply changing the sample processing program.

■

ASSOCIATED CONTENT

* Supporting Information S

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: (510) 642-4192. Fax: (510) 642-3599. Present Addresses †

Department of Mechanical Engineering, Texas Tech University, Lubbock, TX 79409, United States. ‡ HJ Science & Technology, 187 Saratoga Avenue, Santa Clara, CA 95050, United States. § Jet Propulsion Laboratory, 4800 Oak Grove Drive, Pasadena, CA 91109, United States. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Financial support for this work was provided by the Mathies Royalty Fund. Devices were fabricated in the UC Berkeley BioNanotechnology Center (BNC).

■

REFERENCES

(1) Jayarajah, C. N.; Skelley, A. M.; Fortner, A. D.; Mathies, R. A. Anal. Chem. 2007, 79, 8162−8169. 7687

dx.doi.org/10.1021/ac303767m | Anal. Chem. 2013, 85, 7682−7688

Analytical Chemistry

Article

(32) Stockton, A. M.; Chiesl, T. N.; Lowenstein, T. K.; Amashukeli, X.; Grunthaner, F.; Mathies, R. A. Astrobiology 2009, 9, 823−831. (33) Stockton, A. M.; Tjin, C. C.; Huang, G. L.; Benhabib, M.; Chiesl, T. N.; Mathies, R. A. Electrophoresis 2010, 31, 3642−3649. (34) Stockton, A. M.; Tjin, C. C.; Chiesl, T. N.; Mathies, R. A. Astrobiology 2011, 11, 519−528. (35) Mora, M. F.; Stockton, A. M.; Willis, P. A. Electrophoresis 2013, 34, 309−316. (36) Jensen, E.; Stockton, A.; Chiesl, T.; Kim, J.; Bera, A.; Mathies, R. A. Lab Chip 2012, 13, 288−296. (37) Benhabib, M.; Chiesl, T. N.; Stockton, A. M.; Scherer, J. R.; Mathies, R. A. Anal. Chem. 2010, 82, 2372−2379. (38) Mora, M. F.; Greer, F.; Stockton, A. M.; Bryant, S.; Willis, P. A. Anal. Chem. 2011, 83, 8636−8641.

7688

dx.doi.org/10.1021/ac303767m | Anal. Chem. 2013, 85, 7682−7688