Technical Note pubs.acs.org/ac
Integration of Fully Microfabricated, Three-Dimensionally Sharp Electrospray Ionization Tips with Microfluidic Glass Chips Lauri Sainiemi,*,† Tiina Sikanen, and Risto Kostiainen Division of Pharmaceutical Chemistry, Faculty of Pharmacy, University of Helsinki, P.O. Box 56, FI-00014, Helsinki, Finland ABSTRACT: This paper presents parallel microfabrication of three-dimensionally sharp electrospray ionization emitters made out of glass. For the first time, the fabrication of glass emitters relies only on standard microfabrication techniques (i.e., deposition, photolithography, and wet etching), and all manual machining steps are omitted. We also demonstrate a straightforward integration of the three-dimensionally sharp emitter tip with a microfluidic separation channel, which has been one of the major challenges of micro total chemical analysis systems for the past 15 years. As a result, our microfabrication approach provides glass ESI emitters that allow robust performance from run to run and tip to tip and do not suffer from sample spreading at the microchannel outlet. The repeatability of the signal intensity for parallel tips was shown to be within 8.0% RSD (n = 6), and the migration time repeatability for repeated injections was within 6.2% RSD (n = 6). At best, separation plates of up to 2.7 × 105/m were obtained. Since the microfabrication process readily yields three-dimensionally sharp emitter tips, very low ESI voltages (typically 1.4−1.75 kV) suffice for stable ESI, which eventually allows for the use of a variety of different solvent compositions from purely aqueous to high organic content. Here, the advantage of using aqueous conditions is demonstrated in protein analysis.
T
performance.8−13 However, both approaches were prone to large dead volumes at the outlet of the microchannel, which caused band-broadening and loss of the previously achieved separation. The formation of electrospray from a blunt edge appeared inefficient because of sample spreading around the hydrophilic microchannel outlet, and the spray was produced from the sharpest protruding microdefect. Although manual attachment of a sharp, external needle at the end of the microchip improved the stability of the ESI, the needle alignment was laborious, and the benefits of parallel microfabrication were severely compromised. Because fabrication of a sharp ESI tip from glass has proven to be extremely difficult, the emphasis was put on development of silicon-14−17 and polymer18−21-based electrospray chips. Owing to the well-established silicon microfabrication techniques, it was fairly easy to realize sharp ESI tips out of silicon, but their integration with CE appears impossible because silicon is electrically conductive. Instead, attempts have been made to integrate liquid chromatographic (LC) columns with the silicon ESI tips.22,23 Often, polymer microfabrication offers the most comfortable route for monolithic integration of separation units (either LC or CE) with ESI tips.24,25 Although polymer microfabrication techniques at best provide sharp emitters (e.g., micromilling of polyimide24 or SU-8 photolithography25), stable electrospray has also been produced from
he coupling of microchips with mass spectrometry (MS) via electrospray ionization (ESI) is a powerful analytical tool for modern drug and bioanalysis.1−3 MS is a universal detection method that provides superior selectivity and specificity compared with any optical detection technique (e.g., absorbance or fluorescence). The success of ESI is based on its wide applicability, ranging from small inorganic and organic molecules to large and labile biomolecules.4 Combining ESI with microchip technology is also fairly straightforward because the ideal flow rates for electrospray are similar to those used in microfluidic separation chips. In recent years, numerous different approaches have been taken to combine microfluidic chips to ESI/MS detection.1−3 The first electrospray chips were fabricated of glass using microfabrication techniques adopted from electronics industry.5−13 Glass was an obvious material of choice because its surface chemistry is similar to that of conventional silica capillaries. The negative surface charge of glass gives rise to stable electroosmotic flow (EOF), which can be used for fluid transportation. Glass is also electrically insulating, which enables the use of high voltages for capillary electrophoretic (CE) separations and ESI. Unfortunately, glass microfabrication is less explored and more difficult than processing of silicon. Therefore, monolithic integration of sharp ESI tips with glass microchannels has remained challenging. In the first designs, electrospray was generated at the microchannel outlet on a blunt edge of the microchip.5−7 Alternatively, a fused-silica capillary or a nanospray needle was manually attached at the end of the microfluidic channel to improve the electrospray © 2012 American Chemical Society
Received: June 11, 2012 Accepted: October 9, 2012 Published: October 9, 2012 8973
dx.doi.org/10.1021/ac301602b | Anal. Chem. 2012, 84, 8973−8979
Analytical Chemistry
Technical Note
Figure 1. Fabrication of microfluidic glass chips with monolithically integrated three-dimensionally sharp tips. (a) Process started with poly-Si-coated Borofloat wafers purchased from CSEM (channel wafer). (b) Patterning of the poly-Si layer (inlets and ESI tip) on the backside and isotropic etching of glass (2 h 10 min). (c) Patterning of poly-Si layer (inlets, ESI tip, and microchannel network) on the topside of the wafer (1 h 10 min). (d) Completing the through-wafer etching of the inlets and the ESI tip and creating the microchannel network (10 min). (e) Process started with poly-Si-coated Borofloat wafers purchased from CSEM (coverlid). (f) Patterning of the poly-Si layer (ESI tips) on backside of the wafer and isotropic etching through the wafer (2 h 15 min). (g) Removal of poly-Si layers in RIE, wafer cleaning and fusion bonding of the channel wafer and the coverlid, topsides together (3 h 40 min). (h) Schematic 3D illustration of the CE−ESI microchip design with critical lateral dimensions (dimensions not to scale). (i−k) Photographs of an array of nine ESI emitters, front (i) and side (j) views, and of a separation chip (k) with two CE−ESI channels. (k) Schematic view of the voltage distribution of the CE−ESI chip during the MS measurements.
et al. who demonstrated that a two-dimensionally sharp ESI tip could be realized by simply sawing a sharp corner at the end of a glass microchannel with a dicing saw.31−34 Using extremely thin glass wafers (150 μm), the emitter tip became sharp enough in the vertical direction to allow efficient electrospray from the outlet of the microchannel. However, fabrication of the most critical part of the chip (i.e., the emitter tip) relied on fairly harsh mechanical machining instead of parallel microfabrication methods.34 In this study, we present parallel microfabrication of threedimensionally sharp ESI glass emitters that are monolithically integrated with microfluidic channels. Distinct from all previous designs, only standard microfabrication (i.e., deposition, photolithography, and wet etching) and standard wafers are used, which allows mass production of identical glass microchips with sharp ESI emitters. The performance of the fully microfabricated glass emitters is characterized by a series of ESI/MS analyses of proteins and small molecules. Feasibility of the chip design for rapid CE−ESI/MS separation is also demonstrated. The good analytical figures of merit evidence that our design allows efficient ESI and no sample spreading occurs at the microchannel outlet. Because of the universal nature of the microfabrication process and its feasibility for large-scale mass production, we believe this approach is able to replace the commercial nanospray needles by providing equally
blunt-edged polymer chips thanks to the greater hydrophobicity of polymers, which reduces sample spreading at the microchannel outlet.26,27 Unfortunately, the hydrophobicity of polymers often impairs the separation performance because of increased nonspecific interactions between the analytes (proteins in particular) and the walls of the hydrophobic microchannels. For some commercial polymers, the exact material composition is not even known, which makes it very difficult to control the interactions. Therefore, researchers continue to seek alternative ways to realize sharp, fully microfabricated ESI emitters on hydrophilic glass whose surface chemistry is well-known and, thus, easily manipulated toward different applications. Only recently have two different approaches been proposed for fabrication of monolithically integrated CE−ESI chips from glass. Hoffmann et al. integrated a sharp ESI tip with a commercial CE microchip made of borofloat glass.28−30 First, the blunt edge of the glass chip was manually machined to the shape of a cone, after which the sharp ESI tip was realized at the end of the channel using a home-built glass puller. The fabrication scheme circumvents the problems related to the dead volume at the interconnection of the microchannel and the ESI tip. However, each chip has to be separately machined and manually pulled, which is far from parallel high-throughput microfabrication. Another approach was presented by Mellors 8974
dx.doi.org/10.1021/ac301602b | Anal. Chem. 2012, 84, 8973−8979
Analytical Chemistry
Technical Note
robust performance from tip to tip, but more options for online sample preparation prior to ESI/MS.
were created during a continuous, 68-min-long HF/HCl wet etching step (Figure 1f). Last, the photoresist and the poly-Si layers were removed from both wafers using acetone and RIE, respectively. The wafers were RCA-1 cleaned (H2O−NH4OH−H2O2 5:1:1, +80 °C) for 10 min after which the topsides of the wafers were bonded together (Figure 1g). The alignment and prebonding were done in a wafer bonder (AWB-04, Applied Microengineering Ltd., Oxfordshire, U.K.) at +480 °C using 8 kN force after which the bonding was finalized in a furnace at +650 °C. The final microchips were separated from the wafer using a dicing saw. Mass Spectrometry. The microchips were coupled to an Agilent 6330 iontrap mass spectrometer (Agilent Technologies, Santa Clara, CA) using a modified nanospray frame (Proxeon Biosystems, Odense, Denmark) equipped with an x−y−z aligning stage and a CCD camera. The distance between the ESI tip and the MS sampling orifice was typically 5−10 mm. The ion trap was operated in positive ion mode using nitrogen (4.0 L min−1, +70 °C) produced from compressed air by a Parker nitrogen generator (Cleveland, OH) as the drying gas. The MS parameters were optimized in the smart target mode, and the ion optics were as follows: capillary voltage, −1200 V; end plate offset, 500 V. The MS data were acquired averaging two cycles over a mass range of m/z 100−2200 with maximum accumulation time of 200 ms. Data Analysis 3.4 was used for data acquisition and processing. An external high voltage power supply (Micralyne, Edmonton, AB) was used for application of the ESI and separation voltages. Before measurements, a thin sheet of PDMS with 2 mm inlet holes was attached on top of the microchannel inlets to avoid sample spreading on a hydrophilic glass surface. The separation and ESI voltages were applied through platinum wires placed in the microchannel inlets. In direct spray experiments, an ESI voltage between +1400 and +1750 V was applied to the sample inlet of the emitter tip. In CE−ESI/MS analyses, the samples were introduced electrokinetically (20.0 and 30.0 s for small molecule and protein samples, respectively) through the simple cross injector in pinched injection mode. Injection voltages of +600 and +400 V were applied to the sample inlet (SI) and buffer inlet (BI), respectively, while the sample outlet (SO) was grounded (Figure 1h). The sheath liquid inlet (SLI) was left floating during injection (20.0 s). During the separation step, a separation voltage of +4500 V was applied to the BI, and pushback voltages of +4100 V, to the SI and SO (Figure 1k). The ESI voltage (+3500 V), which also served as the counter voltage for the CE separation, was applied to the SLI (Figure 1k). The separation current was typically 30−40 μA, and the electrospray current was below 200 nA. The excess current from the separation channel was grounded through a 50 MΩ resistor coupled in parallel with the ES voltage supply.
■
EXPERIMENTAL SECTION Chemicals and Reagents. Acetic acid, ammonium acetate, and methanol were purchased from Sigma-Aldrich (Steinheim, Germany). Tetraethylammonium iodide and tetrahexylammonium bromide were from Sigma-Aldrich, and tetrapropylammonium iodide, tetrabutylammonium iodide, and tetrapentylammonium iodide were from Fluka (Buchs SG, Switzerland). Cytochrome c from bovine heart was purchased from SigmaAldrich and verapamil hydrochloride was from ICN Biomedicals (Aurora, OH). All reagents and solvents were of analytical or HPLC grade. Water was purified with a Milli-Q water purification system (Millipore, Molsheim, France). Before use, all sample and buffer solutions were filtered (0.2 μm) and degassed by sonication for 10 min. Polydimethylsiloxane (PDMS, Sylgard 184 silicone elastomer, Dow Corning Co., Midland, MI) was used for preparation of support layers on the top and bottom of the glass chip by mixing the base elastomer and the curing agent in a mass ratio of 9:1. Microfabrication. All microchips were composed of two 500-μm-thick borosilicate glass wafers (Borofloat 33, CSEM, Neuchatel, Swizwerland) that were bonded together. The channel wafer incorporated 31-μm-deep microchannels and through-wafer etched inlets and ESI tips (Figures 1a−d). The coverlid wafer, which acts as the bottom of the microchip during analyses, had only through-wafer etched ESI tips (Figures 1e-g). Two different microchip designs were fabricated. The first design was an array of nine separate ESI emitters (Figure 1i, j). Each emitter tip had a separate inlet (2.5 mm in diameter) connected to a straight 10-mm-long and 160μm-wide microfluidic channel that ended at an ESI tip. The second design comprised two separate CE/ESI units with a 20mm-long and 90-μm-wide CE separation channel, a simple cross-injection channel, and an 8-mm-long and 190-μm-wide sheath liquid channel (Figures 1h, k). The width of the joint channel after the intersection of the separation and sheath liquid channels was 210 μm. The fabrication process and the estimated process times are presented in Figure 1. The processing of the channel wafer started with low-pressure chemical vapor deposition (LPCVD) of a 400-nm-thick layer of polycrystalline silicon (poly-Si) on the glass wafer (Figure 1a). The inlets and the tips were patterned on the poly-Si layer to the backside of the wafer using photolithography and sulfur hexafluoride (SF6)-based reactive ion etching (RIE) (Plasmalab 80, Oxford Instruments, Bristol, U.K.). The wafer was etched at room temperature in hydrofluoric acid/hydrochloric acid (HF/HCl, 10:1) solution for 60 min using the poly-Si layers as etch masks. The resulting etch depth was ∼450 μm (Figure 1b). Subsequently, the inlets, the ESI tips and the channels were patterned to the poly-Si layer on the topside of the wafer using double-sided photolithography and RIE (Figure 1c), then the wafer was immersed back in the HF/HCl solution for 4 min to complete the through-wafer etching process and to create the 31-μmdeep microchannel network (Figure 1d). Processing of the coverlid wafer also started with LPCVD of a 400-nm-thick poly-Si layer on another glass wafer (Figure 1e). The ESI emitters were patterned to the poly-Si layer on the backside of the wafer using photolithography and RIE. The tips
■
SAFETY CONSIDERATIONS
HF/HCl solution has to be handled with extreme care due to toxicity and corrosiveness of the solution. Inhalation, skin contact, and ingestion are strictly forbidden. Glass etching has to be carried out in a ventilated fume hood using protective gloves, apron, and goggles. In addition, strong bases and solvents need to be handled with care. 8975
dx.doi.org/10.1021/ac301602b | Anal. Chem. 2012, 84, 8973−8979
Analytical Chemistry
■
Technical Note
RESULTS AND DISCUSSION Microfabrication Aspects. Glass microfabrication is often considered challenging due to the chemical inertness of glass. Deep reactive ion etching (DRIE), which is typically used to create deep and narrow structures in silicon, is not capable of producing deep structures (>100 μm) in glass without extreme masking methods.35 Therefore other techniques such as powder blasting36 and wet etching37,38 have been utilized instead. Our fabrication approach relies on isotropic wet etching of glass in HF/HCl solution using polycrystalline silicon as an etch mask similar to earlier work.38 Typically, dimensional control during isotropic etching is challenging because of severe undercutting. However, the isotropic nature of wet etching can be beneficial when fabricating threedimensional structures. Our process starts with isotropic etching of the inlets and the ESI emitters almost through the wafer from the backside of the channel wafer. Here, undercutting is beneficial because the bulk glass under the emitter tip is removed, forming a curved profile sharpening toward the tip (Figure 1b). However, the throughwafer etching is not completed all at once. Therefore, the microchannel network is easily defined into the poly-Si layer on the topside of the same wafer with the help of a standard UVphotolithography step, followed by RIE of poly-Si (Figure 1c). Finally, the second short isotropic glass wet etching step completes the through-wafer etching of the inlets and the ESI tip and simultaneously defines all microchannels (Figure 1d). As a result of the initial isotropic etching from the backside of the wafer, an extremely sharp overhanging emitter tip is formed at the end of the CE channel (Figure 2a−c). The depth of the
isotropic etching defines the three-dimensionally sharp emitter tips. In the final step, the coverlid and the channel wafer are aligned in a wafer bonder and fusion-bonded in a furnace as described elsewhere.31,38 Compared with the previously reported microfabricated ESI emitters,28,31 our approach provides three-dimensionally sharp emitters similar to commercial nanospray needles, but without the need for serial tip-to-tip processing, such as manual pulling28 or sawing.31 In addition, our process makes use of standard 500-μm-thick borosilicate glass wafers which are easier to handle than the extremely thin (150 μm) wafers that are required if electrospray is produced from a rectangular corner of a glass chip.31−33 Furthermore, our wafer-scale process is flexible regarding microchip layout, and microchannel dimensions can be easily changed without affecting the threedimensional shape of the emitter tip, especially if the last wet etching step (defining the microchannels) is replaced by a DRIE process. To our knowledge, parallel fabrication of multiple glass ESI emitters simultaneously, by using only standard microfabrication processes, has not been possible thus far. As a whole, our process is completely compatible with other parallel microfabrication techniques and, therefore, well-suited for mass production. Thus, we believe this approach resolves the long-time problems and challenges associated to the microfabrication of sharp ESI emitters out of glass. Mass Spectrometry. The analytical performance of the fully microfabricated glass emitters was tested by a series of ESI/MS measurements on small molecule and protein standards. First, the repeatability of analysis was examined by comparing the ion current stabilities and signal intensities between replicate emitters from an array similar to that in Figure 1i, j. For all emitters tested, the ion current stabilities of the protonated test compound verapamil (m/z 455.4) was within 5.3−13.5% RSD, and the repeatability of the verapamil signal intensity was 8.0% RSD (n = 6 replicate emitters). With a single sample application, stable ion current was obtained for as long as >15 min (Figure 3). Thanks to the good chemical stability of glass, the background spectrum was clear, with no interfering ions at higher mass range (Figure 3). Instead, for some polymer-based emitters, PDMS in particular, the background interference is a major problem because of the extraction of unpolymerized monomers by the organic solvents.
Figure 2. Scanning electron microscopy (SEM) images of the bonded glass microchips that reveal the three-dimensional sharpness of the ESI emitter tip: (a) a tilted overview, (b) a tilted close-up view, and (c) a side view of an ESI emitter tip. (d) The cross-section of a microfluidic channel after bonding.
second glass etching step is only a few tens of micrometers, and therefore, it could be done with DRIE as well. The DRIE approach would allow better control over the lateral dimensions of the microchannels than the wet etching approach (Figure 2d). Thanks to the anisotropic and one-sided nature of DRIE, the quality of the ESI emitter could also be improved. The processing of the coverlid wafer also exploits isotropic through-wafer etching. Only the ESI tips are patterned into the poly-Si layer on the backside of the wafer, and the throughwafer etching is done in one continuous process. Again,
Figure 3. The mass spectrum and extracted ion current (EIC) of 5 μM verapamil (m/z 455.4) obtained from a microfabricated glass ESI emitter. The solvent used was methanol−water 80:20 containing 1% acetic acid, and the applied ESI voltage was +1750 V. 8976
dx.doi.org/10.1021/ac301602b | Anal. Chem. 2012, 84, 8973−8979
Analytical Chemistry
Technical Note
sharp emitters. Since the ESI onset voltage scales to the outer dimensions of the emitter tip (in addition to the surface tension), the emitters need to be sharp in all three dimensions to reach low onset voltages and to be able to electrospray aqueous solutions. The main advantage of our emitter design is that the developed glass microfabrication process readily provides three-dimensionally sharp emitters that allow the use of low ESI voltages (typically 1.40−1.75 kV) and, thus, electrospraying of aqueous solutions. On the other hand, the same microfabrication process allows for monolithic integration of the emitter tip with any kind of microfluidic network, so online sample preparation and separation can be performed prior to ESI/MS detection. In addition to direct spray experiments performed with the emitter array design (Figure 1i, j), a series of CE−ESI/MS analyses were carried out with the separation chip design (Figure 1h, k). The potential difference for the CE (typically 300−500 V/cm) was applied between the separation channel inlet and the sheath liquid inlet. The fluidic resistances of the separation and the auxiliary channels were adjusted on the basis of our previous work25 so that fluid actuation in the separation channel was by cathodic EOF (no pressure-induced flow), while the sheath liquid flow in the auxiliary channel was mainly induced by electrospray aspiration. First, the performance of the separation chip was characterized by using a constant sheath liquid composition of methanol−water 80:20 containing 1% acetic acid and changing the methanol content (0−60%) of the separation buffer (20 mM ammonium acetate). Similar to direct spray experiments, all buffer compositions tested provided stable electrospray (5.3−8.9% RSD), evidencing the robustness of the developed design. To demonstrate the feasibility of the separation chip design for high-throughput analysis, CE separation of five tetraalkylammonium halides was performed. Even if the effective separation channel length was only 20 mm, the five ammonium halides were separated from each other within 60 s (Figure 5a). The very narrow peak widths (1.4−2.3 s at half height) and good separation efficiency (1.7 × 105−2.7 × 105 theoretical plates/m) of the ammonium halides evidenced negligible sample spreading at the emitter tip or at the intersection of the separation and auxiliary channels. The repeatability of CE−ESI/MS analysis from run to run was determined from repeated (n = 6) injections of 5 μg/μL cytochrome c solution. Thanks to the high hydrophilicity and the overall negative charge of cytochrome c at the buffer pH (pH 7), the hydrophobic and electrostatic interactions toward the negatively charged glass surface were reduced to a minimum, even if a significantly high protein concentration was used. As a result, very good migration time repeatability of 6.2% RSD (n = 6) was obtained. The protein peaks were also symmetric and narrow (3.1 ± 0.2 s at half height, Figure 5b), similar to those of the tetraalkylammonium halides, which confirmed that no sample spreading occurs at the microchannel outlet. The average number of theoretical separation plates obtained for cytochrome c was 2.1 × 104/m. In all, these results confirm that our parallel microfabrication approach provides sharp ESI emitters that allow robust performance from run to run and tip to tip and do not suffer from sample spreading at the microchannel outlet. Since the microfabrication process readily yields three-dimensionally sharp emitter tips, very low ESI voltages can be used, which eventually allows for the use of a variety of different solvent compositions from purely aqueous to high organic content. The figures of merit (e.g., separation efficiency) for the CE−
The electrospray performance of the microfabricated glass emitters was further examined by analyzing intact cytochrome c (12230 Da; pI 9.5) in solutions containing varying amount of methanol (10−80% v/v, in water). It was observed that the emitters provided equally robust performance, even if the solvent composition varied from nearly aqueous (10% methanol in water) to high organic content (80% methanol in water). Protein spectra of good quality were obtained at both extremes (Figure 4), allowing, for instance, molecular weight
Figure 4. MS spectra of 100 ng/μL cytochrome c dissolved in (a) methanol−water, 10:90, and (b) methanol−water, 80:20, both containing 1% acetic acid. The applied ESI voltage was +1400 V. CYC = cytochrome c. The adduct ion marked with an asterisk (*) represents molecular mass of M + 98 and likely originates from sulfate and phosphate impurities41 present in the commercial cytochrome c sample.
determination from both aqueous and organic solutions with equally high accuracy (12 231.0 ± 0.6 Da, 0.008%). Protein analysis, in particular, greatly benefits from the possibility to use aqueous solutions because this allows better insights into protein interactions under physiological conditions than the use of organic solvents. The model compound, cytochrome c, is a protein that is known to be relatively sensitive to the addition of organic modifiers, and it easily loses its native, folded conformation in the presence of methanol, for example.39 This is illustrated in Figure 4, where drastic change in the protein charge envelope is observed between aqueous (native, folded state) and organic solution (denatured, unfolded state). In its native state (Figure 4a), cytochrome c is less charged, since most of the high proton affinity (PA) residues capable of taking protons are hidden inside the folded structure. Instead, unfolding of the tertiary structure due to addition of methanol reveals more high-PA residues and shifts the charge envelope to the lower mass range (Figure 4b). The purpose of adding organic solvents to the sample solution arises from the need to decrease the ESI onset voltage by lowering the surface tension. Low onset voltage, in turn, reduces the required flow rate and thus enables more efficient ionization and better sensitivity, which are the main advantages of the miniaturization of the ESI source.40 However, most microfabricated ESI emitters thus far published are operated at relatively high ESI voltages, typically 2−3 kV relative to MS.3 This is because almost all microfabrication processes are planar and, thus, produce only two-dimensionally 8977
dx.doi.org/10.1021/ac301602b | Anal. Chem. 2012, 84, 8973−8979
Analytical Chemistry
Technical Note
molecule and protein standards. Future work envisions replacement of the last wet etching step, which defines the microchannels, with plasma etching to allow more freedom in the microchannel cross-sectional dimensions and more complex fluidic designs for more demanding bioanalyses.
■
AUTHOR INFORMATION
Corresponding Author
*Phone: +358 50 5711741. Fax: +358 9 191 59556. E-mail: Lauri.Sainiemi@live.fi. Present Address †
Microsoft Oy, Keilaranta 7, 02150, Espoo, Finland
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by the Academy of Finland (Grant Nos. 138674 and 251629) and the Biocentrum Helsinki collaboration Grant.
■
REFERENCES
(1) Koster, S.; Verpoorte, E. Lab Chip 2007, 7, 1394−1412. (2) Gibson, G. T. T.; Mugo, S. M.; Oleschuk, R. D. Mass Spectrom. Rev. 2009, 28, 918−936. (3) Sikanen, T.; Franssila, S.; Kauppila, T. J.; Kostiainen, R.; Kotiaho, T.; Ketola, R. A. Mass Spectrom. Rev. 2010, 29, 351−391. (4) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Science 1989, 246, 64−71. (5) Xue, Q.; Foret, F.; Dunayevskiy, Y. M.; Zavracky, P. M.; McGruer, N. E.; Karger, B. L. Anal. Chem. 1997, 69, 426−430. (6) Ramsey, R. S.; Ramsey, J. M. Anal. Chem. 1997, 69, 1174−1178. (7) Xue, Q.; Foret, F.; Dunayevskiy, Y.; Foret, Y. M.; Karger, B. L. Rapid Commun. Mass Spectrom. 1997, 11, 1253−1256. (8) Figeys, D.; Ning, Y.; Aebersold, R. Anal. Chem. 1997, 69, 3153− 3160. (9) Zhang, B.; Liu, H.; Karger, B. L.; Foret, F. Anal. Chem. 1999, 71, 3258−3264. (10) Bings, N. H.; Wang, C.; Skinner, C. D.; Colyer, C. L.; Thibault, P.; Harrison, D. J. Anal. Chem. 1999, 71, 3292−3296. (11) Deng, Y.; Henion, J.; Li, J.; Thibault, P.; Wang, C.; Harrison, D. J. Anal. Chem. 2001, 73, 639−646. (12) Akashi, S.; Suzuki, K.; Arai, A.; Yamada, N.; Suzuki, E. I.; Hirayama, K.; Nakamura, S.; Nishimura, Y. Rapid Commun. Mass Spectrom. 2006, 20, 1932−1938. (13) Lazar, I. M.; Ramsey, R. S.; Sundberg, S.; Ramsey, J. M. Anal. Chem. 1999, 71, 3627−3631. (14) Desai, A.; Tai, Y. C.; Davis, M. T.; Lee, T. D. Proceedings of the IEEE International Conference on Solid-State Sensors and Actuators (Transducers’97), Chicago, IL, 16−19 June, 1997, 2, pp 927−930. (15) Schultz, G. A.; Corso, T. N.; Prosser, S. J.; Zhang, S. A. Anal. Chem. 2000, 72, 4058−4063. (16) Arscott, S.; Le Gac, S.; Rolando, C. Sens. Actuators B 2005, 106, 741−749. (17) Sainiemi, L.; Nissilä, T.; Jokinen, V.; Sikanen, T.; Kotiaho, T.; Kostiainen, R.; Ketola, R. A.; Franssila, S. Sens. Actuators B 2008, 132, 380−387. (18) Kim, J. S.; Knapp, D. R. Electrophoresis 2001, 22, 3993−3999. (19) Arscott, S.; Le Gac, S.; Druon, C.; Tabourier, P.; Rolando, C. Sens. Actuators B 2004, 98, 140−147. (20) Tuomikoski, S.; Sikanen, T.; Ketola, R. A.; Kostiainen, R.; Kotiaho, T.; Franssila, S. Electrophoresis 2005, 26, 4691−4702. (21) Nissilä, T.; Sainiemi, L.; Franssila, S.; Ketola, R. A. Sens. Actuators B 2009, 143, 414−420. (22) Mery, E.; Ricoul, F.; Sarrut, N.; Constantin, O.; Delapierre, G.; Garin, J.; Vinet, F. Sens. Actuators B 2008, 134, 438−446.
Figure 5. Extracted ion electropherograms of (a) five tetraalkylammonium halides (each 10 μM) and (b) cytochrome c (5 μg/μL) separated by the glass CE−ESI chip (overlaid electropherograms from three repeated runs). The separation buffer was 20 mM ammonium acetate containing 40% methanol, and the sheath liquid was methanol−water 80:20 containing 1% acetic acid. The electric field strength during CE separation was (a) 300 or (b) 500 V cm−1.
ESI chip design are similar to those of the previously published glass CE−ESI chips,28−34 but the possibility to use only parallel microfabrication techniques and the lack of manual postprocessing of the ESI tips renders our approach superior in terms of mass production and analytical reproducibility.
■
CONCLUSIONS In this work, we present the solution for monolithic integration of three-dimensionally sharp ESI emitters with separation microchips made of glass, which has been one of the major challenges of micro total chemical analysis systems since the beginning. Our approach utilizes only standard and parallel microfabrication techniques and is thus feasible for mass production. The same approach can also be applied to implementation of various different microchannel layouts without much affecting the three-dimensional structure of the ESI tips. The ESI/MS analyses carried out with the six replicate emitters showed robust performance from tip to tip. The repeatability of signal intensity was 8.0% RSD (n = 6). On the other hand, the CE−ESI/MS analyses carried out with the separation chip design confirmed that no sample spreading occurs at the microchannel outlet by showing high separation efficiency (up to 2.7 × 105 theoretical plates/m) with good migration time repeatability from run to run (6.2% RSD, n = 6). In this study, the feasibility of the glass emitters for high throughput CE−ESI/MS analysis was demonstrated with small 8978
dx.doi.org/10.1021/ac301602b | Anal. Chem. 2012, 84, 8973−8979
Analytical Chemistry
Technical Note
(23) Sainiemi, L.; Nissilä, T.; Kostiainen, R.; Franssila, S.; Ketola, R. A. Lab Chip 2012, 12, 325−332. (24) Yin, H.; Killeen, K.; Brennen, R.; Sobek, D.; Werlich, M.; Van de Goor, T. Anal. Chem. 2005, 77, 527−33. (25) Sikanen, T.; Tuomikoski, S.; Ketola, R. A.; Kostiainen, R.; Franssila, S.; Kotiaho, T. Anal. Chem. 2007, 79, 9135−9144. (26) Rohner, T. C.; Rossier, J. S.; Girault, H. H. Anal. Chem. 2001, 73, 5353−5357. (27) Huikko, K.; Ö stman, P.; Grigoras, K.; Tuomikoski, S.; Tiainen, V. M.; Soininen, A.; Puolanne, K.; Manz, A.; Franssila, S.; Kostainen, R.; Kotiaho, T. Lab Chip 2003, 3, 67−72. (28) Hoffmann, P.; Häusig, U.; Schulze, P.; Belder, D. Angew. Chem., Int. Ed. 2007, 46, 4913−4916. (29) Fritzsche, S.; Hoffmann, P.; Belder, D. Lab Chip 2010, 10, 1227−1230. (30) Fritzsche, S.; Ohla, S.; Glaser, P.; Giera, D. S.; Sickert, M.; Schneider, C.; Belder, D. Angew. Chem., Int. Ed. 2011, 50, 9467−9470. (31) Mellors, J. S.; Gorbounov, V.; Ramsey, R. S.; Ramsey, J. M. Anal. Chem. 2008, 80, 6881−6887. (32) Mellors, J. S.; Jorabchi, K.; Smith, L. M.; Ramsey, J. M. Anal. Chem. 2010, 82, 967−973. (33) Chambers, A. G.; Mellors, J. S.; Henley, W. H.; Ramsey, J. M. Anal. Chem. 2011, 83, 842−849. (34) Chambers, A. G.; Ramsey, J. M. Anal. Chem. 2012, 84, 1446− 1451. (35) Li, X.; Abe, T.; Esashi, M. Sens. Actuators A 2001, 87, 139−145. (36) Yamahata, C.; Lacharme, F.; Burri, Y.; Gijs, M. A. M. Sens. Actuators B 2005, 30, 1−7. (37) Iliescu, C.; Jing, J.; Tay, F. E. H.; Miao, J.; Sun, T. Surf. Coat. Technol. 2005, 198, 314−318. (38) Saarela, V.; Haapala, M.; Kostiainen, R.; Kotiaho, T.; Franssila, S. Lab Chip 2007, 7, 644−646. (39) Samalikova, M.; Matecko, I.; Müller, N.; Gandori, R. Anal. Bioanal. Chem. 2004, 378, 1112−1123. (40) Wilm, M.; Mann, M. Anal. Chem. 1996, 68, 1−8. (41) Chowdury, S. K.; Katta, V.; Beavis, R. C.; Chait, B. T. J. Am. Soc. Mass Spectrom. 1990, 1, 382−386.
8979
dx.doi.org/10.1021/ac301602b | Anal. Chem. 2012, 84, 8973−8979
