1 Anal. Chem. 2007, 79, Small-Molecule Analysis with Silicon-Nanoparticle-Assisted Laser Desorption/ Ionization Mass Spectrometry Xiujuan Wen, Shai Da...
1 Letters in Peptide Science, 4 (1997) ESCOM LIPS 165 Rapid semi-on-line monitoring of Fmoc solid-phase peptide synthesis by matrix-assisted laser des...
1 Anal. Chem. 2005, 77, Ordered Silicon Nanocavity Arrays in Surface-Assisted Desorption/Ionization Mass Spectrometry Nancy H. Finkel, Brian G. Prevo,...
1 Anal. Chem. 2008, 80, Nanostructured Diamond-Like Carbon on Digital Versatile Disc as a Matrix-Free Target for Laser Desorption/Ionization Mass Spec...
1 Anal. Chem. 2000, 72, An Algorithm for Automated Bacterial Identification Using Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry Kristi...
1 Anal. Chem. 2004, 76, Electrowetting-Based Microfluidics for Analysis of Peptides and Proteins by Matrix-Assisted Laser Desorption/Ionization Mass S...
1 RAPID COMMUNICATIONS IN MASS SPECTROMETRY Rapid Commun. Mass Spectrom. 2007; 21: Published online in Wiley InterScience ( Adduct simplification in ...
1 Statistical discrimination of black gel pens inks analyzed by laser desorption/ionization mass spectrometry Abstract Pearson correlation coefficient...
1 Anal. Chem. 2007, 79, Analysis of Adenosine Triphosphate and Glutathione through Gold Nanoparticles Assisted Laser Desorption/Ionization Mass Spectr...
1 University of Massachusetts - Amherst Amherst Doctoral Dissertations May current Dissertations and Theses 2017 Analysis of Gold Nanoparticles and Th...
Anal. Chem. 2008, 80, 2973-2981
Dynamic Electrowetting on Nanofilament Silicon for Matrix-Free Laser Desorption/Ionization Mass Spectrometry Chia-Wen Tsao, Parshant Kumar, Jikun Liu, and Don L. DeVoe*
Department of Mechanical Engineering, University of Maryland, College Park Maryland 20742
Dynamic electrowetting on nanostructured silicon surfaces is demonstrated as an effective method for improving detection sensitivity in matrix-free laser desorption/ ionization mass spectrometry. Without electrowetting, silicon surfaces comprising dense fields of oriented nanofilaments are shown to provide efficient ion generation and high spectral peak intensities for deposited peptides bound to the nanofilaments through hydrophobic interactions. By applying an electrical bias to the silicon substrate, the surface energy of the oxidized nanofilaments can be dynamically controlled by electrowetting, thereby allowing aqueous buffer to penetrate deep into the nanofilament matrix. The use of electrowetting is shown to result in enhanced interactions between deposited peptides and the nanofilament silicon surface, with improved signal-to-noise ratio for detected spectral peaks. An essential feature contributing to the observed performance enhancement is the open-cell nature of the nanofilament surfaces, which prevents air from becoming trapped within the pores and limiting solvent penetration during electrowetting. The combination of nanofilament silicon and dynamic electrowetting is shown to provide routine detection limits on the order of several attomoles for a panel of model peptides. Following its introduction in the late 1980s,1 matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) has emerged as a leading method for soft ionization of analyte molecules prior to mass determination, typically by time-of-flight (TOF) MS. The technique has been widely used for a range of biopolymers, which are typically cocrystallized with a UV-absorbing organic matrix to enhance their ionization efficiency. However, MALDI-MS is well-known to suffer from excessive matrix background signal for low molecular weight analytes, and thus its application has been limited to the study of relatively large biopolymers. Furthermore, MALDI-MS sensitivity is generally lower than that of electrospray ionization (ESI)-MS, with typical detection limits around 1 fmol,2 although limits on the order of 10-100 amol can achieved using microstructured targets and optimized target preparation methods.3-5 * Corresponding author. Phone: 301-405-8125. Fax: 301-314-9477. E-mail: [email protected] (1) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60 (20), 2299-2301. (2) Siuzdak, G. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 11290-11297. (3) Keller, B. O.; Li, L. Anal. Chem. 2001, 73 (13), 2929-2936. 10.1021/ac7026029 CCC: $40.75 Published on Web 02/27/2008
In 1999, Wei et al. reported the desorption/ionization on porous silicon (DIOS) technique,6 in which an electrochemically etched porous silicon (pSi) surface serves to efficiently absorb UV laser energy during LDI-MS analysis, allowing effective biomolecular measurements in the absence of organic matrix. As a result, matrix interference typically encountered below ca. 500 m/z in MALDIMS is eliminated, allowing biomolecules within this range to be accessed by the DIOS-MS technology.1 In addition to smallmolecule analysis,7 the technology has been applied to protein characterization.8-10 Because the pSi targets can be readily functionalized, preferential binding of analytes based on affinity11,12 or hydrophobic13 interactions has been demonstrated. The high sensitivity combined with good tolerance to contaminants also makes DIOS-MS an attractive platform for forensics applications.14 Detailed reviews of the DIOS-MS technique, including background and applications, have been presented in recent review papers.15,16 Nanostructured pSi surfaces developed for DIOS-MS are prepared by galvanostatic etching of silicon,6,17,18 resulting in a surface consisting of nanoscale pores which exhibit a closed-cell morphology. For the analysis of biopolymers containing hydrophobic domains, such as peptides, silylation of the pSi surfaces (4) Ekstrom, S.; Ericsson, D.; Onnerfjord, P.; Bengtsson, M.; Nilsson, J.; MarkoVarga, G.; Laurell, T. Anal. Chem. 2001, 73 (2), 214-219. (5) Schuerenbeg, M.; Luebbert, C.; Eickhoff, H.; Kalkum, M.; Lehrach, H.; Nordhoff, E. Anal. Chem. 2000, 72 (15), 3436-3442. (6) Wei, J.; Buriak, J. M.; Siuzdak, G. Nature 1999, 399 (6733), 243-246. (7) Seino, T.; Sato, H.; Torimura, M.; Shimada, K.; Yamamoto, A.; Tao, H. Anal. Sci. 2005, 21 (5), 485-490. (8) Thomas, J. J.; Shen, Z. X.; Crowell, J. E.; Finn, M. G.; Siuzdak, G. Proc. Natl. Acad. Sci. U.S.A. 2001, 98 (9), 4932-4937. (9) Prenni, J. E.; Shen, Z. X.; Trauger, S.; Chen, W.; Siuzdak, G. Spectrosc.: Int. J. 2003, 17 (4), 693-698. (10) Hu, L. G.; Xu, S. Y.; Pan, C. S.; Zou, H. F.; Jiang, G. B. Rapid Commun. Mass Spectrom. 2007, 21 (7), 1277-1281. (11) Meng, J. C.; Siuzdak, G.; Finn, M. G. Chem. Commun. 2004, No. 18, 21082109. (12) Go, E. P.; Uritboonthai, W.; Apon, J. V.; Trauger, S. A.; Nordstrom, A.; O’Maille, G.; Brittain, S. M.; Peters, E. C.; Siuzdak, G. J. Proteome Res. 2007, 6 (4), 1492-1499. (13) Trauger, S. A.; Go, E. P.; Shen, Z. X.; Apon, J. V.; Compton, B. J.; Bouvier, E. S. P.; Finn, M. G.; Siuzdak, G. Anal. Chem. 2004, 76 (15), 4484-4489. (14) Thomas, J. J.; Shen, Z. X.; Blackledge, R.; Siuzdak, G. Anal. Chim. Acta 2001, 442 (2), 183-190. (15) Lewis, W. G.; Shen, Z. X.; Finn, M. G.; Siuzdak, G. Int. J. Mass Spectrom. 2003, 226 (1), 107-116. (16) Peterson, D. S. Mass Spectrom. Rev. 2007, 26 (1), 19-34. (17) Shen, Z. X.; Thomas, J. J.; Averbuj, C.; Broo, K. M.; Engelhard, M.; Crowell, J. E.; Finn, M. G.; Siuzdak, G. Anal. Chem. 2001, 73 (3), 612-619. (18) Bisi, O.; Ossicini, S.; Pavesi, L. Surf. Sci. Rep. 2000, 38 (1-3), 1-126.
Analytical Chemistry, Vol. 80, No. 8, April 15, 2008 2973
using a reagent which presents fluorinated groups can serve to bind analyte molecules through hydrophobic interactions, allowing contaminants in the deposited sample to be removed from the pSi surface while retaining bound analyte molecules. With the use of this approach for peptide analysis, extremely low detection limits have been achieved by DIOS-MS.13 Further improvements in detection limits are desirable for the analysis of limited samples or low-abundance species. In this context, a number of nanostructured surfaces have been recently explored as alternatives to pSi for matrix-free LDI-MS, including porous alumina,19 silicon nanowires,20-22 and carbon nanotubes.23,24 Here we investigate the use of dynamic electrowetting on nanofilament silicon (nSi) as an approach toward increasing the sensitivity of matrix-free LDI-MS. The nSi term is used as a convenient and descriptive notation to refer to surfaces which possess high aspect ratio silicon needles oriented parallel to one another and normal to the plane of the substrate. The nSi surfaces employed in this work were produced using an metal-assisted oxidative etching process termed HOME-HF as first reported by Li and Bohn25 and later explored for the fabrication of LDI-MS target substrates.26 By controlling the surface preparation and etching conditions, nSi surfaces exhibit dense fields of oriented nanofilaments, resulting in a high specific surface area and an open-cell morphology. This morphology differentiates the nSi surfaces from pSi, as their open-cell nature prevents the entrapment of air which could otherwise limit binding capacity by preventing effective interactions between deposited analyte and the full height of the nanofilaments. Another useful feature of nSi surfaces is their superhydrophobic nature. Following the approach described by Trauger et al.,13 the nSi chips are oxidized and silylated to produce a fluorinated surface suitable for peptide analysis. Due to the geometry and morphology of the nanofilaments, both as-fabricated and fluorinated nSi exhibits superhydrophobic behavior with large effective water contact angles. As a result, the contact area between the sample droplet and nanofilament surface remains small, preventing dispersion over the surface and allowing concentration of analyte as solvent within the droplet evaporates. In addition to further evaluating nSi as a matrix-free LDI-MS substrate, the use of dynamic electrowetting on nSi is demonstrated as an effective method for improving LDI-MS detection limits. By applying a bias voltage between the sample droplet and superhydrophobic nSi surface, the nanofilament surface energy is increased, rendering the nSi surface hydrophilic and allowing sample to penetrate deep into the nanofilament matrix, resulting (19) Nayak, R.; Knapp, D. R. Anal. Chem. 2007, 79 (13), 4950-4956. (20) Go, E. P.; Apon, J. V.; Luo, G.; Saghatelian, A.; Daniels, R. H.; Sahi, V.; Dubrow, R.; Cravatt, B. F.; Vertes, A.; Siuzdak, G. Anal. Chem. 2005, 77 (6), 1641-1646. (21) Luo, G. H.; Chen, Y.; Daniels, H.; Dubrow, R.; Vertes, A. J. Phys. Chem. B 2006, 110 (27), 13381-13386. (22) Kang, M. J.; Pyun, J. C.; Lee, J. C.; Choi, Y. J.; Park, J. H.; Park, J. G.; Lee, J. G.; Choi, H. J. Rapid Commun. Mass Spectrom. 2005, 19 (21), 31663170. (23) Xu, S. Y.; Li, Y. F.; Zou, H. F.; Qiu, J. S.; Guo, Z.; Guo, B. C. Anal. Chem. 2003, 75 (22), 6191-6195. (24) Hu, L. G.; Xu, S. Y.; Pan, C. S.; Yuan, C. G.; Zou, H. F.; Jiang, G. B. Environ. Sci. Technol. 2005, 39 (21), 8442-8447. (25) Li, X.; Bohn, P. W. Appl. Phys. Lett. 2000, 77 (16), 2572-2574. (26) Kruse, R.; Li, X.; Bohn, P.; Sweedler, J. V. Anal. Chem. 2001, 73, 36393645.
Analytical Chemistry, Vol. 80, No. 8, April 15, 2008
in increased peptide-surface interactions and higher LDI-MS spectral intensities. Electrowetting on nSi is observed to be a dynamic process, with full wetting of the nanofilament surface occurring on the time scale of several tens of seconds. As a result, the method allows a sample droplet to wet the full depth of the nanofilaments while preventing the droplet from spreading beyond a contact area matched to the LDI-MS laser spot size of several hundred micrometers. EXPERIMENTAL SECTION Materials and Reagents. The 10 cm diameter p-type 〈100〉 silicon wafers with resistivity of 0.005-0.02 Ω cm (Customized Communication Inc., La Center, WA) were used as the substrates for all nSi chips. Ethyl alcohol was purchased from PharmcoAAPER (Shelbyville, KY). The 49% hydrofluoric acid (HF), 37% concentrated hydrochloric acid, 30% hydrogen peroxide, and urea were purchased from J.T. Baker (Phillipsburg, NJ). Buffered oxide etchant (BOE) was purchased from Transene Company (Danvers, MA). Tris base, HPLC grade water, methanol, and acetonitrile were purchased from Fisher Scientific (Pittsburgh, PA). Lyophilized bovine serum albumin (BSA), 0.1% LC-MS grade trifluoroacetic acid (TFA) solution, DL-dithiothreitol, and lyophilized desArg9-bradykinin were purchased form Sigma-Aldrich (St. Louis, MO). MassPREP DIOS target chips and lyophilized model peptides were purchased from Waters (Milford, MA). Mass spectrometry grade trypsin was purchased from Promega (Madison, WI). (Pentafluorophenyl)propyldimethylchlorosilane was purchased from Gelest (Morrisville, PA). Conductive epoxy was purchased from Chemtronic (Kennesaw, GA). Sample Preparation. Lyophilized peptide sample containing of 1 nmol of angiotensin fragment 1-7 (MW 898.7), RASG-1 (MW 1000.5), angiotensin II (MW 1045.5), bradykinin (MW 1059.6), angiotensin I (MW 1295.7), and renin substrate (MW 1758.0) were reconstituted with 0.1% aqueous TFA to an initial concentration of 1 pmol/µL. The peptide solution was diluted 10-, 100-, and 1000fold with HPLC grade water for final peptide concentrations of 100, 10, and 1 fmol/µL, respectively. Lyophilized des-Arg9bradykinin (MW 904.0) sample was prepared by reconstituting in HPLC grade water at 0.11 µg/µL followed by subsequent serial dilution to aliquots of 1 pmol/µL, 10 fmol/µL, 100 amol/µL, and 1 amol/µL sample concentrations. For the preparation of tryptic digest, BSA was dissolved in a solution consisting of 8 M urea, 5 mM DL-dithiothreitol, and 50 mM Tris-HCl buffer (pH 8.2) to a concentration of 5 µg/µL. The solution was heated at 95 °C for 20 min to denature the protein. After cooling to room temperature, 50 mM Tris-HCl buffer was added to decrease the concentration of urea to 0.5 M, and trypsin was added to a final enzyme/substrate ratio of 1:50 (w/w). The digestion was allowed to proceed at 37 °C for 20 h. The digested BSA sample was desalted using a C18 ZipTip pipet tip (Millipore, Billerica, MA) into a final acetonitrile/TFA (7:3 v/v) solution. Target Preparation. Nanofilament substrates were fabricated by the HOME-HF etching process developed by Li and Bohn.25 A 3 nm thick gold layer was deposited onto selected regions of a virgin silicon wafer by e-beam evaporation, using a shadow mask for pattern generation, followed by immersion into a solution of HF, H2O2, and EtOH (1:1:1 v/v/v) in a Teflon beaker for 32 s. Upon removal from the etchant, the substrate was immediately
rinsed with methanol and blown dry in a gentle stream of N2. The etched wafer was then oxidized by a 30 min UV/ozone treatment (PSD-UV, Novascan Technologies, Ames, IA) to form a thin SiO2 layer on the nanofilaments. The oxidized surface was fluorinated by dispensing 15 µL of (pentafluorophenyl)propyldimethylchlorosilane in a glass dish and fixing the nSi substrate on a glass cover to seal the dish and allow the silylating reagent to coat the nanofilaments by evaporation. The vapor deposition process was performed at 65 °C for 15 min on a hot plate. The treated nanofilament substrate was rinsed with methanol and blown dry in a stream of N2. The pSi substrates used for comparison were treated using the identical silylation procedure. Morphological characterization of the resulting surfaces was performed using a field emission scanning electron microscope, Hitachi S-4700 FESEM (Hitachi High Technologies America Inc., Schaumburg, IL). Safety Considerations. HF is a hazardous acid which can result in serious tissue damage if burns are not properly treated. Etching of silicon should be conducted in a well-ventilated fume hood with appropriate safety considerations including a chemical smock, face shield, and double-layered nitrile gloves. Take care not to breath any HF fumes. Etching with HF should only be performed if calcium gluconate ointment is available for burn treatment. If contact with HF is suspected, remove contaminated clothing, flush vigorously with cold water, and massage calcium gluconate into the exposed area to neutralize the HF acid. Immediately seek physician care following any HF exposure. Hydrogen peroxide is an exceptionally strong oxidizer which can cause severe eye, skin, and respiratory burns. As with HF, etching with hydrogen peroxide should only be performed in a well-ventilated fume hood while wearing appropriate chemical safety protection. Sample Deposition and Electrowetting. Sample deposition was performed using a 100 µm i.d. capillary (Upchurch Scientific, Oak Harbor, WA) connected to a syringe pump (PHD2000, Harvard apparatus, Holliston, MA) to generate a 150 nL sample droplet at the cleaved capillary tip. For deposition tests without electrowetting, the capillary tip was positioned to bring the droplet into contact with the nSi surface for 30 s, followed by removal of the capillary tip. When applying sample to the surface without electrowetting, no residual liquid was observed on the nSi chip after capillary removal. For deposition tests employing electrowetting, the nanofilament substrate was first bonded to a stainless steel plate using conductive epoxy (Techni-Tool, Worcester, PA) and cured for 12 h at room temperature. Prior to bonding, the native oxide layer on the backside of the silicon was removed by applying BOE with a cotton swab to ensure good electrical conductivity. The deposition capillary was connected with a platinum electrode through a capillary tee (Upchurch Scientific), and the electrode was connected to a dc power supply (E3612A, Agilent Technologies, Santa Clara, CA) to provide the desired electrowetting bias. A digit multimeter (34401A, Agilent) was used to monitor the electrowetting current. After bringing the 150 nL droplet into contact with the nSi target, the electrowetting bias was applied. After 30 s, the capillary was raised off the chip, depositing a sample spot approximately 0.5 mm in diameter on the nanofilament surface.
The residual liquid was removed by blowing a stream of N2 across the nSi surface. All deposition experiments were performed under an optical goniometer instrument (Cam Plus Micro, Tantec Inc., Schaumburg, IL) for effective control of droplet delivery and to provide for real-time measurements of droplet contact angles. LDI-MS Characterization. Mass spectrometry was performed using a Kratos Amixa MALDI-TOF instrument (Kratos Analytical, Manchester, U.K.). All LDI-MS spectra were recorded in the linear, positive ion mode and averaged over 25-50 laser pulses using a 337 nm nitrogen laser with a 3 ns pulse width. The effective laser spot size was estimated to be around 200 µm based on visual observation during experiments. Deposited sample spot locations could not be readily visualized under standard illumination in the Kratos tool, and individual laser shots were used to determine sample spot locations by searching for maximum signal-to-noise ratio (S/N). After locating a sample spot, laser energy was adjusted for maximum S/N prior to each test. Signal intensity and S/N were obtained by analyzing collected mass spectral data using Kompact software version 2.3.4 (Kratos Analytical). RESULTS AND DISCUSSION Nanofilament Substrate Fabrication. The metal-assisted HF/H2O2/EtOH etching process was found to provide a simple and reproducible technique for fabricating well-oriented silicon nanofilaments with excellent control over the final surface morphology. During the etching process, the color of the goldcoated silicon surface rapidly changes from yellow to brown and, finally, to black after a sufficient etching time. Regions of the silicon surface without gold coating remain unaffected by the etchant. The nSi chips used in the present work contain a 64element array of 1 mm diameter nanofilament pads spaced 2 mm center-to-center, as shown in Figure 1. The morphology of the resulting nanofilaments can be seen from the electron micrographs in Figure 2a. The nSi surface consists of individual nanofilaments 7-12 nm in diameter and 2.8 µm deep. Condensation of the nanofilament tips, presumably due to capillary forces during the postetch drying process, results in the appearance of nanofilament “curtains” and a distribution of pore sizes with characteristic dimensions ranging from about 20 to 80 nm, with a smaller number of larger pores approaching 200 nm. Variations in nanofilament diameter, pore diameter, and pore distribution can be realized by altering the thickness of the initial gold film prior to etching, while nanofilament height can be Analytical Chemistry, Vol. 80, No. 8, April 15, 2008
Figure 2. Electron micrographs of (a) fabricated nSi and (b) commercial pSi substrates.
independently adjusted by varying the etch time. Although not intended as an exhaustive evaluation, superior LDI-MS performance was observed for the particular nanofilament geometry and morphology depicted in Figure 2a. In comparison, micrographs of a commercially available pSi substrate are shown in Figure 2b. It has been reported that smaller pSi pore sizes produce more intense ion signals, with optimal results realized for pore diameters between 70 and 120 nm and pore depths around 200 nm.6,17 Similarly, ion signal from porous alumina has also been shown to increase as the pore dimensions are reduced to 100 nm, with 600 nm optimized thickness,19 although smaller pores were not report due to fabrication challenges. After etching, the hydrogen-terminated nSi surfaces are initially highly hydrophobic, but over time they become increasingly hydrophilic as native oxide forms on the filaments. To avoid this, the chips were modified with (pentafluorophenyl)propyldimethylchlorosilane after etching to generate fluorinated surfaces with superhydrophobic behavior. The hydrophobicity of a solid surface is related to both the chemical groups present on the surface and the micro- or nanoscale surface texture. It is wellknown that if a flat hydrophobic surface with a water contact angle of θ is given a suitable microtexture, an effective contact angle θo > θ will be observed for water droplets placed on the textured 2976
Analytical Chemistry, Vol. 80, No. 8, April 15, 2008
surface. The effective contact angle can be described by the Cassie model,27 which assumes that pockets of air remain in the pores beneath the droplet so that the total interfacial free energy is a function of the fractional contact area φ between the liquid and the solid. On the basis of the Cassie model, the effective contact angle for a superhydrophobic surface is given by
cos θo ) -1 + φ(1 + cos θ)
The fractional contact area of the nanofilament surface shown in Figure 1 was measured at φ ) 58%, and goniometer measurements on the regions far from the nanofilaments present a nominal contact angle of θ ) 82°, resulting in a predicted effective contact angle of θo ) 109°. This value compares favorably with a measured effective contact angle of θo ) 105° for this particular chip. Although nSi chips with lower values of φ were observed to result in substantially higher effective contact angles up to 148°, this increase occurs at the expense of surface area, potentially reducing the performance of these surfaces for matrix-free LDI-MS. LDI-MS on nSi. The performance of nSi as an LDI-MS target was first evaluated in the absence of electrowetting. A dilution (27) Cassie, A. B. D.; Baxter, S. Trans. Faraday Soc. 1944, 40, 546-551.
Figure 3. (a) LDI-MS spectra for des-arg9-bradykinin on nSi and pSi substrates and (b) comparison of signal intensity for pSi and two different nSi substrates. Error bars represent one standard deviation.
study using a single model peptide, des-Arg9-bradykinin, was performed following a report by Trauger et al. in which exception-
ally high sensitivity was realized for this analyte using pSi DIOSMS.13 Typical mass spectra using an optimized nSi chip are shown Analytical Chemistry, Vol. 80, No. 8, April 15, 2008
Figure 4. LDI-MS spectra for a model peptide mixture at varying sample loadings on (a) nSi and (b) pSi substrates.
together with mass spectra acquired from a commercial pSi chip in Figure 3a, and a comparison of average spectral intensities over three individual tests is shown in Figure 3b for nSi chips with different pore sizes in addition to pSi. Noting that the MS detector was saturated for the 150 fmol nSi loading case, the nanofilament chips with 20-200 nm pore dimensions provide nearly double the ion intensity of the nSi surface with smaller pores and 2-4 times the ion intensities of pSi. The former observation is consistent with nSi LDI-MS measurements performed by Kruse et al.,26 which revealed lower ion generation efficiency for smaller pores around 3 nm. At the lowest loading level, 15 amol, S/N for the nSi target with 20-200 nm pores was larger than 100, suggesting a detection limit around 0.5 amol without electrowetting. The nSi targets also performed well when using a mixture of model peptides. Figure 4 shows LDI-MS spectra generated for different loading amounts of a peptide mixture including RASG1, angiotensin fragment 1-7, angiotensin I, angiotensin II, bradykinin, and renin substrate. With the use of the nSi target, peaks from each of these analytes were generated for loadings of 1.5 and 15 fmol, while all but RASG-1 (MW 1000.5) were apparent for the 150 amol loading case. Under identical sample target 2978
Analytical Chemistry, Vol. 80, No. 8, April 15, 2008
preparation conditions, peak intensities from the pSi target were significantly lower, with no peaks seen above noise at 150 amol loading. To observe the performance of nSi as a matrix-free LDI-MS substrate for the analysis of more complex peptide samples, a 700 fmol sample of BSA tryptic digest on nSi was also evaluated. As shown in Figure 5, the spectrum reveals many peaks between 500 and 2300 m/z corresponding to theoretical singly charged digest fragments. With the use of the Mascot search engine (Matrix Science, Boston, MA), the spectrum was used to successfully identify BSA by peptide fingerprinting with 34% sequence coverage. Electrowetting Deposition. After fluorination of the oxidized silicon surface, the nanofilament regions display superhydrophobic behavior, with water droplets deposited on the nanofilaments exhibiting substantially higher effective contact angles than the surrounding untextured surface. As a result of the low surface energy, sample droplets brought into contact with the nSi surface remain pinned to the capillary tip rather than transferring to the nanofilaments. However, application of a voltage between the nSi chip and sample droplet using the apparatus shown in Figure 6a allows the nanofilament surface energy to be increased through
Figure 5. Mass spectrum from 700 fmol of BSA tryptic digest.
Figure 7. Hydrophobic to hydrophilic transition using electrowetting on nSi.
apparent contact angle of a water drop which fully wets the pores is described by32
cos θo ) 1 + φ(cos θ - 1)
Figure 6. Electrowetting deposition setup and optical goniometer projection image shows electrowetting deposition to a superhydrophobic nanofilament substrate.
electrowetting. Selecting a suitable electrowetting bias enables the sample droplet to enter into the porous structure of the nanofilament network, substantially increasing the adhesion force between the droplet and chip surface and allowing complete droplet transfer from the capillary, as depicted in Figure 6b. As with other reports of electrowetting on nanostructured surfaces,28-30 water contact angles on the nanofilament surfaces scale with the inverse of the electrowetting bias. As shown in Figure 7, effective contact angles drop from an initial value of 105° to steady-state values ranging from 46° to 70°, depending on the applied electrowetting bias, over a period of several minutes. The reduction in contact angle is due to liquid wetting the heights of nanofilaments to varying degrees. According to the Cassie model, hydrophobicity is strengthened because air pockets remain trapped below the liquid drop. This is a reasonable assumption for closed-cell porous surfaces such as pSi, since air cannot be readily displaced from within the pores. However, the analysis of electrowetting behavior on open-cell nSi requires a different treatment. According to the Wenzel model,31 hydrophobicity of a rough surface can be geometrically modified due to the increase in effective surface area. For the Wenzel model, air may be partially or fully removed from the pores. In the latter case, the (28) Krupenkin, T.; Taylor, J. A.; Kolodner, P.; Hodes, M. Bell Labs Tech. J. 2005, 10 (3), 161-170. (29) Krupenkin, T. N.; Taylor, J. A.; Schneider, T. M.; Yang, S. Langmuir 2004, 20 (10), 3824-3827.
On the basis of this model, the apparent contact angle for an aqueous droplet which fully wets the pores of an nSi surface is predicted to be 60°. The somewhat lower contact angles observed at high electrowetting biases is likely due to a lower-than-expected value of θ used in eq 2, since θ was determined for a smooth surface without electrowetting. The observed time dependence of the contact angle during electrowetting contrasts with a previous study of electrowetting on nanostructured surfaces which reported instantaneous CassieWenzel transitions with correspondingly rapid decreases in water contact angle upon application of the electrowetting bias.29 Instead, electrowetting on superhydrophobic nSi is found to be a dynamic process, with the time-dependent effective contact angle θ(t) exhibiting a smooth first-order step response upon application of a constant bias voltage, with a characteristic time constant of 22 s. This dynamic behavior allows precise control of contact angle by adjusting the voltage application time. In addition, it appears that the dynamic electrowetting also makes it possible to drive sample deep into the nanofilament matrix without wetting a large region of the chip. As a result, good analyte-surface binding may be achieved without sacrificing local sample concentration during the initial deposition step. Electrowetting LDI-MS. Spectral peak intensities for the model peptides were found to increase linearly with the applied electrowetting voltage used during sample deposition. At the same time, S/N was also found to increase, presumably due to larger amounts of contaminants entering the nanofilament pores at higher electrowetting biases. By optimizing the laser energy to maximize S/N for each deposited sample spot, the average S/N (30) Krupenkin, T. N.; Taylor, J. A.; Wang, E. N.; Kolodner, P.; Hodes, M.; Salamon, T. R. Langmuir 2007, 23 (18), 9128-9133. (31) Wenzel, R. N. Ind. Eng. Chem. 1936, 28, 988-994. (32) Bormashenko, E.; Pogreb, R.; Whyman, G.; Erlich, M. Langmuir 2007, 23, 6501-6503.
Analytical Chemistry, Vol. 80, No. 8, April 15, 2008
Electrowetting-enhanced sample deposition was also performed on the commercial pSi target substrate. With the use of identical deposition and LDI-MS procedures, electrowetting on the pSi targets resulted in an average S/N enhancement factor of 1.8, demonstrating that, although the electrowetting technique can improve the performance of LDI-MS from pSi, the enhancement is not as pronounced as that observed from the nSi surfaces. This difference may be due to the contrasting morphologies of the two surfaces, with more favorable electrowetting of the open-cell nSi pores compared with the closed-cell pSi pores, which can trap air and prevent full wetting of the pore walls. Across all tests, optimal laser energy for the nSi surfaces was ∼80% of the pSi energy levels, compared with only 2% reported by Go et al. for LDI-MS from silicon nanowire surfaces.20 The full utility of the electrowetting technique depends upon the preparation of sample using a solvent with a relatively high aqueous content. Although less polar organic solvents such as acetonitrile or methanol can wet deep into the nanofilament pores without the need for electrowetting, these solvents also readily spread over the target surface, resulting in a substantial reduction in local analyte concentration. Although some charged peptides exhibit low solubility in water, solubility can be readily enhanced by shifting the pH of the aqueous solution or by adding minimal amounts of organic components to the aqueous solvent. This same issue applies to other LDI-MS target technologies which seek to take advantage of hydrophobic regions for on-target sample concentration.
Figure 8. Signal-to-noise ratio of LDI-MS spectra for a peptide mixture spotted on an nSi chip using varying dynamic electrowetting voltages with (a) 15 fmol, (b) 1.5 fmol, and (c) 150 amol loadings. Error bars show standard deviations based on three measurements across different spotting runs.
value for each peptide was found to increase with the electrowetting voltage for loading amounts varying from 15 fmol to 150 amol. In Figure 8, measured S/N ratios for the three peptides with the highest spectral intensities, namely, angiotensin I, angiotensin II, and bradykinin, are shown for each loading level. Overall, the average S/N was enhanced by a factor of 3.5 using the electrowetting approach (Figure 8, Supporting Information). 2980
Analytical Chemistry, Vol. 80, No. 8, April 15, 2008
CONCLUSIONS Nanofilament silicon provides an attractive alternative to electrochemically etched porous silicon for matrix-free LDI-MS as pioneered by Siuzdak’s group. The simple metal-assisted nSi fabrication process developed by Li and Bohn is easy to implement, highly repeatable, and amenable to large-scale production. The resulting nanofilament surfaces are particularly well-suited for use with the dynamic electrowetting technique described here, with dynamic electrowetting on nSi LDI-MS targets substantial improvements in detection limits for a panel of model peptides. Further improvements may be possible by optimizing the nanofilament geometry and target preparation methods. Furthermore, the use of electrowetting for effective target deposition and nanofilament penetration, combined with on-target solvent evaporation for in situ analyte concentration, is an attractive avenue toward higher sensitivity matrix-free LDI-MS. The morphology of the nSi pores appears to be an important factor in providing signal enhancement by allowing analyte to penetrate deep into the nanofilament matrix. The open-cell structure of the pore network prevents the formation of air pockets, ensuring a complete Cassie-Wenzel transition upon application of a sufficiently high electrowetting voltage. Although electrowetting is a particularly efficient method for controlling solvent/analyte penetration, similar results can likely be achieved using other approaches, such as pressure, vibration, heating, or a combination of these methods. Both vibration32 and pressure33 have been used to induce Cassie-Wenzel transitions on microtextured surfaces, although implementing these techniques for (33) Lafuma, A.; Quere, D. Nat. Mater. 2003, 2, 457-460.
the preparation of LDI-MS targets may be more difficult than electrowetting.
SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.
ACKNOWLEDGMENT The authors thank Mr. Timothy Maugel for his assistance with SEM images and Tom Loughran for his technical support of e-beam gold deposition. Support from NIH Grant R01GM072512 is gratefully acknowledged.
Received for review December 21, 2007. Accepted January 23, 2008. AC7026029
Analytical Chemistry, Vol. 80, No. 8, April 15, 2008