Anal. Chem. 2005, 77, 8170-8173
Electrospray from Nanostructured Tungsten Oxide Surfaces with Ultralow Sample Volume Jingyueh Jeng,† Che-Hsin Lin,‡ and Jentaie Shiea*,§
Department of Biotechnology, Chia-Nan University of Pharmacy and Science, Tainan, Taiwan, and Department of Mechanical and Electro-Mechanical Engineering, and Department of Chemistry, National Sun Yat-Sen University, Kaohsiung, Taiwan
This study demonstrated the feasibility of performing protein analysis with ultralow sample volume by combining a tungsten oxide nanowire (TON) fiber with a miniaturized electrospray ionization interface. An increase in wettability of the tugsten surface after growing randomly oriented TON on its surface allows strong adhesion of ∼50 nL of the methanol solution at its tip. Under the influence of a high electric field, electrospray from a Taylor cone on the adhered methanol solution was observed and the multiply charged ions of protein molecules predissolved in the solution were detected. The synthesis of tungsten oxide nanowires (TON) on a tungsten wire can be achieved by simply heating the tungsten wire under an Ar atmosphere.1-3 The TON will be grown on the surface of the wire through the reaction of tungsten with traces of oxygen; no external catalyst or solvent is required. Because of its large surface area and high conductivity, a tungsten fiber with TON on its surface may be used to increase the lateral resolution of scanning tunneling microscopy or behave as a conductive carrier in nanosensors.4-6 Studies on the influence that a high electric field has on the liquid attached on the TON of a tungsten fiber remain rare, however; and no reports describe applying such material to chemical or biochemical analyses. In this study, we describe the generation of electrospray from traces of methanol solution on the TON and the applications of such a tungstenTON fiber to protein analysis by mass spectrometry. EXPERIMENTAL SECTION The preparation of the TON on a tungsten fiber tip followed the procedures described by Gang Gu et al.1 The surface of a * To whom correspondence should be addressed. E-mail: jetea@ mail.nsysu.edu.tw. † Chia-Nan University of Pharmacy and Science. ‡ Department of Mechanical and Electro-Mechanical Engineering, National Sun Yat-Sen University. § Department of Chemistry, National Sun Yat-Sen University. (1) Gu, G.; Zheng, B.; Han, W. Q.; Roth, S.; Liu, J. Nano Lett. 2002, 2, 849851. (2) Zhu, Y. Q.; Hu, W.; Hsu, W. K.; Terrones, M.; Grobert, N.; Hare, J. P.; Kroto, H. W.; Walton, D. R. M.; Terrones, H. Chem. Phys. Lett. 1999, 309, 327-334. (3) Liu, Z. W.; Bando, Y.; Tang, C. C. Chem. Phys. Lett. 2003, 372, 179-182. (4) Chen, J. C. Introduction to Scanning Tunneling Microscopy; Oxford University: Oxford, 1993. (5) Hembacher, S.; Giessibl, F. J.; Mannhart, J. Science 2004, 305, 380-383. (6) Cui, Y.; Wei, Q.; Park, H.; Lieber, C. M. Science 2001, 293, 1289-1292.
8170 Analytical Chemistry, Vol. 77, No. 24, December 15, 2005
tungsten fiber (250-µm diameter) was cleaned using an electrochemical etching process by dipping the fiber into 0.5 N KOH solution and then applying an ac voltage (∼3-10 V) between the fiber and a platinum circle electrode. Subsequently, the fiber was placed into a tube furnace and was heated at 700 °C under an Ar atmosphere (Ar flow rate, 300 mL/min) for 30 min. The fiber was then cooled to room temperature before examining its surface composition using a scanning electron microscope. We found that, after this simple heat treatment process, the surface of the tungsten tip was modified with numerous TON (Figure 1a). The average length of the nanowires was ∼500 nm. The diameters of the nanowires ranged from 30 to 150 nm and remained uniform along their entire length. No tungsten nanoparticles were observed. The investigations into the hydrophilic properties (or wettability) of the tungsten fiber surface were conducted by monitoring their contact angles (i.e., the angle between the water/air and water/solid interfaces at the three-phase boundary). Measurements of contact angle were performed in a sealed glass chamber at 100% humidity by using a microsyringe needle to place 0.5 µL of pure water onto the tungsten tip. An image of the drop was recorded a few seconds after droplet deposition. Each individual experiment was repeated three times to determine the average contact angles of three tungsten fibers. Application of trace sample solution (∼50 nL) to the tungsten fiber tip was simply done by dipping and removing the fiber from the sample solution. To apply more sample solution onto the tungsten fiber surface, 0.2 µL of the sample solution was deposited at the fiber tip by a pipet. The tungsten fiber was then placed ∼8 mm away from the ion-orifice inlet of the mass analyzer for ion detection. Mass spectrometric analysis was performed using a Bruker Bio-Tof Q mass spectrometer (Bruker Daltonics, Billerica, MA). The mass analyzer was scanned from m/z 200 to 2000 at a rate of ∼1 s/scan. RESULTS AND DISCUSSION The results of contact angle measurements showed, without modification, the contact angle of a tungsten fiber was 16.6° (Figure 2a). This indicates that the surface of the tungsten fiber is effectively hydrophilic (cf. the contact angles of most hydrophobic metal surfaces, usually 150°). After surface modification with the TON, the contact angle of the fiber decreased to 6.1° (Figure 2b), which indicates that the surface of the tungstenTON fiber became extremely hydrophilic. We found that traces 10.1021/ac0512960 CCC: $30.25
© 2005 American Chemical Society Published on Web 10/27/2005
Figure 1. Scanning electron microscopy images of tungsten-TON surfaces (a) before and (b) after electrospraying.
Figure 2. Water droplets deposited onto the surfaces of (a) a tungsten fiber, (b) a tungsten-TON fiber, and (c) a tungsten-TON fiber after electrospray. The contact angles of the respective tungsten surfaces are provided at the upper corner of each figure.
of polar solvents, such as water or methanol, strongly adhered onto this hydrophilic surface. Removing the TON from the surface under the influence of a strong electric field caused the contact angle to change to 21.2° (Figure 2c). In general, the wetting of a solid surface is controlled by the surface free energy and the geometrical microstructure of the surface.7-10 The observation of increasing wet behavior of the tungsten-TON surface can be understood in terms of the oxidation of tungsten and the roughness of the nanowires on the tungsten tip surface. Apparently, the oxidation of tungsten on the surface increases the surface free energy of the material and (7) Jennissen, H. P. Macromol. Symp. 2005, 225, 43-69. (8) Chen, W.; Fadeev, A. Y.; Hsieh, M. C.; Oner, D.; Youngblood, J.; McCarthy, T. J. Langmuir 1999, 15, 3395-3399. (9) Gau, H.; Herminghaus, S.; Lenz, P.; Lipowsky, R. Science 1999, 283, 4649. (10) Mittal, K. L. Contact angle, wettability and adhesion; VSP: Utrecht, The Netherlands, 1993.
increases its hydrophilicity.11,12 The presence of numerous randomly oriented TON on the surface of the tungsten fiber also increases the roughness and hydrophilicity of the fiber surface. This phenomenon contrasts the situation found on a well-oriented nanostructured surface (e.g., gold), where a completely hydrophobic surface having self-cleaning capacity is usually encountered.11,12 The ability to collect ultratrace quantities of samples (e.g., single cells and traces of explosives) for adequate chemical analyses has long been a challenging task for analytical chemists.13-17 Miniaturization of analytical instruments seems to be a key step toward the analysis of ultratrace quantities of samples.18-21 Nevertheless, a difficulty in developing miniaturized devices for sample preparation and detection is the low volume requirement associated with these devices, which creates potential problems when transferring a sample from one device to another. In this regard, there is increased demand for new kinds of sample convection microsystems that can fit into micro- or nanoscaled domains. Microfluidic devices, such as nanosprays, combined with mass spectrometry have been developed to perform highly sensitive detection of ultralow volumes of samples because of the short diffusion distances and low velocities per volume flow rate.22-25 There are practical limits, however, to the (11) Furstner, R.; Barthlott, W.; Neinhuis, C.; Walzel, P. Langmuir 2005, 21, 956-961. (12) Abdelsalam, M. E.; Bartlett, P. N.; Kelf, T.; Baumberg, J. Langmuir 2005, 21, 1753-1757. (13) Gilman, S. D.; Ewing, A. G. Anal. Chem. 1995, 67, 58-64. (14) Frohlich, J.; Konig, H. FEMS Microbiol. Rev. 2000, 24, 567-572. (15) Rose, A.; Zhu, Z. G.; Madigan, C. F.; Swager, T. M.; Bulovic, V. Nature 2005, 434, 876-879. (16) Sun, B.; Chiu, D. T. J. Am. Chem. Soc. 2003, 125, 3702-3703. (17) Munce, N. R.; Li, J.; Herman, P. R.; Lilge, L. Anal. Chem. 2004, 76, 49834989. (18) Lion, N.; Reymond, F.; Girault, H. H.; Rossier, J. S. Curr. Opin. Biotechnol. 2004, 15, 31-37. (19) McGlennen, R. C. Clin. Chem. 2001, 47, 393-402. (20) Roper, M. G.; Shackman, J. G.; Dahlgren, G. M.; Kennedy, R. T. Anal. Chem. 2003, 75, 4711-4717. (21) Wachs, T.; Henion, J. Anal. Chem. 2001, 73, 632-638. (22) Yue, G. E.; Roper, M. G.; Jeffery, E. D.; Easley, C. J.; Balchunas, C.; Landers, J. P.; Ferrance, J. P. Lab Chip 2005, 5, 619-627. (23) Krenkova, J.; Foret, F. Electrophoresis 2004, 25, 3550-3563. (24) Koerner, T.; Turck, K.; Brown, L.; Oleschuk, R. D. Anal. Chem. 2004, 76, 6456-6460. (25) Schilling, M.; Nigge, W.; Rudzinski, A.; Neyer, A.; Hergenroder, R. Lab Chip 2004, 4, 220-224.
Analytical Chemistry, Vol. 77, No. 24, December 15, 2005
8171
Figure 3. Low-magnification photographs of trace amounts of methanol solutions deposited onto a tungsten fiber under the influence of a high electric field: (a) 0, (b) 1500, and (c) 3000 V. (d) The total ion currents of five repeated electrospray analyses (∼12 s for each analysis) from trace amounts of methanol solution predeposited onto the surfaces of the tungsten-TON fibers.
trend toward smaller channels; flow cells having channels with volumes in the low-microliter to nanoliter range are costly to fabricate and prone to clogging. Also, an unstable electrospray is usually obtained when spraying with water-rich solutions. Previous research results have shown that increasing hydrophobicity of the hollow capillary surface by chemical modification successfully minimized the spread of liquid over the capillary surface during the electrospray.26 However, in this study, hydrophilicity of the tungsten fiber surface is increased to make the fiber surface highly adhesive to liquid, which is then used to initiate the electrospray process. In a previous study, we demonstrated that a stable electrospray could be generated from a solution placed on a small copper ring that was subjected to a high voltage.27 For a typical analysis, a few microliters of sample solution must be deposited by pipet onto the ring; nevertheless, the sample droplet flew away (taking ∼90% of the sample volume away) before the critical electrospray charge state (i.e., the Rayleigh limit) was reached.28 Modification of a glass rod surface with Nafion increases its wettability; however, proteinNafion adducts might be also generated during analysis.29 The increased wettability and the inert nature of the tungsten surface bearing the TON suggests that a strong retention of trace amounts of water or polar organic solvents on the tungsten-TON tip may allow further chemical or biological analyses. Figure 3a presents a low-magnification photograph of a trace amount of methanol deposited on the surface of a tungsten-TON fiber. The volume of the methanol attached on the surface of the fiber was ∼50 nL, as calculated by subtracting the volume of the tungsten cylinder from the volume of ellipsoid methanol solution in the image. This tungsten-TON tipsupon which was adhered the trace amount of methanol solutionswas then placed in front of the (26) Tojo, H. J. Chromatogr., A 2004, 1056, 223-228. (27) Hong, C. M.; Lee, C. T.; Lee, Y. M.; Kuo, C. P.; Yuan, C. H.; Shiea, J. Rapid Commun. Mass Spectrom. 1999, 13, 21-25. (28) Kuo, C. P.; Yuan, C. H.; Shiea, J. J. Am. Soc. Mass Spectrom. 2000, 11, 464-467. (29) Jeng, J.; Shiea, J. Rapid Commun. Mass Spectrom. 2003, 17, 1709-1713.
8172
Analytical Chemistry, Vol. 77, No. 24, December 15, 2005
orifice inlet of a quadrupole time-of-flight mass analyzer. The voltage between the tungsten-TON tip (at ground all the time) and the ion-sampling orifice of the Q-TOF mass analyzer was varied (from 0 to 3 kV) to examine the effects of applying an electric field to the methanol solution. At 1500 V, in the absence of any external pressure, the surface of the solution became charged and moved toward the direction of the ion entrance; the solution accumulated at the tungsten-TON tip through electrohydrodynamic effects (Figure 3b). Subsequently, when the applied voltage reached the electrospray onset voltage (3000 V), electrospray of the methanol solution from a Taylor cone was observed (Figure 3c). The induced electrical shear stress accrued on the surface of the liquid, which behaved akin to the sheath gas that is used at a conventional ESI source. The electrospray lasted for ∼12 s for each sample application (Figure 3d). For each tungstenTON fiber, the electrospray was reproducible for more than five times, but after that, we found that nearly all of the TON on the tungsten surface had been removed through physical shear forces (Figure 1b) and subsequent adhesion of the sample solution onto the tungsten tip became difficult. Since the surface area of the tungsten fiber is extremely small (and also the amount of TON on the fiber tip), the removal of trace TON during electrospray does not lead to the deposition of TON on the MS interface, and therefore, clogging over time became a serious concern. We further examined the performance of the electrospray from the tungsten-TON tip by applying methanol solutions containing angiotensin I and bovine ubiquitin (1 µM each) for their respective analysis. Figure 4 displays the mass spectra of angiotensin I (Figure 4a) and bovine ubiquitin (Figure 4b). Each analysis required ∼50 nL of the sample solution and consumed a total of 50 fmol of the analyte. The ion patterns observed in the resulting mass spectra of angiotensin I and ubiquitin are nearly identical to those obtained using conventional ESI-MS. Due to the inert nature of tungsten and low ionization efficiency of TON in ESI, no tungsten-protein adduct ions or TON ion signals were observed even after multiple sample analyses. The lowest detec-
the electrospraying time (to ∼25 s) to obtain the MS/MS spectrum. From a Mascot database search, the MS/MS sequence coverage is 100%. It matched a predicted peptide, DRVYIHPF, that is unique to human angiotensin II.
Figure 4. Positive-mode ESI mass spectra of (a) angiotensin I (10-6 M in methanol) and (b) bovine ubiquitin (10-6 M in methanol) obtained using a tungsten-TON fiber as the electrospray emitter. (c) The MS/ MS spectrum of the molecular ion of human angiotensin II (10-6 M in methanol).
10-7
tion limit occurred when using a M angiotensin I solution (i.e., 5 fmol) as the sample for analysis. This detection limit is comparable to conventional ESI-MS. This suggests that absorbing of the proteins to the fiber surface due to an increase of hydrophilicity of the fiber surface is not considered a serious problem. Figure 4c displays the MS/MS spectrum for the fragmentation of the chosen parent ion of human angiotensin II (m/z 1046). In this analysis, the tungsten-TON tip was bedewed with a greater amount of sample solution (∼150 nL) to elongate
CONCLUSION A novel ESI emitter by modifying a tungsten fiber with randomly oriented TON on its surface was developed. The increase in wettability of the tungsten-TON fiber allows the adhesion of an ultralow volume of methanol (in the nanoliter regime) onto the fiber surface. Under the influence of a strong electric field, an electrospray was generated from the adhesive solution and multiply charged ions from protein molecules predissolved in the methanol were detected. The manufacture of the tungsten-TON is extremely simple and low cost; therefore, the preparation of disposable ESI emitters is possible. Because no capillary is used for sample delivery, clogging problems are minimized. Sample application was simply done by dipping the fiber into the solution. ACKNOWLEDGMENT We thank the National Science Council and the University Integration Program (for Cheng Kung and Sun Yat-Sen University) of the Ministry of Education, Taiwan, for supporting this research financially. Received for review July 21, 2005. Accepted September 30, 2005. AC0512960
Analytical Chemistry, Vol. 77, No. 24, December 15, 2005
8173
