Anal. Chem. 2000, 72, 5740-5743
Tantalum-Filament Pulling Device for Fabrication of Submicrometer Fused-Silica Tips Anders Lundqvist, Johan Pihl, and Owe Orwar*
Department of Chemistry, Go¨teborg University, S-412 96 Go¨teborg, Sweden
Nanometer-scale analytical chemistry is a rapidly developing field.1 New methods and techniques are continuously developed for manipulation, separation, probing, and detection of trace-level components in ultra-small-volume elements. For many of these techniques, a narrow-tapered pipet is utilized. The ability to produce such small tips with good reproducibility is therefore a key step for achieving higher resolution and more detailed information. In near-field scanning optical microscopy (NSOM), where the tip is used as a scanning probe, optical resolutions as high as 12 nm has been achieved with 20-nm aperture tips.2 In atomic force microscopy (AFM), a nanotip mounted on a cantilever is used to record forces exerted by the sample.3 For characterizing individual cells and subcellular structures, techniques that can handle small volumes and amounts are required. Using capillary electrophoresis with a tapered injection end, it is possible to analyze the contents of single cells and organelles.4-7 Also other bioanalytical methods are developing toward single-
cell and single-organelle research. For example, improvements of the patch-clamp technique, including patch-clamp pipets made from extremely low electrical noise glass and high-resistance cell membrane seals, made it possible to measure the activities of single ion channels.8 For many of the techniques mentioned above, the probing resolution and detection sensitivity is inversely proportional to the tip size and also dependent on tip geometry. Both practical and theoretical studies have been performed to find conditions resulting in the best properties of the tapered probe.9,10 Fused silica (quartz) offers several advantages over borosilicate or aluminosilicate glasses in many applications. Quartz is stronger and will not break as easy as other glasses when, for example, penetrated through tough tissues. Thin-walled fused silica can be used to form tips that can penetrate cells as can thick-walled borosilicate, yet the thin-walled fused-silica tips have a larger pore size and therefore lower resistance. For use as intracellular voltage and ion-selective microelectrodes, quartz seems to be the best material to use.11 For applications requiring a low-noise glass, such as in single-channel patch-clamp recordings, fused silica has proven to be the most suitable material.12-15 It has a low dielectric constant and a low loss factor16 and contains none of the metals used in conventional glasses to lower their melting point. In optical applications, fused silica gives very little background fluorescence and is transparent in a broad wavelength range. Despite all these favorable properties, fused silica is not employed very often. This is because of its high softening temperature. The glass must be heated to over 1600 °C to be formed or pulled, and this is well above what most metal filaments can bare. For this reason, CO2 laser-based pullers have been developed, producing a hot welldefined spot as filament. Heating filaments based on graphite have also been developed for pulling quartz capillaries in modified commercial capillary pullers.11 In addition, to reach the glass softening temperature, the form of the heating filament is an important parameter for producing reliable tips.17 Another tech-
* Corresponding author: (phone) +46 31 772 2778; (fax) +46 31 772 2785; (e-mail)
[email protected]. (1) Gimzewski, J. K.; Joachim, C. Science 1999, 283, 1683-1688. (2) Betzig, E.; Trautman, J. K.; Harris, T. D.; Weiner, J. S.; Kostelak, R. L. Science 1991, 251, 1468-1470. (3) Hamers, R. J. J. Phys. Chem. 1996, 100, 13103-13120. (4) Gilman, S. D.; Ewing, A. G. Anal. Chem. 1995, 67, 58-64. (5) Chiu, D. T.; Lillard, S. J.; Scheller, R. H.; Zare, R. N.; Rodriguez-Cruz, S. E.; Williams, E. R.; Orwar, O.; Sandberg, M.; Lundqvist, J. A. Science 1998, 279, 1190-1193. (6) Olefirowicz, T. M.; Ewing, A. G. Anal. Chem. 1990, 62, 1872-1876. (7) Lillard, S. J.; Chiu, D. T.; Scheller, R. H.; Zare, R. N.; Rodriguez-Cruz, S. E.; Williams, E. R.; Orwar, O.; Sandberg, M.; Lundqvist, J. A. Anal. Chem. 1998, 70, 3517-3524.
(8) Hamill, O. P.; Marty, A.; Neher, E.; Sakmann, B.; Sigworth, F. J. Pflugers Arch. 1981, 391, 85-100. (9) Valaskovic, G. A.; Holton, M.; Morrison, G. H. Appl. Opt. 1995, 34, 12151228. (10) Brown, K. T.; Flaming, D. G. In Advanced micropipet techniques for cell physiology; Smith, A. D., Ed.; John Wiley & Sons: Chichester, 1986. (11) Munoz, J. L.; Coles, J. A. J. Neurosci. Methods 1987, 22, 57-64. (12) Benndorf, K. In Single-channel recording; Sakmann, B., Neher, E., Eds.; Plenum Press: New York, 1995; pp 129-145. (13) Sakmann, B.; Neher, E. In Single-channel recording; Sakmann, B., Neher, E., Eds.; Plenum Press: New York, 1995; pp 637-650. (14) Rae, J. L.; Levis, R. A. Pflugers Arch. 1992, 420, 618-620. (15) Levis, R. A.; Rae, J. L. Biophys. J. 1993, 65, 1666-1677. (16) Rae, J. L.; Levis, R. A. Methods Enzymol. 1992, 207, 66-92.
A simple and low-cost pulling device for fused-silica capillaries was developed. By using a tantalum heating filament and the self-tension in a bent capillary, tips and constricted regions with outer diameters of ∼1 µm and inner diameters of a few hundred nanometers could be reproducibly pulled from 50-µm-i.d., 375-µm-o.d. capillaries. The tips can be used in different applications such as microinjection, micromanipulation, and single-channel patch-clamp, injection ends for CE or as electrospray tips. Constricted capillaries with optimized dimensions to minimize cylindrical lensing effects and to match the size of a diffraction-limited laser focus can be used as optical detection windows in CE and micro-HPLC. Fused silica has several advantages over other glasses such as high melting temperature and superior optical and mechanical properties.
5740 Analytical Chemistry, Vol. 72, No. 22, November 15, 2000
10.1021/ac000600g CCC: $19.00
© 2000 American Chemical Society Published on Web 10/12/2000
Figure 1. Photograph of the pole plates, the filament, and one of the PTFE disks to the right. The scale bar is 7 mm.
nique to produce tips with desired geometries is by using hydrofluoric acid etching.18,19 Etched tips have been used in many applications such as injection ends for CE,4,6 intracellular microelectrodes,11 NSOM tips,18 and integrated decouplers for CE with electrochemical detection.20 The tip geometries are hard to control when etching the capillary and the hazardous HF requires extremely careful handling. Often the tips are jagged and uneven, which can puncture and damage biological membranes, e.g., during injection. It is also difficult to control the tip inner diameter, even with a flow through the capillary of organic solvents or helium as protection. To produce smoother and rounder tips from fused silica, the capillary can be pulled by hand in a butane flame.21 This process is hard to make reproducible, and it is extremely difficult to control the tip size and geometry. Often the success rate of the pulling process is below 50% and every pulled tip must be investigated in a microscope before use. In this paper, we present a reliable and simple low-cost pulling device as an alternative to expensive laser-based pullers. With this device, submicrometer tips and constricted detection areas can be pulled from commercial fused silica capillaries. EXPERIMENTAL SECTION Materials. Fused-silica capillaries with 50 µm inner diameter and outer diameters of 192 and 375 µm were purchased from Polymicro Technologies LLC (Phoenix, AZ). Quartz capillaries with 1.0 mm outer and 0.70 mm inner diameters were from Sutter Instrument Co. (Novato, CA), and borosilicate capillaries (1.0 and 0.58 mm outer and inner diameters, respectively) were from Clark Electromedical Instruments (Pangbourne, England). Tantalum and tungsten wires, 0.25 and 0.3 mm diameters, respectively, 99.9% purity, annealed were both from Goodfellow Cambridge Lt. (Cambridge, England). Nitrogen gas (Nitrogen Plus) was from (17) Flaming, D. G.; Brown, K. T. J. Neurosci. Methods 1982, 6, 91-102. (18) Sayah, A.; Philipona, C.; Lambelet, P.; Pfeffer, M.; Marquis-Weible, F. Ultramicroscopy 1998, 71, 59-63. (19) Muheim, M. H. Pflugers Arch. 1977, 372, 101-102. (20) Qian, J.; Wu, Y.; Yang, H.; Michael, A. C. Anal. Chem. 1999, 71, 44864492. (21) Chiu, D. T.; Hsiao, A.; Gaggar, A.; Garza-Lopez, R. A.; Orwar, O.; Zare, R. N. Anal. Chem. 1997, 69, 1801-1807.
Figure 2. Schematic front view of the capillary puller device and expanded view of a capillary positioned close to the filament: (1) fused-silica capillary, (2) capillary holder with mounting nuts, (3) pole plates with the filament placed in between, (4) PTFE disks over which a capillary is stretched, (5) adjustable capillary tensioner, (6) plastic plate around which the tensioner is sliding, (7) fixing nuts for capillary, and (8) tantalum filament.
AGA Gas AB (Sundbyberg, Sweden). All chemicals were purchased from Sigma-Aldrich (St. Louis, MO). The power supplies used were from Alpha Elettronica S.n.c. (model AL 370/S; Parma, Italy) and from Einhell AG (model Chargemaster 20A; Landau, Germany). The tips and the constricted capillaries were filled with the desired liquid by a HPLC pump (PU 980, Jasco Inc., Easton, MD) with a 300 × 4 mm, packed with 5 µm C-18 particles column (Mikropak MCH 10) as a pressure holder and filter. The filament temperature was measured with an optical pyrometer from Chino Pyrostar (model IR-U), and the pulling forces were estimated from calculations by replacing the self-tension with different weights. RESULTS AND DISCUSSION The high softening temperature of quartz (1665 °C) excludes a lot of metals as filament. Platinum has a melting temperature of 1772 °C, which is too low to be used for melting quartz. Tungsten is a metal with a high melting point (3410 °C), but it oxidizes very easily at high temperatures and becomes very brittle. We observed a destructive sputtering from W filaments at high temperatures, and both the capillaries and the W filaments were covered with a yellowish dust after usage. This led us to discontinue the use of such filaments. Tungsten also has a low electrical resistivity (5.4 µΩ‚×cm), which requires a more powerful power supply. Tantalum has almost as high melting temperature as tungsten (2996 °C) but has a higher electrical resistivity (13.5 µΩ‚×cm). Favorably, no sputtering behavior was observed with tantalum filaments at high temperatures. Therefore, the feasibility of using filaments made from this metal for pulling quartz capillaries was investigated. To form a filament, a piece of 0.25 mm diameter Ta wire was folded once and wound hard together with the help of a rotating Analytical Chemistry, Vol. 72, No. 22, November 15, 2000
5741
Figure 4. Instrumental setup for CE-LIF analysis: (1) argon ion laser, (2) 488 nm band-pass filter and lenses, (3) dichroic mirror, (4) pinhole (50 µm), (5) single photon avalanche diode (SPAD), (6) oil immersion objective (100× NA ) 1.3), (7) capillary with constricted detection window, (8) personal computer with multichannel scaler, and (9) lens and fluorescence filter. Open arrows are excitation light, and filled arrows are fluorescence light. The scale in the box is not the same as in the rest of the drawing.
Figure 3. Electron micrographs of two fused-silica capillary tips. The tips were pulled with a double tantalum filament (outer and inner diameter of 2.7 and 1.1 mm, respectively) from capillaries with 375 µm outer diameter and 50 µm inner diameter. The power applied for heating the filament was 132 W. The scale bar is 1 µm.
lathe (1500 rpm) to a helix with about 20-25 turns/cm. The wire helix was cut into pieces of ∼6.5 cm and wound 10-12 turns around a glass tube (outer diameter 1 mm) that then was removed. The filament was finally bent to a U-shape and mounted between two pole plates (Figure 1). For making double filaments, two twinned wires were wound together with a lathe to a wire thickness of 0.8 mm. Otherwise the procedure used was the same as for thin filaments. The double filaments produced by this method have outer and inner diameters of 2.7 and 1.1 mm, respectively. The distance between the arms of the U-shaped filaments was between 1.7 and 2.2 mm. The pulling device was placed in an airtight Plexiglas box (9 L) with a gas inlet at one short end and a sealable gas outlet in the other short end. The box was flushed with N2 gas for a few minutes to remove as much oxygen as possible before the outlet was closed and the pulling process started. The airtight Plexiglas box made it possible to create an inert atmosphere with a gas that prolongs the filament lifetime and helps to keep the humidity at constant level to increase the reproducibility of the pulling process. The box also keeps the air turbulence near zero, which prevents uneven heat dissipation, which otherwise can lead to uneven and bent tips. To facilitate the handling, in particular of the fragile constricted detection windows, the capillary was inserted in a plastic holder and vertically mounted with the capillary as close as possible to the filament. The plastic holder can be modified according to the requirements needed for a specific experiment. The capillary was 5742 Analytical Chemistry, Vol. 72, No. 22, November 15, 2000
firmly stretched at 90° angle around two PTFE disks (10-mm diameter), positioned just above and below the plastic holder, and secured in both ends (Figure 2). A power supply variable between 6 and 12 V delivering up to 24 A was connected to the filament poles, and when the fused silica reached its softening temperature, the forces from the 90° stretch pulled the capillary in a fast process. Thin filaments reached a temperature of 1850 °C, and thick filaments reached a temperature of 2150 °C, which is enough for the applied force of ∼80 mN to pull the capillary. It takes ∼5 s from when the power is applied until the capillary is pulled. Figure 3 shows electron micrographs of ∼1µm outer and 200 nm inner diameter quartz capillary tips produced by the prototype puller device. Tips were produced with good accuracy and reproducibility, both with regard to tip size and tip geometry from 192 and 375 µm outer diameter capillaries. The produced tips are smooth and do not need to be polished in a flame afterward. The tip size can be manipulated by regulating the stretching of the capillary or the starting diameter. The lifetime of the filament was strongly dependent on the atmosphere in the Plexiglas box. The best result was achieved when the box was flushed with N2 a few minutes before starting the pulling process. Without the inert atmosphere, filaments lasted for only one or two pulls, but with the N2 present, ∼25 pulls could be made. The thicker filaments had longer lifetime and improved physical properties compared to the thinner filaments. Capillaries were also pulled in a one-step procedure to produce constricted optical detection channels. Ideally, the detection windows are made in the same size as the diffraction-limited laser focus, which forces all molecules present in the sample to pass the probe volume and to be registered. The narrow detection channel also reduces the Brownian motion of analyte molecules and thereby increases the observation time.22 Furthermore, the (22) Lyon, W. A.; Nie, S. Anal. Chem 1997, 69, 3400-3405.
in CE with laser-induced fluorescence for detection of analytes at ultralow concentrations (Figure 4). As an example we show in Figure 5 an electrophoretic separation of two dyes, fluorescein and 5-carboxyfluorescein. The detection channel has a fairly large inner diameter (12 µm) compared with the laser focus, but the o.d./i.d. ratio is almost optimal, resulting in a concentration limit of detection at low-picomolar levels for these two dyes. Using the thicker filament it was also possible to pull capillaries up to 1.0 mm outer and 0.70 mm inner diameter with good results (data not shown). The puller design was slightly modified to pull the thick-walled capillaries in order to obtain a sufficient pulling force. A power supply delivering more power was used to create enough heat for the thick wall to soften. The puller was also tried on ordinary borosilicate capillaries with good results. The melting temperature for borosilicate glass is lower (710-830 °C, depending on manufacturer) than for quartz so the pulling procedure did not require so much power and was faster than for quartz capillaries. Figure 5. Separation of fluorescein and 5-carboxyfluorescein by CE-LIF with a constricted detection window. Capillary length, 50 cm; i.d./o.d., 50/192 µm with detection window pulled to 12/60 µm; applied voltage, 17.5 kV. Injection was made by 8.5-cm siphoning for 10 s. The first peak is fluorescein and the second peak is 5-carboxyfluorescein (300 pM each). A 30 mM borate buffer (pH 9.3) was used as separation buffer. The signal dips in the electropherogram are due to closing the input to the detector when the laser focus is checked. Only the latest part of the run is shown.
ratio between the inner and outer diameter of the detection window has a very high impact on the signal-to-noise ratio, and the right geometry can minimize the backscattering light and maintain the shape of the laser focus.23,24 The diameter of the constricted detection channel can be varied by changing the stretch over the PTFE disks and the amount of heat applied. We have pulled capillaries with detection channels ranging from 12 to 0.5 µm inner diameter. The detection channels have been used (23) Enderlein, J. Opt. Commun. 1999, 160, 201-206. (24) Maystre, F.; Bruno, A. E. Anal. Chem. 1992, 64, 2885-2887.
CONCLUSIONS We demonstrate here a simple and inexpensive device for pulling submicrometer tips and constricted detection channels from commercial fused-silica or quartz capillaries. The heating source is a tantalum filament and the pulling process, which is performed in inert atmosphere, is due to the self-tension in a bent capillary. ACKNOWLEDGMENT We thank Dr. Anders Kvist at the Department of Experimental Physics, Go¨teborg University, for the electron micrographs. The work was supported by the Swedish Research Council for Engineering Sciences (TFR) and the Swedish Foundation for Strategic Research (SSF). Received for review May 23, 2000. Accepted August 31, 2000. AC000600G
Analytical Chemistry, Vol. 72, No. 22, November 15, 2000
5743
