FIELD OF THE INVENTION
The invention relates generally to mechanical probe tips such as those used in atomic force microscopy. In particular, the invention relates to a carbon nanotube grown directly on a pointed end of a probe.
Atomic force microscopes (AFMs) have been recently developed for mechanically profiling small features, for example, determining critical dimensions (CDs) of via holes in semiconductor integrated circuits. Such holes have depths of about 1 μm and widths which are being pushed to 180 nm and below. For detailed measurement of the feature, an exceedingly fine probe tip is disposed on the end of a cantilever overlying the feature. In the pixel mode of operation, the probe tip is successively positioned at points on a line above and traversing the feature being probed. The cantilever lowers the probe tip until it encounters the surface, and both the horizontal position and the vertical position at which the probe meets the surface are recorded. A series of such measurements provide the desired microscopic profile. An example of such an atomic force microscope is the Stylus Nanoprobe SNP available from Surface/Interface, Inc. of Sunnyvale, Calif. It employs technology similar to the rocking balanced beam probe disclosed by Griffith et al. in U.S. Pat. No. 5,307,693 and by Bryson et al. in U.S. Pat. No. 5,756,887.
Such a tool is schematically illustrated in the side view of FIG. 1. A few more details are found in U.S. patent application Ser. No. 09/354,528, filed Jul. 15, 1999 and incorporated herein by reference in its entirety. A wafer 10 or other sample to be is supported on a support surface 12 supported successively on a tilt stage 14, an x-slide 16, and a y-slide 18, all of which are movable along their respective axes so as to provide horizontal two-dimensional and tilt control of the wafer 10. Although these mechanical stages provide a relatively great range of motion, their resolutions are relatively coarse compared to the resolution sought in the probing. The bottom y-slide 18 rests on a heavy granite slab 20 providing vibrational stability. A gantry 22 is supported on the granite slab 20. A probe head 24 hangs in the vertical z-direction from the gantry 22 through an intermediate piezoelectric actuator 26 providing about 10 μm of motion in (x, y, z) by voltages applied across electrodes attached to the walls of a piezoelectric tube. A probe assembly with a tiny attached probe tip 28 projects downwardly from the probe head 24 to selectively engage the probe tip 28 with the top surface of the wafer 10 and to thereby determine its vertical and horizontal dimensions.
Principal parts of the probe head 24 of FIG. 2 are illustrated in the side view of FIG. 2. A dielectric support 30 fixed to the bottom of the piezoelectric actuator 26 includes on its top side, with respect to the view of FIG. 1, a magnet 32. On the bottom of the dielectric support 30 are deposited two isolated capacitor plates 34, 36 and two interconnected contact pads 38.
A beam 40 is medially fixed on its two lateral sides and is also electrically connected to two metallic and ferromagnetic ball bearings 42. The beam 40 is preferably composed of heavily doped silicon so as to be electrically conductive, and a thin silver layer is deposited on it to make good electrical contacts to the ball bearings. The two ball bearings 42 are placed on respective ones of the two contact pads 38 and generally between the capacitor plates 34, 36, and the magnet 32 holds the ferromagnetic bearings 42 and the attached beam 40 to the dielectric support 30. The attached beam 40 is held in a position generally parallel to the dielectric support 40 with a balanced vertical gap 46 of about 25 μm between the capacitor plates 34, 36 and the beam 40. Unbalancing of the vertical gap allows a rocking motion of about 25 μm. The beam 40 holds on its distal end a glass tab 48 to which is fixed a stylus 50 having the probe tip 52 projecting downwardly to selectively engage the top of the wafer 10 being probed.
Two capacitors are formed between the respective capacitor plates 34, 36 and the conductive beam 40. The capacitor plates 34, 36 and the two contact pads 38, commonly electrically connected to the conductive beam 40, are separately connected by three unillustrated electrical lines to three terminals of external measurement and control circuitry This servo system both measures the two capacitances and applies differential voltage to the two capacitor plates 34, 36 to keep them in the balanced position. When the piezoelectric actuator 26 lowers the stylus 50 to the point that it encounters the feature being probed, the beam 40 rocks upon contact of the probe tip 52 with the wafer 10. The difference in capacitance between the plates 34, 36 is detected, and the servo circuit attempts to rebalance the beam 40 by applying different voltages across the two capacitors, which amounts to a net force that the stylus 50 is applying to the wafer 10. When the force exceeds a threshold, the vertical position of the piezoelectric actuator 26 is used as an indication of the depth or height of the feature.
This and other types of AFMs have control and sensing elements more than adequate for the degree of precision for profiling a 1180 nm×1 μm hole. However, the probe tip presents a challenge for profiling the highly anisotropic holes desired in semiconductor fabrication as well as for other uses such as measuring DNA strands and the like. The probe tip needs to be long, narrow, and stiff. Its length needs to at least equal the depth of the hole being probed, and its width throughout this length needs to be less than the width of the hole. A fairly stiff probe tip reduces the biasing introduced by probe tips being deflected by a sloping surface.
One popular type of probe tip is a shaped silica tip, such as disclosed by Marchman in U.S. Pat. Nos. 5,395,741 and 5,480,049 and by Filas and Marchman in U.S. Pat. No. 5,703,979. A thin silica fiber has its end projecting downwardly into an etching solution. The etching forms a tapered portion near the surface of the fiber, and, with careful timing, the deeper portion of the fiber is etched to a cylinder of a much smaller diameter. The tip manufacturing is relatively straightforward, and the larger fiber away from the tip provides good mechanical support for the small tip. However, it is difficult to obtain the more desirable cylindrical probe tip by the progressive etching method rather than the tapered portion alone. Furthermore, silica is relatively soft so that its lifetime is limited because it is continually being forced against a relatively hard substrate.
One promising technology for AFM probe tips involves carbon nanotubes which can be made to spontaneously grow normal to a surface of an insulator such as glass covered with a thin layer of a catalyzing metal such as nickel. Carbon nanotubes can be grown to diameters ranging down to 5 to 20 nm and with lengths of significantly more than 1 μm. Nanotubes can form as single-wall nanotubes or as multiple-wall nanotubes. A single wall is an cylindrically shaped atomically thin sheet of carbon atoms arranged in an hexagonal crystalline structure with a graphitic type of bonding. Multiple walls bond to each other with a tetrahedral bonding structure, which is exceedingly robust. The modulus of elasticity for carbon nanotubes is significantly greater than that for silica. Thus, nanotubes offer a very stiff and very narrow probe tip well suited for atomic force. microscopy. Furthermore, carbon nanotubes are electrically conductive so that they are well suited for scanning tunneling microscopy and other forms of probing relying upon passing a current through the probe tip. Dai et al. describe the manual fabrication of a nanotube probe tip in “Nanotubes as nanoprobes in scanning probe microscopy,” Nature, vol. 384, 14 November 1996, pp. 147-150.
Typically, nanotubes suffer from the disadvantage that a large number of them simultaneously form on a surface producing either a tangle or a forest of such tubes, as is clearly illustrated by Ren et al. in “Synthesis of large arrays of well-aligned carbon nanotubes on glass,” Science, vol. 282, 6 November 1998, pp. 1105-1107. The task then remains to affix one nanotube to a somewhat small probe tip support. Dai et al. disclose an assembly method in which they coat the apex of a silicon pyramid at the probe end with adhesive. The coated silicon tip was then brushed against a bundle of nanotubes, and a single nanotube can be pulled from the bundle. This method is nonetheless considered expensive and tedious requiring both optical and electron microscopes. Additionally, there is little control over the final orientation of the nanotube, certainly not to the precision needed to analyze semiconductor features. Cheung et al. describe another method of growing and transferring nanotubes in “Growth and fabrication with single-walled carbon nanotube probe microscopy tips,” Applied Physics Letters, vol. 76, no. 21, 22 May 2000, pp. 3136-3138. However, they either produce poor directional control with a very narrow, single nanotube or require a complex transfer mechanism with nanotube bundles.
Ren et al. describe a method of growing isolated nanotubes in “Growth of a single freestanding multiwall carbon nanotube on each nanonickel dot,” Applied Physics Letters, vol. 75, no. 8, 23 August 1999, pp. 1086-1088. They deposit 15 nm of nickel on silicon and pattern it into a grid of nickel dots having sizes of somewhat more than 100 nm. Plasma-enhanced chemical vapor deposition using acetylene and ammonia produces a single nanotube on each dot having an obelisk shape with a base diameter of about 150 nm and a sharpened tip. However, Ren et al. do not address the difficult problem of transferring such a nanotube, which they describe as being tightly bonded to the nickel, from the nickel-plated substrate to a probe end.
Cheung et al. disclose another method of growing isolated nanotubes in “Carbon nanotube atomic force microscopy tips: Direct growth by chemical vapor deposition and appliaction to high-resolution imaging,” Proceedings of the National Academy of Sciences, vol. 97, no. 8, 11 April 2000, pp. 3809-3813. They etch aniostropic holes in a silicon tip and deposit the catalyzing iron or iron oxide in the bottom of the holes. The carbon nanotubes grow out of the holes. However, growth in such restricted geometries is considered to be disadvantageous and to favor single-wall rather than multiple-wall nanotubes. Further, this method provides only limited control over the number and size of the nanotubes being grown.
Accordingly, a more efficient method is desired for forming a probe tip having a single carbon nanotube. Furthermore, the structure of the probe end and probe tip should facilitate assembly of the probe and contribute to its robustness.
SUMMARY OF THE INVENTION
A probe end is shaped to have sloping sides and a generally flat end, that is, in the shape of sloping mesa. The diameter of the mesa top is preferably in the range of 20 to 300 nm. Nickel or other material that catalyzes the growth of carbon nanotubes is directionally deposited onto the probe end. Because of the geometry, the thickness of the deposited nickel, as measured from the underlying surface, is greater on the mesa top than on the mesa sides. The nickel is then isotropically etched for a time sufficient to remove the nickel from the mesa sides but to leave sufficient nickel on the mesa top to catalyze the growth of a single carbon nanotube. Typically, the nanotube grows with a bottom diameter approximately equal to that of the nickel dot on top of the mesa.
The milling produces a shaped tip 64′, illustrated in the cross-sectional view of FIG. 5, having a flat end 70 and sloping sidewalls 72. Then, as illustrated in the cross-sectional view of FIG. 6, a film 76 of nickel or other catalyzing metal is then directionally deposited onto the probe tip 64′, preferably by sputtering metal atoms along the longitudinal axis of the shaped tip 64′. The thickness of the deposition, as measured along the longitudinal axis, is substantially constant between the area of the flat end 70 and the sloping sidewalls 72 of the shaped tip 64′. However, the thickness, as measured at a perpendicular to the underlying surface, is substantially thicker in the area overlying the flat 70 than in the areas overlying the sloping sidewalls 72. The effect is primarily geometric. If the probe tip has a tip angle 2θ and the deposition is totally aniostropic, then the sidewall thickness is sinθ times the end thickness. For example, if 2θ=31.3°, then the sidewall thickness is 27% of the end thickness. The sputtering may be performed in an ion sputtering system using a nickel target. Such a system is the Model 681 High Resolution Ion Coater from Gatan of Pleasanton, Calif. Other types of deposition are possible, such as molecular beam techniques usually associated with molecular beam epitaxy.