CA2054334C

CA2054334C - Near field scanning optical microscope and applications thereof

Info

Publication number: CA2054334C
Application number: CA002054334A
Authority: CA
Inventors: Robert Eric Betzig; Ernst Michael Gyorgy; Frances Hellman; Jay Kenneth Trautman; Raymond Wolfe
Original assignee: American Telephone and Telegraph Co Inc
Current assignee: AT&T Corp
Priority date: 1990-11-19
Filing date: 1991-10-28
Publication date: 1995-11-28
Anticipated expiration: 2011-10-28
Also published as: CA2054334A1; JPH04291310A; JP3267992B2

Abstract

An optical system useful, e.g., for near-field scanning optical microscopy is provided. The system incorporates a probe 20 having improved properties. In one embodiment, the probe comprises a tapered and partially metallized portion of a single-mode optical fiber 70.

Description

~Q~433~
.

NEAR FIELD SCANNING OPTICAL MICROSCOPE
AND APPLICATIONS THEREOF

Field of the Invention This invention relates to optical microscopes in which a small aperture, 5 typically smaller than an optical wavelength, is positioned within the optical near field of a specimen, i.e., within a distance of less than about an optical wavelength of the specimen, and the aperture is scanned in raster fashion over the surface of the specimen to produce a time-varying optical signal that is detected and reconstructed to produce an image having extremely high resolution. The invention further relates 10 to methods of use of such microscopes for inspecting workpieces in the course of manufacturing processes.
Art Back~round A number of researchers have investigated the use of optical scanning to circumvent the inherent limitations of conventional optical imaging systems. That 15 is, in so-called near-field scanning optical microscopy (NSOM), an aperture having a m.oter that is smaller than an optical wavelength is positioned in close proximity to the surface of a specimen, and scanned over the surface. In one scheme, the specimen is reflectively or tr~n~mi~sively ilhlmin~ted by an external source. A
portion of the reflected or transmitted light is collected by the aperture and relayed to 20 a photodetector by, for example, an optical fiber. In an alternate scheme, light is relayed by an optical fiber to the aperture, which itself functions as miniscule light source for reflective or tr~n~micsive illllmin~tion of the specimen. In that case, conventional means are used to collect and detect the selected or tr~n~mitte~l light.
In either case, the detected optical signal is reconstructed to provide image 25 information.
Thus, for example, U.S. Patent No. 4,604,520, issued to W. D. Pohl on August 5, 1986, describes an NSOM system using a probe made from a pyramidal, optically transparent crystal. An opaque metal coating is applied to the crystal. At the apex of the crystal, both the tip of the crystal and the metal coating overlying the 30 tip are removed to create the aperture, which is essentially square and has a side length less than 100 nm.
Also described in U.S. Patent No. 4,604,520 is an alternative aperture made from a single-mode optical fiber. One planar end of the fiber is met~lli7ecl, and a coaxial hole is formed in the coating so as to just expose the core of the fiber.

20~433~

In a somewhat different approach, U.S. Patent No. 4,917,462, issued to A. L. Lewis, et al. on April 17, 1990, describes a probe formed from a pipette, i.e., a glass tube that is drawn down to a fine tip and coated with an opaque metal layer.
After drawing, the pipette retains a hollow bore, which emerges through both the5 glass and the overlying metal layer at the tip. The resulting metal annulus defines the aperture. The aperture may be smaller than the bore defined in the glass, as a result of radially inward growth of the metal layer.
One drawback of the above-described methods is that light is tr~n~mitte~l through the probe with relatively low efficiency. As a consequence,10 signal levels are relatively low. In some cases, a~e~ es must be made larger in order to compensate for low signal levels. This measure is undesirable because it results in lower resolution.
For example, when light is tr~n~mitte~l from a source to the aperture through an uncoated pipette, the optical field has a substantial amplitude at the outer 15 walls of the pipette. In order to confine radiation, it is necessary to coat the walls with metal. However, attentuation occurs as a result of absorption in the metal coating. Moreover, metal coatings are prone to imperfections, such as pinholes, that permit optical leakage. When this tendency is countered by increasing the metal thickness, the length (i.e., the thickness in the axial direction of the pipette) and outer 20 diameter of the metal annulus ~lefining the aperture are also increased. As a result, optical losses due to absorption and evanescence in the metal annulus are increased and the size of the tip is increased. Enlarging the tip makes it more difficult to scan narrow concave topographical features of the specimen while m~int~ining close ~ro,~ Jity to the specimen surface. (Significantly, the problem of excessive tip size 25 due to metal deposition also applies to constant-diameter optical fiber probes of the type having an aperture defined by a hole in a metal layer coating the end of the fiber.) Problems that occur when the probe is a single-mode fiber having a sharpened (e.g., by etching) conical tip and no metal coating are described, e.g., in 30 C. Girard and M. Spajer, "Model for reflection near field optical microscopy,"
Applied Optics, 29 (1990) pp. 3726-3733. One problem is that a portion of the light passing through the fiber toward the tip may be reflected by, and then tr~nsmitte~l through, the sides of the conical taper. A second problem is that the sides of the taper may capture undesired optical signals that can propagate through the fiber, 35 resulting in an increased noise level at the detector.

-3- ~o5~334 In view of the foregoing discussion, it is appalent that investigators have hitherto been unsuccessful in providing an NSOM probe that combines efficient transmission of light (i.e., transmission that is relatively free of attenuation due to optical interactions with the walls of the probe) with relatively small tip dimensions, 5 high resolution, and high reliability.
Summary of the Invention In accordance with one aspect of the invention there is provided an optical system, comprising: a probe, at least a portion of which is optically transmissive at least at a given wavelength, the probe having a distal end; an optical aperture defined 10 in the distal end, the aperture having a diameter smaller than the given wavelength;
and means for displacing the probe relative to an object, CHARACTERIZED IN
THAT a) the probe comprises a portion of an optical fiber having a core and a cladding, the cladding having an outer surface, there being associated with the fiber at least one guided dielectric mode for radiation of the given wavelength; b) the fiber has 15 a taper region that is adiabatically tapered, at least a portion of the taper region being capable of guiding light of at least the given wavelength; c) the taper region terminates in a substantially flat end face oriented in a plane substantially perpendicular to the fiber, the aperture being defined in the end face; d) the cladding outer surface in the taper region is substantially smooth; and e) at least a portion of the cladding outer 20 surface in the taper region is coated with an opaque material which has a relatively small penetration depth for electromagnetic radiation of the given wavelength, defining a metallic waveguide portion capable of guiding a metallic mode.
In accordance with another aspect of the invention there is provided a method for e~mining a specimen, COlllpliSi~lg the steps of producing a magnified25 image of at least a portion of the specimen; and visually e~rnining the image, CHARACTERIZED IN THAT the image-producing step comprises situating a probe adjacent a first surface of the specimen and distant from the first surface by no more than a given wavelength, impinging electromagnetic radiation on the specimen, collecting electromagnetic radiation emitted by the first surface, and detecting at least a 30 portion of the collected radiation, the impinging radiation being em~n~ted from the probe and being of at least the given wavelength, or the collected radiation being - - 3a- 205433~

collected by the probe and being of at least the given wavelength, wherein: a) the probe comprises a portion of an optical fiber having a core and a cl~-ling, the cladding having an outer surface, there being associated with the fiber at least one guided dielectric mode for radiation of the given wavelength; b) the fiber has a taper S region that is adiabatically tapered, at least a portion of the taper region being capable of guiding light of at least the given wavelength; c) the taper region terminates in a substantially flat end face oriented in a plane substantially perpendicular to the fiber, the aperture being defined in the end face; d) the cladding outer surface in the taper region is substantially smooth; and e) at least a portion of the cladding outer surface in 10 the taper region is coated with an opaque material which has a relatively small penetration depth for electromagnetic radiation of the given wavelength, defining a metallic waveguide portion capable of guiding a metallic mode.
In accordance with yet another aspect of the invention there is provided a method for paLL~ ling a surface, COlll~l;Si~lg the steps of: a) providing a workpiece 15 having a surface to be pa~ cd; and b) impinging light of at least a given wavelength on the surface such that at least a portion of the impinged light is incident from, or reflected or tr~n~mitt~od to, an optical aperture, CHARACTERIZED ~N THAT the impinging step further comprises: c) situating a probe adjacent the surface, the probe having an end face and the optical a~ u,e defined in the end face, the distance 20 between the aperture and the surface being at most about one given wavelength; and wherein: d) the probe comprises a portion of an optical fiber having a core and a cl~ ing, the cladding having an outer surface, there being associated with the fiber at least one guided dielectric mode for radiation of the given wavelength; e) the fiber has a taper region that is adiabatically tapered, at least a portion of the taper region being 25 capable of guiding light of at least the given wavelength; f) the taper region terrnin~tes in a substantially flat end face oriented in a plane substantially perpendicular to the fiber, the aperture being defined in the end face; g) the cladding outer surface in the taper region is substantially smooth; and h) at least a portion of the cladding outer surface in the taper region is coated with an opaque material which has a relatively 30 small penetration depth for electromagnetic radiation of the given wavelength, defining a metallic waveguide portion capable of guiding a metallic mode.

A
..~

- - 3b - 20~ 334 Brief D~3_.;"lion of the Drawin~
FIG. 1 is a schematic drawing of an exemplary optical system useful for near-field sc~nning optical microscopy.
FIG. 2 is a schematic drawing of an alternative, exemplary optical system S useful for near-field sc~nning optical microscopy.
FIG. 3 is a schematic drawing of a prior-art optical fiber probe.
FIG. 4 is a schematic drawing of a prior-art optical fiber probe.
FIG. 5 is a schematic drawing of an optical fiber probe according to one inventive embodiment.
FIG. 6 depicts an exemplary method for met~ ing an optical fiber probe.

. ~ . ~

4 2~S~334 FIG. 7 is a schematic depiction of an optical aperture according to one inventive embodiment.
FIG. 8 is a schematic depiction of an alternative optical ap~l ~ur~.
FIG. 9 is a flowchart illustrating a manufacturing process according to 5 one inventive embodiment.
Detailed Description In one embodiment, the invention involves an optical system. Turning to FIG. 1, such an optical system includes a light source 10, a probe 20, displacement means 30 for displacing the probe relative to an object 40 disposed, exemplarily on a 10 stage 50, adjacent the probe tip 60. The optical system further comprises means for optically coupling light source 10 to probe 20. In the example illustrated in FIG. 1, the optical coupling is provided by a single-mode optical fiber 70 e~cten-ling between light source 10 and probe 20. (Fiber 70 may, in fact, be integral with probe 20.) Light source 10 is exemplarily a laser. Light from source 10 is readily injected into 15 the optical fiber by way, e.g., of a single-mode coupler 80, which includes amicroscope objective 90 and a fiber positioner 100. A mode stripper 110 is also optionally included in order to insure that only the single mode in the core is propag~ted to the probe, and not other modes in the cl~ lin~- The displacement means 30 may, for example, be a piezoelectric tube adapted for moving the probe 20 vertically as well as in two orthogonal lateral tlim~nsions. Alternatively, the displacement means may be mechanical or piezoelectric means for moving the stagerather than the probe, or some combination of stage-displacement and probe-displacement means.
One possible use for an optical system as described is for direct writing.
25 That is, the sample surface proximate the probe tip may be coated with a photosensitive layer that is capable of being exposed by light emitted from the light source. An exposure pattern is created in the photosensitive layer by displacing the probe relative to the sample, while light from the light source is continuously or in~ Lently emitted from the probe tip.
A second possible use for an optical system as described is for imaging of the sample in a so-called "illl]min~tion" mode. According to this application, light from the probe tip is tr~n.~mitte-l through the sample and collected below the stage (as shown in FIG. 1) by microscope objective 120. (Illustrated is an illllmin~tion-transmission mode; an illllmin~tion-reflection mode is also readily 35 practiced.) The collected light is directed into detector 130, which is exemplarily a photomultiplier tube. For visual positioning of the probe, it is also desirable to 2~5Q3~4 include a beamsplitter 140, which directs a portion of the collected light into an eyepiece 150. Significantly, when the sample is scanned by a raster-like displacement of the probe, the signals from detector 130 can be reconstructed toproduce an image of the sample portion that has been scanned. Such scanning 5 methods are employed in near-field scanning optical microscopy (NSOM), in which the probe tip is brought to within a very small distance of the sample surface, typically less than a wavelength of the light emitted by the light source. NSOM
provides very high optical resolution by also employing an aperture in the probe tip that is very small, also typically less than one wavelength. NSOM apparatus is 10 well-known in the art, and is described, for example, in U.S. Patent No. 4,604,520, issued to W. D. Pohl on August 5, 1986, and in U.S. Patent No. 4,917,462, issued to A. Lewis, et al. on April 17, 1990.
Yet a third possible use for an optical system as described is shown in FIG. 2. In the arrangement of FIG. 2, the probe tip serves as a collector of light 15 rather than as an emitter of light. Such an arrangement is useful, e.g., for NSOM
imaging in a so-called "collection" mode. (Illustrated is a collection-reflection mode. A collection-tr:~n~mi~ion mode is also readily practiced.) Light from light source 10 is directed via tilted mirror 160 and tilted annular mirror 170 to annular objective lens 180. Lens 180 focuses the light onto the sample surface. Light 20 reflected or emitted from the surface is collected by the probe tip and directed via fiber 70 and objective 120 to detector 130. Reflection mode NSOM is described, for example, in U.S. Patent No. 4,917,462, cited above.
The detector (or, more generally, the tr~n~ cer)130 converts the detected light to electrical signals. These signals are readily used to create a two-25 climen.~ nal image on a video display device such as a cathode-ray tube. For such purpose, a scan generator is used to control the displacement of the probe relative to the object and to provide a reference signal for constructing the displayed image.
The electrical signals generated by tr~n~ c~rl30 are typically analog signals.
These are optionally converted to digital signals before they are displayed. In such a 30 case, a digital memory is optionally provided for storing the digitized signals, and a digital processor is optionally provided to process the digitized signals (for, e.g., image enhancement) before they are displayed.
One possible form of probe 20 is a single-mode optical fiber. Optical fiber probes have, in fact, been disclosed in the prior art. F~Gs. 3 and 4 show examples of such fiber probes. FIG. 3 shows an untapered, single-mode optical fiber 190, having an annular cap of opaque m~teri~l, such as metal, deposite,d so as to - 6 ~ 5 4 3 34 cover the tip. The orifice 210 at the center of the annulus defines the optical a~GlLulG
of the probe. FIG. 4 shows a bare optical fiber 220 that is tapered, for example by chemic~31 etching.
We have discovered that an improved probe 230 is readily made by 5 heating a single-mode optical fiber to soften it, and drawing the softened fiber to form a tapered fiber. After drawing, at least a portion of the tapered fiber must be coated with an opaque m~t~ri~l 280, exemplarily metal. With reference to FIG. 5,the tip 240 of such a drawn fiber is tapered at an angle ,~, and terminates in an end flat 250.
The optical fiber comprises a cladding 260 and a core 270. Although specific cladding and core compositions are not essential to the invention, an exemplary cladding composition is fused silica, and an exemplary core composition is doped, fused silica having a higher refractive index than the cl~lding Although specific cladding and core ~limen~ions are also not essential to the invention, an 15 exemplary core diameter is about 3 ~m, and an exemplary cladding outer diameter is about 125 ,um. One guided mode, the filnrl~mental or HE 1l mode, is associated with a corresponding single-mode fiber that has not been tapered. Such a mode is characteristic of a cylindrical dielectric waveguide, and for that reason is hereafter referred to as a dielectric mode.
When an optical fiber is tapered by heating and drawing, both the core diameter and the cladding outer ~ meter are decreased. The fractional change in the core diameter is ap~lv,~imately equal to the fractional change in the cladding outer diameter. (In other words, the cross section of the fiber changes in scale only.) Significantly, the angle at which the core is tapered is substantially smaller than ,~.
25 For example, in a linearly tapered fiber having the exemplary dimensions, thetangent of the core taper angle is only 3/125, or 2.4%, times the tangent of taper angle ,B. For this reason, even for relatively large values of ,~, for example values of 30 or even more, the core will have an ~ b:~tic taper, as discussed below.
In the untapered fiber, the electric field of the dielectric mode is largely 30 confined to the core, and it falls to a very small amplitude, typically less than 10- 10 times the peak amplitude, near the cladding outer surface. That is not necessarily the case in a tapered fiber. As a guided light wave propagates into the taper region, it encounters a progressively narrowing core. Eventually, the core becomes too small to substantially confine the guided mode. Tn~te~d, the light wave is guided by the 35 interface between the cladding and the surrounding material, which may be air, or a metal such as alul~inulll. As discussed above, the core will generally be tapered at a 2~ 433 ~

small enough angle for the mode to be atli~batic. By a~ bati~ is meant that substantially all of the energy of the initial HE ll mode remains concentrated in a single mode, and is not coupled into other modes, particularly radiation modes, which are capable of causing optical losses by radiation.
Initially, the guided mode that escapes from the core subst~nti~lly retains the properties of a HE ll mode. In particular, the amplitude of the guided electromagnetic field is relatively small at the cladding outer surface. However, as the fiber diameter continues to decrease, the field amplitude at the cladding outer surface increases relative to the peak amplitude within the fiber. Eventually, the field 10 attains a relatively large amplitude at the cladding outer surface. Such a mode could be guided, for example, by the interface between the cladding and surrounding air.
However, such an arrangement is undesirable.
Because the electric field extends significantly outside of the cladding, a significant amount of optical leakage can be expected. This is undesirable because it 15 reduces the efficiency with which light is channeled both to and from the probe tip.
Moreover, light leaking from the cladding walls can llltim~tely fall upon portions of the sample lying relatively far from the probe tip, resulting in unintended exposure of the sample, or in increased background levels at the detector. Furthermore, if the cladding outer walls are surrounded only by air, stray light can enter the fiber, again 20 resulting in increased background levels at the detector (if, for example, the probe is being used to collect light from the sample).
For the above reasons, it is desirable to coat at least the t~rmin~l portion of the taper, from which optical leakage is a significant factor, with an opaquematerial, exemplarily a metal such as alu,llinun.. Significantly, the guided mode in 25 such a metsllli7sd portion has characteristics typical of a guided mode in a metal, rather than a dielectric, waveguide. Thus, for example, an initial HE 1l dielectric mode may be converted to a TE ll m-~t~llic mode as it approaches and enters the metallized region. Significantly, in a waveguide having an adiabatically tapered core, the HE 1 l mode can couple with relatively high efficiency (typically, greater than 30 10% efficiency) to the TE l l mode. That portion of the fiber in which, if the fiber were bare, the guided mode would have a significant amplitude at the fiber outersurface, but would still retain substantial HE l l character is here referred to as the "transition region." In view of the foregoing discussion, it is clearly desirable (although not absolutely required) to metal-coat the fiber throughout the transition 35 region, and from the transition region to the distal end of the fiber. It should be noted in this regard that where a significant amount of optical energy is present in 205433~

radiation modes, relatively thick metal coatings are required in order to subst~nti~lly elimin~te the possibility of pinhole leaks and, in view of the finite conductivity of any actual coating, the possibility of penetration of the metal by the electromagnetic field. This is undesirable both because the coating thickness may hinder extremely 5 close approach of the probe tip to the sample, and because very thick metal coatings may develop granularity that tends to promote, rather than inhibit, the occurrence of pinhole leaks. By contrast, it is generally expected here that relatively little energy will couple into radiation modes, and therefore relatively thin metal coatings, O O
typically 750 -1500 A, and preferably less than about 1250 A will generally suffice.
Generally, the guided metallic mode will initially be a propagating mode. However, as the fiber diameter continues to decrease, the guided mode may eventually be transformed to an evanescent mode exhibiting relatively strong attenuation in the propagation direction. Such a transition may be associated with a characteristic quantity here referred to as the "cutoff ~ mete~r~ The cutoff diameter 15 is the cladding outer diameter of a given tapered waveguide at which the transition from a propag~ting to an evanescent mode would occur if the waveguide were coated with an infinitely conductive metal. Generally, for the TE 1l mode, the cutoff 1i~mlot~r iS roughly equal to one-half the guided wavelength. That portion of the fiber extending from the cutoff rli~meter to the probe tip is here referred to as the 20 "evanescent region."
The cutoff diameter of a TE 1 l mode in, e.g., a circularly cylin-lrieal waveguide is readily predicted from the theory of metallic waveguides as discussed, for example, in J. D. Jackson, Classical Electrodynamics, 2d Edition, John Wileyand Sons, Inc., New York, 1975, p. 356. It is equal to the quantity X / kOn, where ko 25 is the free-space wavenumber of the guided light, n is the refractive index of the waveguide (with reference to the guided mode and wavelength), and X is a quantity related to the specific mode that is guided. For a TE 1l mode, X is equal to 1.841.
Values of % corresponding to other guided modes will be readily apparent to the ordinarily skilled practitioner.
It should be noted in this regard that the transformation of the filn-l~m-~ntal dielectric mode to the TEll metallic mode may be incomplete. Somefinite portion of the energy of the dielectric mode may be coupled into metallicmodes other than the TE 1l metallic mode. However, the TE 1l mode generally experiences less attenuation, in the evanescent region, than any of the other metallic 35 modes. For that reason, a light wave that has traversed the evanescent regionsubstantially contains only the TE 1l metallic mode. (It should further be noted that 2~5~334 because of the finite conductivity of actual metal coatings, a perturbed TE 1l mode is expected; i.e., a mode in which the electric field has a small longitll-lin~l colllpollent.) Because substantial attenuation takes place in the evanescent region, it is 5 desirable to make that region as short as possible. However, other factors, to be discussed below, may militate for differing evanescent lengths in differing applications.
It is desirable for the cladding outer surface in the tapered portion of the fiber to be substantially smooth in order to reduce sc~tt~ring of light from the HE 1l 10 mode near the cladding outer surface, and in order to receive a metal coating that is relatively thin (preferably, less than about 1500 A thick) and also subst~nti~lly free of defects capable of leaking optical radiation. A cladding surface is here considered to be substantially smooth if no surface texture is appalent on a scale of more than about 50 A as observed in a scanning electron micrograph (SEM). Surfaces of such15 desirable smoothness are readily produced by heating and drawing the fiber.
It is desirable for the end flat to be subst~nti~lly planar, and to be oriented substantially perpendicular to the axial direction of the fiber. The end flat is here considered to be subst~nti~lly planar if, over the surface of the end flat,e7~min~tion by SEM reveals no deviation from planarity greater than about 100 A
20 by any feature of any lateral extent.
It is desirable for the edges of the end flat to be relatively sharply defined. The edges are here considered to be sharply defined if ex~min~tion by SEM
reveals an average radius of curvature, at the edge, of less than about 100 A. End flats of such desirable planarity, and having edges of such desirable sharpness, are 25 also readily produced by heating and drawing the fiber.
One method for metal-coating the termin~l portion of the fiber (preferably including at least the transition region, as noted above) involves the use of an evaporation source of, e.g., aluminum As shown in FIG. 6, the fiber 230 is not oriented end-on toward the source 290 during evaporation and deposition of the 30 aluminum. Instead, the probe end of the fiber is pointed away from the source, such that the end flat lies in a shadow relative to the direction of incidence of thebombarding metal. Typically, the fiber axis is inclined at an angle 0 of about 75 relative to a line drawn from the source to the fiber tip. During deposition, the fiber is rotated about its own axis in order to uniformly coat all sides of the fiber. When 35 such a method is employed, a coating is readily produced that smoothly covers the cladding outer surface, but leaves the end flat substantially bare of any metal. With 2~S433~

reference to FM. 7, the above-described method is useful for producing a probe in which the optical aperture 300 corresponds to the entire surface of the end flat.
Because the spatial resolution of the probe is determined mainly by the aperturediameter, it is desirable, for high-resolution applications, to make the end flat 5 diameter quite small. Thus, the end flat diameter in such a probe would generally be made smaller than the cutoff diameter. For example, the cutoff diameter for 5145 A
light from an argon ion laser is typically about 2000 A in a fiber of the exemplary ~limensions discussed above. A typical, corresponding end flat diameter is about 500 - 1000 A, and end flats as small as about 200 A, or even smaller, are readily made by 10 heating and drawing the fiber. In general, the average thickness of the aluminllm layer is preferably not less than about 750 A, because substantially thinner layers may be subject to excessive optical leakage. (If the probe is used for both illnmin~ting the object and collecting light from the object, an even smaller thickness may be acceptable.) The thickness is preferably not greater than about15 1500 A, because it is desirable to minimi7e the total diameter of the probe tip and to make the coating as smooth as possible. The total diameter is the sum of the end flat diameter and the metal thicknesses bounding the end flat at diametrically opposite positions. It is desirable to minimi7e this total ~ m~ t~r in order to provide a probe tip that can be inserted within relatively narrow cavities or crevices in the sample 20 surface, and that can, more generally, penetrate as deeply as possible into the near field of the sample.
Significantly, a probe of the above-described type has an evanescent region extending from the cutoff diameter to the end flat. Such an evanescent region is relatively long, by comparison to an alternate design to be described below. For 25 example, with a cutoff diameter of 2000 A, an end flat ~ m~ter of 500 A, and a taper angle of 15, the evanescent length is 5600 A. In such a distance, a significantamount of attenuation may take place. Consideration of the attenuation milit~tes for making the taper angle as large as is feasible in order to shorten the evanescent length. However, it must also be considered that a narrower taper angle makes it30 easier to penetrate crevices and to approach step-like surface features more closely.
According to a preferred method for tapering the fiber, the fiber is first mounted in a commercial pipette puller. The fiber is heated prior to and during drawing. An exemplary heat source is a carbon dioxide laser. The fiber is drawn until it breaks. Controllable parameters include the incident illnmin~tion intensity 35 (which determines the heating rate), the pulling force applied to the fiber ends, and the number of individual pulling steps that are applied. Generally, the fiber is finally 2~5~33~

broken by a "hard pull," that is, a pulling step conducted with sharply increased force. The velocity of the fiber ends at the moment of application of the hard pull is also controllable, as is the time delay between cessation of heating and application of the hard pull.
S We have discovered that a given set of process pa~ lel~l s produces probes having highly reproducible characteristics. The taper angle is readily increased by any selection or combination of reducing the heating rate, reducing the pulling force, pulling in multiple steps, or reducing the extent of the heated region.
The end flat diameter is readily increased by increasing the heating rate, or 10 increasing the pulling force, or both. In addition, the use of a fiber having a higher temperature glass composition provides a bigger taper angle and a smaller tip.
Desirable combinations of process parameters will be readily apparent to the ordinarily skilled practitioner after minim~l experiment~tion.
An ~ltern~tive type of probe is illustrated in FIG. 8. In this type of 15 probe, the metal coating is deposited over the end flat as well as on the cladding outer surface. As discussed above, the metal thickness on the cl~ ling outer surface is desirably about 750-1500 A thick. The metal layer overlying the end flat is desirably about 250 - 500 A thick because it needs to be thick enough to substantially elimin~te stray light, but also as thin as practicable in order to minimi7e 20 the evanescent length. The optical aperture 300 in this case corresponds, not to the entire end flat, but only to a portion of the end flat. The metal layer overlying the end flat is formed into an annulus having an opening which defines the optical a~ lule. Significantly, the opening may be centrally located on the end flat, or, alternatively, it may be displaced from the center of the flat. A non-central aperture 25 may be desirable where multiple modes are expected to be present in the tip region.
Because not all modes will couple efficiently to a centrally located aperture, a non-central aperture in such a case may provide more efficient output coupling. In particular, a non-central aperture is expected to give the optimal output coupling for the TE 1l mode.
The central opening is formed, for example, by initially providing a fiber that incorporates an embedded rod of glass that is more rapidly etched than the host glass of the fiber. (An example is borosilicate glass embedded in fused silica.) The end of the fiber is exposed to a chemical etchant after pulling but before the fiber is met~lli7ed The etchant produces a cavity in the end flat. The cavity is shadowed35 during metallization of the end flat, resulting in an opening in the metal layer.

205~33~

A typical aperture made by this method is about 200-2000 A in diameter, and the same method is readily employed to form apc;l~ulc~s even larger or smaller.
The metal coating is exemplarily deposited from an evaporation source, as discussed above in connection with the bare-tipped probe. However, two separate deposition steps are employed: one for coating the the cladding outer surface, as described above, and a second step at a different fiber orientation for coating the end flat.
Because in this type of probe, the aperture is not determined by the end 10 flat ~ m~ter, it is not generally necessary for the end flat ~ m~ter to be smaller than the cutoff diameter, even for high-resolution applications. Accordingly, in order to minimi7e attenuation, it is desirable for the end flat diameter to be, approximately, at least the cutoff diameter. However, as discussed above, it is desirable to minimi7e the total probe tip diameter in order to be able to penetrate crevices. For this reason, the end flat f~ meter is preferably also not substantially greater than the cutoff diameter.
It should be noted in this regard that the ~tillule diameter is generally smaller than the cutoff diameter, and therefore the central opening in the metal layer overlying the end flat defines a metallic waveguide in which the optical field is evanescent. However, because the coating is typically only about 500 A thick or less, the evanescent length is much smaller than it is in the case of the bare-tipped probe ~ cllsse~l previously. Clearly, in order to minimi7e attenuation, it is desirable to make the metal coating on the end flat as thin as is practicable. An acceptable range for metal thicknesses on the end flat is about 250- 500 A, and, as noted, 500 A
25 is typical.
We have calculated the ideal power attenuation expected in the evanescent portion of a bare-tipped probe. The corresponding power tr~n~mission coefficient is conveniendy expressed in terms of %, which, as noted above, is a characteristic of the guided mode, the taper angle ,~, and a dimensionless parameter 0 a, which is defined by a = %/nkOa, where a is the aperture diameter, and as noted above, ko is the free-space wavenumber of the guided light, and n is the relevant value of the waveguide refractive index. It is helpful to note that a is equal to the ratio of the cutoff diameter to the aperture diameter. When expressed in decibels, the power transmission coefficient corresponding to ideal evanescent attenuation is given by:

2~433~

TeV (db) = 8.68 (%/tan,B) ~ -ln (a + ~a2-l ) We have sllcGeedçd in making probes, having various values of a ranging between 1 and 10, and various values of ,~ ranging between 5 and 20, in which the overall measured power tr~n~mi~sion coefficient falls below TeV by no more than 10 dB.
At least for values of a ranging between 2 and 8, T eV may be roughly approxim~ted by -2xaltan ,B dB. Accordingly, we have succeeded in making probes for which the overall measured power tr:~nsmicsion coefficient, denoted T, satisfies T (db) ~ -(2%a/tan ~) - 10.
The inventive optical system is useful, inter alia, as a manufacturing 10 tool. For example, the probe is readily used to conduct actinic radiation from a light source to a small region of a workpiece surface sitll~tçd ~dj~cent the optical aperture of the probe. Thus, for example, a workpiece surface may be coated with a photosensitive layer such as a photoresist, and a pattern formed in the layer bymoving the probe relative to the layer while transmitting light through the probe 15 such that light exits the aperture and falls on the layer, thereby exposing the layer.
Additional steps are then readily performed, leading, for example, to the completion of an article of manufacture. Exemplarily, the article of manufacture is a semiconductor integrated circuit, and the additional steps include developing the photoresist and subjecting the resist-coated surface to an etchant such that a layer 20 underlying the photoresist is paLL~.Ilcd. Significantly, not only semiconductor wafers, but also glass plates intended for use as photolithographic masks, are readily patterned in this fashion.
In an alternative process, actinic radiation from the probe is readily directed to a workpiece surface in order to cause deposition of m~teri~l on the 25 surface. For example, the surface may be exposed to an organometallic vapor or solution. Actinic radiation impinging on the surface may excite local decomposition of one or more constituents of the vapor or solution, producing, e.g., a metallic residue that adheres to the surface.
The inventive optical system is also useful as an inspection device on a 30 manufacturing line. For example, with reference to FIG. 9, in the manufacture of semiconductor integrated circuits, a semiconductor wafer is often provided (Step A), having a surface that is patterned at one or more stages of the manufacturing process.
The patterns that are formed have characteristic dimensions, sometimes referred to as "line widths," that must generally be kept within close tolerances. The inventive 2~334 optical system is readily used for measuring (Step D) line widths, such as the width of metallic conductors on a wafer, or the length of gates formed in metal-oxide-semiconductor (MOS) structures on a wafer. Such line widths can then be coll~aled (Step E) with desired values. Process parameters, exemplarily lithographic exposure 5 times or etching times, which have been initially set (Step B) are readily adjusted (Step H) to bring the measured dimensions within the desired tolerances. Additional steps (Step G) toward completion of a manufactured article are then performed. The measuring step includes situating the probe adjacent the patterned surface (Step I) and impinging (Step J) light upon, and detecting (Step K) light from the surface such 10 that the light source and detector are optically coupled via the inventive probe, which is substantially as described above.
The inventive optical system is also useful for inspecting bit patterns in magnetic digital storage media such as magnetic disks. Because magnetic storage media generally exhibit a Faraday rotation of polarized light, the bit patterns, which 15 are characterized by modulated magnetization directions, are readily visualized by inspection using, e.g., crossed polarizers. Thus for example, the inventive optical system may be used for reflection-mode imaging of the me~ m, using a polarized light source and including a polarizing filter in front of the detector. In a manufacturing process that involves impressing such media with bit patterns having 20 predetermine~l properties, relevant process parameters may be adjusted to bring the detected bit patterns into conformity with the desired patterns.
In at least some cases, the laser light source is itself a source of polarized light. In other cases, it may be desirable to pass light from the source through a linearly polarizing film. Before such light is coupled into optical fiber 70 25 (see FIG. 1), it is often useful to pass it through a half-wave plate and then a quarter-wave plate. The orientations of these plates are readily adjusted in order to compensate for birefringence in the optical fiber. (A polarization preserving fiber may be useful, but is not necessary.) The linearly polarized component of light em~,lgillg from the fiber is optimized by, e.g., visually detecting it through a second 30 linearly polarizing film while adjusting the half-wave and quarter-wave plates.
Other applications of the inventive optical system for manufacturing and inspection will be readily apparent to the skilled practitioner.
For example, the inventive optical system is also useful for impressing digital data for storage on an optical or magnetic storage medium, and for reading 35 such stored data from the storage medium. For example, it is well known that data can be recorded on magnetizeable metallic films in the form of patterns of spots 2~A 334 .

having local magnetization that differs in direction from the magnetization of surrounding film portions. This data storage technology is described, for example, in R.J. Gambino, "Optical Storage Disk Technology," MRS Bulletin 15 (April 1990) pp. 20-24, and in F.J.A.M. Greidanus and W.B. Zeper, "Magneto-Optical Storage 5 Materials," MRS Bulletin 15 (April 1990) pp. 31-39.
A typical magnetic storage medium is a layer which comprises an amorphous alloy of one or more of the rare-earths and one or more of the transition metals. (Alternative magnetic storage materials include cobalt-pl~tinllm or cobalt-palladium multilayer films, and magnetic oxide materials such as ferrites and 10 garnets.) A spot representing, e.g., a bit of digital data, is written by exposing the medium to a magnetic field and optically heating the spot area above the Curie temperature or above the compensation point of the medium. (In some cases, the internal demagnetizing field of the medium is sufficient to cause local reversal of magnetization, and an external field need not be applied.) Such spots are 15 conventionally made about 1 ~m in diameter. However, the inventive optical system is readily employed to make smaller spots, exemplarily spots having diamG~el~ ofabout 0.2 - 0.5 ,um, and even as small as 0.06 llm or less. Such smaller spots can also be read by the inventive optical system.
In a currently preferred embodiment, the magnetic storage medium is a 20 thin, amorphous film of an alloy comprising at least one rare earth and at least one transition metal. An exemplary such alloy is ~ell,iulll-iron. The aperture of the near-field probe is ~in1atefl less than about one illumination wavelength from the surface of the medium. (If larger spots are desired, the probe is readily positioned more than one wavelength from the medium.) The illllmin~tion wavelength is 25 selected to give adequate heating to the medium. For example, a terbium-iron film is readily heated by a dye laser which is pumped by a YAG laser and emits roughly nanosecond pulses at a wavelength of roughly 600 nm. Light from the laser is directed into the inventive probe via an optical fiber, and is impinged from the probe tip onto the recording m~lillm A typical, local temperature change required to write 30 a spot is about 150C. Although writing can be achieved using either a continuous or pulsed laser for illumination, a pulsed laser is preferable in order to relax the average power re(luilGlllen~ of the laser while heating a relatively small spot.Significantly, the invention is inten(led to encompass not only magnetic film recording media, but also other media that are capable of being written by an 35 illllmin~tion source. Such media include, for example, polycrystalline films (of, e.g., tellurium-doped indium antimonide) that are locally heated to, e.g., a temperature ~4334 -which exceeds the melting point of the film by a laser pulse and rapidly quenched to an amorphous state. The cooling rate can be influenced by, e.g., a~r~liately shaping the time dependence of the laser pulse.
Spots representing recorded data are typically written in a track in the 5 recording medium. Such a track extends, e.g., ~iircull~l~;ntially on a rotating disk.
The diameters of spots written in accordance with the instant invention are typically much smaller than the width of such a track. Accordingly, multiple such spots are advantageously written in bands extending substantially transversely across the track. One advantage of the invention is that the spots in such bands can be read 10 ~imlllt~neously by a linear array of inventive near-field probes.
In a currently preferred method of reading a pattern impressed on a magnetic film recording m~ lm, linearly polarized light emitted by the inventiveprobe is tr~n~mitte-l through the medium, and conventional means are used to collect a portion of such tr~n.cmitte~ light and to polarization analyze it. For reading, 15 plerelled wavelengths are those having a maximal optical response (i.e., m~xim~l rotation of the direction of polarization of the light as it traverses the magnetic medium). For transition-metal-rare-earth media, such wavelengths will typically lie in the near-infrared or visible spectrum.
In an ~ltern~te embodiment, the probe is used both to impinge polarized 20 light onto the surface of the medium, and to collect a portion of such light that is reflected from the surface. As a result of either tr~nsmi~sion through, or reflection from, a magnetized spot in the medium, a polarization rotation of, typically, about 0.5 is produced. The effect of such rotation is to modulate the intensity of light tr~n~mitte~l through an analyzer and subsequently detected. Such modulation is 25 readily decoded to reproduce the information recorded in the medium. As is well known, such information may, for example, be recorded sound, images, text, or digital data.
In alternate embodiments, such as those involving phase changes, modulation is typically provided by changes in reflectivity, rather than by 30 polarization rotation. Such changes are also readily detectable using the inventive probe to collect reflected light.
The inventive microscope is also advantageously employed for imaging applications in biological research and clinical medicine. In particular, the inventive microscope overcomes, in at least some cases, the problem of low signal level that 35 often detracts from the usefulness of prior-art NSOM systems for medical and biological imaging. Thus, for example, the inventive microscope is readily used to 4 33~

image sectioned samples of biological tissue in order to find and identify physical pathologies in the tissue. In a similar manner, the inventive microscope is readily used to examine the distribution in sectioned tissue samples of m~t.ori~l~ that are detectable by their inherent appearance or fluorescence, as well as m:~t~ri~l~ that are 5 labeled by, e.g., fluorescent dyes.
The inventive microscope is also useful, in genetic clinical and research applications, for imaging chromosomes while they are in, e.g., the metaphase state.
In particular, chromosomes or portions of chromosomes that are labeled with fluorescent m~teri~l are readily identified using the inventive microscope. Methods 10 of fluorescent labeling, which generally involve reacting cell nuclear material with material that includes fluorophores, are well-known in the art and need not be described in detail herein.
Fluorescent imaging is readily achieved using the microscope in either an illllmin~tion mode (FIG. 1) or a collection mode (FIG. 2). In the former, 15 electromagnetic radiation capable of inducing the sample to emit fluorescent light impinges on the sample from the probe and the resulting fluorescence is collected by, e.g., a conventional microscope objective. In the latter, the exciting radiation is conventionally impinged on the sample, and the fluorescent light is collected by the probe.
As noted, a preferred probe 20 (see FIG. 1) is made by tapering a single-mode optical fiber. It should be noted that in at least some cases a useful probe can also be made by tapering and coating a multiple mode fiber. Depending on the dimensions of a multiple mode fiber (relative to the guided wavelength), a greater or lesser number of modes will be guided. In general, it is expected that the 25 fewer the number of guided modes (i.e., the more closely the fiber resembles a single-mode fiber), the greater will be the signal-to-noise ratio achieved for a given aperture.
Also as noted, an exemplary opaque coating for the probe is a metal such as aluminum. More generally, an appl~liate coating is one consisting of a 30 material in which the guided radiation has a low penetration (skin) depth.
Aluminum is a plc;f~ d m~tçri~l when the radiation lies in, e.g., the visible spectrum, but semiconductors such as silicon may be preferable where, e.g., ultraviolet radiation is to be guided.

- 18~ 34334 EXAMPLE
A 3-~1m, single-mode fiber (FS-VS-2211) having a 450 nm cutoff, was drawn in a Mod. P-~7 micrupipeue puller m~nllf~ctured by Sutter Instruments, while heating the fiber with a 25-watt, 3-mm spot from a 25-watt carbon dioxide laser.5 The micropipette puller was programmed to provide a hard pull at a setting of 75 (range 0 - 255), a "velocity at pull" at a setting of 4 (range 0 - 255), and a time delay of 1 (range 0 - 255). The tip properties that were obtained were: a 12 taper angle, a 670-A-diameter end flat, and a value for a of about 3. One fiber end was placed in a rotator and evaporation coated with about 1260 A of alulllinulll at a base pressure of 10 about 10-6 torr. The tapered end was then mounted in a piezoelectric tube in the optical arrangement of FIG. 1. One milliwatt of 514.5-nm light from an argon ionlaser was coupled into the fiber. The optical power output at the fiber tip was measured to be about 1.1 nanowatt, corresponding to an overall power tr:ln~mi~sion coefficient of about -60 dB. When used to form an image of a sample surface, the15 probe provided a spatial resolution of about 25 nm.

Claims

1. An optical system, comprising:
a probe, at least a portion of which is optically transmissive at least at a given wavelength, the probe having a distal end;
an optical aperture defined in the distal end, the aperture having a diameter smaller than the given wavelength; and means for displacing the probe relative to an object, CHARACTERIZED IN THAT
a) the probe comprises a portion of an optical fiber having a core and a cladding, the cladding having an outer surface, there being associated with the fiber at least one guided dielectric mode for radiation of the given wavelength;
b) the fiber has a taper region that is adiabatically tapered, at least a portion of the taper region being capable of guiding light of at least the givenwavelength;
c) the taper region terminates in a substantially flat end face oriented in a plane substantially perpendicular to the fiber, the aperture being defined in the end face;
d) the cladding outer surface in the taper region is substantially smooth;
and e) at least a portion of the cladding outer surface in the taper region is coated with an opaque material which has a relatively small penetration depth for electromagnetic radiation of the given wavelength, defining a metallic waveguideportion capable of guiding a metallic mode.

2. The optical system of claim 1, further comprising means for optically coupling a light source to the probe such that at least some light emitted by the source, at least at the given wavelength, enters or exits the probe through the aperture.

3. The optical system of claim 2, further comprising:
a source of electromagnetic radiation, to be referred to as a "light source";
a scan generator for driving the displacement means such that the probe is displaced in a raster pattern adjacent a portion of a surface of the sample;

transducive means for detecting at least a portion of the light from the light source which enters or exits the probe, and for generating an electrical signal in response to the detected light;
far-field microscopic means for viewing at least a portion of the probe and at least a portion of the object; and video display means, in signal-receiving relationship to the transducive means, for displaying a two-dimensional image which relates to the amount of detected light at least at some displacements of the probe relative to the sample, the displacements being part of the raster pattern.

4. The optical system of claim 3, wherein the transducive means are adapted to generate analog electrical signals, and the system further comprises means for converting the analog electrical signals to digital signals and transmitting the digital signals to the video display means.

5. The optical system of claim 4, further comprising storage means for digitally recording at least a portion of the digital signals.

6. The optical system of claim 5, further comprising means for digitally processing the digital signals, the digital processing means being in receiving relationship with the analog-to-digital conversion means and in transmissive relationship to the video display means.

7. The optical system of claim 1, further comprising:
a source of electromagnetic radiation capable of stimulating at least a portion of said object to emit fluorescent light, said source optically coupled to the probe such that radiation is transmitted from the source, through the probe, to said object; and means for detecting at least a portion of fluorescent light emitted by said object.

8. The optical system of claim 1, further comprising:
a source of electromagnetic radiation capable of stimulating an object to emit fluorescent light at least at the given wavelength; and means for supporting the object relative to the radiation source and the probe such that at least a portion of the radiation emitted by the source impinges on the object and at least a portion of the fluorescent light emitted by the object enters the probe through the optical aperture.

9. The optical system of claim 1, wherein a cutoff diameter is defined for the metallic mode, and the end face has a diameter that is less than or equal to the cutoff diameter.

10. The optical system of claim 9, wherein:
the end face is substantially free of overlying opaque material;
the fiber has, at each longitudinal position, a cladding outer diameter;
in at least a portion of the metallic waveguide portion, referred to as the "evanescent region", the cladding outer diameter is smaller than the cutoff diameter;
and the aperture substantially coincides with the end face.

11. The optical system of claim 9, wherein:
the end face diameter is approximately equal to the cutoff diameter; and the probe further comprises an annular layer of opaque material overlying the end face, such that the annular layer surrounds, and thereby defines, the aperture.

12. The optical system of claim 9, wherein:
the ratio of the cutoff diameter to the aperture diameter is denoted .alpha.;
the metallic mode is the TE11 mode;
the taper angle is denoted .beta.;
.alpha. is at least about 2 and at most about 8; and the probe has a transmission coefficient, denoted T, which, when expressed in decibels, satisfies the relation T>-(3.68.alpha./tan.beta.)- 10

13. The optical system of claim 11, wherein the aperture has a center, the end face has a center, and the center of the aperture is displaced from the center of the end face.

14. A method for examining a specimen, comprising the steps of producing a magnified image of at least a portion of the specimen; and visually examining the image, CHARACTERIZED IN THAT
the image-producing step comprises situating a probe adjacent a first surface of the specimen and distant from the first surface by no more than a given wavelength, impinging electromagnetic radiation on the specimen, collecting electromagnetic radiation emitted by the first surface, and detecting at least a portion of the collected radiation, the impinging radiation being emanated from the probe and being of at least the given wavelength, or the collected radiation being collected by the probe and being of at least the given wavelength, wherein:
a) the probe comprises a portion of an optical fiber having a core and a cladding, the cladding having an outer surface, there being associated with the fiber at least one guided dielectric mode for radiation of the given wavelength;
b) the fiber has a taper region that is adiabatically tapered, at least a portion of the taper region being capable of guiding light of at least the givenwavelength;
c) the taper region terminates in a substantially flat end face oriented in a plane substantially perpendicular to the fiber, the aperture being defined in the end face;
d) the cladding outer surface in the taper region is substantially smooth;
and e) at least a portion of the cladding outer surface in the taper region is coated with an opaque material which has a relatively small penetration depth for electromagnetic radiation of the given wavelength, defining a metallic waveguideportion capable of guiding a metallic mode.

15. The method of claim 14, wherein the specimen has a second surface, the impinging step comprises impinging radiation on the second surface, and the radiation emitted by the first surface is radiation that has been transmitted through the specimen.

16. The method of claim 14, wherein the impinging step comprises impinging radiation on the first surface, and the collecting step comprises collecting reflected radiation.

17. The method of claim 14, wherein the impinging step comprises impinging radiation on the first surface, the impinged radiation comprises radiation of at least one wavelength capable of exciting fluorescence in at least a portion of the specimen, and the collecting step comprises collecting fluorescent light.

18. A method for patterning a surface, comprising the steps of:
a) providing a workpiece having a surface to be patterned; and b) impinging light of at least a given wavelength on the surface such that at least a portion of the impinged light is incident from, or reflected or transmitted to, an optical aperture, CHARACTERIZED IN THAT the impinging step further comprises:
c) situating a probe adjacent the surface, the probe having an end face and the optical aperture defined in the end face, the distance between the aperture and the surface being at most about one given wavelength; and wherein:
d) the probe comprises a portion of an optical fiber having a core and a cladding, the cladding having an outer surface, there being associated with the fiber at least one guided dielectric mode for radiation of the given wavelength;
e) the fiber has a taper region that is adiabatically tapered, at least a portion of the taper region being capable of guiding light of at least the givenwavelength;
f) the taper region terminates in a substantially flat end face oriented in a plane substantially perpendicular to the fiber, the aperture being defined in the end face;
g) the cladding outer surface in the taper region is substantially smooth;
and h) at least a portion of the cladding outer surface in the taper region is coated with an opaque material which has a relatively small penetration depth for electromagnetic radiation of the given wavelength, defining a metallic waveguideportion capable of guiding a metallic mode.

19. The method of claim 18, wherein the surface comprises a recording medium upon which a bit pattern can be impressed; in the impinging step, light is incident on the medium from the optical aperture; and the impinging step results in a localized physical change in the medium such that an optical signal can be modulated by the physical change, the modulation carrying at least one bit of information.

20. The method of claim 19, wherein the medium is a magnetic medium, and the impinging step results in local heating of the medium above its Curie temperature or above its compensation point, resulting in the formation of a spot where the local magnetization is different from the magnetization of adjacent regions surrounding the spot.

21. The method of claim 20, wherein the impinging step is performed in the presence of an external magnetic field.

22. The method of claim 20, wherein the medium is a magnetic alloy comprising at least one rare earth and at least one transition metal.

23. The method of claim 19, wherein the medium is substantially polycrystalline, and the impinging step results in local heating of a portion of the medium followed by rapid cooling, such that a local, substantially amorphous spot is formed.

24. The method of claim 18, wherein the providing step comprises:
providing a multiplicity of semiconductor wafers, each wafer having a surface to be patterned;
the method further comprises:
a) setting at least one process parameter;
b) processing at least one first wafer according to the process parameter such that a pattern is formed on the surface of the wafer, the pattern having a characteristic dimension;
c) measuring the characteristic dimension on the first wafer, d) comparing the characteristic dimension to a predetermined range of values;
e) if the characteristic dimension lies outside the predetermined range of values, changing the process parameter to bring the characteristic dimension within the predetermined range of values;
f) after (e), processing at least one second wafer according to the process parameter; and g) performing, on at least the second wafer, at least one additional step toward completion of the article; and h) the impinging step comprises, during the measuring step, situating the probe adjacent the patterned surface of the first wafer, impinging light on the surface such that at least a portion of the light is reflected from the surface, the impinging light being emanated from the probe or a portion of the reflected light being collected by the probe, and detecting at least a portion of the reflected light.

25. The method of claim 18, wherein the providing step comprises providing a multiplicity of substrates, each substrate having a surface layer of magnetic material to be impressed with a bit pattern;
the method further comprises:
a) setting at least one process parameter;
b) processing at least a firse substrate according to the process parameter such that a bit pattern is formed in the surface of the substrate;
c) sensing the bit pattern in at least the first substrate;
d) comparing the sensed bit pattern to a predetermined bit pattern;
e) if the sensed bit pattern is different from the predetermined bit pattern, changing the process parameter to bring the sensed and predetermined bit patterns into conformity;
f) after (e), processing at least a second substrate according to the process parameter, and g) performing, on at least the second substrate, at least one additional step toward completion of the article; and h) the impinging step comprises, during the sensing step, situating the probe adjacent the patterned surface of the first wafer, impinging light on the surface such that at least a portion of the light is reflected from the surface, the impinging light being emanated from the probe or a portion of the reflected light being collected by the probe, and detecting at least a portion of the reflected light.

26. The method of claim 18, wherein:
the surface comprises a photoresist;
the impinging step comprises impinging actinic, electromagnetic radiation on the surface from the optical aperture such that a latent image is formed in the photoresist; and the method further comprises the steps of:

developing the photoresist; and exposing the workpiece to an etchant such that a pattern is formed in the surface.

27. The method of claim 18, further comprising the step of exposing the surface to a chemically reactive liquid or vapor, and wherein the impinging stepcomprises impinging actinic, electromagnetic radiation on the surface from the optical aperture such that material is formed by the chemical decomposition of the liquid or vapor, the material depositing on and adhering to the surface, whereby the surface is patterned.