WO2000079566A1

WO2000079566A1 - Alignment of low-energy electron beams using magnetic fields

Info

Publication number: WO2000079566A1
Application number: PCT/US2000/017334
Authority: WO
Inventors: Andres Fernandez; Lawrence P. Muray; James P. Spallas; Tai-Hon Philip Chang
Original assignee: Etec Systems, Inc.
Priority date: 1999-06-22
Filing date: 2000-06-22
Publication date: 2000-12-28
Also published as: AU5762100A

Abstract

For wafer or mask fabrication using electron beam scanning lithography, a registration mark which is a small magnetic field generator is located on the surface of the substrate which is being subject to processing. The magnetic field generator, which is typically a small area of ferromagnetic particles which are magnetically aligned, forms a magnetic field which penetrates through any overlying resist or oxide layer. The magnetic field of the registration mark is strong enough to alter the trajectory of secondary electrons produced by the incident low-energy electron beam. A suitably configured detector detects this alteration in the secondary electron beam caused by local presence of the magnetic fields. An example of a suitable registration mark is a small rectangular area of ferromagnetic particles.

Description

ALIGNMENT OF LOW-ENERGY ELECTRON BEAMS

USING MAGNETIC FIELDS

FIELD OF THE INVENTION This invention relates to lithography alignment, and particular to detection of alignment marks for low-energy charged particle beam lithography, including electron beam lithography.

BACKGROUND

The use of registration or fiducial marks in lithography and especially in lithography for semiconductor processing is well known. Registration marks are used to align one pattern layer of metal, insulator or semiconductor material on a substrate with another pattern layer to insure that features of the successive layers bear the correct spatial relationship to one another. The registration (alignment) marks are typically used to align the substrate with the lithographic writing tool being used, such as optical or electron beam lithography. It is to be understood that such lithography is typically used both for fabrication of semiconductor wafers and for fabrication of masks used for photolithography to fabricate semiconductor wafers. In optical lithography, the registration mark is typically observed with an optical scanner. Although this method may also be used with direct writing electron beam lithography (scanning lithography) , if the registration mark is under a layer, for instance of resist, the registration mark is conventionally "observed" by detecting the back scattered electron current generated when the electron beam contacts the registration mark.

It is to be understood an optical registration mark is typically a raised or lowered area, having a predetermined shape, on the principal surface of the substrate. In the case of electron beam lithography, a registration mark is typically a small conductive area of material different from that of the substrate, or alternatively a raised or lowered area on the substrate. The raised or lowered registration marks are formed, for instance, by etching to define either a depression or to remove surrounding material to define a raised area, and have shapes such as crosses or rectangles which are readily detected.

A conventional electron beam used in direct writing lithography typically has a high energy level in excess of 10 keV and up to 50-100 keV. Such a high energy electron beam easily penetrates a layer of resist (resist is the sensitive material coated onto a substrate which is exposed by the electron beam) having a thickness of 2,000 A to 2 μm and contacts the underlying registration mark. As the electron beam penetrates the resist layer, back scattered electrons are produced. By detecting the contrast in the back scattered electron current (signal) caused when the electron beam contacts the registration mark underlying the layer of resist, the location of the registration mark is determined. The electron beam is then aligned accordingly. This alignment is also suitable, for instance, for marks buried under a layer of oxide or other material, so long as the material is transmissive to the incident electron beam. The back scattered electrons are typically detected by an electron detector. As the incident electron beam is scanned across the resist layer, the contrast in the back scattered electron beam' s current detected by the detector indicates the location of a mark.

Thus to detect underlying (buried) registration marks, conventional electron beams operate at an energy level sufficient to penetrate the resist layer. If the electron beam does not have sufficient energy to penetrate the resist layer and contact the buried registration mark, there is no contrast in the back scattered electrons to indicate the location of the registration mark and hence this approach is inoperable. An example of an electron beam that is unlikely to have sufficient energy to penetrate a typical layer of resist or, for instance, oxide, is a miniature electron beam microcolumn

("microcolumn") . Microcolumns are physically short electron beam columns that output low-energy electron beams, currently 1-2 keV, and thus may have difficulty detecting registration marks underlying a layer of resist having a thickness greater than, for instance, 1,000 A. (The mean free path of electrons having an energy of 1-2 keV in resist is on ^• the order of only 60 nm. ) Microcolumns are well known structures based on microfabricated electron "optical" components and field emission sources and may be used for direct writing lithography. Microcolumns are well known; see, for instance, "Electron-Beam Microcolumns for Lithography and Related Applications" by T.H.P. Chang, et al., Journal- of Vacuum Science Technology Bulletin 14(6), pp. 3774-81, November/ December, 1996, incorporated herein by reference.

Thus, to detect a conventional registration mark (whether conductive or raised/lowered) , the electron beam must operate at an energy level sufficient to penetrate the overlying resist or oxide layer. If the resist or oxide layer has a thickness greater than the penetration depth of the incident electron beam, one method to permit the electron beam to contact the registration mark and to generate the contrasting back scattered electron current signal is simply to remove the resist in the area directly over the registration mark. The registration marks are made accessible by first exposing a large blanket region over each mark, and then developing these regions, thus eliminating the portion of the resist overlying the marks. A second exposure step (using the low-energy electrons from the microcolumn) is then used to pattern the desired lithographic pattern. Disadvantageously this approach adds costly extra process steps and significantly lowers throughput. Hence it is not very desirable.

For another approach see co-pending and commonly owned U.S. Patent Application Ser. No. 09/060,496, filed April 14, 1998, entitled "Detecting Registration Marks With a Low-energy Electron Beam," Tai-Hon Philip Chang et al . , incorporated herein by reference. This discloses detecting a buried registration mark by applying an AC voltage to the registration mark to capacitively induce a voltage potential on the surface of the resist layer. The incident electron beam thus generates low-energy secondary electrons that are affected by the voltage potential locally present on the surface of the layer of resist immediately overlying the registration mark. The detected contrast in the secondary electron signal indicates the location of the registration mark. This approach does require, however, provision of suitable local voltage to the target substrate in order to generate the local voltage potential in the area of the mark.

SUMMARY In accordance with this invention, an approach somewhat related to that of the above-referenced U.S. patent application is used. However, instead of applying a local voltage, the alignment marks in accordance with this invention are of a type that each generate a local magnetic field. For instance, each mark is a small permanent magnet area on the target substrate. These local magnetic fields alter the trajectory of the secondary electrons produced by the incident primary electron beam. A suitably configured electron detector located over the target substrate (and thus over the resist or oxide layer) detects the change in the secondary electron beam emission, thereby allowing location of the magnetic registration marks. This approach of providing magnetic field generators on the substrate is not limited to low-energy electron beams (microcolumns) , but is generally applicable to all types of charged particle beams, including ion beams. Therefore, located on the target substrate, for instance a semiconductor wafer or mask blank, are small magnetized (e.g. permanently magnetized) areas which provide registration of the electron beam. The detector, sensitive to the effects of magnetic fields on the trajectory of the secondary electrons, allows detection of the buried alignment marks. This arrangement advantageously allows production of two dimensional alignment images that convey information about the location of the local magnetic fields.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 shows a perspective view of a portion of a target substrate having (schematically depicted) magnetic alignment marks located thereon.

Figure 2 is a diagram showing detection of the alignment marks of Figure 1 by a detector.

Figure 3 shows a detector for magnetic alignment marks.

DETAILED DESCRIPTION

This invention is directed to registration (alignment) of low-energy electron beams for beam writing but is not so limited. As described above, alignment marks are fabricated on the target substrate. These alignment marks each generate a magnetic field. This magnetic field alters the trajectory of secondary electrons that are produced by the incident electron (or other charged particle) beam. A suitably configured detector, located above the target substrate, detects the resulting local change in the secondary beam trajectory. The magnetic field is produced by fixed magnets or may be generated dynamically by electromagnets. Figure 1 depicts (schematically) suitable alignment marks. (This is not an actual pictorial representation nor is it to scale.) Figure 1 depicts a substrate 20 which is, for instance, a semiconductor wafer or reticle (mask) blank on which are located a plurality of spaced apart registration marks 24a, 24b, 24c, 24d. It is to be emphasized that these marks are not depicted to scale and are meant to illustrate individual magnetic field generators. Each mark 24a-24d is a magnetic field generator fabricated so that it is oriented having a north and south pole, thus being a permanent, although small, magnet. The size of each alignment mark is such that a uniformly magnetized area is the only energetically favored magnetic configuration.

This means that if the magnet is small enough, it will spontaneously form a north (N) and south (S) pole. It does not need to be magnetized. However, this is not a limitation.

Depending upon the type of magnetic material chosen, one could envision "magnetizing" the marks with an external magnet so that N and S poles are produced prior to lithography. In this magnetic state, a strong external magnetic field is produced through the magnetic poles at the ends of each alignment mark. Moreover, there may be only one such alignment mark or several (as shown in Figure 1) located in whatever arrangement is desired. They need not all have the same N-S pole orientation. The overlying resist and/or oxide layers are not depicted in Figure 1 for purposes of simplicity; these would conventionally overlie the principal surface 20a of substrate 20, and thus also overlie the alignment marks 24a-24d. Also, in other embodiments the shapes in the plane defined by surface 20a of the alignment marks would not be rectangular as shown but might be, for instance, crosses or other readily detectable shapes .

Each alignment mark 24a, etc., shown in Figure 1 in one embodiment consists of an array of small ferromagnetic (e.g., cobalt) particles, arranged in the shape of a rectangle. Such ferromagnetic particles are commercially available and typically are adhered to the surface 20a of the substrate 20 and patterned by, e.g., optical lithography and etching. For example, each alignment mark 20a is a rectangular area having dimensions of 0.4 μm long by 0.2 μm wide by 0.2 μm thick, thus producing a magnetic field of 0.2 T (Tesla) at a distance of

0.1 μm above the individual marks which is detectable above the resist. The important characteristic of the magnetic field generated by each mark is that it extends vertically through any overlying layers such as resist or oxide or polycrystalline silicon (polysilicon) . Hence the strength of the magnetic field is determined by the intended thickness of the overlying resist and/or oxide or other layers. Of course, such a method is not operative if the overlying layer is a non-magnetic metal, due to the metal interfering with the magnetic field. Each mark 24a is composed of a large number of small particles. Fabrication of such marks could use a technique, for instance, similar to that used to fabricate magnetic recording media, such as videotape or disks for hard disk drives.

Use of the Figure 1 structures is illustrated in a side view in Figure 2 where structures identical to those of Figure 1 are identically labeled. In Figure 2, only a small portion of the substrate 20 is shown with the single alignment mark 24a, overlain by the resist layer 28. The magnetic field (curved lines) from mark 24a extends vertically through the resist layer 28. A conventional microcolumn 30 which includes a thermionic source 32 and suitable deflection 34 and focusing (lens) structures 36 outputs a focused electron beam 38 which is directed towards the substrate (target) 20 and is scanned thereover. Beam 38 impinges on the resist 28 and as a result secondary emissions (electrons) travel upwards. The presence of each of the alignment marks, for instance mark 24a, is detected due to the resulting changes in the trajectories of the secondary electrons emitted from the resist 28 surface in the vicinity of the mark 24a. The trajectories, of course, are altered by the presence of the magnetic field which has the well known effect (due to the Lorentz force) upon the secondary electron emissions of slightly bending the trajectory.

Figure 2 depicts one way to detect this effect of the magnetic fields using a split detector 40, which is shown (also in side view) as being coaxial to the incident electron beam 38 and defining a central hole for passage of the electron beam 38, with the central hole being small enough so that the secondary electrons 42 are readily detected.

In the Figure 2 embodiment, the detector 40, which is typically circular in plan view, is split into two halves 40a and 40b, which generate output signals respectively 12 and II, which are coupled to a differential amplifier 46. Amplifier 46 determines any difference between signals II and 12 resulting in an output difference signal. The detector 40 is centered about the optical axis of the primary electron beam 38 and is located either internal or external to the microcolumn.

If the beam 38 is located distant from all of the marks, a symmetric angular distribution of secondary electrons is emitted, resulting in a null output signal from amplifier 46. As the primary electron beam 38 is directly over a particular mark 24a, the angular distribution of the secondary electrons is shifted towards one or the other of the detector halves 40a, 40b due to the Lorentz force exerted by the magnetic field. This results in an output signal from amplifier 46 that to the first order is proportional to the magnitude of the secondary electrons deflection.

To estimate this effect in one example, assume that alignment mark 24a is a cobalt ferromagnetic particle area as described above. If this area is buried under a layer of resist 28 having a thickness of 0.1 μm, and the secondary electrons leave the resist surface in a perpendicular direction with an energy of 5 electron volts, then the secondary electrons are deflected by an angle of 2^»10^~3 radians. This would cause the output signal from detector 46 to change by a value of approximately 0.1%. Since the magnetic field produced by the mark 24a is perfectly registered to the mark, that is the magnetic field extends vertically from the mark as shown in Figure 2, the primary electron beam 38 can be registered to the mark 24a by mapping the output signal of the detector 46 as a function of the electron scan voltage. (This is the signal voltage applied to the column which controls the scanning of beam 38.) A detector as in Figure 2 split into two halves allows registration, of course, in only one direction, depicted as the x direction in Figure 2. By splitting the detector into four quadrants, and using an appropriate arrangement of summing and difference amplifiers, deflections in both x and y directions are simultaneously monitored as shown in Figure 3.

Figure 3 thus shows a quad (quadrant) detector 40' which is a modified version of detector 40 of Figure 2 and having four quadrants (segments) respectively 40a, 40b, 40c and 40d, operative as shown in both the x and y directions. ^' It is understood that Figure 3 is a plan view of the detector, in contrast to the side view in Figure 2. The electron beam to which detector 40' is coaxial is shown at 38. Each detector quadrant 40a, 40b, 40d, 40c outputs current signals II, 12, 13 and 14, respectively, and has an associated current to voltage converter 50a-50d, respectively. The output signals of these converters are each connected to two of the differential amplifiers, 54, 56, 58 and 60, the output signals of which are coupled to summing amplifier 68, 70 and 72 to provide the indicated output signals ΔI_Y, ΔI_X and I_γ + I_x. Appropriate signal processing of these output signals therefore results in detection in both the x and y dimensions of the location of each magnetic mark. It is to be understood that other types of detectors and signal processing for the detector signals are also within the scope of this invention. The material of the detector quadrants 40a, etc., is conventional and, for instance, is similar to the secondary electron detector of commonly owned U.S. Patent Application Ser. No. 08/726,449, filed October 4, 1996, Lee H. Veneklasen, et al . , now U.S. Patent No. 5,900,667 issued May 4, 1999, incorporated herein by reference. This discloses an annular shaped solid state detector of material well known in the charged particle beam field, typically consisting of a planar diode structure on a thin substrate of high purity doped silicon. Each detector quadrant is a thin doped layer covered with an even thinner conductive overcoating. A bias voltage is applied to the opposing back surface creating a depletion layer in the silicon. The secondary electrons penetrate into the depletion region, forming many electron hole pairs that, when extracted, are an amplified version of the incident electron current signal. Of course, the detector is scaled to be of the appropriate size and signal processing capabilities depending on the expected energy of the primary electron beam and the resulting energy and current of the secondary electrons.

This disclosure is illustrative and not limiting; further modifications will be apparent to one skilled in the art in light of this disclosure and are intended to fall within the scope of the appended claims .

Claims

CLAIMS:We claim:

1. A method of aligning a charged particle beam to a target on which the beam is incident, comprising the acts of: providing a magnetic field generator on the target; directing the charged particle beam onto the target; and detecting secondary charged particles emitted from the target as a result of the incident beam, whereby a trajectory of the secondary charged particles is influenced by the magnetic field.

2. The method of claim 1, wherein the act of detecting comprises locating a detector coaxial to the beam.

3. The method of Claim 2, wherein the act of providing comprises fixing ferromagnetic particles to a principal surface of the target.

4. The method of Claim 3, wherein the particles are arranged in a predetermined shape in a plane defined by the principal surface of the target.

5. The method of Claim 3, wherein the particles are overlain on the target by a layer of material.

6. The method of Claim 5, wherein the magnetic field extends at least to an exposed surface of the layer of material .

7. The method of Claim 5, wherein the material is resist or oxide or polysilicon.

8. The method of Claim 1, wherein a plurality of spaced apart magnetic field generators are provided on the target.

9. The method of Claim 8, wherein each magnetic field generator defines a magnetic north and south pole.

10. The method of Claim 2, wherein the detector includes a plurality of detection regions.

11. The method of Claim 1, wherein the charged particle beam is an electron beam having a voltage less than about 2 keV.

12. A charged particle beam system comprising: a source of a beam of the charged particles; a lens coaxial to the beam and which focuses the beam; a support for a target onto which the focused beam is incident; and a detector coaxial to the beam and located above the target, wherein the detector detects an influence on a trajectory of secondary charged particles emitted from the target by local magnetic fields on the target.

13. The method of Claim 12, wherein the source and lens are incorporated into a microcolumn.

14. The system of Claim 12, wherein the detector includes a plurality of detection areas.

15. The system of Claim 12, wherein the charged particle beam is an election beam having a voltage less than about 2

KeV.

16. An article of manufacture comprising: a semiconductor wafer; a plurality of spaced apart ferromagnetic areas on a principal surface of the wafer, each area defining a north and a south magnetic pole; and a layer of resist or oxide or polysilicon over the areas, wherein a magnetic field of each area extends through the layer.

17. An article of manufacture comprising: a reticle blank carrying a layer of material opaque to light; a plurality of spaced apart ferromagnetic areas on the wafer, each area defining a north and a south magnetic pole; and a layer of resist or oxide or polysilicon over the areas, wherein a magnetic field of each area extends through the resist or oxide or polysilicon layer.