GB2235148A - Fabrication of electronic devices. - Google Patents

Abstract

A method of fabrication of electronic devices comprises bombardment of a metal-containing substrate using a narrow beam of electrons to leave a pattern of said metal in predetermined positions on said substrate.

Description

Fabrication of electronic devices This invention relates to the fabrication of electronic devices.

One embodiment of the invention is concerned with making very fine wires. With a high enough electron beam current a hole formed. In some materials, such as Al F3, a pad of metal is left in the hole and, by causing the beam to scan, a wire can be formed in situ. Such a wire may be full of voids, but the whole specimen can be heated treated and, since the matrix is a ceramic, a very high melting point ceramic, the wire can be melted and allowed to resolidify.

If the metal is magnetic it can form the basis for magnetic storage.

Another embodiment of the invention is concerned with the fabrication of field emission cathodes for ultra small vacuum tubes. For normal field emission, thousands of volts potential differences are required. With conventional photolithography, field emission tips on the micron scale are feasible and the voltages fall to about 10 volts. With the present invention, we can make field emitters which are on a nanometre scale and it becomes practicable to have devices with low voltage field emitters.

Another embodiment of the invention is concerned with free electron lasers. In a free electron laser there is a periodic array of magnets. An electron beam passes through them and is caused to oscillate. Oscillating charges radiate. The frequency of the radiation is just the periodicity of the electron beams. Such a laser would be useful, for say, crystallography.

According to the present invention there is provided a method of fabrication of electronic devices comprising bombardment of a metal-containing substrate using a narrow beam of electrons to leave a pattern of said metal in predetermined positions on said substrate.

An embodiment of the invention will now be described by way of example with reference to the accompanying drawings in which: Figure 1 is a schematic drawing of one form of apparatus used in the invention; Figure 3 is a micrograph of a crystal of MgO having holes drilled therein; and Figure 3 is an electron micrograph of a pattern produced in a specimen of Al F3.

Referring now to the drawings, Figure 1 shows a scanning electron microscope. The electron source is a field electron gun FEG with a cathode C incorporating an orientated single crystal of tungsten, thinned to a very small diameter to act as a very small electron source. The advantage of such a source is that it permits very high brightness with a very small diameter probe. The gun is positioned at the base and is fed by a high tension cable HT. The accelerating voltage used is typically 100KV, but a wide range of voltages can be used.

The apparatus is evacuated to a high vacuum and requires to be baked from time to time. The working vacuum is of the order to l0#11mBar. Lenses L focus the beam on to the specimen S.

The specimen is positioned above the lenses in a vacuum of about 10'10mar. As the beam is scanned, the various signals which are generated are picked up on detectors positioned around the microscope. Included amongst the detectors is a photomultiplier PM which picks up electrons which are elastically scattered and produces a dark-field image; the straight transmitted beam goes through a spectrometer SP, bent through 900 and hits another detector DS to give a bright field image. This spectrometer is for energy-filtered electrons. By looking at electrons which have lost energy as they pass through the specimen, it is possible to identify the chemical composition of the material.

There is also an x-ray detector which can be used to collect x-rays generated when the beam hits the specimen - the particular x-ray detector is positioned very close to the specimen, has a very high solid angle of detection and, with a 2nm probe, it is possible to determine the chemical composition from an area containing about 100 atoms or so. It is a windowless detector, which makes it possible to detect carbon, oxygen, and other light elements, as well as all the hearer elements. Adjacent to the basic column is an ultra-high vacuum preparation chamber into which a specimen may be introduced via an air lock. It is also possible to attach an ion gun which could be used for wiping clean a surface or for evaporating materials.

The whole instrument is pumped by diffusion pumps, which are provided with cold traps. In addition, the column is pumped by means of a titanium sublimation pump and the gun is pumped by an ion pump. Another feature of the apparatus is that it is provided with a cold stage cooled by liquid nitrogen so that the specimen can be run at liquid nitrogen temperatures.

The control electronics for the microscope includes a multi-channel analyser for displaying the x-ray information.

The electron beam scan is controlled by a computer. Information is generated on a keyboard and written to the computer. It is then stored in an array of 512x512 points. It is converted to a binary image and then that image data is replayed on to the specimen via a digital scan which is generated by the computer.

Alternatively, a video camera may be used to transfer an image to a frame store. These data can then be used to write on to a specimen. The data may, for example, be a circuit layout or simply information to be stored.

The control apparatus and software also displays the x-ray data and the energy loss data and can be used for acquiring images from the microscope which can then be processed by modification of the contrast, etc.

A vacuum rack monitors the vacuum in various parts of the apparatus.

Magnesium oxide specimens are made by burning magnesium in air and passing a grid through the smoke. This results in the deposition of very small cubes of magnesium oxide of side 70nm.

Figure 2 shows a specimen magnified about 3-5 million times.

When drilling holes in magnesium oxide a circular cross-section electron beam produces square cross-sectional holes. We have found there is a small mark on the electron entrance surface, but the actual creation of the hole starts on the exit surface and then spreads back through the material.

The profile is trumpet-shaped, probably due to beam broadening by scattering as the beam penetrates. The hole progresses in steps. As it reduces in diameter, the steps are probably half lattice spacing or full lattice spacing. The sides of the hole are < 100 > surfaces, which are energetically favourable and, in fact the < 100 > planes are so energetically favourable over all others that is probably the reason for the formation of facetted holes. In most other materials they are round. The smallest hole we have created is about 8A, and, using this technique, it is possible to write a circuit with a line width of 8A.To drill such a hole (it depends on the thickness of the crystal) takes from 2 to 5 minutes, and the greater the thickness of the crystal, the greater the line width that is generated The current density profile of the beam is a Gaussian function which depends on the electron optical conditions. It is very highly peaked with at least 70%, a substantial proportion, in the central region, the central l.Onm diameter, but we have shown that, under other conditions, the wings of the beam step back to 5 or lOnm. The electron beam current density is typically 107Am~2 or more.

The reproducibility of the beam under the same conditions, using the same apertures is very good. The positional accuracy is also very high. The stage which holds the specimen is very stable and there is practically no drift once an initial period has elapsed. It will remain on a single point and will stay there for five or ten minutes, even with holes as small as 8A.

Figure 3 shows an example of the accuracy with which the beam can be controlled. The specimen in this case is A1F3, which has been evaporated as a thin layer on to a carbon support film. An image was created using the control computer. This was converted to a binary image from which the computer generated the scanning electron beam and drove it in this pattern. Where there is a character in the raster, then it will dwell for a specified dwell time; where there is no data, it just skips quickly. For this writing AlF3 is drilled very quickly requiring about lOms per picture point. It took only about l5mins to write the image shown. At this density of data storage, it is possible to store the whole of the Encyclopaedia Britannica on approximately 4mm2, all 29 volumes.This size was chosen to give the optimum sharpness. By sacrificing optimum sharpness, it is possible to write significantly smaller. The actual individual holes are similar to the beam diameter in their dimensions.

In information storage another electron microscope is required as a reader. The system may be used for archival storage. An advantage is that the medium is a high melting point ceramic which can survive many fires. For instance, if the engraving is on sapphire, only really bad fires would damage it. The material is also unaffected by magnetic fields.

A further application is producing apertures for scanning optical microscopes. These require very small holes to be produced in a material which is not transparent to light.

Currently these use glass syringe needles which are drawn down and then coated with aluminium. The problem with these is that they are not completely opaque - they do transmit some light.

We can produce a 3mm disc of through aluminium which finds immediate application.

We can scribe parallel lines containing a very small hole which could be used for diffraction grating work for x-ray diffraction. We can put lines and slots through the material.

These are 250A apart and about 10 or 20A in diameter through a thickness of 500A and over a whole array we can make l0#m x lOpm square of grating which give rise to these Molr fringes. It is possible to go right down to 20A spacing and to cover an area of many cm2, given a suitable stage.

We also fabricated arrays of dots to display diffraction effects.

There are several classes of materials which behave differently. There are amorphous, inorganic films such as Awl 203, AlFa - all the fluorides - that drill very rapidly - in milliseconds, where there are probably electronic interactions taking place. Ionisation occurs in the bulk of the material underneath the electron beam. The cations and anlons, then appear to migrate, forming either bubbles of oxygen (for oxides), or fluorine (for fluorides), in this case, underneath the beam. The bubbles eventually coalesce and burst or rupture the surfaces and this leaves a hole with, in this case, aluminium around it, at the edge of the hole and through the whole length of the specimen. Crystalline materials drill much slower, e.g.MgO and sapphire, drilling in minutes rather than milliseconds. The holes drill from both sides, progressing slowly. ~Amorphous alumina, made by anodising aluminium, on the other hand, drills very quickly. Many other materials, such as silicon and metals have a current threshold below which they do not drill at all. We have made holes 2 or 3nm in diameter, 200nm thick, in layers of Al or Si. The beam energies are below the threshold for creating displacements in the bulk and they don't produce the holes in those materials by sputtering. In Al the holes are produced from the electron entrance surface.All of these damage conditions are probably linked to the fact that we have very highly controlled conditions, a very high current density electron beam, very high temperature gradients, even though the temperature of the specimen under the beam is not raised by very much, not more than a fraction of a degree.

There are, however, very high thermal gradients (-1010m-1) and also very high electric fields set up around the beams.

This finds application with high Tc ceramic superconductor devices. With, for example, Josephson devices, interelectrode spacing are orders of magnitude lower than conventional superconductors and, in consequence, much smaller width junctions are required.

With superconductor materials, when investigating segregation of the elements at the grain boundaries, we have found that you can drill the amorphous phase in 1-2-3 yttrium-barium-copper superconductors and they can then be processed into crystalline superconducting phase, with the superconductivity locally destroyed on a nanometre scale by the electron beam, thus making a Josephson junction. We have also found that barium-containing superconductors damage quite rapidly as well.

It is possible, by tilting the specimen, to drill a crystal in other planes and other directions to create a lattice of holes It is also possible to machine materials. Just by moving the beam across it is possible to cut of a corner. This may be used in applications such as the production of silicon motors about lpm in size.

We have found that silver halide particles just disintegrate under the beam although they only disintegrate above 1000K. At low temperatures, however, they are completely stable.

A development of the present apparatus has a beam diameter of about 1.5A, so about 0.lem, making it possible to remove single atoms. There is the possibility of taking a structure such as yttrium barium copper oxide, putting the beam on and just removing a single column of atoms.

We have demonstrated that chemical effects arising from condensed gases can be used to enhance drilling rates, or even to drill materials which are otherwise undamaged by irradiation. Direct electron beam writing on carbon support films is not possible. However, when electron beam writing is performed on an ice layer deposited on carbon, and the ice layer is then removed by allowing the specimen to warm up, the pattern is replicated in the carbon support film. The potential therefore exists for reactive electron beam etching on a nanometre scale using condensed gases, the composition of which is tailored to specific materials.

If a beam of circular cross-section is incident upon most inorganic materials that drill, circular holes result. In a few materials, however, we have observed some evidence of facetting. In the case of MgO this facetting is extremely pronounced. Approximately square cross-section holes with (100) sides cut in a MgO cube with the electron beam (of circular cross-section) incident along a < 100 > direction. Damage has also occurred to the MgO cube in the raster scanning used for imaging, as is particularly apparent at the cube edges and faces. Thickness fringes formed from a tilted cube exhibit square cut serrations, illustrating that regions of partial mass loss due to the scanning beam are bounded by crystallographic faces. The damage in MgO thus has a very marked crystallographic dependence, and the holes have an approximately square cross-section over a 2 to lOnm size range. An array of such holes might be used for a microporous filter, filtering molecules by both size and shape. Unlike sodium ss-alumina, or amorphous alumina, drilling in MgO proceeds from one surface.

All surfaces contain surface steps on an atomic scale.

However, by scanning our intense electron beam across an area of an MgO crystal, it is possible to 'plane' the electron entrance and exit surfaces, leaving them completely free of steps and completely smooth on an atomic scale. Our ability to tailor surfaces has application to various physical and chemical processes, e.g. catalysis, in which the surface steps are believed to play an important role.

Claims (12)

Claims

1. A method of fabrication of electronic devices comprising bombardment of a metal-containing substrate using a narrow beam of electrons to leave a pattern of said metal in predetermined positions on said substrate.

2. A method of fabrication of electronic devices as claimed in claim 1 wherein the current density profile of said narrow beam of electrons is peaked so that a substantial proportion of the charge carriers are concentrated in the central region of said beam.

3. A method of fabrication of electronic devices as claimed in claim 2 wherein the diameter of said central region is of the order of l.Onm. or less.

4. A method of fabrication of electronic devices as claimed in claim 3 wherein the diameter of said central region is of the order of 0.15nm. or less.

-

5. A method of fabrication of electronic devices as claimed in any one of the preceding claims wherein the current density to bombard said substrate is of the order of l07Am#2 or greater.

6. A method of fabrication of electronic devices as claimed in any one of the preceding claims wherein said substrate is bombarded by said electron beam to create a thermal gradient of the order of 1010m'l adjacent the region of impact.

7. A method of fabrication of electronic devices as claimed in any one of the preceding claims wherein said beam of electrons is scanned over the surface of said substrate to leave a wire-like deposit of metal on said substrate.

8. A method of fabrication of electronic devices as claimed in claim 7 wherein said wire-like depositit is subsequently heat treated to remove voids therein.

9. An electronic device made by a method as claimed in any one of the preceding claims.

10. An electronic device including a field-emission cathode made by a method as claimed in any one of the preceding claims 1 to 8.

11. A magnetic storage electronic device made by a method as claimed in any one of the preceding claims 1 to 8.

12. A free electron laser made by a method as claimed in any one of the preceding claims 1 to 8.