KR101275943B1 - Subsurface Imaging Using an Electron Beam - Google Patents

Abstract

A method of navigating and determining interruption points for an ultrafine structure by subsurface imaging using an electron beam having sufficient energy to penetrate the surface and produce an image below the surface. In the method of determining the interruption point, the image below the surface is relatively clear at a particular electronic energy, and the user can see that the embedded feature is reached. In the search method, the image since the surface may be formed with reference points or other features to determine the position of the beam on the device.

Description

Subsurface Imaging Using an Electron Beam

1 shows a dual beam (ion bar and electron bar) system to be used to implement a preferred embodiment of the present invention.

2A-2D show images showing metal layers embedded using various electron beam voltages.

3A-3B show images of the embedded metal layer of FIGS. 2A-2D obtained at different system parameters.

4 is a flowchart showing a preferred embodiment used in the semiconductor device.

5 illustrates a device operated using the steps of FIG. 4.

6 is a flow chart illustrating a preferred embodiment of associating system coordinates with workpiece coordinates.

FIG. 7 shows a wafer operated by the method shown in FIG. 6.

The present invention is a method for navigation including interruption point determination using ultrafine features embedded below the surface of a workpiece.

Modern integrated circuits consist of multiple layers of conductors, insulators and semiconductors. Many modern integrated circuits are assembled using "flip chip" technology, in which the circuit is inverted and mounted on a carrier. After the chip is mounted, it is necessary to access the circuit from the back to examine or alter the inner layers of such circuit. It is common for semiconductor wafers to have a thickness of several hundred microns, so it is necessary to remove a sufficient amount of material from the back side of the circuit before reaching the circuit. When accessing the circuit from the back, there is no reference point for navigation. That is, it is not easy to determine exactly where a particular feature in which the circuit is located. Therefore, in order to access the circuit on a flip chip, it is necessary to determine where the material is removed from the backside to expose the circuit and when it is necessary to stop removing the material to avoid damaging the circuit. Determining when to stop milling is called "end pointing".

Removing the backside material is typically carried out over several steps. The first step typically involves a process such as mechanochemical polishing that quickly thins the entire chip while leaving enough material to provide the mechanical strength to handle the chip. The next step involves making a large hole in the material located midway on the temporary position of the circuit feature of interest. Such a process is typically carried out using a laser or ion beam. In the process of quickly removing material, it is impossible to stop at the correct depth, so that the back hole reaches the circuit, and another more accurate process is usually used.

The slow approach from the back of the circuit uses ion beams, with an "end pointing" technique that indicates when adjacent or approaching the feature to be exposed. In one break point determination technique, light is seen inside the hole, and as the hole reaches the transistor region of the circuit, the light induces a current. As the current induced by the optical beam increases, the user finds closer to the transistor region of the circuit.

Another stop point determination technique is Lundquist's U.S. There is what is described in JP 2002/0074494, which uses ion beam milling focused to reach the active transistor region of the circuit from the back. As the ion beam approaches the circuit, charge carriers are generated that cause a leakage current through the transistor. The ion beam is modulated and a frequency responsive amplifier amplifies the power supply leakage current at the modulation frequency. When the current acquires a certain level, the user assumes that the ion beam is very close to the active transistor region. By this method it is possible to know when the user approaches the active transistor region, but no information is provided about where the ion beam is applied to the surface, i.e. near the active transistor region.

Either on the back side of the flip chip or on the front side of a conventional circuit, the general technique for determining when to stop milling is to observe an image of the circuit as the layer is milled. Although an optical microscope can be used to form the image, although the resolution of the optical microscope is about 0.5 μm, it is insufficient to observe circuit features, and may be about 0.1 μm. A more suitable method for observing ultrafine devices is to use charged particle beam imaging, such as using scanning ion microscopy or using scanning electron microscopy.

Charged particle beams, such as focused ion beams or electron beams, scan across the surface. The impact of the charged particle beam causes various particles to be injected, including secondary electrons, scattered electrons, and ions. The number of particles emitted at each point corresponds to the composition and topography at that point. With the brightness of each point in the image corresponding to the number of particles emitted from the surface at the corresponding point, an image is formed on the video monitor. If the image is associated with known information about the circuit, the image may provide information for navigation.

The workpiece is typically supported by the stage. The stage can move in three dimensions ("X", "Y", "Z"), and the movement of the stage and beam are specified and controlled using system coordinates. The workpiece typically has unique system coordinates used by the designer to specify where the various features are formed. By finding a registration mark called a "reference point" contained in a workpiece, it becomes possible to associate the workpiece coordinates with system coordinates so that the user can use the workpiece coordinates to specify a location on the workpiece, The system can move the stage and direct the beam to the particular location (ie navigation). This interconnection is called registration. During milling at the back of the chip, the reference point is not visible, so it is difficult to register the workpiece and find the desired position.

Although imaging techniques are useful for navigation in plane, these imaging techniques have the disadvantage of determining stop points. When using imaging to determine when a layer is exposed, the layer can be damaged before the break point is determined. In addition, in order to find the reference point in the image to determine where the beam is located, it is necessary to undergo trial and error in a relatively large area, potentially damaging each exposed area.

Zaluzec's U.S. "Scanning Confocal Electron Microscope" in Patent No. 6,548,810 describes an electron microscope capable of imaging substrate features. However, because the system uses electrons that are transmitted, the substrate must be relatively thin, and the system must be set up to detect transmission electrons, so it cannot be easily used in existing SEMs.

There is a need for an improved method for navigation or interruption point determination on semiconductor chips or other workpieces with embedded substrate features.

It is an object of the present invention to determine the location of embedded microscopic features, for example to associate the coordinates of a physical system with the coordinates (or computer design information) of an image of the system, or to perform a milling operation that reaches the embedded features. Provides a subsurface observation method to determine when to stop.

When the charged particle beam strikes the surface, secondary electrons and back scattered electrons are generated. The number and amount of secondary electrons and backscattered electrons provide information about the surface. When the ion beam impinges on the surface, it is common for the emitted electrons to provide information about a few nanometers above the surface. When an electron beam with a relatively high energy is directed towards the surface, electrons penetrate to a certain depth at the surface. The specific depth depends on the electron energy, so that the electrons emitted can represent features below the surface.

In a preferred embodiment, an electron beam with energy high enough to penetrate the surface is directed towards the substrate and an image of the subsurface features is formed. The user uses the subsurface image to determine the location of the collision of the beam and direct the beam to the desired subsurface feature. Unlike focused ion beam imaging, where secondary electrons are generated at several nanometers of surface, electrons with sufficient energy can penetrate more than one micron of the surface to provide information about subsurface features. For example, an example of a subsurface feature may be an orientation mark, such as a reference point on an integrated circuit, or any other feature. By observing the reference point, the user can register or correlate the physical surface with a map of the substrate (e.g., CAD data of the integrated circuit), allowing the user to use the map to locate the exact location on the real surface of the beam. Can

Since the depth of observation is determined by the electron energy, the depth below the surface of the observed feature is determined. Therefore, electron beam observation below the surface can also be used to determine the stop point, ie when to stop milling.

Subsurface observation for alignment is particularly useful in back search without exposed features to establish orientation. It can also be used in front alignment when a reference point or other mark is covered by a layer.

1 depicts a dual system 100 useful for implementing the present invention. For example, one suitable system is Model Strata 400 from FEI Company, the applicant of the present invention. The present invention can be implemented using any electron beam system, signal detector that can produce an electron beam with sufficient beam energy, and the resolution required for a particular application.

In the illustrated embodiment, the electron beam column 102 and the ion beam column 104 are oriented at a particular angle to each other, with each column acting on the same spot 106 on the substrate 108 to produce a beam. In other embodiments, the impact points are separate and the stage moves the substrate exactly between the beam impact positions. In such an embodiment, the beams may be oriented at a certain angle to each other to reduce the stage travel distance. Or the beams may be parallel. In another embodiment, U.S. As described in Publication No. 20040108458, the ion beam and the electron beam may be coaxial. The detector 112 detects secondary electrons emitted from the target by colliding with the ion beam or electron beam. Alternatively, backscattered electron detectors, through-the-lens detectors or other detectors may be used.

Those skilled in the art will appreciate that the system 100 may include a number of additional means (eg, a gas injection system 116 for particle beam deposition or reinforcement etching). Although the present invention can be implemented in a low vacuum system, such as a scanning electron microscope (Mancuso et al. US Pat. No. 4785182), the substrate 108 is for example about 10 ⁻⁵ mbar (.001 N / m ² ). It is common to remain in the high vacuum state.

In a preferred embodiment, a focused ion beam is included to deform the workpiece, but the workpiece can also be deformed by laser or electron beam using appropriate etch-assisted chemistry. Therefore, not all embodiments include the FIB column.

Aspects of the present invention include using an electron beam having a high energy sufficient to form an image under the surface (ie, an image of a feature covered by another material). The electron energy used in the present invention is typically larger than the energy used in the scanning electron microscope and less than the energy used in the transmission electron microscope. Preferred electron energies vary with the type and thickness of the material of the layer being covered. In various embodiments, electrons with energy greater than 5 keV, electrons with energy greater than 10 keV, electrons with energy greater than 15 keV, electrons with energy greater than 25 keV, electrons with energy greater than 30 keV, Electrons with energies greater than 50 keV may be preferred. The present invention is not limited by this particular electron energy, where lower energy will be useful for thinner layers and larger energy for thicker layers.

2A-2D show images for observing the inside of a trench created by FIB milling using xenon difluoride as an etch enhancing gas, produced using electron beams of various energies and secondary electron detectors. The substrates shown in FIGS. 2A-2D include metal lines embedded under silicon of about 1 to 2 μm, with FIB-deposited relatively transparent about 1 μm of silicon dioxide present on the silicon. FIG. 2A shows that the electron beam forming the image has an acceleration voltage of 5 kV, showing no detail below the surface of the metal layer. 2B, where the electron beam forming the image has an acceleration voltage of 15 kV, begins to show some circuit details in a portion of the image, perhaps because the silicon layer is thinner in the portion of the image or below the portion. This is due to the charge accumulated in the circuit part. 2C, where the electron beam forming the image has an acceleration voltage of 20 kV, shows more circuit details. 2D, in which the electron beam forming the image has an acceleration voltage of 30 kV, provides sufficient circuit details to navigate the surface associated with the surface or computer-aided design (CAD) data, the optical map of the surface or other indications. Shows.

In order to produce usable images, electron beam process parameters may vary depending on the application. 3A and 3B are the same substrates as in FIGS. 2A-2D and show images by varying the pressure and operating distance of the sample chamber, ie the distance between the electron lens and the workpiece, using an electron beam of 30 kV. Represents the formed image. FIG. 3A shows an image taken under a high vacuum condition of operating distance of 27.7 mm, 10 ⁻⁵ (.0013 N / m ² ), while FIG. 3B shows pressure 10 ⁻¹ mbar (.0013 N / m ² ), 4.9 An image taken at an operating distance of mm is shown.

In accordance with one aspect of the present invention, subsurface imaging may be used to look around features below the surface and to determine when to stop milling to avoid damaging the substrate. . 4 is a flow chart illustrating a method of correlating design data with a physical surface of a device to enable navigation of the device. 5 illustrates device 500 in which the steps of FIG. 4 are performed. Device 500 includes a buried circuit comprising a metal layer 502. In step 400, device 500 is thinned from backside 504 until it has a thickness of about 200 microns by chemical mechanical polishing (CMP). In step 402, the location of the region of interest estimated to include the desired subsurface features is identified, and holes 506 of 200 μm × 200 μm size and about 10 to 500 μm deep are milled in the device, where This hole is centered at the point where it is supposed to carry the circuit. In step 404, a 1 μm × 1 μm hole 512 is milled at the bottom of the hole 506. Periodically, milling is stopped and the bottom of the hole 512 is inspected using an energy electron beam with sufficient energy to observe the subsurface features in step 406. The electrons in the beam preferably have an energy greater than 15 keV, more preferably greater than 20 keV, even more preferably greater than 25 keV, most preferably near 30 keV or more, or near 50 keV or more. The electron energy used will depend on how far the user is looking from below the surface and the capability of the electron column.

When the thickness of the covering material is thin enough for imaging below the surface, the electron beam image shows a noticeable contrast between the metal layer, the insulating layer and the semiconductor layer. The contrast between different kinds of semiconductors is not as sharp. The present invention therefore facilitates the observation of the metal below the surface, which is useful for orientation in the substrate.

As estimated from the milling speed and the thickness of the material on the metal layer, as the bottom of the hole 512 approaches the metal layer 502, the user stops milling and in step 406 the electron beam of sufficient energy Acquire an image below the surface. First, if the user carefully stops milling before reaching the metal layer, the electron beam image will typically not show the metal layer. This is because the semiconductor material on the metal layer is so thick that the electron beam cannot penetrate it. As shown in decision block 410, if the metal layer 502 is not visible, the user continues milling at step 412. As the material is further removed, the user periodically acquires an image below the surface, repeating from step 406 to step 412. As the bottom of the hole 512 reaches the metal layer 502, the user will begin to see the faint shape of the metal lines in the subsurface image for the first time.

As the lower portion of the hole 512 gets closer, the image of the metal layer 502 becomes clearer, and depending on the electron beam energy, the image of the metal line buried below 1 μm to 2 μm in the material, allows the user to It may be clear enough to determine which of the beams should be directed. At step 420, the user can visually navigate the feature 508 with a beam. Optionally, at step 422, the user can correlate an image of the physical circuit with a known map of the circuit, and at step 420, assist in locating the desired feature 508 on the circuit. Then, at step 424, the user can handle the exact feature or desired position without damaging other areas due to the effort to find the beam position. The sharpness of the image below the surface also provides information about the depth of the metal layer present below the surface, so that the user can stop milling before the circuit is unintentionally damaged. Thus, the present invention is useful both in searching for planes and in searching for three dimensions for determining stop points.

In another aspect of the present invention, a user may use a subsurface imaging to observe a buried reference mark, thereby referring to a physical specimen (eg, an optical image taken at the device in the process step when CAD data or features are exposed). Align with the image. In some processes performed during assembly of an integrated circuit, a layer is deposited covering the reference point used for alignment and search on the device. Although the reference point may appear to be an area protruding from the covering layer, the signal may be sharper when the surface is flattened, i.e., polished to produce a smooth surface to prepare the next processing layer.

6 is a flow chart showing the preferred steps for another embodiment of the present invention. 7 shows a wafer 700 comprising multiple elements or circuits 702, each circuit comprising multiple reference points 704 created on the wafer by a photolithographic pattern supplied to the wafer to create the circuit. do. The circuit and the reference point are covered by a material, for example a metal or an insulating material, which material is deposited on the wafer in a previous assembly step. In step 600, the user searches an area estimated to include the first reference point 704, and fires a relatively high energy electron beam over a wide area at low magnification to find the reference point. The electron beam energy is "relatively high". That is, typically greater than the energy of the electron beam used only in the image surface features. In step 602, the user searches for the area of the reference point found in step 600. In step 604, the user can determine the position more accurately by observing the reference point at a higher magnification and the system can determine the location of the reference point on the system coordinates. Based on the determination of the position and orientation of the first reference point, in step 608, the user can navigate the general second reference point area and locate the embedded reference point using a relatively high energy electron beam at a low magnification. have. In step 610, the second reference point is observed at a higher magnification, and the system records the coordinates in the system suppository system. Typically, at step 614, the user searches for a buried third reference point and observes the vicinity of the reference point using a relatively high energy electron beam at a magnification low enough to locate the reference point. The third reference point is observed at a higher magnification, and at step 618 the system records its coordinates. In step 620, the position of the reference point of the system coordinate system is associated with the workpiece coordinate system as data from a computer aided design (CAD) database. Or the reference point position is associated with an inspection system such as another microscope that provides accurate stage placement and position readings that can be used for coordinate determination. In step 622, the user can navigate to any point on the workpiece, which is done using the CAD coordinates of the point, which translates into system coordinates, which allows the workpiece stage to be moved and the beam to be fired. Can be.

The process shown in the flowchart of FIG. 6 can be easily automated using, for example, image recognition software from Cognex Corporation (Machachusetts, Natick).

“Navigate” herein refers to determining the location, moving the X and Y directions on the substrate, as well as determining the stopping point (ie, as the user approaches the embedded feature, milling). Determining a vertical position to assist). The present invention is useful in any multilayer substrate that includes, but is not limited to being used on integrated circuits, ultrafine features covered with other materials.

While the advantages of the invention have been described in detail, various modifications are possible and alternatives and modifications are possible within the spirit and scope of the invention. For example, the present invention is not limited only to the integrated circuit described above, and any kind of ultrafine element having embedded features is useful. In addition, the scope of the present invention is not limited to specific embodiments such as processes, machines, manufacturing, and the like. Those skilled in the art will understand the processes, machines, preparations, compositions, methods, means, steps, etc. of the present invention.

Subsurface observation to align is particularly useful in backside searching without exposed features to locate. It is also useful in front alignment when the reference point or other marks are covered by layers.

Claims (20)

A method of processing a semiconductor device using ion beam milling and electron beam imaging, the method comprising Directing the focused ion beam towards the semiconductor device, milling a hole reaching the buried metal layer; Directing a beam of electrons containing electrons with energy to obtain a subsurface image towards the bottom of the aperture and collecting electrons emitted from the aperture to pass through a silicon layer of Observing the image below the surface, Using an image of the metal layer to determine when to stop milling using the focused ion beam Semiconductor device processing method comprising a. The method of claim 1, Using the subsurface image to identify subsurface features for orientation or processing A semiconductor device processing method further comprising. 3. The semiconductor device of claim 2, wherein using the subsurface image, identifying subsurface features for orientation or processing includes correlating the subsurface image with a map of the semiconductor device. Treatment method. 2. The method of claim 1 wherein the electrons in the electron beam have an average energy of greater than 15 kW. The method of claim 1, wherein the electrons in the electron beam have an average energy of greater than 25 Hz. 2. The method of claim 1 wherein the electrons in the electron beam have an average energy of greater than 30 Hz. delete delete delete delete delete delete delete delete delete A method of using a focused beam to make a hole in a substrate to expose a microscopic subsurface feature in the substrate, the method comprising: A subsurface image of the submicrosurface feature with a beam of electrons having energy capable of producing an image of the submicrosurface feature from the secondary or scattering electrons, towards the surface covering the submicrosurface feature. determining when the beam attempts to expose features below the ultra-fine surface And using the focused beam to make holes in the substrate, the method comprising: a. 17. The method of claim 16, wherein the thickness of the material covering the features below the ultrafine surface is greater than 1.0 [mu] m. The method of claim 16, wherein the thickness of the material covering the features below the ultrafine surface is between 1.0 μm and 2.0 μm. 17. The method of claim 16, wherein the sub-microscopic features comprise a metal or metal layer. 20. The method of claim 19, Stopping the use of the focused beam to make a hole when the beam attempts to expose features below the ultra-fine surface The method of using a focused beam, characterized in that further comprises.

Images