JP2009300202A - Method and system for inspecting semiconductor device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method and a system for inspecting a semiconductor device which enable pinpointing of a faulty spot in a short time. <P>SOLUTION: A plurality of spots in an observation object area of the semiconductor device are irradiated locally by laser light as the light is made to scan the area, while the semiconductor device is made to operate for testing at the same time, and the result of this test is made to correspond to the position of irradiation of the laser light to produce a quality determination image made a two-dimensional image, so that the faulty spot may be pinpointed. In the case of producing the quality determination image with the maximum number of pixels allowed for one screen, in this method for inspecting the semiconductor device, the laser light irradiation is performed with the number of points of irradiation lessened and irradiation point intervals widened, while a range of the laser light scanning is left unvaried. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a semiconductor device inspection method and a semiconductor device inspection apparatus.

  With the recent high integration and high performance of LSIs, there is an increasing need for function failure analysis with LSI testers connected. As the analysis method, there is a laser irradiation dynamic analysis method (see, for example, Non-Patent Document 1). When a test pattern is input to a device from an LSI tester and irradiated with laser light, the operating state changes due to laser heating in a wiring defect location such as a void or a transistor with poor characteristics, and the device passes / fails. The state (good or bad state) changes. By capturing this change as a signal and imaging it, it becomes possible to identify a time delay defect or a marginal defect location related to the operation state.

  In the laser irradiation dynamic analysis method, a function test is performed simultaneously with a tester while laser irradiation is performed on a region to be observed of a semiconductor chip to determine whether the semiconductor chip is good or bad. By matching this pass / fail judgment result with the location where the laser is irradiated and displaying each pixel corresponding to each laser irradiation point in light or dark or pseudo color, it is possible to narrow down the marginal failure locations. In order to adopt such a method, it is necessary to complete at least one test pattern for one loop while irradiating laser light to one place, and to end one test.

In other words, the laser irradiation position cannot be moved until at least one test (good / bad determination) is completed for each irradiation point, and the total time of failure analysis is greatly influenced by the test time. As an actual observation sequence, first, a relatively wide area is observed at a low magnification, and after roughly narrowing the suspected failure area, the narrowed area is observed for detailed location identification at a high magnification. In some cases, it was necessary to repeat observation from low magnification to intermediate magnification several times, and the total analysis time tended to be long.
Futagawa Kiyoshi, “All about LSI failure analysis technology”, Industrial Research Institute, Inc., November 2007, p. 103-104

  The present invention provides a semiconductor device inspection method and a semiconductor device inspection device that can identify a failure location in a short time.

  According to one aspect of the present invention, a laser beam is scanned locally on a region to be observed of a semiconductor device, and a plurality of locations are irradiated locally. At the same time, the semiconductor device is subjected to a test operation. A method for inspecting a semiconductor device that generates a pass / fail judgment image that is associated with an irradiation position to form a two-dimensional image and identifies a fault location, and generates the pass / fail judgment image with the maximum number of pixels allowed per screen. In contrast, there is provided a semiconductor device inspection method characterized in that laser beam irradiation is performed by reducing the number of irradiation points and increasing the interval between irradiation points while keeping the same laser beam scanning range.

  According to another aspect of the present invention, a laser light source, and a scanning unit that locally irradiates a plurality of locations while scanning the observed region of the semiconductor device with the laser light output from the laser light source, A tester for performing a test operation of the semiconductor device simultaneously with the laser light irradiation, and an image processing unit for generating a pass / fail judgment image in which the test result of the semiconductor device and the irradiation position of the laser light are associated with each other to form a two-dimensional image In contrast to the case where the pass / fail judgment image is generated with the maximum number of pixels allowed per screen, the laser beam irradiation is performed by reducing the number of irradiation points and widening the irradiation point interval while keeping the laser beam scanning range the same. And a control unit for controlling the scanning unit so that the inspection is performed.

  According to the present invention, there are provided a semiconductor device inspection method and a semiconductor device inspection device capable of specifying a failure location in a short time.

  As a method for shortening the analysis time in the laser irradiation dynamic analysis method, it is conceivable to shorten the test time by shortening the test pattern, or to reduce the number of laser light irradiation points.

  In order to shorten the test pattern, an engineer who has the skill to understand the circuit operation in the target semiconductor device and to create a pattern for testing the circuit when analyzing the failure is required. However, it is difficult to always secure such engineers for actual failure analysis. Also, in recent semiconductor devices with a large circuit scale, test patterns are usually created automatically by CAD (Computer Aided Design) tools, and even when creating short test patterns, it takes time to analyze them. In some cases, a test pattern with a proven track record of detecting a failure is often used as it is. Therefore, in reality, it is almost impossible to shorten the test time itself.

  If the number of laser light irradiation points (corresponding to the number of pixels in the obtained pass / fail judgment image) is reduced, the time required for acquiring the pass / fail judgment image for one screen can be shortened. However, for example, as shown in FIG. 5, the number of irradiation points is simply set as shown in FIG. 5 when the laser beam irradiation and the test are performed on (irradiation number 512 × column number 512) irradiation points. For example, when the number is reduced to (number of rows 128 × number of columns 128), the laser beam scanning range, that is, the range for determining pass / fail is narrowed.

  If it is possible to narrow down in advance a certain range of the location where the failure is suspected, it is possible to omit observing a relatively wide range at a low magnification or intermediate magnification, and at a high magnification for the immediately narrowed range. The observation time can be shortened by only irradiating and testing the laser beam only in a narrow range where the number of irradiation points is further reduced.

  However, Filter generally marginal defects are difficult to narrow the scope or advance where is, or can be a even failure analysis is often asked to issue a short time results, without taking the preliminary preparation period sufficiently In many cases, analysis must be performed without any problems. When the laser beam scanning range is narrowed to reduce the number of laser beam irradiation points, if the suspicious point of failure is not narrowed down, the short range of observations in the narrow range will be repeated many times, which is ultimately necessary in total. Therefore, the observation time will not be shortened.

  In view of the above circumstances, in the actual failure analysis scene, a method that can shorten the observation time without prior narrowing down of failure locations is desired.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  In general, a product that does not satisfy the function during the manufacturing process or at the time when the manufacturing is completed is called a defective product, and the phenomenon is called a defective product. A phenomenon in which a non-defective product at the time of completion of manufacture becomes defective afterwards is called a failure. A faulty product is called a faulty product or a defective product. In this specification, the term “failure” is used to indicate a defect.

  FIG. 1 is a schematic diagram showing a configuration of a semiconductor device inspection apparatus according to an embodiment of the present invention.

  The semiconductor device inspection apparatus according to the present embodiment includes laser light sources 11 and 12, a scanning unit 13, a tester 14, a processing device 15, and a display device 18 as main components.

  The laser light source 11 and the laser light source 12 have different wavelengths of laser light to be output. The laser light source 11 outputs infrared laser light having a wavelength of 1.3 μm, for example. The laser light source 12 outputs laser light having a wavelength (for example, 1.06 μm) capable of generating a photoexcitation current in an active region (a region including a pn junction) of a transistor formed in the semiconductor device 10 to be inspected. The infrared laser light from the laser light source 11 does not generate a photoexcitation current in the semiconductor device 10. Note that one laser light source capable of switching and outputting two different wavelength laser beams may be used.

  Laser light output from the laser light source 11 or the laser light source 12 is converged finely by a microscope (not shown) or the like, and irradiated to an observation region in the semiconductor device (semiconductor chip) 10 in which integrated circuits such as transistors and wirings are formed. Is done. This laser light is irradiated locally at a plurality of locations in the observed region of the semiconductor device 10 while scanning a predetermined range by the scanning unit 13.

  The tester 14 inputs a test pattern to the semiconductor device 10 receiving the laser beam irradiation, and causes the semiconductor device 10 to perform a test operation. The “test operation” here is not limited to a state in which a test pattern is input, but also includes a state in which a constant voltage or a constant current is applied.

  In addition, the tester 14 detects a change in the state of the semiconductor device 10 that has been irradiated with the laser light and is in a test operation, and performs pass / fail determination based on the change in the state. “State change” means resistance change, voltage change, current change, change in Pass / Fail state (good or bad state), and the like. While one irradiation point is being irradiated with laser light, at least one test pattern is completed once around one loop, and a quality determination result corresponding to each irradiation point is obtained. In other words, one target irradiation point continues to receive laser light irradiation until one test is completed.

  The processing device 15 includes an image processing unit 17 and a control unit 16. The image processing unit 17 receives the above pass / fail determination result from the tester 14 and associates the pass / fail determination result with the irradiation point position irradiated with the laser beam when the pass / fail determination result is obtained to obtain a pass / fail image. A determination image is generated and displayed on the display device 18. The control unit 16 controls the scanning unit 13 to control a laser light scanning step (irradiation point interval, scanning line interval, etc.).

  According to the method for inspecting a semiconductor device according to the embodiment of the present invention performed using the above-described device, a failure location of the semiconductor device can be specified. A specific example of specifying a marginal failure will be described below. The marginal failure is a failure that results in a failure (failure) depending on the power supply voltage and timing even within the range of the power supply voltage and timing that should normally operate normally.

  For example, the laser light source 11 irradiates an observed region of the semiconductor device 10 with an infrared laser beam, and at the same time, a test pattern set by a power supply voltage and timing conditions in the vicinity of the boundary between Pass (good) and Fail (bad). 14 to the semiconductor device 10. The resistance of the portion that has been irradiated with the laser light changes due to heating, and the Pass / Fail state may change due to the resistance change at the failure location.

  The tester 14 performs pass / fail determination for each irradiation point from the change in the Pass / Fail state, and the image processing unit 17 generates a pass / fail determination image by associating the pass / fail determination result with the position of the laser light irradiation point to form a two-dimensional image. . In the pass / fail judgment image, the failed part is displayed as a luminance (brightness / darkness) or a color difference with respect to the normal part.

  The above pass / fail judgment image is an image obtained by linking the laser light irradiation and the pass / fail judgment result by the tester 14. However, in the apparatus according to the present embodiment, the unit that performs laser light irradiation must be linked to the tester 14. Also, it functions as a laser scanning microscope that obtains a simple optical image (integrated circuit pattern image) of the observed region of the semiconductor device 10.

  That is, the laser beam irradiation unit further includes a photodetector that detects reflected light of the laser beam that is scanned and irradiated onto the observation region of the semiconductor device 10, and the output signal of this photodetector is the image processing unit 17. To generate an optical image of the observed region.

  Further, the image processing unit 17 generates an image obtained by superimposing the optical image and the quality determination image. The pass / fail judgment image is overlaid with the optical image having the same field of view and the same magnification. From this superimposed image, it is possible to specify where in the integrated circuit the failure location is.

  When identifying the failure location, narrow down the observation range in stages and narrow down the failure location. First, a pass / fail judgment image is acquired at a low magnification (for example, 1 ×) or an intermediate magnification (for example, 5 × to 20 ×) over the entire observation area or a relatively wide range, and the range where a failure point is suspected is narrowed to some extent. In the above, a pass / fail determination image is acquired at a high magnification (for example, 50 to 200 times) for the narrowed narrow range, and the failure location is specified in more detail.

  In high-magnification observation for observing a narrow, narrow range, the observation range is narrow, so the number of laser light irradiation points (corresponding to the number of pixels in the pass / fail judgment image) can be reduced, and the pass / fail judgment image is displayed for one screen. It doesn't take much time to get.

  However, in the case of observing a relatively wide range at a low magnification or an intermediate magnification, the number of laser light irradiation points is large and it takes time to acquire a pass / fail judgment image.

  For example, it takes 10 milliseconds to perform one test while laser light irradiation is being performed for one irradiation point, and the irradiation of laser light is 512 rows per screen × 512 irradiation points (number of pixels) per screen. If this is done, the acquisition time of one screen of the pass / fail judgment image will take 512 × 512 × 10 milliseconds≈2600 seconds≈43 minutes.

  On the other hand, in the embodiment of the present invention, the number of irradiation points is reduced and the irradiation point interval is kept while the laser beam scanning range is kept the same as when the pass / fail judgment image is generated with the maximum number of pixels allowed per screen. Irradiate with laser light. This can be realized by controlling the scanning unit 13 by the control unit 16.

  FIG. 4 shows an irradiation example with the maximum number of pixels, and FIG. 2 shows an irradiation example with the number of irradiation points reduced according to the embodiment of the present invention. 2 and 4, the laser beam scanning range 20 corresponding to one screen of the obtained pass / fail judgment image is the same range (area). In each figure, the irradiation point of the laser beam is schematically shown by a black circle.

  If the maximum number of pixels allowed per screen is, for example, as shown in FIG. 4 (number of rows 512 × number of columns 512), in this embodiment, for example, as shown in FIG. 128) Reduce to pixels. However, while the laser beam scanning range 20 is the same as in the case of the maximum number of pixels (FIG. 4), the number of irradiation points can be reduced without narrowing the observation range by widening the interval between irradiation points as compared with the case of the maximum number of pixels. Can do.

  In the example shown in FIG. 2, the distance between the columns of the respective irradiation points and the distance between the rows are expanded as compared with the case of the maximum number of pixels, and the distance between the respective irradiation points is made equal in the column direction and the row direction. The points are arranged without deviation over the entire scanning range 20.

  Since the region having no irradiation point is a region that is not observed, it is desirable that the irradiation point be present throughout the laser beam scanning range 20 without being biased in order to avoid substantial reduction of the observation region. Further, the number of irradiation points is not limited to 128 × 128 as long as it is smaller than the maximum number of pixels.

  As described above, according to the present embodiment, the number of irradiation points (number of pixels) can be reduced with respect to the maximum number of pixels. For example, in the case of the above specific example, since the number of irradiation points can be reduced from (512 × 512) to (128 × 128), laser beam irradiation is performed in order to obtain a single pass / fail judgment image for the same scanning range 20 as the maximum number of pixels. In addition, the number of points to be tested can be reduced to 1/16 of the maximum number of pixels, and the time for obtaining one screen can be reduced to 1/16. In addition, since the scanning range, that is, the observation range is not reduced, an increase in the number of pass / fail judgment images to be acquired in total, that is, an increase in the entire observation time can be suppressed.

  Such a method of the present embodiment is particularly effective for observation at a low magnification (for example, 1 ×) or an intermediate magnification (for example, 5 × to 20 ×) for performing observation over a relatively wide range. Therefore, it is possible to significantly reduce the observation time at a low magnification or an intermediate magnification without taking time for changing the operating conditions (test pattern) of the semiconductor device or performing prior analysis. As a result, the overall time required for failure analysis can be reduced.

  In addition, the above-described embodiment of the present invention can be realized only by a software approach such as how to control the scanning unit 13 with respect to an existing apparatus for performing a laser irradiation dynamic analysis method. In terms of hardware, it is not necessary to make a change, or even if a change is made, it can be realized with a very small change, so that the cost increase associated with the implementation of the embodiment of the present invention is minimized. Is possible.

  When the laser light source 11 that only heats the irradiated portion is used, when an abnormal portion (for example, an abnormal via having a poor continuity) is heated by being directly irradiated with laser light, a test result different from that when heating the normal portion is obtained. The obtained quality can be determined. However, if the laser beam does not hit the abnormal part, a test result peculiar to heating of the abnormal part cannot be obtained, and the abnormal part cannot be detected. In particular, when the observation is performed with the interval between the irradiation points being increased with respect to the maximum number of pixels as in the above-described embodiment, there is a possibility that the abnormal part is overlooked without being irradiated with the laser beam.

  Therefore, it is effective to use a laser light source 12 that outputs a laser beam having a wavelength capable of generating an optical excitation current (OBIC: Optical Beam Induced Current).

  FIG. 3A shows a part of the observed region of the semiconductor device 10. In the region shown here, active regions (regions including a pn junction) 5a, 5b of the transistor and wirings 6 connecting these active regions 5a, 5b are formed. Further, it is assumed that there is an abnormal via 9 in the middle of the wiring 6. In addition, the irradiation point of the laser beam is schematically shown by a black circle in FIG.

  When the active regions 5a and 5b are irradiated with laser light from the laser light source 12, electron-hole pairs are generated by photoexcitation, and a photoexcitation current is generated if a potential gradient is set at the generation location. . The photoexcitation current generated in the active regions 5 a and 5 b affects the propagation time of the signal propagating through the wiring 6. Therefore, the degree of influence of the signal propagating through the wiring 6 from the photoexcitation current differs depending on whether or not the wiring 6 has the abnormal via 9, and the propagation time of the signal propagating through the wiring 6 is detected by the tester 14 to thereby detect the abnormal via. Even if the laser beam is not directly applied to 9, the abnormality in the wiring 6 connecting the active regions 5a and 5b can be determined.

  Even when the laser light source 12 capable of generating a photoexcitation current is used, the portion irradiated with the laser light is heated. Therefore, if the abnormal via 9 is directly irradiated with the laser light, a resistance change due to heating or a Pass / Fail state is generated. It is possible to make a pass / fail judgment based on the change in.

  FIG. 3B shows a pass / fail judgment image obtained as a result of observing the region of FIG. Pixels P are generated in one-to-one correspondence with laser beam irradiation points.

  In the above-described observation using the photoexcitation current caused by the laser beam not hitting the abnormal via 9 and the laser beam hitting the active regions 5a and 5b, the propagation of the signal propagating between the active regions 5a and 5b via the wiring 6 is performed. Since the pass / fail judgment is based on time, the position of the abnormal via 9 itself cannot be pinpointed.

  Therefore, in the pass / fail judgment image shown in FIG. 3B, the pixels for displaying the active areas 5a and 5b are displayed with different brightness (brightness and darkness) or different colors from other pixels. . This indicates that the active regions 5a and 5b or the wiring 6 connecting them are defective.

  If the range of the fault location is narrowed to some extent by the above method, the position of the abnormal via 9 can be specified by observing the narrowed narrow range at a high magnification.

  According to the observation using the photoexcitation current as described above, the laser beam does not directly hit the failure point in the relatively coarse laser beam scanning at the low magnification or the intermediate magnification, that is, in the laser beam scanning with the relatively wide irradiation point interval. However, it becomes possible to narrow down the suspected part of the failure in a short time.

  In general, the area of the active region is an order of magnitude larger than that of the via, and even if the irradiation point interval is increased, the active region is more likely to hit the active region than the via. However, in order to ensure that the active region is irradiated with laser light, it is desirable that the interval between the irradiation points be smaller than the plane size of the active region.

  For example, in the example shown in FIG. 3A, the row interval of irradiation points indicated by black circles is made smaller than the vertical (row) direction dimension Y of the active regions 5a and 5b, and the horizontal (column) of the active regions 5a and 5b. If the column spacing of the irradiation points is made smaller than the directional dimension X, the laser beam can be reliably applied to the active regions 5a and 5b to generate the photoexcitation current.

  The embodiments of the present invention have been described above with reference to specific examples. However, the present invention is not limited to them, and various modifications can be made based on the technical idea of the present invention.

  It is not limited to performing pass / fail determination based on the Pass / Fail state, but simply detecting a change in current or voltage due to a change in the resistance of the irradiation point by laser light heating, and determining pass / fail from the magnitude or degree of the change. You may make it perform.

  In addition, specific numerical values such as the number of irradiation points (number of pixels) and the laser light wavelength described in the above embodiment are merely examples, and are not limited thereto.

The schematic diagram which shows the structure of the inspection apparatus of the semiconductor device which concerns on embodiment of this invention. The schematic diagram which shows the example of laser beam irradiation which reduced the number of irradiation points by embodiment of this invention. (A) shows the example of irradiation using the laser beam of the wavelength which can generate | occur | produce photoexcitation current in embodiment of this invention, (b) is a schematic diagram which shows the quality determination image of the area | region corresponding to (a). The schematic diagram which shows the example of laser beam irradiation in the maximum number of irradiation points (the number of pixels). The schematic diagram which shows the example of laser beam irradiation which reduced the number of irradiation points by a comparative example.

Explanation of symbols

  5a, 5b ... active region, 6 ... wiring, 9 ... abnormal via, 10 ... semiconductor device, 11 ... laser light source, 12 ... laser light source, 13 ... scanning unit, 14 ... tester, 15 ... processing device, 16 ... control unit, 17 ... Image processing unit, 18 ... Display device

Claims (5)

A laser beam is scanned over a region to be observed of a semiconductor device to irradiate a plurality of locations at the same time, and the semiconductor device is tested at the same time, and the test result and the irradiation position of the laser beam are associated with each other in two dimensions. A method for inspecting a semiconductor device that generates an image of a pass / fail judgment image and identifies a fault location,
Compared to the case where the pass / fail judgment image is generated with the maximum number of pixels allowed per screen, the number of irradiation points is reduced and the irradiation point interval is widened while the laser beam scanning range remains the same, and laser beam irradiation is performed. A method for inspecting a semiconductor device.   2. The method for inspecting a semiconductor device according to claim 1, wherein the wavelength of the laser beam is set to a wavelength capable of generating a photoexcitation current in the semiconductor device.   3. The method for inspecting a semiconductor device according to claim 2, wherein the irradiation point interval is made smaller than a planar dimension of an active region in the semiconductor device capable of generating electron-hole pairs by photoexcitation of the laser beam. A laser light source;
A scanning unit for locally irradiating a plurality of locations while scanning the observed region of the semiconductor device with the laser light output from the laser light source;
A tester for performing a test operation of the semiconductor device simultaneously with the irradiation of the laser beam;
An image processing unit that generates a pass / fail judgment image in which a test result of the semiconductor device and an irradiation position of the laser beam are associated with each other to form a two-dimensional image;
When the pass / fail judgment image is generated with the maximum number of pixels allowed per screen, the number of irradiation points is reduced while the laser beam scanning range is the same, and the irradiation of the laser beam is performed with a wider interval between irradiation points. A control unit for controlling the scanning unit as shown in FIG.
An inspection apparatus for a semiconductor device, comprising:   5. The semiconductor device inspection apparatus according to claim 4, wherein the laser light source is capable of outputting a laser beam having a wavelength capable of generating a photoexcitation current to the semiconductor device.

Images