JP5965952B2 - Defect inspection method and defect inspection apparatus - Google Patents

Description

  The present invention relates to a surface defect inspection method and an inspection apparatus for inspecting minute defects existing on a sample surface with high sensitivity and high speed.

  In a production line for semiconductor substrates, thin film substrates, and the like, in order to maintain and improve product yield, inspection of defects existing on the surface of semiconductor substrates, thin film substrates, and the like is performed. As prior arts, JP-A-9-304289 (Patent Document 1) and JP-A 2000-162141 (Patent Document 2) are known. In order to detect minute defects, a laser beam condensed to several tens of μm is irradiated on the sample surface to collect and detect scattered light from the defects.

JP-A-9-304289 Japanese Unexamined Patent Publication No. 2000-162141

  In recent years, LSI wiring has been rapidly miniaturized, and the size of defects to be detected is approaching the detection limit of optical inspection. According to the semiconductor roadmap, as of 2007, mass production of 65nm node LSIs is about to begin, and the ability to detect defects that are about half the size of DRAM 1/2 pitch is required.

  It is known that the magnitude I of scattered light generated when a defect is illuminated with a laser has a relationship of I∝d ^ 6 where d is the particle size of the defect. That is, as the defect size decreases, the generated scattered light decreases rapidly. Methods for increasing the generated scattered light include shortening the illumination wavelength, increasing the output of the laser, and reducing the laser illumination spot. The detection sensitivity improvement by shortening the wavelength will be described. If the illumination wavelength is λ, there is a relationship of I∝λ ^ (− 4) with the size I of the scattered light. That is, by shortening the illumination wavelength, the generated scattered light can be increased, which is effective in improving detection sensitivity. However, shortening the illumination wavelength generally increases the absorption coefficient of the object and increases the temperature rise on the sample surface.

  The improvement of detection sensitivity by increasing the output of the laser will be described. The magnitude of the scattered light is approximately proportional to the magnitude of the laser output, and the scattered light can be increased by increasing the output of the laser. However, the temperature rise at the sample surface increases in the same way as when the illumination is shortened, and the detection sensitivity cannot be expected to increase beyond the current level at higher output.

  The detection sensitivity improvement by reducing the illumination spot will be described. By reducing the illumination spot, the scattered light from the wafer roughness (small surface irregularities) can be reduced, and the detection sensitivity can be improved from the viewpoint of reducing noise. However, since the laser illuminance per unit area increases due to the reduction of the beam spot, the temperature rise on the sample surface increases.

  As described above, in the extension of the conventional method, damage to the sample due to temperature rise becomes a bottleneck, and further improvement in detection sensitivity is severe. An object of the present invention is to provide a defect inspection method and apparatus for improving detection sensitivity by suppressing a temperature rise on a sample surface.

In order to solve the above-described problems, the present invention illuminates a sample with linear illumination, and lengthens the illumination field length with respect to the feed amount of the stage, so that the substantially same region of the sample to be inspected can be performed multiple times in one inspection. It is proposed to improve detection sensitivity by illuminating and adding the plurality of scattered lights.
By performing linear illumination, the number of cases where two or more defects are simultaneously present in the illumination range may increase. By using a sensor having a plurality of pixels and dividing and detecting the illumination range, each defect can be detected independently.
By applying appropriate processing such as amplification and noise removal to the multiple scattered lights obtained in one inspection by an analog circuit, the scattered light generated from roughly the same area of the sample to be inspected is added by the signal processing unit. Improve defect detection sensitivity.

The following is a brief description of an outline of typical inventions disclosed in the present application.
(1) A defect inspection method for a sample surface, the step of irradiating the same region of the sample surface with a laser beam a plurality of times, the step of detecting scattered light from the same region at each time, and the detection And a step of adding a plurality of signals. Thereby, the detection sensitivity can be improved while suppressing the temperature rise of the sample surface.
(2) The defect inspection method according to (1), wherein in the step of irradiating the same region of the sample surface with the laser beam a plurality of times, the laser beam is illuminated so as to be linear on the sample surface, and the line Defect inspection, wherein the same area is irradiated multiple times by moving the linear illumination area at a pitch shorter than the length of the linear illumination area in the longitudinal direction of the linear illumination area Is the method. This makes it possible to irradiate and detect the same region a plurality of times while maintaining good throughput, and to improve detection sensitivity.
(3) A defect inspection apparatus for inspecting a sample surface, the stage holding the sample, an illumination optical system for linearly illuminating the sample with a laser beam, and scattered from a linear illumination region in the sample A detection optical system that detects light, and a detection unit that converts scattered light detected by the detection optical system into an electrical signal, and the illumination optical system has a longitudinal direction of a linear illumination region on the sample surface. In addition, the defect is configured to irradiate the same region on the sample surface a plurality of times by moving the laser light at a pitch shorter than the length of the linear illumination region in the longitudinal direction. Inspection equipment.
(4) The defect inspection apparatus according to (3), further including a signal processing unit that adds light scattered from the same region on the sample surface.

  ADVANTAGE OF THE INVENTION According to this invention, the defect inspection method and apparatus which suppress a temperature rise of a sample surface and improve detection sensitivity can be provided.

It is a schematic block diagram of the sample surface inspection apparatus which is this invention. It is a detailed explanatory view of an imaging system. It is a schematic block diagram in case an illumination optical system and a detection optical system exist in a different elevation angle direction. It is a schematic block diagram in case a detection optical system exists in a different azimuth direction. It is explanatory drawing of the test | inspection method which performs illumination in multiple times with respect to the same defect. It is explanatory drawing of the definition of SN ratio. It is explanatory drawing of the example which isolate | separates and detects the scattered light from a defect by using a photodiode array when a some defect exists in an illumination field. It is explanatory drawing about the relationship between the position where a defect passes, and the magnitude | size of the scattered light to generate | occur | produce when illumination intensity distribution is Gaussian distribution. It is explanatory drawing about the relationship between the position through which a defect passes, and the magnitude | size of the scattered light to generate | occur | produce when illumination intensity distribution is uniform distribution. It is explanatory drawing of the illumination optical system in the case of performing long illumination by dividing | segmenting the laser beam oscillated from one light source into plurality, and arranging and illuminating them in the radial direction. It is explanatory drawing of the signal addition method when a photodiode array is used for the detection optical system installed in the some azimuth direction. It is explanatory drawing of the signal addition method when a photodiode array and PMT are used for the detection optical system installed in the some azimuth direction. It is explanatory drawing of the detected light quantity when the sensitivity of a light-receiving part differs. It is explanatory drawing of the method of correct | amending sensor sensitivity based on the detection information before one rotation. It is explanatory drawing of the coordinate merging method of the signal detected in multiple times with respect to the same defect. It is a figure explaining the difference of the scanning method in a prior art, and the scanning method in this invention. It is a defect detection processing flow in this inspection apparatus. It is a figure which shows an example of GUI.

  An example of an embodiment of the present invention will be described with reference to FIG. 1 schematically includes an illumination optical system 101, a detection optical system 102, a wafer stage 103, and a circuit / signal processing unit. The illumination optical system 101 includes a laser light source 2, a beam expander 3, a homogenizer 4, mirrors 5 and 6, and a cylindrical lens 7. The laser beam 100 emitted from the laser light source 2 is adjusted to a desired beam diameter by the beam expander 3, converted to a uniform illuminance distribution by the homogenizer 4, and linearly formed in the inspection area of the wafer 1 by the cylindrical lens 7. Illuminate.

Here, the laser light source 2 may be a laser light source that oscillates an ultraviolet or vacuum ultraviolet laser beam.
Further, the homogenizer 4 is used for the purpose of making the illumination intensity uniform, but the illuminance distribution may be made uniform by using, for example, a diffractive optical element or a fly-eye lens. Further, illumination may be performed without using the homogenizer 4. By omitting the homogenizer, the attenuation of the laser beam intensity can be suppressed, and illumination can be performed with strong illuminance.
The cylindrical lens 7 is used for linear illumination. For example, the cylindrical lens 7 uses an anamorphic optical system composed of a plurality of prisms, and uses a beam in only one direction in a plane perpendicular to the optical axis. The sample may be linearly illuminated by changing the diameter and using a condensing lens. Use of an anamorphic optical system is effective in that the optical axis can be easily adjusted.

  The detection optical system 102 includes an imaging system 8 and a photodiode array 9. FIG. 2 shows the detection optical system 102 in detail. The detection optical system 102 includes a condenser lens 21, an image intensifier 22, an imaging lens 23, and a photodiode array 9. The light scattered from the illumination field 20 is condensed by the condenser lens 21, and the image intensifier is collected. The scattered light is amplified by the fire 22 and imaged on the photodiode array 9 via the imaging lens 23.

Here, the image intensifier 22 is used for the purpose of amplifying the scattered light so that the weak scattered light can be detected, but without using the image intensifier, for example, EM-CCD, multi-anode PMT, etc. A sensor having a high amplification factor may be used. Use of these is effective in that the apparatus can be slimmed.
The photodiode array 9 is used for receiving scattered light and performing photoelectric conversion, and a TV camera, a CCD linear sensor, TDI, a photodiode array, a multi-anode PMT, or the like may be used. For example, by using a two-dimensional sensor, a wide area can be inspected at a time.

  The photodiode array 9 generates an electrical signal corresponding to the amount of received light, and the electrical signal is subjected to amplification, noise processing, and analog-digital conversion required by the analog circuit 51, and is scattered from substantially the same region by the signal processing unit 52. Addition of the plurality of optical signals and defect determination are performed, and the map output unit 54 displays the defect map via the CPU 53.

  The wafer stage 103 includes a chuck (not shown) that holds the wafer 1, a rotary stage 10 for rotating the wafer, and a translation stage 11 for moving the wafer in the radial direction. The wafer stage 103 illuminates the entire surface of the sample spirally by performing rotational scanning and translational scanning. Further, the rotation speed and translation speed are controlled by the stage control unit 55 so that a desired area can be illuminated.

  As described above, one of the features of the present invention is that linear illumination is performed on the surface of the sample, and the sample is moved in the same direction as the illumination longitudinal direction while rotating the sample, so that the entire surface of the sample is illuminated spirally and detected. Inspecting surface defects based on scattered light.

  In FIG. 1, the explanation has been made with an example in which there is one illumination optical system and one detection optical system. However, as shown in FIG. 3, the oblique illumination optical system 101a that illuminates the sample from a low elevation angle and is approximately perpendicular to the sample. Vertical illumination optical system 101b that illuminates from the direction, low angle detection optical system 102a that detects the sample at a low elevation angle, and high angle detection optical system 102b that detects the sample at a higher elevation angle than the low angle detection optical system There may be a plurality of illumination optical systems and detection optical systems.

The oblique illumination optical system 101a is composed of a laser light source 2, a beam expander 3, a homogenizer 4, mirrors 5 and 6a, and a cylindrical lens 7a. , A mirror 6b, and a cylindrical lens 7b. Needless to say, these configurations can be similarly omitted and replaced in the embodiment shown in FIG.
Here, the mirror 5 can change the path direction of the laser beam 100, and can switch between the oblique illumination optical system 101a and the vertical illumination optical system 102b as necessary.

  With regard to the use of oblique illumination optical system and vertical illumination optical system, detection sensitivity can be improved by using oblique illumination optical system, and defect classification performance is improved by using this vertical illumination optical system. Therefore, it may be used as appropriate depending on the application.

  The low angle detection optical system 102a is composed of an imaging system 8a and a photodiode array 9a, and the high angle detection optical system 102b is similarly composed of an imaging system 8b and a photodiode array 9b. The imaging system 8a includes a condenser lens, an image intensifier, and an imaging lens (not shown), and the imaging system 8b is the same. Needless to say, these configurations can be similarly omitted and replaced in the embodiment shown in FIG.

  Each of the photodiode arrays 9a and 9b generates an electrical signal corresponding to the amount of received light, and the electrical signal is subjected to necessary amplification, noise processing, and analog-to-digital conversion in the analog circuits 51a and 51b. Addition of a plurality of optical signals scattered from the same region and defect determination are performed, and a defect map is displayed by the map output unit 54 via the CPU 53.

As for the proper use of the low angle detection optical system and the high angle detection optical system, the same location is detected at different elevation angles at the same time, so the sensitivity of each sensor is adjusted and used in the same inspection to detect the detected particle size. The dynamic range can be expanded.
In addition, the accuracy of defect classification can be improved by using a combination of these illumination optical system and detection optical system. For example, it is possible to detect large scattered light with a low-angle detection optical system when illuminating from an oblique direction for convex defects, and high when illuminating from a vertical direction for concave defects. Large scattered light can be detected by the angle detection optical system.

  Although FIG. 3 has been described with an example in which detection optical systems exist in different elevation angles, a plurality of detection optical systems may exist in different azimuth directions as shown in FIG. That is, FIG. 4 is a diagram showing the embodiment of the present invention from the viewpoint from above, and shows the wafer 1, the illumination optical system 101, and the detection optical systems 102c to 102h. The detection optical systems 102c to 102h are constituted by imaging systems 8c to 8h and photodiode arrays 9c to 9h, respectively. The detection signal is subjected to amplification, noise processing, and analog-to-digital conversion required by the analog circuit, and the signal processing unit adds a plurality of optical signals scattered from the same region and determines the defect, and maps it via the CPU. A defect map is displayed on the output unit (not shown). Here, regarding the configuration of the detection optical system, the imaging systems 8c to 8h are each configured by a condenser lens, an image intensifier, and an imaging lens (not shown).

  In this way, when using a detection optical system that exists at multiple azimuth angles, when the angle characteristics of the scattered light change depending on the size, shape, sample film type, and surface roughness of the defect, the noise is small and the defect Since it is possible to select and inspect a detection optical system that can detect a large amount of scattered light from the light, it is possible to improve detection sensitivity.

  As for the arrangement of the detection optical systems, FIG. 4 shows an example in which six detection optical systems are arranged in different azimuth angle directions, but the number of detection optical systems does not have to be six, and the azimuth angles to be arranged are arranged. There are no restrictions on the direction. Further, it is not necessary that a plurality of detection optical systems be arranged at substantially the same elevation angle. Furthermore, it is not necessary that the detectors are arranged at substantially the same azimuth angle.

  One feature of the present invention is that the same defect is illuminated a plurality of times by performing linear illumination and scanning. First, the conventional inspection method will be described, and then the inspection method according to the present invention will be described.

  The stage translates at a substantially constant speed in the radial direction (R direction) while rotating, and the distance traveled in the radial direction at the time of approximately one revolution is called the feed pitch. By rotating and translating, the entire surface of the specimen is scanned in a spiral, but in the conventional technology, the length of the illumination field in the radial direction is roughly the same as the feed pitch length, and only one defect is illuminated once. Is often not performed.

In the present invention, linear illumination is performed, and illumination is performed a plurality of times for the same defect by making the illumination field length longer than the feed pitch length.
FIG. 5 is an explanatory diagram in the case where the length of the illumination field 20 is four times the feed pitch 26, and the defect 25 is illuminated four times, and this is used to explain the illumination multiple times. . First illumination is performed on the defect 25 at time t1. At time t2, the wafer rotates approximately once, the illumination field advances in the radial direction by the distance of the approximate feed pitch 26, and the defect 25 is illuminated again. Thereafter, at time t3 and time t4, the wafer rotates approximately once, and the defect 25 is illuminated. That is, in the case of FIG. 5, the defect 25 can be illuminated four times, and the detected light is subjected to addition processing by an analog circuit or a signal processing unit. Note that the number of times of illumination need not be four, and any number of times may be used as long as illumination is performed a plurality of times.

The reason why the detection sensitivity is improved by the addition of a plurality of scattered lights performed in the present invention will be described.
The SN ratio is defined in FIG. 6, and the reason why the detection sensitivity is improved by using it is explained. The vertical axis of the graph in FIG. 6 indicates the detected light amount, and the horizontal axis indicates time. Even if there is no defect in the illumination field, since the sample surface is always illuminated, the scattered light 30 generated according to the unevenness of the wafer roughness is continuously detected. If the average value of the scattered light from the wafer roughness is N ₀ , the detected light quantity fluctuates with an amplitude of √N ₀ due to fluctuations generated by photoelectric conversion on the sensor light receiving surface, which becomes noise. When a defect enters the illumination field, scattered light 31 is generated from the defect. When the magnitude of the scattered light of the N ₀ from defects in the case of a reference to S _0, SN ratio is defined by "S ₀ / √N _0".
By adding the scattered light from the same defect n times, the scattered light from the defect increases from S ₀ → n × S ₀ , and the scattered light from the wafer roughness increases from N ₀ → n × N ₀ . That is, the SN ratio is “n × S ₀ / √ (n × N ₀ )”, and the detection sensitivity is improved by a factor of √n.

  As for the illumination field length, it is effective to make the illumination defect a plurality of times by making it longer with respect to the feed pitch, that is, by generating linear illumination. Although the throughput is reduced, it is possible to illuminate the same defect a plurality of times by shortening the feed pitch without increasing the illumination field length.

As shown in Fig. 7, by using a photodiode array to divide and detect the illumination field, even when two or more defects are illuminated simultaneously, each scattered light can be detected independently. . The case where there are four light receiving portions 35a, 35b, 35c, and 35d in the photodiode array will be described.
When the defect 25a and the defect 25b are simultaneously illuminated with the illumination field 20, the imaging system 8 can detect and divide the light by the light receiving portions 35a and 35d of the photodiode array 9. Moreover, by dividing the illumination field and detecting it, it becomes possible to reduce the noise from the wafer roughness, and it can be expected to improve the detection sensitivity.

  As described above, the inspection method according to the present invention has been described. The main feature of the inspection method is to illuminate a substantially same area multiple times with linear illumination and to improve detection sensitivity by adding a plurality of scattered lights. It is. The detection sensitivity can be improved without shortening the illumination wavelength. Similarly, the detection sensitivity can be improved without using techniques such as increasing the laser output and reducing the illumination field. That is, according to the present embodiment, it is possible to suppress damage to the sample and improve detection sensitivity.

Here, regarding the illumination optical system, the intensity distribution of the laser beam generally has a Gaussian distribution, but in the present invention, illumination may be performed with a uniform intensity distribution.
For example, when the intensity distribution of the illumination field 20a has a Gaussian distribution 40a as shown in FIG. 8, when the defect 25a is illuminated with the illumination field 20a, scattered light 41a is generated, and when the defect 25b is illuminated, the scattered light 41a Larger scattered light 41b is generated. If the intensity distribution of the illumination field 20b has a uniform distribution 40b as shown in FIG. 9, the scattered light 41c having approximately the same size is emitted regardless of whether the illumination field 20b illuminates the defect 25a or the defect 25b. Occurs.
The wafer stage rotates at a high speed during inspection, and also generates vibrations in the height and radial directions, so the frequency of sample height fluctuations and waviness is high, and the positional relationship between the defect and the illumination field. However, if the illumination intensity is a uniform distribution, fluctuations in the amount of detected scattered light due to deviations in the illumination position can be reduced, and reproducibility and stability can be improved with regard to defect detection sensitivity and coordinate accuracy. Conceivable.

Regarding the illumination optical system, in addition to generating linear illumination using a beam expander and cylindrical lens, the Wollaston prism is used to divide the laser beam and illuminate the divided laser beams side by side in the radial direction. It is also possible to generate a long illumination field and illuminate the sample surface.
A laser beam splitting method will be described with reference to FIG. Since the laser beam 100 oscillated from the laser light source 2 is generally linearly polarized light, after passing through the beam expander 3 and the homogenizer 4, it is circularly polarized by the quarter wavelength plate 42a, and then by the Wollaston prism 43. Split into two linearly polarized beams orthogonal to each other. The divided laser beam is again circularly polarized by the quarter-wave plate 42b, and the sample surface is illuminated by the illumination fields 20c and 20d by the condenser lens 44. In this way, a plurality of illumination fields may be generated, and the long illumination fields described above may be generated by illuminating the illumination fields side by side.

The distance between the two illumination fields divided in this way can be freely adjusted. The illumination may be performed by overlapping the beams or may be performed separately. Thereby, it is possible to adjust the number of times of illuminating substantially the same region.
The intensity of the divided laser beam can be adjusted by adjusting the angle formed by the oscillation direction of the linearly polarized light of the laser beam 100 and the slow axis of the quarter wave plate 42a to change the ellipticity of elliptically polarized light and the azimuth angle of the elliptical long axis. The illumination intensity of the illumination fields 20c and 20d divided by the Wollaston prism 43 can be arbitrarily adjusted by adjusting the ellipticity and the ellipse major axis azimuth. This makes it possible to expand the dynamic range of detectable defects (described later in this specification). The intensity of the illumination fields 20c and 20d may be approximately the same or different.

The number of divisions of the laser beam has been described with reference to the example in which the laser beam is divided into two in FIG. 10, but the Wollaston prism and the quarter-wave plate are placed in front of the quarter-wave plate 42b and the condenser lens 44. A plurality of the combinations may be arranged, and the laser beam may be divided into four, eight, or more beams. Further, the length of the illumination range can be adjusted by adjusting the distance between the plurality of illumination fields, and the intensity of the plurality of laser beams can be freely adjusted.
Further, not only the divided laser beams may be arranged and illuminated from substantially the same direction, but also illumination may be performed simultaneously using, for example, an oblique illumination optical system and a vertical illumination optical system, and two illumination fields may be arranged. As a result, the same defect is illuminated substantially vertically and obliquely in one inspection, and defect classification performance can be improved by utilizing the difference between the detected elevation angle and detected azimuth direction.

Next, an example in which the detection optical system is in different azimuth directions will be described with respect to the scattered light addition method of the detection optical system. A case where two detection optical systems 102c and 102d exist as shown in FIG. 11 (a) will be described as an example.
FIG. 11B is an enlarged view of the illumination field, the photodiode array light receiving unit, and the analog circuit. The illumination field 20 is divided and detected by the photodiode 9c, and the detection signals at the respective light receiving units are amplified and noise-removed by the circuits 45a to 45d. Similarly, the illumination field 20 is divided and detected by the photodiode 9d, and the detection signals at the respective light receiving units are amplified and noise-removed by the circuits 46a to 46d. In the addition units 47a to 47d, the outputs of the light receiving units that detect scattered light from substantially the same region of the illuminated unit are added together. Since the output of the circuit 45a and the output of the circuit 46a are signals in substantially the same region in the illuminated part, they are added by the adder 47a. Similarly, since the output of the circuit 45b and the output of the circuit 46b, the output of the circuit 45c and the output of the circuit 46c, the output of the circuit 45d and the output of the circuit 46d are signals in substantially the same region in the illuminated part, the adder 47b Add at ~ 47d to improve detection sensitivity.
Here, the case where the number of light receiving portions of the photodiode array is four has been described, but the present invention is not limited to this, and any number may be provided. Further, the number of light receiving portions of the photodiode array 9c and the photodiode array 9d need not be the same. When the number of light receiving parts is different, the signals of the light receiving parts that detect substantially the same region are added.
Furthermore, the number of detection optical systems is not necessarily the two exemplified here, and a large number may exist in a plurality of azimuth and elevation directions. When there are a plurality of detection optical systems, the signals of the light receiving sections that detect substantially the same region in each detection optical system are added.

Next, with respect to the scattered light addition method when the detection optical system exists in different azimuth directions, detection using a photomultiplier tube (PMT) as a sensor like the detection optical system 104 in FIG. An example in the case where an optical system is present is shown. The detection optical system 104 includes a condenser lens 60, a pinhole 61, and a PMT 62. By using a pinhole 61 in the detection optical system 104, the detection range is reduced and noise is reduced.
FIG. 12 (b) is an enlarged view of the illumination field, photodiode array, PMT, and analog circuit. The detection signal from the PMT 62 is amplified and noise-removed by the circuit 63. Since the output of the circuit 63 and the output of the circuit 45c are signals in substantially the same region of the illuminated part, the addition is performed by the adder 64 to improve detection sensitivity. The signals of the circuits 45a, 45b, and 45d are used without being added.

Here, PMT has a high response speed and is a point sensor, so it has a small amount of data, and when high detection sensitivity is not required, partially use PMT to improve inspection speed It becomes possible.
In addition, the number of detection optical systems need not be the two exemplified here, and a large number may exist in a plurality of azimuth and elevation directions. Further, the ratio of the detection optical system using the photodiode array and the detection optical system using the PMT does not need to be 1: 1.
Furthermore, regarding the arrangement of the detection optical system, there is no restriction on the azimuth angle in which the detection optical system using the photodiode array and the detection optical system using the PMT are arranged, but at least one of the photodiode arrays is illuminated. It is desirable to arrange at a position substantially parallel to the direction.

Next, regarding the inspection method, in the present invention, since the same defect is illuminated a plurality of times, the previous detection signal can be fed back when the same defect is illuminated for the second time or later. As an example, a case where the dynamic range is expanded by correcting the sensor sensitivity will be described.
FIGS. 13 (a) and 13 (b) are graphs with the detected light amount on the vertical axis and the time on the horizontal axis. In the case of FIG. 13 (a), the detected light quantity 70 can be detected without being saturated. In the case of FIG. 13 (b), the detected light quantity 71 is saturated, and an accurate scattered light quantity cannot be measured. Since the size determination in the defect determination is performed based on the magnitude of the detected light amount, it is important to detect the scattered light without being saturated.
A method for preventing saturation of the detected light amount will be described with reference to FIGS. 14 (a) and 14 (b). Scattered light generated when the illumination field 20 is illuminated for the first time with respect to the defect 25 is detected by the light receiving unit 35a of the photodiode array 9 via the imaging system 8. If the detected light quantity at that time is saturated as shown in FIG. 13B, the sensor sensitivity is lowered and the second illumination is performed. In the second illumination, the scattered light from the defect 25 is detected by the light receiving part 35b of the photodiode array 9, and can be detected without saturating the scattered light as shown in FIG. 13 (a).

  Here, with regard to the sensor sensitivity correction method, for example, the sensor sensitivity can be adjusted by changing the applied voltage of the image intensifier or multi-anode PMT, changing the sensor accumulation time, or changing the illumination intensity. Can be changed. The correction of the sensor sensitivity is not limited to an example in which the sensitivity is appropriately adjusted during the inspection. For example, the sensitivity is changed for each light receiving unit from the start of the inspection, or sensors having different sensor sensitivity are arranged in an array. It may be set in advance by using. When a Wollaston prism is used to divide a laser beam and the divided laser beams are arranged and illuminated, the laser beam is divided into laser beams having different intensities, and the divided laser beams are arranged in a pseudo manner. It is possible to change the detection sensitivity for each light receiving unit by generating linear illumination with different illumination intensity distribution and performing linear illumination. Sensor sensitivity correction may be performed using this method.

Next, a method for adding scattered light in the signal processing unit will be described with reference to FIG. In FIG. 15, the vertical axis indicates the R (radius) direction, the horizontal axis indicates the θ (rotation) direction, and the coordinates 112a to 112d indicate the coordinates at which the same defect is detected. In the present invention, since the same defect is illuminated a plurality of times, one defect is detected approximately the same number of times as the number of illuminations. For example, the defect coordinates 112a are coordinates detected at the first illumination, 112b is the second, 112c is the third, and 112d is the coordinates at which the same defect is detected at the fourth illumination. Although 112a to 112d are coordinates of the same defect, since errors such as a shift in stage height and illumination position occur each time illumination is performed, it is considered that the coordinates at which the defect is detected vary. Therefore, the fixed region 111 is divided by 110a and 110b in the R direction and 110c and 110d in the θ direction, and signals detected within the range are determined to be signals from the same defect and integrated and processed. As an example of the integration process, the final defect coordinates 113 are obtained by taking the centroid of the defect coordinates 112a to 112d. Other than setting the center-of-gravity coordinates to the final defect coordinates, weighting may be performed to obtain the final defect coordinates. By averaging the coordinate shifts that occur at each detection, it is possible to expect improved coordinate reproducibility.
As for the method of adding scattered light in the signal processing unit, the final detected light amount is obtained by adding a plurality of detected light amounts detected within a certain range 111 in FIG. 15, and the defect size is determined based on this detected light amount. To do. In addition to adding the plurality of detected light amounts to obtain the final detected light amount, an average value of the plurality of detected light amounts may be taken and the average value of the detected light amounts may be used as the final detected light amount. When the detected light quantity is large, addition is not performed, and averaging is performed to reduce variation in the detected light quantity, thereby improving the reproducibility and stability of the defect size determination. Here, FIG. 15 shows an example in which a defect is detected only four times, but any number of defects may be detected.

Next, an example of the inspection method of the present invention will be described with reference to FIG.
In the prior art, scanning is performed spirally as shown in FIG. 16 (a), and the inspection of the sample ends when the illumination field reaches the outermost periphery. In the present invention, when the illumination field reaches the outermost periphery as shown in FIG. 16 (b), the stage movement in the radial direction is stopped, and scanning is performed concentrically. This makes it possible to illuminate substantially the same region a plurality of times in the outer peripheral portion as in the inner peripheral portion, and to prevent a decrease in detection sensitivity in the outer peripheral portion.
Here, the number of illuminations can be freely set because concentric scanning is performed on the outer peripheral portion. In the inner periphery, the number of times that the same area can be illuminated is determined by the relationship between the illumination field length and the feed pitch length, but the number of illuminations in the outer periphery need not be the same as the number of illuminations in the inner periphery, You may increase more than the frequency | count of illumination in an inner peripheral part.

  Next, the defect detection processing flow will be described with reference to FIG. First, inspection conditions such as illumination direction and sensor sensitivity are set in recipe setting (step 120). It includes setting the processing method for the length of the illumination field, the feed pitch, and the detected scattered light. Wafer scanning is started (step 121), and signal processing set in the recipe is performed on the detected scattered light (step 122). Defect determination is performed based on the processed signal (step 123), and a defect map is displayed (step 124).

  FIG. 18 shows an example of a GUI. The defect map 130 displayed after the inspection is completed, and the sub-window for setting the inspection mode before the inspection are used as the configuration requirements. The defect map is displayed based on the defect signal and coordinates taken at the time of inspection. The inspection mode 131 can be selected by direct input or pull-down selection. It is not necessary that the number of times of illumination of the same defect is the same during one inspection. For example, the inspection mode can be set to the standard mode at the inner periphery of the sample (132), and the high sensitivity mode can be set at the outer periphery of the sample (133). . For example, in the high sensitivity mode, the detection sensitivity is improved by reducing the feed pitch and increasing the number of times of illumination.

  As described above, according to the embodiment of the present invention, it is possible to improve detection sensitivity by illuminating the same defect a plurality of times in one inspection and adding scattered light generated a plurality of times. Further, by using a photodiode array having a plurality of pixels, inspection can be performed without reducing the throughput. Further, according to the embodiment of the present invention, it is possible to realize an inspection method and an inspection apparatus capable of achieving both improvement in detection sensitivity and high throughput. In addition, by using information that is detected a plurality of times and detected a plurality of times, it is possible to expand the dynamic range and improve the accuracy of coordinate accuracy and defect size determination.

1 wafer, 2 laser light source, 3 beam expander, 4 homogenizer, 5 switching mirror, 6 · 6a · 6b mirror, 7 · 7a · 7b cylindrical lens, 8 · 8a to 8h imaging system, 9 · 9 a to 9h Photodiode array, 10 rotation stage, 11 translation stage, 20 / 20a to 20d illumination field, 21 condenser lens, 22 image intensifier, 23 imaging lens, 25 / 25a / 25b defect, 26 feed pitch, 30 wafer roughness Scattered light from 31, scattered light from defects, 35a to 35d light receiver, 40a Gaussian distribution, 40b uniform distribution, 41a to 41c Detected light intensity, 42a / 42b 1/4 wavelength plate, 43 Wollaston prism, 44 Condenser lens, 45a to 45d circuit, 46a to 46d circuit, 47a to 47d adder, 51 / 51a / 51b analog circuit, 52 signal processor, 53 CPU, 54 map output unit, 55 stage controller, 60 condenser lens , 61 pinhole, 62 photomultiplier tube, 63 Path, 64 adder, 70 unsaturated detected light intensity, 71 saturated detected light intensity, 100 laser beam, 101 / 101a / 101b illumination optical system, 102 / 102a to 102h detection optical system, 103 wafer stage, 104 PMT used Detection optical system, 110a-110d region boundary, 111 setting region, 112a-112d defect coordinate, 113 integrated defect coordinate, 120-124 processing flow, 130 defect map, 131 inspection mode, 132 inner peripheral inspection mode, 133 Inspection mode at the outer periphery

Claims (10)

When inspecting one sample, the sample is irradiated with a beam whose detection sensitivity is changed when the defect is detected for the second time or later based on the prior inspection information when the defect on the sample is inspected in advance. An illumination optical system,
A detection optical system for detecting scattered light from the sample scattered by irradiation by the illumination optical system;
A processing system for processing a signal based on the scattered light detected by the detection optical system;
A defect inspection apparatus comprising: The defect inspection apparatus according to claim 1,
The illumination optical system irradiates the sample with a beam whose irradiation intensity is changed based on the preliminary inspection information. The defect inspection apparatus according to claim 1,
The defect inspection apparatus, wherein the illumination optical system irradiates the sample with a beam whose irradiation intensity is changed based on an intensity of a prior detection signal when the defect is inspected in advance. The defect inspection apparatus according to claim 1,
2. The defect inspection apparatus according to claim 1, wherein the illumination optical system changes an application voltage or an accumulation time of an image intensifier of the detection optical system , a multi-anode PMT, or a photodiode array based on the preliminary inspection information. The defect inspection apparatus according to claim 1,
The illumination optical system irradiates the substantially same region of the sample a plurality of times, and the predetermined intensity is determined based on the intensity of the signal based on the scattered light detected by the detection optical system by the predetermined irradiation. A defect inspection apparatus that irradiates after the first time. When inspecting one sample, the sample is irradiated with a beam whose detection sensitivity is changed when the defect is detected for the second time or later based on the prior inspection information when the defect on the sample is inspected in advance. Lighting process to
A detection step of detecting scattered light from the sample scattered by irradiation by the illumination step;
A processing step of processing a signal based on the scattered light detected in the detection step;
A defect inspection method comprising: The defect inspection method according to claim 6,
The defect inspecting method, wherein the illumination step irradiates the sample with a beam whose irradiation intensity is changed based on the preliminary inspection information. The defect inspection method according to claim 6,
The defect inspecting method, wherein the illumination step irradiates the sample with a beam whose irradiation intensity is changed based on an intensity of a prior detection signal when the defect is inspected in advance. The defect inspection method according to claim 6,
In the defect inspection, the illumination process may change an application voltage of an image intensifier , a multi-anode PMT or a photodiode array used for detection in the detection process or change an accumulation time based on the preliminary inspection information. Method. The defect inspection method according to claim 6,
The illuminating step irradiates the substantially same region of the sample a plurality of times, and the irradiation intensity determined based on the intensity of the signal based on the scattered light detected in the detection step by the predetermined irradiation, the predetermined time and thereafter A defect inspection method characterized by irradiating a laser beam.

Images