JP7723640B2 - Defect Inspection Equipment - Google Patents

Description

The present invention relates to a defect inspection device that uses light to detect defects on a sample.

As hard disk (HD) data capacity continues to increase, the HD's surface condition, defect size, and shape have a significant impact on performance and yield. To maintain and improve yield, it is necessary to inspect for low-aspect-ratio, low-step defects (heights of a few nanometers, widths of a few tens of micrometers) and waviness defects (periods of a few tens of micrometers, heights of a few nanometers) caused by surface polishing or foreign matter contamination. Low-aspect-ratio defects are low-step defects that generate almost no scattered light. Interferometric measurements, such as differential interference contrast (DIC) inspection, which uses the principles of differential interference microscopy, are used to detect low-step defects. Scattered light inspection is used, which can detect waviness defects as surface roughness (haze) without including defect height or period information.

Patent Document 1 below discloses an optical system that simultaneously performs scattered light inspection and DIC inspection by scanning a wafer surface with laser light illumination. The document describes the following technology: "An inspection apparatus for simultaneous dark field (DF) and differential interference contrast (DIC) inspection includes an illumination source and a sample stage configured to hold a sample. The inspection apparatus includes a first sensor, a second sensor, and an optical subsystem. The optical subsystem includes an objective lens and one or more optical elements arranged to direct illumination from one or more illumination sources to the surface of the sample via the objective lens. The objective lens is configured to collect signals from the surface of the sample, and the collected signals include signals based on scattering from the sample and/or signals based on phase. The inspection apparatus includes one or more separation optical elements arranged to spatially separate the collected signals into DF signals and DIC signals by directing the DF signals and DIC signals along DF paths and DIC paths, respectively." (See Abstract.)

Special table 2017-531162 publication

Due to surface polishing during the hard disk manufacturing process, random-period undulation defects exist on the hard disk surface, which are difficult to detect using conventional scattered light detection. As hard disk capacity increases, the distance between the magnetic head and the hard disk surface becomes shorter, creating a need to inspect for undulation defects, which were previously outside the scope of control. Furthermore, there is also a similar need to detect low-level step defects that do not have a periodicity. High-sensitivity, high-throughput inspection is required for both undulation and low-level step defects.

DIC-based inspection equipment is suitable for detecting low-level step defects. Therefore, it would be ideal if DIC-based inspection equipment could also be used to detect waviness defects. However, when detecting waviness defects using a DIC-based inspection equipment, the DIC signal strength varies depending on the period of the waviness defects (the interval between defects: pitch), and in particular, the signal amount for a period of shear amount x 1/n is zero or very weak, making it difficult to detect. Therefore, it is difficult to detect waviness defects with an arbitrary period simply using a DIC-based optical defect inspection equipment.

The present invention was made in consideration of the above-mentioned problems, and aims to provide a defect inspection device that can simultaneously inspect the height and pitch of low-level step defects and waviness defects on a sample surface with high sensitivity and high throughput.

The defect inspection device according to the present invention includes a first sensor that detects a differential interference signal and a second sensor that detects a pupil plane image of the illumination optical system. The differential interference signal is used to detect step defects, and the signals detected by the first and second sensors are used to detect waviness defects.

The defect inspection device according to the present invention can simultaneously inspect the height and pitch of low-level step defects and waviness defects on a sample surface with high sensitivity and high throughput. Issues, configurations, and advantages other than those mentioned above will become clear from the description of the embodiments below.

1 is a schematic configuration diagram of a defect inspection device 1 according to a first embodiment. 10A and 10B are diagrams illustrating a rotation inspection of the surface of the disk 2. FIG. The arrangement of the detection elements of the sensors 14 and 17 is shown. 1A and 1B are cross-sectional and top views of a bump defect 40 and a waviness defect 42, respectively. 10A and 10B are diagrams illustrating the relationship between a wavy defect along the θ direction and a height signal. 10 shows the intensity ratio between the height signal 61 and the height signal 62 obtained by simulation with respect to the period of the undulation defect. FIG. 1 is a functional block diagram of a processor 105. 10A and 10B are diagrams illustrating a method for determining the height of a waviness defect in the θ direction. 3A and 3B are diagrams illustrating data output by the defect inspection device 1. FIG. 1 is a system block diagram of a defect inspection device 1. FIG. 10 is a configuration diagram of a defect inspection device 1 according to a second embodiment. FIG. 10 is a functional block diagram of a processor 105 according to a second embodiment. 10 shows the configuration of a detection element of a sensor 17 in embodiment 3. 1A and 1B are a cross-sectional view and a top view of an R-direction waviness defect 142. FIG. 10 is a functional block diagram of a processor 105 according to a third embodiment.

First Embodiment
FIG. 1 is a schematic diagram of a defect inspection apparatus 1 according to a first embodiment of the present invention. The defect inspection apparatus 1 is an apparatus that detects step defects and waviness defects present on the surface of a sample by irradiating the sample with light. Here, as an example of a sample, the case of inspecting surface defects of a hard disk will be described. Each of the components of the defect inspection apparatus 1 shown in FIG. 1 will be described below.

The light source 3 is composed of a coherent single-wavelength laser. A longer wavelength allows for a wider detection range. A shorter wavelength increases sensitivity to minute heights. The illumination optical system 4 is positioned to optimize the beam shape irradiated onto the disc 2. For example, to achieve a thin linear beam spot as described below, the illumination optical system 4 can be realized by using two cylindrical lenses and an anamorphic prism. The mirror 5 is assumed to be a general-purpose reflecting mirror. It is preferable to use a multilayer film designed to prevent phase change within the plane of the beam as the mirror 5.

Beam splitters 6 and 7 are beam splitters that split the beams. Beam splitter 6 transmits light from light source 3 to disc 2 and reflects light reflected from disc 2 toward sensor 17. Beam splitter 7 transmits light from light source 3 to disc 2 and reflects light reflected from disc 2 toward sensor 14 and also transmits it toward beam splitter 6. It is desirable for each beam splitter to distribute the intensity appropriately so that the optimal amount of light is obtained at sensor 14 and sensor 17, respectively. Here, we assume that there is a sufficient amount of light from light source 3 and assume a beam splitter that splits the intensity at 1:1.

The quarter-wave plate 8 is provided to adjust the light intensity to 1:1 when the polarization separation element 9 separates the beam into P-polarized and S-polarized light. The polarization separation element 9 separates the incident beam into P-polarized and S-polarized light. For example, the polarization separation element 9 can be realized using a Nomarski prism or a Wollaston prism.

The focusing optical system 10 is an optical system that focuses the light separated by the polarization separation element 9 onto the disc 2. To obtain a differential interference signal, the point at which the light is separated by the polarization separation element 9 and the point at which the light is focused on the disc 2 are set to be focal points. As a result, the P-polarized and S-polarized light reflected from the disc 2 are recombined by the polarization separation element 9. For implementation reasons, using a relay optical system for the focusing optical system makes it easier to design.

The detection optical system 11 is an optical system configured so that the beam spot on the disk 2 is imaged on the sensor 14. It is assumed that the magnification will be converted to optimize the size of the sensor 14. It is desirable to position the sensor 14 at a position conjugate with the beam spot on the disk 2. This allows for accurate detection of differential interference signals.

The detection optical system 16 is an optical system arranged so that the intensity of the image on the pupil plane of the focusing optical system 10 can be detected by the sensor 17. It is desirable to set the pupil position of the focusing optical system 10 and the sensor 17 so that they are conjugate, but it is sufficient if the pupil image of the focusing optical system 10 can be detected by the sensor 17. The detection optical system 16 may also be magnified to match the size of the sensor 17.

Light output from light source 3 is directed toward disc 2. The beam is split by polarization splitter 9 into two beams of S-polarized light and P-polarized light, which are directed toward the surface of disc 2. Light reflected from disc 2 is detected by sensors 14 and 17. Sensor 14 detects differential interference contrast (DIC) signals from the two beams of S-polarized light and P-polarized light. Sensor 17 detects the image at the pupil plane of focusing optical system 10.

Processor 105 uses the DIC signal detected by sensor 14 to detect minute, low-level defects on the surface of disk 2. Furthermore, processor 105 uses the pupil plane beam intensity detected by sensor 17 and the DIC signal detected by sensor 14 to detect waviness defects on the surface of disk 2, based on the principle described below.

The pupil plane intensity has a slight tilt in the ray direction of light reflected from undulation defects on the surface of disk 2 compared to light reflected from a flat surface. This tilt is so small that it cannot be detected on the imaging plane, but on the pupil plane it can be calculated as an intensity ratio of a few percent from the center. This intensity ratio depends on the height and period (pitch) of the undulation defects. Therefore, by determining the intensity ratio on the pupil plane, processor 105 can calculate the height and period of the undulation defects.

Figure 2 is a diagram illustrating rotational inspection of the surface of disk 2. During the inspection of disk 2, disk 2 is rotated in the θ direction by stage 103 (described below). Light beam 20 and light beam 21 are irradiated onto the surface of disk 2. Light beams 20 and 21 are S-polarized and P-polarized, respectively, as shown in Figure 1. The detected signal remains unchanged regardless of whether light beam 20 or light beam 21 is S-polarized or P-polarized, and these can be in any order. Light beams 20 and 21 are long in the radial (R) direction 22 and short in the θ direction 23 (rotation direction) (thin line shape). By making them long in the radial (R) direction, surface defects can be detected over a wide area at once, enabling high-throughput full-surface inspection.

When inspecting the surface of disk 2, light beams 20 and 21 are stationary, and disk 2 is rotated along the θ direction arrow 23 while moving along the radial direction arrow 22, inspecting the entire surface of disk 2. By increasing the beam diameter in the R direction, it is possible to inspect the entire surface without overlooking even small defects.

Figure 3 shows the arrangement of detection elements in sensors 14 and 17. Sensors 14 and 17 have detection elements arranged on the detection surface as shown in Figure 3, depending on the properties of the light beam to be detected.

The surface of the sensor 14 is conjugate with the disc 2, and imaging light 30 of the area being observed is incident on it (the condensed light forms an image on the sensor surface). The sensor 14 has N pixels (N detection elements) arranged in the R direction, from sensor pixels S1_1 to S1_N, and can detect DIC signals for each pixel. This allows DIC signals to be detected over a wide range at once without reducing detection sensitivity for each pixel. This means that even small step defects on the disc 2 can be detected without being overlooked.

Sensor 17 detects the pupil plane intensity of the focusing optical system 10. Because the NA (Numerical Aperture) is larger in the thin line direction (θ direction) on the surface of the disc 2, the size of the beam spot 31 on the pupil plane is larger in the θ direction. Therefore, sensor 17 has two pixel detection elements arranged along the θ direction. Processor 105 (or sensor 17 itself) detects the difference in signal intensity between pixels S2_1 and S2_2.

Figure 4 shows a cross-sectional view and a top view of a bump defect 40 and a wavy defect 42. Low-level defects include protruding defects on the surface of the disk 2, such as bump defect 40, and recessed defects (not shown). Low-level defects are assumed to have heights and surface sizes of 0.1 μm to several μm. A fine beam is used to ensure these defects are not overlooked. As shown in the figure, a wavy defect 42 is a wavy surface with small steps (heights of about 1 nm). In hard disks, wavy defects with a wavy period 43 (wavy pitch) in the range of about 20 μm to 100 μm are said to cause bad sectors.

The DIC signal can detect a phase step 41 between the irradiation positions of the light beams 21 and 20. Detecting the phase step 41 makes it possible to detect bump defects 40 and waviness defects 42. However, in the case of waviness defects 42, it is important to note that the detected phase step changes not only depending on the height of the waviness defect but also on the waviness period 43. This will be explained next.

Figure 5 is a diagram explaining the relationship between undulation defects and height signals along the θ direction. The upper part of Figure 5 shows the relationship between P/S polarized light beams 20 and 21 and the waviness. The disk 2 is shown in cross section. The horizontal direction in the figure corresponds to the θ direction. The lower part of Figure 5 shows the waveforms in the θ direction of a height signal 50 obtained from the DIC and a height signal 51 obtained from the pupil intensity distribution.

When the waviness period is out of phase with the spacing (shear amount) between light beam 20 and light beam 21 (left side of the figure), when light beam 20 is irradiated onto the bottom of the waviness, light beam 21 is irradiated onto the top of the waviness, as shown in the figure. In this case, light beam 20 and light beam 21 are at different heights on disk 2, so a DIC signal is generated. In other words, height signal 50 generates a signal corresponding to the disk position, as shown in the figure. On the other hand, the pupil intensity distribution obtained from light beam 20 and the pupil intensity distribution obtained from light beam 21 are also out of phase with each other. After reflecting off disk 2, light beam 20 and light beam 21 are combined by polarization separation element 9, so they cancel each other out and no amplitude is generated for height signal 51, as shown in the figure.

When the waviness period is in phase with the spacing (shear amount) between light beam 20 and light beam 21 (right side of the figure), when light beam 20 is irradiated onto the top of the waviness, light beam 21 is irradiated onto the top of the waviness, as shown in the figure. In this case, light beam 20 and light beam 21 are at the same height on disk 2, so the DIC signal is zero. In other words, height signal 50 does not generate a signal corresponding to the disk position, as shown in the figure. On the other hand, the pupil intensity distributions obtained from light beam 20 and light beam 21 are also in phase with each other. After reflecting off disk 2, light beam 20 and light beam 21 are combined by polarization separation element 9, but because they are in phase, they do not cancel each other out. Therefore, height signal 51 becomes a signal corresponding to the disk position, as shown in the figure.

As described above, the DIC height signal 50 and the pupil intensity distribution height signal 51 have a complementary relationship with respect to the waviness period. The present invention utilizes this feature to provide a waviness signal detection method that is independent of the waviness period. Note that the amplitudes of the height signals 50 and 51 depend on the waviness height and period, and can therefore be converted into height information based on the correction method described in Figure 8.

Figure 6 shows the intensity ratios of height signals 61 and 62 obtained by simulation with respect to the period of the undulation defect. Height signal 61 is a DIC signal, and height signal 62 is an optical pupil plane intensity distribution signal. In this simulation, the distance (shear amount) between light beam 20 and light beam 21 is set to 175 μm, and the undulation defect has the same height but a different period. Height signals 61 and 62 are normalized with the maximum value of each detection signal set to 1.

The height signal 61 (DIC signal) is maximum at undulation defect period 63 and minimum at undulation defect period 64. The undulation defect period 63 is in the opposite phase state described in FIG. 5 and is equal to shear amount × 2/(2n + 1) (n is a natural number). The undulation defect period 64 is in the same phase state described in FIG. 5 and is equal to shear amount × 1/n (n is a natural number). The simulation results show that the height signal 61 (DIC signal) differs for each undulation defect period. In contrast, the height signal 62 (optical pupil plane intensity signal) is minimum at undulation period 63 because it is in the opposite phase. The height signal 62 is maximum at undulation period 64 because it is in the same phase. This shows that by detecting height signals 61 and 62, all undulation defects with periods between 20 and 100 μm can be detected.

The relative intensity of the height signals 50 and 51 detected by the inspection device changes for each period, but the relative intensity ratio for each period of the waviness defect obtained by this simulation can be used to correct for the height signal of the actual defect.

Figure 7 is a functional block diagram of processor 105. The current signals from the light-receiving surfaces of sensors 14 and 17 are converted into digital signals by I/V converter 70 and A/D converter 71. Since the I/V converter 70 and A/D converter 71 are common technologies, details will not be given here, but in order to improve inspection speed, it is advisable to design them with attention to frequency characteristics.

The DIC signal converted by the A/D converter 71 has amplitude that varies at a certain intensity level, like the height signal 50 shown in Figure 5. The offset removal unit 72 removes this intensity level and obtains only the amplitude.

The unevenness evaluation unit 73 evaluates micro defects using the acquired amplitude-only signal. The unevenness evaluation unit 73 extracts signals that have changed and determines whether the amplitude is positive or negative to determine whether they are uneven. Furthermore, by measuring the length of the amplitude in the θ direction, it is possible to obtain defect length information. Furthermore, by measuring how many light-receiving surfaces have changed, it is possible to obtain defect length information in the R direction. The unevenness evaluation unit 73 sends amplitude information along with length information of the micro defects in the R and θ directions to the unevenness height evaluation unit 74.

The unevenness height evaluation unit 74 acquires height information based on the amplitude and defect size. Height information acquired using DIC can be determined using standard technology, so details will be omitted. As described above, the sensor 14 can detect the unevenness, size, and height of minute step defects.

The signal addition processing unit 75 adds all of the light receiving surface information from which the intensity levels have been removed by the offset removal unit 72. Here, an example is described in which all spot signals in the R direction on the disc 2 are detected together, but spot signals may also be evaluated individually. Evaluating each signal individually allows for a more detailed evaluation of the waviness period in the R direction.

The digital signal obtained by the sensor 17 is used by the signal calculation unit 77 to perform the following calculation, which outputs a signal S corresponding to the waviness height in the θ direction from the pupil plane intensity distribution: S = (S2_1) - (S2_2).

The FFT processing unit 76 performs a Fourier transform on the signals generated by the signal addition processing unit 75 and the signal calculation unit 77 using a rotation speed corresponding to the R-direction position of the disc 2, and calculates the modulation degree (signal level) corresponding to the waviness period. The frequency evaluation unit 78 extracts waviness periods with large modulation degrees.

The signal selection unit 79 selects the larger of the extracted waviness periods: the modulation degree obtained from the DIC (sensor 14) or the modulation degree obtained from the pupil intensity distribution (sensor 17). The θ-direction waviness height evaluation unit 80 converts the modulation degree into height information by comparing the signal selected by the signal selection unit 79 with the ideal height curve described below. As a result, the waviness period and height in the θ direction at the specified R position can be determined.

The processor 105 can distinguish between micro-step defects and waviness defects by separately outputting information on the unevenness, size, and height of micro-step defects and information on the period and height of waviness defects.

Figure 8 is a diagram illustrating a method for determining the height of a waviness defect in the θ direction. The top of Figure 8 shows the modulation depth of the DIC signal after FFT for three cases. The bottom of Figure 8 shows the modulation depth of the height signal obtained from the pupil intensity distribution for the same three cases. In each graph, the horizontal axis represents the waviness frequency (pitch), and the vertical axis represents the modulation depth. The modulation depth obtained by DIC is height signal 50, and the modulation depth obtained from the pupil plane intensity distribution is height signal 51.

Case 1 is when the swell period in the θ direction is out of phase with the spacing between light beams 20 and 21 (left side of Figure 5). Case 2 is when the swell period in the θ direction is in phase with the spacing between light beams 20 and 21 (right side of Figure 5). Case 3 is an intermediate case between the two.

Ideal height curve 81 and ideal height curve 82 each show the ideal height signal obtained at a specified height. Each ideal height curve describes the ratio of signal level to defect height for each pitch value of the waviness defect. Curve 81 is for the DIC signal, and curve 82 is for the pupil plane intensity signal. The height evaluation unit 80 can calculate the waviness defect height by comparing the signal level with the ideal height curve. Data describing each ideal height curve can be stored in advance in a storage device provided in the defect inspection device 1.

In Case 1, the DIC signal has a large modulation depth due to the reversed phase. Therefore, the signal selection unit 79 selects the DIC signal. The height evaluation unit 80 determines the height from the ideal height curve 81 and the height signal 50 using the following formula: Height = Hs1/Hi.

In case 2, the signals are in phase, so the modulation depth of the pupil intensity distribution signal is large. Therefore, the signal selection unit 79 selects the pupil intensity distribution signal. The height evaluation unit 80 determines the height from the ideal height curve 82 and the height signal 51 using the following formula: Height = Hs2/Hi.

In case 3, the signal selection unit 79 selects the larger of the DIC signal and the pupil plane intensity signal. In the example of Figure 8, the modulation depth of the DIC signal is larger. Therefore, the signal selection unit 79 selects the DIC signal. The height evaluation unit 80 determines the height from the ideal height curve 81 and the height signal 50 using the following formula: Height = Hs1/Hi.

As described above, by using the signal with the larger modulation depth between the DIC signal and the pupil plane intensity signal, it is possible to easily determine the height and period of the waviness defect regardless of its period.

Figure 9 is a diagram explaining the data output by the defect inspection device 1. The information indicated by this data can be provided, for example, on a GUI (graphical user interface) displayed on a monitor by the processor 105. The left side of Figure 9 is a defect map, and the right side of Figure 9 is the waviness distribution in the θ direction.

The defect inspection device 1 can display defect information corresponding to height levels on the disk 2, indicating where on the disk 2 the defect information is located. Figure 9 shows an example showing three levels: a first defect level 92, a second defect level 93, and a third defect level 94, depending on height and size. For example, an acceptable defect level may be preset, and only defects at levels outside the acceptable range may be displayed. Although the disk reference line 91 is not on the disk 2, displaying it on the screen makes it possible to visually display the disk position, defect size, and height in an easy-to-understand manner. If the disk reference line 91 cannot be set, the defect size and height relative to the disk radial position may be displayed. For waviness in the θ direction, it is advisable to plot the disk radius on the horizontal axis and the waviness height on the vertical axis. A judgment level 95 may also be set, and an alarm may be issued if the waviness height 96 exceeds this level.

The position of a defect on the surface of disk 2 can be identified based on the coordinates at which light beams 20 and 21 are irradiated. Processor 105 can display the defect position on the screen by reflecting those coordinates on the GUI of Figure 9. Numerical values of the defect coordinates may also be displayed.

Figure 10 is a system block diagram of the defect inspection device 1. The defect inspection device 1 includes a control panel 101, a controller 102, a stage 103, an optical system 104, a light source 3, sensors 14 and 17, a processor 105, and a monitor 106.

The control panel 101 corresponds to the GUI that the host 100 operates when it wants to use the device. When the host 100 wants to refer to the results of an inspection, it can do so by presenting the GUI on the monitor 106. Based on the commands received by the control panel 101, it sends the information necessary for the inspection to the controller 102.

The controller 102 drives the stage 103 to rotate the disk 2 and turn on the light source 3. The stage 103 has the functions of flipping the disk 2 over, exchanging the disk 2, and transporting it, allowing the front and back of multiple disks to be inspected consecutively. This makes it possible to inspect the front and back of multiple disks 2 at high speed and detect minute, low-level defects and surface waviness.

Light source 3 may continuously emit light in CW (continuous wave) or, to reduce power consumption, may emit light only when inspecting disc 2. Optical system 104 is one of the optical systems shown in Figure 1. The beam emitted from light source 3 is focused onto disc 2 by optical system 104, and defect information on disc 2 is detected by sensors 14 and 17. The detection signals detected by sensors 14 and 17 are processed by processor 105 to obtain the defect information. The procedure for obtaining the defect information is as described above. The defect information is displayed on monitor 106.

<First Embodiment: Summary>
In the defect inspection apparatus 1 according to the first embodiment, for a waviness defect pitch that cannot be detected by the DIC signal detected by the sensor 14, the pupil plane intensity difference signal detected by the sensor 17, which is installed so as to be conjugate with the pupil plane, is the largest. For a waviness defect pitch at which the DIC signal intensity is the largest, it cannot be detected by the pupil plane intensity signal. By combining such a DIC signal and a pupil plane intensity distribution signal in a mutually complementary manner, the defect inspection apparatus 1 can detect step defects as well as waviness defects of all periods.

<Second Embodiment>
11 is a configuration diagram of a defect inspection device 1 according to a second embodiment of the present invention. Compared to the first embodiment, a half-wave plate 12, a polarizing beam splitter 13, and a sensor 15 are newly provided. The other configurations are the same as those of the first embodiment.

The half-wave plate 12 converts the incident polarized light into light polarized at a 45° angle relative to the polarizing beam splitter 13. The polarizing beam splitter 13 transmits P-polarized light and reflects S-polarized light. In this second embodiment, the DIC signal is detected by two sensors, sensors 14 and 15. Specifically, the reflected light from the disc 2 is focused into one beam by the polarization separation element 9, and then split again into two beams by changing the phase of the polarization for detection.

Sensor 15 is the same type of sensor as sensor 14, with the detection element divided into N parts as shown in Figure 3. A signal is generated by taking the difference and sum of the sensor signals from sensors 14 and 15, and then dividing the difference by the sum. This signal processing makes it possible to detect height information that is subject to strong external disturbances and is not dependent on changes in laser intensity.

Figure 12 is a functional block diagram of the processor 105 in embodiment 2. Compared to Figure 7, a new block has been added that processes the detection signal from sensor 15. The signal calculation unit 120 performs the following calculation, where S1 is the signal from sensor 14 and S3 is the signal from sensor 15: S = (S1 - S3) / (S1 + S3). At this time, multiple signals are obtained for each light-receiving surface. Because the signal calculation unit 120 obtains the difference using the above formula, it can remove the intensity level in the same way as the offset removal unit 72. Furthermore, because the signals relative to height are generated in opposite phases between sensors 14 and 15, the amplitude is effectively doubled. Furthermore, because division is performed by the signal obtained by adding S1 and S3, laser intensity information can be removed, and a signal that is not dependent on laser intensity fluctuations can be obtained. In this way, embodiment 2 can detect a DIC signal that is more resistant to external disturbances than embodiment 1.

<Third Embodiment>
Fig. 13 shows the configuration of the detection element of sensor 17 in embodiment 3 of the present invention. In the element configuration shown in Fig. 13, the detection element is divided into four parts in the R direction as well as the θ direction. The height and period signals of a waviness defect in the R direction are detected by calculating the sum signal of sensor pixels S4_1 and S4_4 and the sum signal of S4_3 and S4_2, respectively, and then calculating the difference between these signals. At the same time, the height and period signals of a waviness defect in the θ direction are detected by calculating the sum signal of sensor pixels S4_1 and S4_2 and the sum signal of S4_4 and S4_3, respectively, and then calculating the difference between these signals. This makes it possible to detect the period and height information of a waviness defect in the R direction as well as the waviness defect in the θ direction.

Figure 14 shows a cross-sectional view and a top view of an R-direction undulation defect 142. Similar to undulation defect 42, R-direction undulation defect 142 has a undulating surface with small steps (approximately 1 nm). In hard disks, undulation defects with a undulation period 143 in the range of approximately 20 μm to 100 μm are said to cause bad sectors.

Because the beam spots on the surface of the disk 2 are spaced apart in the θ direction (see the middle of Figure 14), the height and period of defects undulating along the θ direction can be obtained using the configuration of embodiment 1. On the other hand, for defects undulating along the R direction, there is no difference in signal level between the beam spots (see the middle of Figure 14). Therefore, for the R-direction waviness defect 142, the phase difference between the irradiation positions of the light beam 20 and the light beam 21 is zero. In other words, the height and period signals of the R-direction waviness defect 142 cannot be detected using the height signal 140. In contrast, the height signal 141 calculated from (S4_1 + S4_4) - (S4_3 + S4_2) changes in the R direction, making it possible to detect the height and period information of the defect.

Figure 15 is a functional block diagram of the processor 105 in embodiment 3. Compared to Figure 7, it newly includes signal calculation units 150 and 151 and an R-direction waviness height evaluation unit 152. The signal output by the signal calculation unit 150 is equivalent to that of the signal calculation unit 77, and the FFT processing unit 76 and subsequent units are the same as those in Figure 7. The signal output by the signal calculation unit 151 is used by the R-direction height evaluation unit 152 to determine the period and height of waviness in the R direction. For waviness defects in the R direction, the period and height can be detected from the signal intensity, so the period and height signals can be determined by the R-direction height evaluation unit 152 without performing FFT signal processing.

<Modifications of the present invention>
The present invention is not limited to the above-described embodiments and includes various modifications. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to those including all of the described configurations. Furthermore, it is possible to replace part of the configuration of one embodiment with the configuration of another embodiment, or to add the configuration of another embodiment to the configuration of one embodiment. Furthermore, it is possible to add, delete, or replace part of the configuration of each embodiment with other configurations.

In the above embodiments, the controller 102 and processor 105 can be configured using hardware such as a circuit device that implements these functions, or they can be configured by a computing device such as a CPU (Central Processing Unit) executing software that implements these functions.

In the above embodiment, the present invention has been described as being applied to an inspection device used for defect inspections performed in the manufacturing process of semiconductors and the like. However, the present invention is not limited to this application, and can be applied to other defect inspection devices that optically detect both minute step defects and waviness defects.

1: Defect inspection device 9: Polarization separation element 10: Light collecting optical system 14: Sensor 17: Sensor 105: Processor

Claims (15)

A defect inspection apparatus that inspects a sample using light, comprising:
a light source that emits a light beam;
a polarization splitter element for splitting the light beam into a first beam and a second beam having orthogonal polarized light;
an optical system for irradiating the sample with the first beam and the second beam in a narrow linear shape;
a first sensor arranged at a position conjugate with the irradiation positions of the first beam and the second beam on the sample;
a second sensor arranged at a position capable of detecting an image of a pupil plane of the optical system;
a processor that detects defects on the specimen using the signal detected by the first sensor and the signal detected by the second sensor;
Equipped with
the processor obtains a differential interference signal of the first beam and the second beam using the signal detected by the first sensor, and detects a step defect on the sample using the differential interference signal;
The defect inspection apparatus according to claim 1, wherein the processor detects a waviness defect on the sample using a signal detected by the first sensor and a signal detected by the second sensor. the first sensor includes two or more first detection elements arranged along a longitudinal direction of a beam spot having the thin line shape of the light beam reflected from the sample;
2. The defect inspection device according to claim 1, wherein the processor simultaneously detects one or more step defects included in one of the beam spots by detecting the step defect for each of the first detection elements. the second sensor includes two second detection elements arranged along a short direction of the narrow linear beam spot of the light beam irradiated onto the sample,
2. The defect inspection device according to claim 1, wherein the processor detects the height and period of the waviness defect based on the signal detected by each of the second detection elements. the processor extracts a first signal detected by the first sensor and a second signal detected by the second sensor for each period of the waviness defect;
the processor identifies the larger of the extracted first signal and the extracted second signal as a height signal;
The defect inspection device according to claim 1 , wherein the processor detects the height of the waviness defect using the identified height signal. the processor obtains ideal height curve data describing a relationship between the height of the wavy defect and the intensity per period of the first signal or a relationship between the height of the wavy defect and the intensity per period of the second signal;
5. The defect inspection device according to claim 4, wherein the processor calculates the height of the waviness defect by using the intensity of the first signal or the intensity of the second signal and referring to the relationship described by the ideal height curve data. the defect inspection apparatus further includes a first beam splitter that transmits the light beam from the light source toward the sample and reflects the reflected light from the sample toward the first sensor;
2. The defect inspection apparatus according to claim 1, further comprising a second beam splitter that transmits the light beam from the light source toward the sample and reflects the reflected light from the sample toward the second sensor. the defect inspection apparatus further includes a rotation mechanism that rotates the sample,
the optical system irradiates the first beam and the second beam onto the sample at intervals along a rotation direction of the sample;
2. The defect inspection device according to claim 1, wherein the optical system forms the thin line shape by extending the first beam and the second beam in a direction perpendicular to the rotation direction. The defect inspection device according to claim 7 , wherein the processor detects the waviness defect having periodicity along the rotation direction. the defect inspection apparatus further includes a third beam splitter that splits the light beam directed toward the first sensor into a first polarization component and a second polarization component that are orthogonal to each other;
the first sensor detects the first polarization component output by the third beam splitter;
the defect inspection apparatus further includes a third sensor that detects the second polarized light component output by the third beam splitter;
2. The defect inspection device according to claim 1, wherein the processor detects the step defect using a result obtained by dividing a difference between the first polarization component and the second polarization component by a sum of the first polarization component and the second polarization component. the second sensor includes four detection elements arranged along the short-side direction and the long-side direction of the thin linear beam spot of the light beam irradiated onto the sample,
4. The defect inspection device according to claim 3, wherein the processor detects the height and period of the waviness defect based on the signal detected by each of the detection elements. the processor detects heights of the waviness defects having periodicity along the short-side direction using two of the detection elements arranged along the short-side direction,
11. The defect inspection device according to claim 10, wherein the processor detects the height of the waviness defect having periodicity along the longitudinal direction by using two of the detection elements arranged along the longitudinal direction. the second sensor includes a first detection element and a second detection element arranged along the longitudinal direction,
the second sensor includes a third detection element adjacent to the second detection element in the short-side direction,
the second sensor includes a fourth detection element adjacent to the first detection element in the short-side direction,
the processor detects the waviness defect having periodicity in the short-side direction by using a result obtained by subtracting a sum of a third detection signal output by the third detection element and a fourth detection signal output by the fourth detection element from a sum of a first detection signal output by the first detection element and a second detection signal output by the second detection element;
12. The defect inspection device according to claim 11, wherein the processor detects the waviness defect having periodicity in the longitudinal direction by using a result obtained by subtracting a sum of the second detection signal and the third detection signal from a sum of the first detection signal and the fourth detection signal. the defect inspection apparatus further includes a rotation mechanism that rotates the sample,
the optical system irradiates the first beam and the second beam onto the sample at intervals along a rotation direction of the sample;
the optical system extends the first beam and the second beam along a direction perpendicular to the rotation direction, thereby forming the thin line shape whose short-side direction is the rotation direction and whose long-side direction is the perpendicular direction;
When detecting the height of the waviness defect along the rotation direction, the processor extracts a signal intensity corresponding to the period of the waviness defect by performing a Fourier transform using the number of rotations for each position along the orthogonal direction;
12. The defect inspection device according to claim 11, wherein the processor detects the period and height of the undulation defect without performing the Fourier transform when detecting the height of the undulation defect along the orthogonal direction. 2. The defect inspection device according to claim 1, wherein the processor identifies the coordinates of each of the detected waviness defects and step defects and outputs the results. 2. The defect inspection device according to claim 1, wherein the processor discriminates between the period and height information of the detected undulation defect and the height information of the detected step defect, and outputs the results.