TWI908100B - Charged particle beam device - Google Patents

Abstract

This invention provides a charged particle beam device that can avoid shadows or distortions caused by the difference in charge of each pattern and improve the processing capacity. The charged particle beam apparatus comprises: a scanning deflector that scans a charged particle beam emitted from a charged particle source; a signal electron deflector that deflects the orbit of signal electrons emitted from a sample; a detector that detects signal electrons obtained by scanning the charged particle beam; and a booster electrode for boosting the signal electrons to the detector; and further comprises: a computing unit that, when capturing an area containing a plurality of patterns, determines imaging conditions such as uniform distribution of the length measurements of the plurality of patterns based on at least one or a combination of two control parameters, namely, irradiation current density, booster electric field, accelerating voltage, magnification, scanning method, and scanning rotation.

Description

Charged particle beam device

This invention relates to a charged particle beam device.

As semiconductor patterns become miniaturized and highly integrated, even minute shape variations affect device operation, increasing the need for shape management. Consequently, scanning electron microscopes (SEMs) used for semiconductor inspection and measurement require higher sensitivity and precision than before. A scanning electron microscope is a device that detects electrons emitted from a sample, generating signal waveforms by detecting these electrons, for example, measuring the dimensions between peaks (pattern edges).

In recent years, progress has been made in the introduction of EUV (Extreme Ultraviolet) lithography, a technology for forming fine patterns below 10nm on wafers. However, in EUV lithography, the identification of randomly generated defects, termed random defects, has become a problem. This increases the demand for full-surface wafer inspection, requiring inspection equipment with higher throughput.

To improve inspection efficiency (processing capacity), a large-area inspection is considered using low-magnification imaging with high current to examine a single area. Low-magnification observation requires capturing multiple patterns at once, but it is known that the sample charge is different for each pattern. The effect of this charge is significant, especially causing orbital deflection of signal electrons generated from the sample. This results in various phenomena that reduce inspection accuracy, such as image distortion, shadows (uneven brightness), and abnormal contrast. Furthermore, low-magnification imaging produces instances where multiple patterns are captured within a single field of view. If the patterns are different, the sample charge is also different, so shadows or distortions caused by the charge difference between each pattern become a problem. To suppress the influence of sample charge on the signal electron orbit, several charge control methods are proposed.

For example, Patent 1 describes a method for irradiating charged particles with a beam. For samples containing multiple different materials within the irradiation area, or where the density of patterns within the irradiation area varies by location, the method divides the irradiation area into multiple regions and applies charges using beams with different irradiation conditions to suppress uneven charging. This technique risks reducing throughput due to the need for region segmentation. Furthermore, no method for imaging without pre-irradiation is described.

Furthermore, in Patent 2, when acquiring an image of the field of view (FOV), beam irradiation points with gaps are set. A key feature is the control of the deflector to scan the charged particle beam at a higher speed than when irradiating the positions on the sample corresponding to each irradiation point (corresponding to the positions on the sample for signal detection pixels), by irradiating the positions on the sample between irradiation points. This describes a method for mitigating or controlling the charge effects in small areas within the FOV. While Patent 2 describes the charge difference in small areas, it does not describe imaging involving multiple patterns.

[Previous Technical Documents]

Patent Document 1: Japanese Patent Application Publication No. 2011-210509

Patent Document 2: International Publication No. 2015-045498

In charged particle beam devices, shadows or distortions caused by the charge difference between each pattern become a problem in imaging containing multiple patterns. For example, Patent 1 proposes a method for creating images by segmenting areas under different imaging conditions, but it does not disclose a method for capturing different patterns under conditions where a charge difference is not generated simultaneously. Similarly, Patent 2 proposes a method for avoiding the charging of small areas by adjusting the illumination point and scanning speed, but it does not disclose the case of multiple patterns. Based on the above, it is desirable to capture images containing multiple patterns without prior illumination or area segmentation; therefore, simultaneous control of the charge of multiple patterns is necessary.

Therefore, this invention provides a charged particle beam device that avoids shadows or distortions caused by differences in the charge of each pattern, and increases processing capacity.

To solve the above problems, the charged particle beam device of the present invention comprises: a scanning deflector that scans a charged particle beam emitted from a charged particle source; a signal electron deflector that deflects the orbits of signal electrons emitted from a sample; a detector that detects signal electrons obtained based on the scanning of the charged particle beam; and a lifting electrode for lifting the signal electrons to the detector; and further comprises: a computing unit that, when capturing an area containing a plurality of patterns, determines imaging conditions such as uniform distribution of the length measurements of the plurality of patterns based on at least one or a combination of two control parameters, namely, irradiation current density, lifting electric field, accelerating voltage, magnification, scanning method, and scanning rotation.

According to the present invention, a charged particle beam device can be provided that avoids shadows or distortions caused by differences in the charge of each pattern, and increases the throughput.

Other issues, components, and effects not described above will be clarified based on the following description of the implementation.

100: Scanning Electron Microscopy (SEM)

101: Electrocution Gun

102: Electron Beam

103: Focusing Lens

104:1 electron deflector

105: Object Mirror

106:Sample

107: Signal Electron Deflector

108: Focusing Lens (Aperture Angle Adjustable Lens)

109: Detector

110: Signal Electron Iris

111: Signal Electron Deflector

112: Energy Filter

113: Detector

114: Operations Department

115: Memory Section

1100: Image display area

1101: Location Information Input Area

1102: Pattern Input Area

1103: Control Parameter Input Area

1104: Application Buttons

1105: Optimal output range

1300: Multibeam SEM

1301: Beam Splitter

1302: Multiple Detectors

A: Pattern

B: Pattern

C: Pattern

1400: Area containing pattern A with fine lines and space

1401: Area containing pattern B with thick lines and space

1402: Area containing pattern C without a pattern

S300~S302: Steps

S500~S502: Steps

Figure 1 is a schematic diagram showing the configuration of the scanning electron microscope (SEM) of Embodiment 1 of the present invention.

Figure 2 shows the distribution of the length measurements for each charged element.

Figure 3 is a flowchart illustrating how the experimental conditions determine the imaging process.

Figure 4 is an example of a database displaying the distribution of control parameters and measurement values.

Figure 5 is a flowchart illustrating how the database determines the imaging conditions.

Figure 6 is an example illustrating the relationship between illumination current and charged potential.

Figure 7 is an example illustrating the relationship between the illumination current and the charged potential when the electric field is changed.

Figure 8 is an example illustrating the relationship between the boosted electric field and the charged potential.

Figure 9 is an example illustrating the relationship between the increased electric field and the charged potential when the irradiation current is changed.

Figure 10 is an example of a database showing control parameters and charged potentials.

Figure 11 shows an example of a GUI (Graphical User Interface).

Figure 12 shows an example of three different patterns.

Figure 13 shows the relationship between the irradiation current and the charged potential for each pattern before adjustment.

Figure 14 shows the relationship between the irradiation current and the charged potential for each pattern after adjusting the accelerating voltage.

Figure 15 shows the relationship between the irradiation current and the charged potential for each pattern after adjusting the accelerating voltage and the boosting electric field.

Figure 16 is an example of the distribution of 2D pattern shapes from a re-examination SEM of Embodiment 2 of the present invention.

Figure 17 is a schematic diagram showing the configuration of the multi-beam SEM of Embodiment 3 of the present invention.

This manual describes charged particle beam devices such as SEMs and focused ion beam (FIB) devices, but SEMs will be used as an example in this explanation.

Furthermore, in this specification, secondary electrons (SE) and back-scattered electrons (BSE) generated from the sample are referred to as signal electrons generated from the sample.

The following uses diagrams to illustrate embodiments of the present invention.

Implementation Example 1

This embodiment illustrates an example assuming length measurement using SME (CD-SEM, Critical Dimension-Scanning Electron Microscope).

Figure 1 is a schematic diagram showing the configuration of the charged particle beam apparatus, i.e., the scanning electron microscope (SEM) 100 of this embodiment.

As shown in Figure 1, the scanning electron microscope 100, as its main components, includes an electron gun 101, a condenser lens 103, a primary electron deflector (scanning deflector) 104, an objective lens 105, a signal electron deflector 107, a condenser lens (aperture angle adjustment lens) 108, a detector 109, a signal electron aperture 110, a signal electron deflector 111, a detector 113, a computing unit 114, and a memory unit 115. Furthermore, although not shown, the scanning electron microscope 100 also has a display unit for receiving user input, various parameters, and observation patterns. Here, the computing unit 114 is implemented using, for example, a processor such as a CPU (Central Processing Unit) not shown, ROM (Read Only Memory) storing various programs, RAM (Random Access Memory) temporarily storing data from the computing process, and external memory devices. The CPU or other processor reads and executes the various programs stored in the ROM, and stores the execution results, i.e., the computing results, in cloud storage via RAM, external memory devices, or network connections.

As shown in Figure 1, the scanning electron microscope 100 focuses the electron beam (primary electron beam) 102 generated by the electron gun 101 onto the sample 106 using a condenser lens 103 and an objective lens 105. The aperture angle of the electron beam (primary electron beam) 102 can be adjusted using a condenser lens (aperture angle adjustment lens) 108. A primary electron deflector (scanning deflector) 104 causes the electron beam (primary electron beam) 102 to scan the electron beam scanning area of the sample 106. An electron beam (primary electron beam) 102 is used to excite the sample 106 by 2D scanning. Detectors 109 and 113 detect the signal electrons emitted from the sample 106, and the processing unit 114 converts the detected signal into an image, thereby obtaining an observation image of the sample 106. The signal electrons emitted from the sample 106 pass through the signal electron deflector 107 and are divided into electrons that pass through the signal electron aperture 110 and electrons that collide with the signal electron aperture 110. Electrons that collide with the signal electron aperture 110 generate tertiary electrons, which are detected by the detector 109. Electrons that pass through the signal electron aperture 110 are deflected by the signal electron deflector 111 and detected by the detector 113. As shown in Figure 1, in some scanning electron microscopes, an energy filter 112 is provided before the detector 113 to distinguish signal electrons with energy. The detector 113 detects the electrons passing through the energy filter 112. The change in the signal quantity when the voltage applied to the energy filter 112 is automatically adjusted to estimate the charge state of the sample 106. However, the charge measurement of the energy filter 112 is time-consuming, which is unrealistic as it is not feasible to aim for high throughput measurements of 1 ^cm² /hr or higher in the future. The computation unit 114 performs the following operations: control of the various optical elements of the scanning electron microscope 100; control of the voltage applied to the energy filter 112; control of the deflection amount of the signal electron deflector 107; and calculation of the synthesis ratio of the signals detected by the detectors 109 and 113. Furthermore, the computation unit 114 uses the detection signals of the signal electrons detected by the detectors 109 and 113 to create an observation image of the sample 106. The memory unit 115 is a memory device for the data used by the computation unit 114. For example, it can store a database (DB) containing the observation conditions and charged potentials of each pattern as shown in Figure 10. The computing unit 114 determines imaging conditions such as illumination current density, boosting electric field, accelerating voltage, scanning method, and scanning rotation based on a database (DB) stored in the image memory. Alternatively, other computing devices can be installed to replace the computing unit 114 as the component for determining imaging conditions. Furthermore, imaging conditions can be experimentally determined even without a database (DB).

First, the method for experimentally determining observation conditions without using a database (DB) will be explained. Figure 2 shows the distribution of length measurements for each charged pattern. As shown in Figure 2, the length measurement distribution differs depending on the charge. Specifically, in the case of a positively charged pattern shown in the left image of Figure 2, the length measurements at the center of the field of view are coarser than those at the outer edge of the field of view. Conversely, in the case of a negatively charged pattern shown in the right image of Figure 2, the length measurements at the center of the field of view are finer. This is due to the difference in charge between the positive and negative values at the center and the outer edge of the field of view. On the other hand, in the case of zero charge shown in the central image of Figure 2, no length measurement distribution is generated. Utilizing this feature, by selecting an observation condition where the distribution of all length measurements for multiple patterns becomes flat, the charge of multiple patterns can be controlled to zero.

The flowchart shown in Figure 3 illustrates the method for determining the shooting conditions. First, in step S300, the area to be photographed and the patterns contained within that area are selected. This time, the case of a combination of patterns A and B is considered.

Next, in step S301, control parameters and control ranges are set. For example, the control parameters are set to the irradiation current and the boosting electric field, and the control ranges for each parameter are set to 8pA~500pA (irradiation current) and 2kV/mm~4kV/mm (boosting electric field). Within the control parameters and control ranges specified in step S301, the condition with the smallest deviation in the distribution of length measurement values within the field of view is selected to determine the imaging conditions (step S302). At this time, the length measurement value distribution can be used to determine how to change each parameter appropriately. For example, in the positively charged example shown in the right figure of Figure 2, a flat distribution is obtained by increasing the irradiation current or decreasing the boosting electric field. Conversely, in the negatively charged example shown in the right-hand diagram of Figure 2, a flat distribution is achieved by reducing the irradiation current or increasing the boost electric field. The reason for this is explained below using Figures 6-9.

Next, the method for determining imaging conditions based on a database (DB) will be explained. A database (DB) for the distribution of measurement values for the control parameters, as shown in Figure 4, is created in advance. Here, the control parameters are represented as a 2D color graph with the illumination current and the boosting electric field as axes, and the measurement values at the center of the field of view (FOV) / the measurement values at the edge of the FOV as color bars. However, if there are 3 control parameters, it is 3D; if there are 4, it is 4D. The database (DB) used varies depending on the number of control parameters. The flowchart shown in Figure 5 will be used to explain the method for determining imaging conditions. When creating the database (DB), it is not necessary to include multiple patterns in the FOV simultaneously for shooting; images can be shot individually or multiple patterns can be captured collectively. In step S500, first select the area to be photographed and the patterns contained within that area. Similar to Figure 3 above, consider the case of a mixture of patterns A and B. Next, in step S501, set the control parameters and control range. Referring to the database (DB) shown in Figure 4, it suggests a shooting condition where the length values of multiple patterns are evenly distributed (step S502). At this point, it is not necessary to confirm the changes in observation conditions and length value distribution as described above; the intersection of the contour lines under the condition of even length values for each pattern is the optimal condition.

As shown in Figure 1, when an energy filter is mounted near the detector, the surface potential of the sample 106 can also be measured using the energy filter 112, selecting a condition where the charge is zero, as shown in several patterns. Figure 6 shows the charge potential when the irradiation current is set as a control parameter. If the irradiation current is continuously increased, the number of "return electrons" that return to the sample after one release increases, and the self-positive charge reverses to a negative charge. Since the rate of change varies depending on the sample, for example, if patterns A and B are drawn overlaid, an intersection point will inevitably be formed at some point. As shown in Figure 6, when the intersection point is positively charged, by increasing the boost electric field, as shown in Figure 7, the intersection point can be made to overlap with the uncharged point. If the conditions for the boosting electric field and illumination current are selected at this time, multiple patterns can be simultaneously uncharged, suppressing uneven brightness or distortion. Figure 8 shows an example of selecting the boosting electric field as the control parameter. It can be seen that if the boosting electric field is increased, the number of returning electrons decreases, and the positive charge increases monotonically. If patterns A and B are overlaid, a crossover point will inevitably form due to the difference in sensitivity to the boosting electric field. For example, in the case where the crossover point is positively charged, as shown in Figure 9, by increasing the illumination current, the crossover point overlaps with the uncharged state.

Figure 10 is an example of a database showing control parameters and charged potentials. As shown in Figure 10, the charged potentials can also be used as colored bars to create a database (DB). When the irradiation current and the boosting electric field are selected as control parameters, uncharged contour lines are formed in each graph. The intersections of these contour lines represent the optimal conditions. Since obtaining data for all points is time-consuming when creating a database (DB), data can be obtained from a few points, and the remaining data can be calculated and generated from the acquired data. Because the sample potential changes linearly relative to the boosting electric field, if at least two points are used, the slope can be extrapolated. Furthermore, simulation can also be used for estimation.

Figure 11 is an example of a GUI display. The display unit (not shown) constituting the scanning electron microscope (SEM) 100 is shown in Figure 11. On the display screen, the position to be observed is input in the position information input area 1101, and the observed pattern is displayed in the image display area 1100. The observed pattern is input in the pattern input area 1102, and the parameter to be controlled and its range are input in the control parameter input area 1103. The application button 1104 is clicked. While changing the imaging conditions, the distribution of length values or charged potentials is measured, and the points where intersections and uncharged points overlap are displayed in the optimal condition output area 1105. The calculation unit 114 executes these actions. Furthermore, in the control parameter input area 1103, the user has sufficient margin to set the illumination current (Ip [pA]), magnification (Magnification), boost voltage (Vb [kV]), and acceleration voltage (Vacc [kV]). If the application button 1104 is clicked, the calculation unit 114 will output the optimal values for the illumination current (Ip [pA]), magnification (Magnification), boost voltage (Vb [kV]), and acceleration voltage (Vacc [kV]) in the optimal condition output area 1105.

As explained, controlling the charge of multiple patterns requires optimizing multiple parameters, such as the illumination current or the boosting electric field. When the boosting voltage also acts as a deceleration voltage for primary electrons, changing the boosting electric field alters the conditions of the primary optical system, potentially affecting the magnification or rotation. In such cases, feedback is sent to the primary electron deflector 104 to readjust the primary optical conditions.

Figure 12 shows examples of three different patterns. As shown in Figure 12, the three different patterns include: region 1400 containing pattern A with thin lines and space; region 1401 containing pattern B with thick lines and space; and region 1402 containing pattern C without a pattern. Figure 13 shows the relationship between the illumination current and the charged potential for each pattern before adjustment. The order of the illumination current values from smallest to largest when the charged potential reverses from positive to negative is pattern A < pattern B < pattern C. The reason is that the more patterns there are, the easier it is to form a positive charge locally, and the easier it is to generate return electrons. As shown in Figure 13, the intersections of pattern A and pattern B, pattern B and pattern C, and pattern C and pattern A are implemented with different current values. In this state, no matter how the irradiation current is adjusted, it is not possible to achieve multiple patterns simultaneously without charging. Therefore, in addition to the irradiation current, the boosting electric field and the accelerating voltage also need to be optimized for control parameters. It is important to note here that, for example, even if the scanning speed is changed, charging can be controlled, but this is equivalent to changing the irradiation current density and is not an independent parameter. For the simultaneous charging control of the three patterns, three independent charging control parameters need to be optimized. Figure 14 shows the relationship between the irradiation current and the charging potential of each pattern after the accelerating voltage is adjusted. By adjusting the accelerating voltage, the intersection points of the three patterns (patterns A, B, and C) are close, but the uncharged state and the intersection points do not overlap. As shown in Figure 15, further adjustments to the boosting electric field are made to ensure that the intersection points overlap with the uncharged state, thus achieving simultaneous uncharged operation for all three patterns. Furthermore, when there are four patterns, four independent charging control parameters are required. One such independent charging control parameter, separate from the accelerating voltage, irradiation current, and boosting electric field, is the observation magnification.

As described above, according to this embodiment, a charged particle beam device can be provided that avoids shadows or distortions caused by differences in the charge of each pattern, and increases the throughput.

Implementation Example 2

This embodiment illustrates an implementation of SEM (re-examination SEM) for hypothetical verification.

The apparatus for re-examining SEM is basically the same as that shown in Figure 1, but the inspection process differs from that of length measurement SEM (CD-SEM). It uses an optical inspection device to roughly identify specific defective areas, followed by detailed inspection with an electron microscope (SEM). Therefore, by pre-setting the inspection to a condition where all the charges of the multiple patterns contained in the wafer under inspection (sample 106) are zero, rather than simultaneously capturing multiple images, inspection omissions caused by uneven brightness or distortion can be reduced.

During the SEM review, data is matched against the design drawings, and any discrepancies are identified as defects. Therefore, the imaging conditions need to be determined based on the 2D pattern size distribution, rather than simply the measured length. Figure 16 shows an example of the 2D pattern shape distribution. As shown in the right image of Figure 16, when uncharged, the pattern consists of all equally sized perfect circles. However, as shown in the left image of Figure 16 when positively charged, the outer circles become elliptical. Experimentally, when determining imaging conditions, the optical conditions can be set to ensure the same pattern shape at all locations.

Alternatively, the pattern shape distribution can be used as a database, pre-created online, and the resulting database can be used to determine the optical conditions.

According to this embodiment, the same effect as that of Embodiment 1 described above can be achieved.

Implementation Example 3

In this embodiment, an embodiment assuming a multi-beam SEM will be described. Figure 17 is a schematic diagram showing the configuration of the charged particle beam device of this embodiment, namely the multi-beam SEM 1300. Hereinafter, components identical to those shown in Figure 1 will be labeled with the same symbols, and repeated descriptions will be omitted.

As shown in Figure 17, although many parts are similar to those in Figure 1, the multi-beam SEM 1300 has a beam splitter 1301 positioned midway along the trajectory of the primary electrons to divide the primary beam (electron beam 102). Furthermore, since it is necessary to image the signals from each field of view separately, the number of fields of view (FOV) prepared by the multi-detector 1302 is increased. Because the multi-beam SEM 1300 can capture a wide area at once, it is believed that there are more instances of multiple patterns being combined.

In multi-beam SEM1300, a key feature is that the secondary electron optical conditions significantly affect image quality. These secondary electron optical conditions are controlled by the signal electron deflector 107 and need to be adjusted to ensure that the signals from each field of view (FOV) accurately enter each detector 1302. Therefore, the signal electron deflector 107 is more important than in length measurement SEMs or re-examination SEMs.

According to this embodiment, the processing capacity can be further increased compared to Embodiment 1.

Example 1 above has already illustrated how to change the boosted electric field. However, the change in the boosted electric field is related to the energy change of the secondary electrons, therefore the secondary optical system needs further adjustment.

Furthermore, this invention is not limited to the above embodiments and includes various modifications. For example, the above embodiments are described in detail for the purpose of easy understanding of this invention and are not necessarily limited to all the described components. Also, a part of the composition of one embodiment may be replaced with the composition of another embodiment, and the composition of another embodiment may be added to the composition of one embodiment.

Claims (11)

A charged particle beam device is characterized by comprising: a scanning deflector for scanning a charged particle beam emitted from a charged particle source; a signal electron deflector for deflecting the orbits of signal electrons emitted from a sample; a detector for detecting the signal electrons obtained based on the scanning of the charged particle beam; and a boosting electrode for boosting the signal electrons to the detector; and further comprising: a computing unit that, when capturing an area comprising a plurality of patterns, determines imaging conditions such that the length values of the plurality of patterns are uniformly distributed based on at least one or a combination of two control parameters, namely, irradiation current density, boosting electric field, accelerating voltage, magnification, scanning method, and scanning rotation. As in the charged particle beam device of claim 1, the aforementioned computing unit measures the length distribution of a plurality of patterns and selects imaging conditions such that the length distribution is constant in the plane, thereby determining the aforementioned imaging conditions. For example, the charged particle beam device in claim 1 has a pre-made database. The above-mentioned computing unit refers to the above-mentioned database and selects imaging conditions such as a plurality of patterns with a certain distribution of length measurement values, thereby determining the above-mentioned imaging conditions. As in the charged particle beam device of claim 3, the database stores the distribution of the measured values relative to the control parameters. As in the charged particle beam device of claim 2, the aforementioned computing unit determines the imaging conditions by selecting imaging conditions in a manner that the measured charged potentials of a plurality of patterns are within a user-specified charged range. For example, in the charged particle beam device of Request 3, the above-mentioned database maintains the relationship between control parameters and charged potentials, and the above-mentioned charged potentials are taken as colored bars. As in the charged particle beam device of claim 6, the aforementioned computing unit selects the imaging conditions based on the aforementioned database, in a manner that the charged potentials of the measured plurality of patterns are within the charged range specified by the user, thereby determining the aforementioned imaging conditions. The charged particle beam device of claim 1 includes a display unit, on which the display unit has: a position information input area for inputting the position to be observed; an image display area for displaying the observed pattern; and a control parameter input area for inputting the parameters to be controlled and their range; and the calculation unit displays the determined imaging conditions in the optimal condition output area. The charged particle beam apparatus of claim 2 includes a display unit and a memory unit that stores the length distribution of a plurality of measured patterns and the relationship between the measured length values and control parameters such as irradiation current density, boosting electric field, accelerating voltage, magnification, scanning method, and scanning rotation; the computing unit displays the imaging conditions, such as the difference between the length distributions of the plurality of patterns being zero, on the display unit. As in the charged particle beam device of claim 1, the scanning deflector performs an optical condition readjustment once based on the imaging conditions determined by the computing unit. As in the charged particle beam device of claim 1, the aforementioned signal electron deflector performs two optical condition readjustments based on the imaging conditions determined by the aforementioned computing unit.