Scanning electron microscope with time constant measurement function mounted

The present invention relates to an electron microscope for observing a sample using an electron beam, and relates to a scanning electron microscope for detecting electrons generated from the sample by electron beam irradiation and measuring or inspecting the shape and composition of the sample.

ULSI (Ultra-Large Scale Integration) devices are being rapidly miniaturized and highly integrated, and devices with machining dimensions of several tens of nanometers are being machined. In addition, multi-layering is progressing using various new materials such as adoption of low dielectric constant film and metal gate film for high speed, and 3-layer resist for enhancing etching resistance. Therefore, the demand for dimensional accuracy (CD) control during ULSI processing is becoming stricter.

Insulators such as resists, insulating films, and low-k materials are often used in semiconductor processing processes, and the surface of this insulator (sample) is charged by electron beam irradiation. When charged, it changes the amount of secondary electrons that try to escape from the sample surface and bends the trajectory of the primary electron beam, which distorts the image of the scanning electron microscope. As a result, it is difficult to measure the true machining dimensions and shapes. For example, with ArF resist, it is not possible to determine whether line edge roughness (LER) has occurred in the etching process or whether it is an erroneous measurement of dimensions due to charging with an electron microscope. Further, in high-aspect contact hole observation, there is a problem that the shape of the contact hole is distorted and it is difficult to distinguish the upper diameter and the lower diameter of the hole. In addition to spatial changes due to electron movement and diffusion, charging changes spatially / temporally, such as attenuation due to holes and electron recombination. In addition, it may be "positively charged" or "negatively charged" depending on the observation conditions (energy, magnification, current amount, scanning method of electrons incident on the sample surface). Therefore, it is important to control the charge.

Conventionally, the following methods have been considered as charge control. For example, in Patent Document 1, primary electrons are irradiated in advance at a low magnification different from the magnification at the time of observation (predose), electrons are irradiated to a wider range than the observation area to be positively charged, and the observation area is positively charged. It is described that a stable image can be obtained by uniformizing the charge in. Further, in Patent Document 2, by pre-irradiating an electron beam having an energy different from the energy of the primary electron beam at the time of observation, positive and negative charges are canceled out or alleviated, and the influence of charges is reduced. There is. Further, according to Patent Document 3, which discloses a similar technique, the observation region and a wider region are flooded (irradiated) with a positive charge during a plurality of frame cycles, so that the scanning region and the surrounding region are flooded (irradiated). The image of the observation area is brightened and observed by reducing the voltage difference between the two and then scanning the observation area individually. Further, Patent Document 4 discloses a scanning technique for reducing the influence of surface charge when scanning a hole having a substantially circular shape. Further, there is also a method in which the center position of the hole is specified at a low magnification, and the electron beam is repeatedly scanned from the center of the hole outward in a radial direction to a point beyond the edge of the hole at a high magnification. With these technologies, the influence of charging is suppressed and accurate stable image acquisition is achieved.

Further, as a method for measuring the charging potential, Patent Document 5 discloses a technique of obtaining the time constant of the brightness from the change in the brightness of a plurality of images and outputting these time constants as potential information.

<patcit num="1"><text>Japanese Unexamined Patent Publication No. 5-151927</text></patcit><patcit num="2"><text>Japanese Unexamined Patent Publication No. 2000-20057</text></patcit><patcit num="3"><text>Special Table 2001-508592</text></patcit><patcit num="4"><text>Japanese Patent No. 3238705</text></patcit><patcit num="5"><text>Japanese Patent Application Laid-Open No. 62-52841</text></patcit>

<p> In general, it is known that the amount and distribution of charge on the surface of an insulator sample largely depends on the following factors.</p><p> a) Changes that depend on the energy of the primary electrons, current density, magnification and secondary electron / backscattered electron yield of the sample. b) Temporal changes depending on electron movement and diffusion, annihilation due to electron-hole recombination, etc. c) Scanning method dependent changes.</p><p> As described above, Patent Document 1 charges a region wider than the observation region by predose so that the potential in the observation region can be regarded as uniform, and stabilizes the image. However, in this case, even if the influence of the charge can be regarded as uniform, the time change of the charge in b) and the charge in c) do not take into consideration that the electron orbit at the time of the next scan is deflected. There is a problem that the image changes or is distorted depending on the scanning interval and the number of scannings.</p><p> Further, Patent Document 2 relieves positive and negative charging by irradiating in advance an electron beam having an energy different from that of the primary electron beam at the time of observation. In this case, for example, when electrons are irradiated with energy that is positively charged by pre-irradiation and negatively charged during observation, the charge in the observation area is relaxed, but the charge amount and distribution change with the elapsed time, and the scanning location. The distribution changes even in places where it is not considered. That is, since the phenomena b) and c) are not taken into consideration, there is a problem that there is no guarantee that the image obtained by scanning is a'true'image, and that the image changes with time and number of times. Further, Patent Document 3 has the same problem.</p><p> Patent Document 4 is a method of scanning an electron beam from the center of a circular hole to the outside, but an electric charge is accumulated at the bottom of the hole and the electron beam irradiated to the center is deflected at the next scanning. It will be. Since this changes depending on the time required for scanning and the number of scannings, it becomes difficult to stably acquire a true image.</p><p> Further, Patent Document 5 is a method of displaying the potential distribution from the brightness change of a plurality of images. The main factor of the brightness change at this time is that low-energy emitted electrons are pulled back to the sample due to the influence of charging. This is because the orbit is deflected and becomes undetectable. Therefore, the output potential information largely depends on the energy distribution of the emitted electrons and the width of the charged region, and it is difficult to accurately measure the change in the charged potential.</p><p> The present invention has been made in view of such a situation so that a scanning electron microscope can obtain a stable image by determining an optimum scanning method in which the amount of deflection of primary electron beams and secondary electrons is reduced. It is something to do.</p>

<p> In order to solve the above problems, the present invention provides a method of measuring the charged state of a sample and using the measurement result to determine the scanning order of electron beams to reduce the influence of charging.</p><p> In general, the yield of secondary electrons / backscattered electrons differs depending on the electron acceleration energy for each material used for ULSI, and the charge distribution and potential also differ depending on the current density and observation magnification of the electron beam. Therefore, it is necessary to monitor the charge that fluctuates with time for each material. Therefore, in the present invention, an energy filter is used to discriminate the energy of emitted electrons, and attention is paid to the behavior of electrons having a specific energy (particularly, electrons emitted from a sample with high energy).</p><p> That is, the present invention is a scanning electron microscope that irradiates a sample with a primary electron beam to acquire an image of the sample, and has energy for extracting electrons having an energy equal to or higher than a predetermined value among the electrons emitted from the sample. A filter, a time constant extraction means for extracting the time constant of a time change curve showing the relationship between the signal intensity of electrons extracted by the energy filter and the electron beam irradiation time, and an electron beam for a sample based on the extracted time constant. A scanning order determining means for determining the scanning order to be scanned is provided, and the sample is scanned and observed in the determined scanning order.</p><p> Here, the scanning order determining means arranges scanning points in a time-distance space based on the extracted time constant, the scanning time of one line, and the scanning conditions defined by the scanning interval, and of each scanning point. The scanning order may be determined so that the distance is maximized. Then, the scanning order is determined so that the arrangement of the scanning points is a triangle, a quadrangle, or a hexagon. Further, the scanning order determining means changes the aspect ratio of the scanning points arranged in the time-distance space according to the comparison result with respect to the time constant which is the reference of the time constant of the potential fluctuation caused by the electron beam irradiation. It may be. The time constant extraction means fixes the filter potential of the energy filter and extracts the time constant of the potential fluctuation from the time change of the brightness when the sample is irradiated with the primary electron beam. Then, the time constant extraction means controls the electron beam irradiation means so as to irradiate the primary electron beam to a plurality of positions on the sample while moving the irradiation position of the primary electron beam, and the time change of the brightness obtained at the plurality of positions. The time constant of the potential fluctuation is extracted by using the average value as the time change curve.</p><p> The scanning order determination method according to the present invention is a scanning order determination method for the sample in a scanning electron microscope that irradiates the sample with a primary electron beam to obtain an image of the sample, and among the electrons emitted from the sample, a predetermined one. A time constant extraction step for extracting the time constant of a time change curve showing the relationship between the signal intensity of the electrons extracted by the energy filter for extracting electrons having energy equal to or higher than the value and the electron beam irradiation time, and the extracted time constant. It is characterized by including a scanning order determining step of determining the scanning order of scanning the sample with the electron beam based on the above.</p><p> Another form of the scanning electron microscope according to the present invention is a scanning electron microscope that irradiates a sample with a primary electron beam to obtain an image of the sample, and has an energy equal to or higher than a predetermined value among the electrons emitted from the sample. An energy filter for extracting electrons with, and a time constant for recording the time constant of a time change curve showing the relationship between the signal intensity of the electrons extracted by the energy filter and the electron beam irradiation time for each sample observed in advance. The recording means, the sample designation means for designating the sample to be observed according to the instruction input, and the time constant corresponding to the designated sample are acquired from the time constant recording means, and the sample to be observed based on this time constant is obtained. A scanning order determining means for determining a scanning order when scanning with an electron beam is provided, and the sample is scanned and observed in the determined scanning order.</p><p> Another form of the scanning electron microscope according to the present invention is a scanning electron microscope that irradiates a sample with a primary electron beam to acquire an image of the sample, and the sample is subjected to an electron beam for each sample to be observed in advance. The scanning order storage means for recording the scanning order at the time of scanning, the sample designation means for designating the sample to be observed according to the instruction input, and the scanning order corresponding to the designated sample are acquired from the scanning order storage means. A scanning means for scanning a sample with the primary electron beam based on the scanning order is provided, and the scanning order includes electrons having an energy equal to or higher than a predetermined value among the electrons emitted from the sample for each sample. The time constant of the time change curve showing the relationship between the signal intensity of the electrons extracted by the energy filter for extracting the electron beam and the electron beam irradiation time is extracted and determined based on the extracted time constant.</p><p> Further features of the present invention will be clarified below by the best mode and accompanying drawings for carrying out the present invention.</p>

<p> According to the present invention, the optimum scanning order (which is least affected by the charge generated by electron beam irradiation) can be determined, and if the sample is scanned according to the scanning order, an accurate and stable image can be obtained. it can.</p>

Hereinafter, embodiments according to the present invention will be described in detail with reference to the accompanying drawings. It should be noted that the embodiment according to the present invention is merely an example for realizing the present invention, and the present invention is not limited thereto.

&lt;Device Configuration&gt; FIG. 1 is a diagram showing a schematic configuration of an SEM (Scanning Electron Microscope) used in an embodiment of the present invention. A primary electron beam 1 is generated by applying an extraction voltage 13 between the field emission cathode 11 and the extraction electrode 12. The primary electron beam 1 is subjected to scanning deflection by the condenser lens 14, the upper scanning deflector 21, and the lower scanning deflector 22. The deflection intensities of the upper scanning deflector 21 and the lower scanning deflector 22 are adjusted so as to scan the sample 23 two-dimensionally with the lens center of the objective lens 17 as a fulcrum. The deflected primary electron beam 1 is further accelerated by the subsequent acceleration voltage 19 by the acceleration cylinder 18 provided in the passage of the objective lens 17. The primary electron beam 1 accelerated in the subsequent stage is finely focused on the sample 23 by the lens action of the objective lens 17.

When the primary electron beam 1 irradiates the sample 23, secondary signals 2 to 3 are generated. The secondary signals considered here are secondary electrons 2 with low energy and secondary electrons 3 with high energy. Since the electric field created between the objective lens 17 and the sample 23 acts as an accelerating electric field for the generated secondary signal, it is attracted into the passage of the objective lens 17 and acts as a lens by the magnetic field of the objective lens 17. Ascend while receiving. The secondary signals 2 to 3 that have passed through the objective lens 17 pass through the scanning deflectors 22 and 21 and enter the energy filter 30, and the low-energy secondary electrons 2 cannot pass through the energy filter 30 and have high energy. Secondary electrons 3 pass through the energy filter 30. When the sample 23 is continuously irradiated with the primary electron beam 1, the sample 23 is charged and the sample potential Vs fluctuates. Here, the sample potential Vs is the sum of the charging potential ΔVs formed by electron beam irradiation and the retarding potential 25 applied to the sample holder 24. When the sample potential Vs (= Vr + ΔVs) increases due to charging, the number of secondary electrons that can pass through the energy filter 30 decreases, so that the signal electrons 4 detected by the detector 7 decrease.

Then, as shown in FIG. 2, which is a configuration diagram from the time constant extraction to the determination of the scanning order, the output of the detector 7 is supplied to the time constant extraction device 110 as the input D1. The time constant extraction device 110 extracts two charging time constants, τ1 and τ2 (see FIG. 7), from the input data D1. The extracted time constant τ1 or τ2 is stored in the time constant recording device 120, and the scanning order is determined by the scanning order determining device 130 using the saved time constant and the scanning condition D2 (scanning interval Δl, number of scans n). Then, the electron beam is scanned against the sample in the determined scanning order, and the sample is observed.

&lt;Processing Sequence&gt; FIG. 3 is a flowchart for explaining the processing sequence of the SEM according to the present embodiment.

In step S100, the time constants of the same type of sample or the same wafer are previously measured, and it is determined whether or not they are stored in the time constant recording device 120, and the necessity of the time constant measurement is determined. To measure the time constant, move the electron beam to the measurement location of the time constant in step S101. It is desirable that the measurement point of the time constant is a flat part of the sample, an observation part, or a pattern part equivalent to the observation part.

Subsequently, in step S102, the blocking potential applied to the energy filter 30 is determined, and the filter potential is set. This filter potential is a potential for taking in high-energy secondary electrons that are not affected by the local charge distribution on the sample. Then, in steps S103 and S104, the sample 23 is irradiated with the primary electron beam 1 to acquire the relationship between the secondary signal 4 and time (input D1: see FIG. 7).

In step S105, the input D1 acquired in step S104 is input to the time constant extraction device 110, and the time constant τ of the observation sample is extracted. The details of the extraction process will be described later. In step S106, the extracted time constant τ is stored in the time constant recording device 120.

In step S107, the optimum scanning order is determined based on the extracted time constant and the scanning condition (input D2). The details of the scanning order determination method will be described later. Here, the scanning conditions mean the scanning interval Δl, the number of scans n, and the scanning time Δt of 1 line.

In step S108, the sample is actually observed using the scanning order output from the scanning order determining device. Further, if the scanning order output here is saved in the recording device, the optimum scanning is performed without the need to measure the time constant or optimize the scanning order when observing the sample whose materials, structures or patterns can be regarded as equivalent thereafter. You can get SEM images in order.

If it is determined in step S100 that the measurement of the time constant is unnecessary, the process proceeds to step S110. In step S110, when the time constant has been measured in the past, the time constant is read from the time constant recording device 120. If there is no measurement experience in the past, the time constant is estimated from the material of the observation sample.

&lt;Regarding the energy filter and the process of acquiring the relationship between the secondary signal and time (step S104)&gt; FIG. 4 is a diagram illustrating the operation of the energy filter 30. As shown in FIG. 4, when the potentials of Vr (25: for example 0V) and Vf (45: for example 5V) are applied to the sample 23 and the energy filter 30, respectively, the blocking potentials Vrf (= Vr-Vf) and the potentials of the sample 23 are obtained. On the other hand, how low the potential is set) is formed. At this time, among the electrons emitted from the sample 23, those having an energy higher than the blocking potential Vrf can be detected, but the electrons having a lower energy cannot be detected.

Here, consider the case where charging is generated by electron beam irradiation and the potential on the sample rises by ΔVs (see FIG. 5). At this time, the sample potential Vs becomes Vs = Vr + ΔVs, and the blocking potential Vrf (= Vr + ΔVs-Vf) rises by ΔVs, so that the number of electrons taken out by the energy filter and detected decreases. Therefore, the brightness of the image is reduced. Using the above principle, the charging potential ΔVs (= fluctuation amount based on the part on the sample not irradiated with the electron beam) can be obtained from the brightness of the image (shaded part in FIG. 5). Further, if the voltage Vf applied to the energy filter 30 is adjusted and only the change in the yield of the electrons emitted from the sample 23 with high energy is focused on, the emission electron trajectory changes due to the sample charging. Accurate potential measurement can be performed.

As a specific method, a voltage Vf is applied to the energy filter 30, and the time change of the brightness of the image due to the electrons passing through the voltage Vf is measured. As a method of measuring the time change, the time change of the image brightness due to the electrons passing through the energy filter is measured at the observation point or the flat part in the observation wafer. At this time, the scanning method of the primary electrons may be arbitrarily set according to the purpose (single point irradiation, line scanning, TV scanning, random scanning, etc.). Further, from the viewpoint of increasing the S / N ratio, the filter potential is fixed, and the above is carried out at a plurality of locations in the observation wafer.

Further, if the measurement is performed at different locations in the wafer with different materials and pattern shapes, the potential change can be measured at any location. From these measurement results, the time change curve S (t) of the signal strength S (see FIG. 7) is obtained.

In order to determine the scanning order, it is necessary to measure the fluctuation of the sample potential of several V with a high time resolution on the order of μs and extract the time constant. There is a trade-off between the time resolution of potential measurement using an energy filter and the accuracy of the measurement potential, and when the time resolution improves, the S / N of the detected secondary signal 4 decreases and the error of the measurement potential increases. To do. Therefore, the secondary signal 4 is measured a plurality of times by moving a plurality of measurement points on the wafer. If the average value of the measured signal values is used, the S / N can be improved, and both the measurement accuracy of the potential and the time resolution can be achieved. Figure 6 shows the time constant measurement sequence.

In FIG. 6, in step S201, the number of measurements M for obtaining the average value of the signal values is determined. The value of the number of measurements M is determined by first measuring the signal value once and then observing the S / N of the measurement result. That is, in order to obtain the time change curve of the signal as shown in FIG. 7, it is necessary to acquire the relationship between the time and the amount of secondary electron signals, but it is not possible to obtain a good signal by only one measurement. Therefore, in order to improve the S / N, it is decided to measure M times on a flat place.

In step S202, the electron beam is irradiated to acquire the time change of the secondary signal 4. Then, the measurement point is moved in step S203. In step S204, the measurement is repeated until the number of measurements reaches M.

Finally, in step S205, the average curve of the measured M signal values is obtained and passed to the time constant extraction device 110 as the input D1.

FIG. 7 shows the relationship (example) between the signal electron 4 obtained as described above and the irradiation time of the electron beam.

&lt;About the time constant extraction process (step S105)&gt; Next, the time constant extraction process executed by the time constant extraction device 110 will be described in detail.

When the time change of the signal strength is expressed as S = S0 · exp (-t / τ) with the coefficient S0 and the time constant τ, the time constant τ can be obtained by fitting with S (t) in FIG. That is, when the sample is positively charged by the primary electron beam, the signal strength changes as shown in FIG. The first rapid change is a state in which positive charges are accumulated by the electron beam, and the brightness becomes dark. This time constant is expressed as τ1 (time constant that characterizes the process in which positive charges are accumulated by electron beam irradiation). What changes slowly after that is the difference between the process of charge disappearance due to the dissipation of electrons due to movement / diffusion and the recombination of electrons / holes, and the supply of positive charges by the primary electron beam, and this time constant is τ2. Two of τ1 and τ2 are extracted as the charging time constants. Under normal SEM observation conditions, τ1 &lt;τ2.

Here, the relationship between the observation time and the SEM image when observing a sample whose potential fluctuates with a certain time constant τ will be considered.

1) If the observation time is sufficiently smaller than the time constant τ, the SEM image is not distorted because it is not affected by charging. 2) When the observation time and the time constant τ are about the same, the effect of charging changes during observation, so the SEM image is distorted unevenly. 3) If the observation time is sufficiently longer than the time constant, the magnification will fluctuate, that is, the image will be uniformly distorted because it will continue to be affected by the same magnitude of charge during observation.

Therefore, if you want to minimize the amount of distortion for each line, you can select a time constant that is close to the acquisition time of one line and determine the scanning order. If you want to minimize the amount of distortion for each frame, you can select the time constant that is close to the acquisition time of one frame. A constant may be selected to determine the scanning order.

In this way, which of τ1 and τ2 is used as the time constant is determined by the time required for scanning and the characteristic time of the phenomenon of interest. That is, if the scanning order is determined by paying attention to the time required for scanning (time for line scanning), the difference in the amount of distortion within one frame is minimized. In addition, the scanning order pays attention to the characteristic time of decrease, that is, the time required for observation (acquisition time of several pixels, several lines, one frame, one image, several images, several tens of images, etc.). When determined, the amount of distortion with respect to the time required for observation is minimized.

&lt;Process of Scanning Order Determination (Step S107)&gt; Next, the scanning order determination process executed by the scanning order determination device 130 will be described in detail.

The effect of the charge accumulated by the electron beam irradiation can be eliminated by performing the next scan as far as possible from the place where the electron beam is irradiated, or by leaving as much time as possible until the next irradiation. I know it. However, simply increasing the scanning distance and time interval reduces the image acquisition throughput.

Therefore, in order to optimize the scanning method, the relationship between the distance and time exerted by electron beam irradiation will be quantified as follows.

Here, the quantification method is shown by taking the case of scanning an electron beam for each line as an example. Let Δl be the distance from the current scanning line to the next scanning line, Δt be the time required for the next scanning (scanning time), and Q be the amount of charged charge. Here, the charge amount Q is expressed as Q = Q0exp (-Δt / τ) from the time change of the above signal strength. At this time, the force F received by the electron beam when scanning a position l away from the previously scanned line is expressed as F = Q / l, so the deflection amount ΔL of the electron beam is ΔL = with a as a proportional constant. It is expressed as a · Q0exp (-Δt / τ) / l (1). Further, assuming that the number of scans in the observation region is n, the deflection amount L of the primary electron in the entire observation region is expressed as L = ΣΔL = Σa · Q0exp (-Δt / τ) / l (2).

Now, as an example, FIG. 8 shows the arrangement of scanning points when line scanning is performed 9 times (scanning time Δt for one line, scanning interval Δl). Figure 8 (a) shows the arrangement of scanning points when scanning in order from the top of the image. Since the first scan scans the top line, the scan point is placed at A. Since the second scan scans the next line after Δt seconds of the first scan, a scan point is placed at B. During the second scan, the primary electrons are deflected under the influence of the charge formed during the first scan. The deflection amount ΔL at this time is a, Q0, exp (-Δt / τ) / Δl from Eq. (2). When scanning in order from the top in this way, since the distance between the scanning points is short, the influence of the charge formed in the previous scanning is greatly affected.

Therefore, in order to minimize the image distortion due to charging, the scanning points may be arranged so as to minimize the total deflection amount L = ΣΔL represented by Eq. (2). This is equivalent to solving the problem of embedding n scan points in a distance-time space in (Δl, Δt) increments. As a method for that, the scanning points should be arranged evenly in the distance-time space without overlapping, and the scanning points should be arranged so as to be the maximum in the time-distance space. Become. As a result of the above consideration, it can be seen that the most efficient arrangement (equal arrangement) is to fill the scanning points with triangles, quadrangles, or hexagons.

When the scanning time Δt and the scanning interval Δl of one line have an equivalent effect on the deflection amount of the primary electron beam 1, the arrangement is as shown in FIG. 8 (b). According to Fig. 8 (b), the scanning order is line # 4 # 7 # 2 # 9 # 5 # 1 # 8 # 3 # 6, and all 9 lines are scanned in this order. Will be done.

If the time constant τ of the sample is larger than that in FIG. 8 (b), the shape of the scanning point may be stretched horizontally. If the time constant τ of the sample is smaller than that in FIG. 8 (b), the shape of the scanning point may be stretched vertically. Therefore, the shape is as shown in FIG. 8 (c) or FIG. 8 (d). The arrangement method shown in FIGS. 8 (b) to 8 (d) is an example in which the scanning points are set to a quadrangle.

In this way, if the time constant τ of the time change of the amount of electric charge is obtained, it can be seen that the scanning method that minimizes the equation (2) can be uniquely determined.

&lt;Other Embodiments&gt; In the present invention, the time constant of the time change curve showing the relationship between the signal intensity of electrons extracted by the energy filter for each of a plurality of samples in advance and the electron beam irradiation time is recorded in the time constant recording device 120. However, it may be used to determine the optimum scanning order of the sample. That is, the sample to be observed by the user is specified by an input means (keyboard, mouse, etc.) (not shown). Then, the time constant corresponding to the designated sample is read from the time constant recording device 120, and the scanning order when scanning the sample to be observed with the electron beam is determined based on this time constant. The scanning order is determined in the same manner as described above.

Further, for each sample to be observed in advance, the optimum scanning order when scanning the sample with an electron beam may be stored in a memory (not shown), and the sample may be observed using this. That is, the sample to be observed by the user is specified by an input means (keyboard, mouse, etc.) (not shown). Then, the scanning order corresponding to the designated sample is read from the memory, and the sample is scanned by the primary electron beam based on the scanning order. The optimum scanning order stored in the memory is determined for each sample by the method described above.

&lt;Summary&gt; In the present embodiment, among the electrons emitted from the sample, the electrons having energy equal to or higher than a predetermined value are taken out by the energy filter. Then, the time constant of the time change curve showing the relationship between the signal intensity of the extracted electrons and the electron beam irradiation time is extracted, and the scanning order for scanning the sample with the electron beam is determined based on this time constant. In this way, the energy filter is used to remove the low-energy electrons from the electrons emitted when the sample is charged, and to extract the high-energy electrons that are not easily affected by the charge, so that the charge can be performed more accurately. The potential can be estimated. Therefore, since the extracted time constant can be measured with a time resolution on the order of μs, the optimum scanning order can be determined using it.

In addition, when determining the scanning order, scanning points are arranged in a time-distance space based on the extracted time constant, the scanning time of one line, and the scanning conditions defined by the scanning interval, and each scanning is performed. The scanning order is determined so that the distance between the points is maximized. By doing so, the influence of the charge formed during the previous scan can be minimized. If the scanning order is determined so that the scanning points are arranged in a triangle, a quadrangle, or a hexagon, the distance between the scanning points can be increased, so that the charges formed by each scanning affect each other. The most efficient arrangement can be realized without giving a large amount.

Further, the aspect ratio of the scanning points arranged in the time-distance space may be changed according to the comparison result with respect to the time constant which is the reference of the time constant of the potential fluctuation caused by the electron beam irradiation. That is, when the time constant τ is larger than the time constant τ at that time with reference to FIG. 8 (b), the shape of the figure (for example, a quadrangle) composed of scanning points is stretched horizontally (Fig. 8 (c)). When the constant τ is small, the shape is stretched vertically (Fig. 8 (d)). By determining the scanning order in this way, the optimum scanning order can be uniquely determined.

When extracting the time constant, the filter potential of the energy filter is fixed, and the time constant of the potential fluctuation is extracted from the time change of the brightness when the sample is irradiated with the primary electron beam. Since the potential can be measured with high accuracy by paying attention only to the change in the yield of the detected electrons, a more accurate time change curve of the signal intensity can be obtained, and therefore an accurate time constant can be extracted. In order to improve the S / N, it is advisable to measure while moving the irradiation position of the primary electron beam, and to use the time-varying average value of the brightness obtained at a plurality of positions as the time-changing curve.

<figref num="1">It is a figure which shows the whole schematic structure of the SEM of this invention.</figref><figref num="2">It is a device configuration diagram from the time constant extraction to the scanning order determination according to the embodiment of the present invention.</figref><figref num="3">It is a flowchart for demonstrating the process of the scanning order determination by Embodiment of this invention.</figref><figref num="4">It is operation explanatory drawing of the blocking potential type energy filter.</figref><figref num="5">It is a figure which shows the energy characteristic when charge is generated by electron beam irradiation, and the potential on a sample rises by ΔVs.</figref><figref num="6">It is a flowchart which shows the process for acquiring the time change curve which shows the relationship between a quadratic signal and time.</figref><figref num="7">It is a relationship diagram of the amount of secondary electron signals and the irradiation time of an electron beam.</figref><figref num="8">It is explanatory drawing of the scanning order determination method.</figref>

Code description

1: Primary electron beam, 2: Low energy secondary electron, 3: High energy secondary electron, 4: Signal electron, 5: Scintillator, 6: Light guide, 7: Detector, 11: Electroelectric emission cathode, 12 : Extraction electrode, 13: Extraction voltage, 14: Condenser lens, 17: Objective lens, 18: Acceleration cylinder, 19: Post-stage acceleration voltage, 21: Upper scanning deflector, 22: Lower scanning deflector, 23: Sample, 24: Holder, 25: retarding voltage, 26: electrostatic deflector, 27: electrostatic deflector (mesh), 28: deflection coil, 29: reflector,