JP4305421B2 - Electron beam adjustment method, charged particle optical system controller, and scanning electron microscope - Google Patents

Description

  The present invention relates to a charged particle beam apparatus, and more particularly to a measurement method and apparatus using an apparatus for measuring or inspecting the size and shape of a pattern formed on a sample.

  In recent years, with the high integration and miniaturization of semiconductor elements, a wide variety of patterns are formed on a wafer, and the evaluation and measurement of their shapes and dimensions are becoming increasingly important. In order to process a large number of measurement points automatically and at high speed, it is important how fast the detection of the measurement points can be performed. For this purpose, when moving to the measurement point, the pattern is focused on. In addition, it is essential that a desired observation magnification is set.

  In the charged particle optical system, the focus condition on the wafer is determined from the acceleration voltage of the charged particle source, the voltage applied to the wafer, and the height of the wafer. For example, in Patent Document 1, a wafer is irradiated with laser light, the height of the wafer is detected using the reflected light, and the obtained height information is used as an objective that is one of control devices for the charged particle optical system. A method is used in which feedback is applied to the lens control system and necessary excitation is applied to the objective lens at the same time as the movement to the measurement point is completed.

JP-A-11-126573 JP 7-176285 A JP 2001-52642 A

  However, in recent years, wafers with a fixed charge that remain even after grounding have been scattered. The cause of this fixed charging is said to be that the polar substance in the resist is polarized by the friction at the time of applying the resist by a spin coater and the potential is fixed, or that the charging is caused by the etching process using plasma. .

  Such charge remaining on the sample causes a focus shift of the charged particle beam, and further causes a variation in magnification and an error factor of measurement by the charged particle beam apparatus. In order to solve the problem of defocus, for example, as disclosed in Patent Document 2, a method for storing a focus offset value for each measurement point of a scanning electron microscope so that automatic measurement is not hindered by defocus There is. Also, as described in Patent Document 3, a method has been devised in which a plurality of electrometers are installed near a sample in a vacuum and a value based on the measurement result is fed back to a retarding voltage. ing.

  However, the technique disclosed in Patent Document 2 has the following problems. Since the charging voltage of the wafer depends on the temperature, humidity, resist state, and plasma intensity in the manufacturing process, it is not constant even through the same manufacturing process. Therefore, even if the focus shift is stored in the automatic measurement file, it is necessary to update for each wafer. Therefore, the time required for measuring the wafer becomes longer and the productivity is lowered. In addition, since the charged potential of the wafer remains the same, the acceleration voltage that is actually desired to be used differs from the actual acceleration voltage, so the fine structure and contrast that appear in the secondary charged particle image that is formed are different, or the magnification is controlled. Problems such as errors are not solved.

  Further, as disclosed in Patent Document 3, when an electrometer installed in a vacuum is used, the potential cannot be measured unless it is moved to a measurement point, so that the time required for measurement per point is long. In addition, when a failure occurs, it is necessary to open the charged particle optical system and stage, which are vacuum vessels, to the atmosphere once, so that maintenance is poor, and it is necessary to adjust so that a plurality of electrometers always use the same output. There are some problems.

  A first object of the present invention is to provide a method and an apparatus for detecting a charge state specific to a sample without performing charge measurement for each measurement point.

  A second object of the present invention is to provide a sample size measuring method, a magnification adjusting method, and an apparatus for realizing them, which are suitable for reducing or eliminating magnification fluctuations and measurement errors due to charging.

  In order to achieve the first object, in the present invention, the potential distribution on the sample is measured by an electrostatic potentiometer that measures the potential on the sample during the passage of the sample carried by the charged particle beam carrying-in mechanism. We propose a method to do this.

  In order to achieve the second object, the present invention proposes a method for measuring the charge at a specific location on a sample and measuring the charge amount separately from the charge amount. As another method for achieving the second object, a charge function at a specific location is irradiated under at least two charged particle irradiation conditions, and a fitting function representing a change in the charging voltage with respect to the change in the irradiation condition is obtained. A method for correcting the pattern dimension based on the formed fitting function is proposed.

  The specific configuration of the present invention will be described in more detail in the section of the best mode for carrying out the invention.

  According to the example of the present invention, it is possible to appropriately correct the focus shift based on the charge inherent to the sample.

Example 1
Hereinafter, embodiments of the present invention will be described with reference to the drawings. In this embodiment, a scanning electron microscope (SEM) will be described as an example. However, the present invention is not limited to this, and can be applied to other charged particle beam apparatuses such as other ion beam irradiation apparatuses. Is possible. In addition, this embodiment describes an example of detecting secondary electrons and / or reflected electrons that are one of the charged particles, but the present invention is not limited to this, and other charged particles such as secondary ions, for example. May be detected.

  FIG. 1 shows the overall configuration of the present invention. The overall control unit 42 passes through the charged particle optical system control device 41, the stage control device 40, and the wafer transfer device 28 based on the acceleration voltage, wafer information, observation position information, etc. of the charged particles input by the operator from the user interface 43. Thus, the entire apparatus is controlled.

  Upon receiving the command from the overall control unit, the wafer transfer device 28 takes out the wafer from the wafer cassette 29 by the transfer arm 30, and opens the gate valve 26b that separates the sample exchange chamber 25 held in vacuum from the outside that is the atmosphere. Wafer is introduced into the sample exchange chamber. The wafer that has entered the sample exchange chamber is sent to the sample chamber 24 through the gate valve 26 a and is fixed on the sample stage 21.

  The charged particle optical system control device 41 is a high voltage control device 34, a retarding control unit 33, a condenser lens control unit 35, an amplifier 36, an alignment control unit 37, a deflection signal control unit 44, an objective lens in accordance with a command from the overall control unit 42. The control unit 39 is controlled. The primary charged particle beam 13 extracted from the charged particle source 11 by the extraction electrode 12 is converged by the condenser lens 14 and the objective lens 18 and irradiated onto the wafer 19. The trajectory of the charged particle beam is adjusted by the alignment coil 16 and is scanned two-dimensionally on the wafer by the deflection coil 17 that receives a signal from the deflection signal control unit via the deflection signal amplifier 38. In the following description, a configuration for sending a signal for changing an optical condition of a charged particle beam to each optical element and performing a necessary calculation may be called a control unit, a control device, or a control processor.

  A retarding voltage (a negative voltage in the case of an electron microscope) for decelerating the charged particle beam is applied to the wafer from the retarding control unit 33. The secondary charged particle 20 emitted from the wafer due to the irradiation of the primary charged particle beam 13 to the wafer 19 is captured by the secondary charged particle detector 15, and the secondary charged particle image display device is passed through the amplifier. 46 luminance signals are used. Further, since the deflection signal of the secondary charged particle image display device and the deflection signal of the deflection coil are synchronized, the pattern shape on the wafer is faithfully reproduced on the secondary charged particle image display device.

  In order to inspect and observe the pattern on the wafer at high speed, it is necessary to detect the wafer height when the sample stage moves to the desired observation point and to adjust the focus of the objective lens according to the height. . For this reason, a wafer height detection function using light is provided. The sample stage position is detected by the sample stage position detector 32, and the height detection laser emitter 22 emits light toward the wafer from the time when the sample stage approaches the vicinity of a predetermined position. The light is received at 23, and the height of the wafer is detected from the light receiving position. A focus amount corresponding to the detected height is fed back to the objective lens. As a result, when the sample stage reaches a predetermined position, the focus is already set, and the pattern can be automatically detected without any operation by the operator.

  When the wafer is not charged, the excitation current of the objective lens necessary for focusing is generally expressed by a function as shown in Expression (1).

I _obj = F (V _o , V _r , Z) (1)
Here, I _obj is the excitation current of the objective lens when the wafer is not charged, F is a function for calculating the excitation current of the objective lens, V _o is the voltage of the charged particle source, and V _r is the potential of the wafer ( Retarding voltage applied to the wafer), Z is the height of the wafer. The function F can be derived by electron optical simulation or actual measurement. Usually, the potential of the uncharged wafer is the same as the retarding voltage applied to the wafer, and the relationship shown in Expression (1) is established, so that predetermined focus control is possible. However, the excitation current of the objective lens required when the wafer itself is charged has a value as shown in the equation (2), and the focus current differs between when the wafer is not charged and when it is charged.

I _obj ′ = F (V _o , V _g ′, Z) (2)
Therefore, no matter how high the height can be detected, the secondary charged particle image fails to detect the blurred observation point because the focus is not achieved, and automatic measurement becomes impossible. I _obj ′ is the exciting current of the objective lens when the wafer is charged, and V _g ′ is the total voltage V _g = V _r + ΔV _g of the retarding voltage V _r and the wafer charging voltage ΔV _g .

  The charge of these wafers varies depending on the resist and the material of the base, but in many cases, it is charged concentrically. The present invention measures the charge amount of the wafer charged concentrically in this way and feeds back the potential. The wafer stored in the wafer cassette is taken out by the transfer arm 30 (transfer mechanism) and measured by the probe 31 while being transferred into the sample exchange chamber. The measured value is transmitted to the charged particle optical system controller via the electrostatic electrometer 45.

  In the present embodiment, an example in which the probe for measuring the potential on the sample is arranged on the sample moving trajectory transported by the transport mechanism and at a position separated from the sample will be described. There is nothing. For example, the probe may be arranged on a transport track by a mechanism for delivering the sample between the sample chamber and the preliminary exhaust chamber or a mechanism for introducing the sample from the outside to the preliminary exhaust chamber.

  As described above, since the wafer tends to be charged concentrically, if the potential distribution is measured linearly including the center position on the wafer surface, the potential distribution of the entire sample can be roughly grasped. The following description shows an embodiment particularly effective for a scanning electron microscope for measuring a sample such as a semiconductor wafer exhibiting such a potential distribution.

  FIG. 2 shows the relationship between the wafer cassette, the transfer arm, which is an element of the sample carry-in mechanism, the wafer, the electrostatic electrometer, and the sample exchange chamber. The wafer is taken out from the wafer cassette 29 by the transfer arm 30 and transferred into the sample exchange chamber 25. The probe 31 of the electrostatic electrometer is fixed by a fixing base 53 so that the center line 52 of the probe 31 is aligned with the center line 51 of the wafer on the wafer conveyance path. Since the probe of the electrostatic electrometer measures the potential of both the wafer and the transfer arm that is grounded, the wafer potential can be calibrated based on the ground potential of the transfer arm, so that a more accurate measurement value is obtained. Further, the passing position of the wafer is always constant, and since the probe is fixed by the fixing base, the positional relationship between these two does not change, so that stable measurement can always be performed. In addition, since the probe is out of vacuum, it is easy to repair or replace even if the probe fails.

  In this way, in this embodiment, the probe is arranged outside the vacuum in consideration of the ease of handling. However, the probe is not limited to this, and may be arranged somewhere along the path through which the wafer passes. good. In this embodiment, the wafer is moved so that the center of the probe coincides with the center line of the wafer. However, the present invention is not limited to this. As described above, the wafer is often charged concentrically. When this charge distribution is formed in a so-called mountain shape in which the charge amount decreases as it goes from the wafer center to the edge of the wafer, the center of the probe slightly deviates from the wafer center. However, since the potential distribution can be roughly grasped, the entire potential distribution may be grasped from the linear potential distribution shifted from the wafer center.

  FIG. 3 shows a procedure for expressing the charging voltage as a distribution function on the wafer surface from the measurement of the surface potential and feeding it back to the retarding. In general, since the wafer transfer arm does not move at a constant speed, even if the potential measurement cycle is constant in time, the coordinates on the wafer are not constant. Therefore, if the coordinates on the wafer at the time of potential measurement are calculated from the velocity pattern of the transfer arm, the potential corresponding to the exact coordinates can be obtained. A potential distribution function is created based on the obtained data. First, an approximate expression is created using an even function (a quartic function in FIG. 3) using all the obtained data.

  Next, the difference from the approximate expression is calculated at each measurement point. An error is included in the measured value of the potential, and those having a difference value larger than the set threshold value are excluded as having a large measured value error. Create an approximate expression again, excluding the excluded data. This procedure is repeated several times, and the process ends when the difference between all values finally becomes smaller than the threshold value. The function created here is a function of the electric potential with respect to the distance from the center of the wafer.

  The potential to be corrected is calculated from this function and the stage coordinates obtained from the stage controller, and the corrected potential is supplied to the wafer via the retarding control unit shown in FIG. Data acquisition is performed each time a wafer to be observed is sent to the sample exchange chamber, and is valid until the observation of the wafer is completed and a command for returning to the original wafer cassette is generated.

  The above is the description of the embodiment of the present invention. In the embodiment of the present invention, a method of feeding back the measured charging potential of the wafer to the retarding voltage as it is is shown, but the charging potential may be converted into an excitation current of the objective lens and fed back. However, in this case, it is necessary to add the retarding voltage and the charged potential of the wafer so as not to exceed the potential of the charged particle source. For example, when the voltage of the charged particle source is −2000 V and the potential of charged particles irradiated on the sample is to be −300 V, the retarding voltage applied to the wafer needs to be −1700 V.

  Considering the case of observing a wafer charged up to −290V under these conditions, even if a potential for correcting −290V is applied to the retarding, or the potential is converted into an excitation current and applied to the objective lens. Even the primary charged particle beam can reach the sample. However, in a wafer charged at a maximum of −310 V, the sum of the retarding voltage and the charging voltage is −2010 V, which exceeds the potential of the charged particle source.

  In this case, the primary charged particle beam cannot reach the sample and is reflected. In this case, a voltage 310V for correcting -310V must be applied to the retarding voltage. Further, the measured potential may be fed back to the potential of the charged particle source instead of the retarding voltage. Further, in the embodiment of the present invention, the feedback target is not a magnetic lens having a large inductance and difficult to control at high speed, but a retarding voltage. However, the objective lens is an electrostatic lens or an electrostatic lens other than the magnetic lens. And a focus correction value obtained based on the charged potential may be fed back to these electrostatic lenses.

  In addition, in the case of an SEM using a so-called boosting method in which a cylindrical electrode to which a positive voltage is applied is arranged in the objective lens, the focus may be adjusted by adjusting the applied positive voltage. good. Furthermore, it is possible to apply all other techniques for adjusting the focus of an electron beam.

  Although the present invention has been described with respect to the case where one probe of the electrostatic electrometer is placed so as to coincide with the center of the wafer, a plurality of probes may be provided. FIG. 4 is one of the configuration diagrams in the case where a plurality of probes are arranged on the wafer conveyance path to measure the entire wafer surface. A plurality of probes 31 are arranged in a matrix on a fixed base. In this case, the wafer 19 is temporarily stopped at a predetermined position during conveyance, and charging at each location is measured. Since the transfer arm is stopped, there is an advantage that there is no need to consider the relationship between the speed and the coordinates, or that a distribution function can be created even when the charge does not have an axisymmetric distribution. In addition, since a scanning electron microscope that inspects semiconductor pattern widths and defects in a fully automatic manner has a predetermined measurement location, a charge amount at the measurement location or in the vicinity thereof is selectively detected and feedback is applied. You may do it.

  In this embodiment, the feedback value of the retarding voltage may be obtained by superimposing other information as well as the feedback based on the simple charge amount. In addition, if the electrostatic potential meter malfunctions for some reason, if feedback to the retarding voltage is performed, the focus value may be shifted. When the focus evaluation value shows an abnormality, a means for diagnosing a malfunction of the electrostatic potential meter, stopping the focus feedback process based on the charged potential measurement, or warning the operator that there is an abnormality It may be provided.

  As described above, according to the present invention, it is possible to correct the charged potential even on a wafer that has been defocused due to charging so far and the success rate of pattern detection during automatic measurement has decreased. Therefore, automatic measurement can be performed in the same manner as an uncharged wafer. Further, since the charging voltage is measured for each wafer, there is an advantage that the file does not depend on the measurement file and the file need not be corrected depending on the presence or absence of the charge.

(Example 2)
In the examples shown below, it is difficult to accurately inspect and measure because different charging phenomena are mixed in the sample (semiconductor wafer, etc.). The present invention relates to a method and a device for enabling proper inspection and measurement.

In the charged particle beam device, as described above, the output information of the secondary charged particle detector is reproduced on the image display device in synchronization with the scanning of the charged particle beam. The ratio of the distance A between two points of the scanned image on the cathode ray tube (image display device) to the distance a between the two points on the sample is the observation magnification _MSEM .

M _SEM = A / a (3)
Usually, since the screen size on the cathode ray tube is fixed, the distance a between two points on the sample is inversely proportional to the observation magnification _MSEM . Therefore, by measuring the distance A between the two points of the scanned image on the cathode ray tube and dividing by the observation magnification M _SEM , the wiring dimension on the sample such as a = A / M _SEM can be derived.

  In recent years, as the miniaturization of the semiconductor industry has progressed, SEM is used in the process of manufacturing semiconductor devices or inspection after completion of the process (for example, dimensional measurement using an electron beam or inspection of electrical operation) instead of an optical microscope. Became. In a sample (wafer) in the semiconductor industry in which an insulator is used, charging of the insulator fluctuates with time due to irradiation of a primary electron beam, and a scanned image deteriorates.

  As a representative technique for avoiding this problem, the SEM described in JP-A-5-151927 has a pre-dose method in which a primary electron beam is irradiated at a magnification different from the magnification at the time of observation and the sample surface is positively charged. ing. Subsequently, a retarding method and a boosting method described in JP-A-9-171791 were developed, the retarding voltage applied to the sample was adjusted, and a primary electron beam with a low acceleration voltage of 1 kV or less that positively charges the insulator. By observing with, a stable surface charge with reproducibility can be generated, and a high resolution of about 3 nm can be realized.

  After that, as in the SEM described in Japanese Patent Application Laid-Open No. 2000-200509, a method of generating a surface charge more positively by irradiating an electron beam having energy different from the primary electron beam during normal observation in advance has been developed. By these methods, a stable high surface charging voltage can be easily generated, and a potential contrast can be observed based on the residual film on the bottom surface of the contact hole having a high aspect ratio and the difference in charging voltage.

However, when observed under the condition of the surface charging voltage, it has been found that the dimensional measurement value fluctuates by several percent as the surface charging is increased. This dimensional measurement value fluctuation amount reaches an allowable limit in a more miniaturized manufacturing process. The cause of this defect is that the observation magnification _MSEM varies with surface charging.

FIG. 9 shows the relationship between the observation magnification M _SEM and the optical magnification M _obj of the objective lens 106 and the coil current I ₇ of the scanning deflector 107 in an electron optical system composed of a scanning deflector, an objective lens and a sample. The primary electron beam 101 emitted radially from one point on the crossover surface is focused on one point on the wafer 108 surface. When the scanning deflector 107 is used to move the virtual primary electron emission point away from the central axis by 1, the sample surface is separated by M _obj . If the conversion factor and coil current of the scanning deflector 107 are K and I ₇ , respectively, the distance a between two points on the sample can be calculated by the following equation.

a = KM _obj I ₇ (4)
If the conversion coefficient of the cathode ray tube is L, the distance A between the two points of the scanned image on the cathode ray tube is expressed by the following equation.

A = LI ₇ (5)
Here, 'considering the case where the change in current for scanning the distance between two points a on the samples I ₇ from I _7' optical magnification M _obj from M _obj changed to scan images on a cathode ray tube The distance between the two points changes from A to A ′.

a = KM _obj I ₇ ′ (6)
A ′ = LI ₇ ′ (7)
Eventually, the observation magnification varies from _MSEM to _MSEM '.

M _SEM ′ = (M _obj / M _obj ′) M _SEM (8)
If the following formula is used, the dimensions can be measured correctly even if the observation magnification varies.

a = A ′ / M _SEM ′ (9)
Eventually, for high-accuracy dimension measurement, it is sufficient that the optical magnifications M _obj and M _obj ′ can be accurately calculated regardless of the presence or absence of charging.

FIG. 10 illustrates the principle of surface charging on the wafer. A retarding voltage V _r is applied to the wafer substrate. The wafer may be inherently charged as shown in (a) before observing with a SEM because of friction during resist coating by a spin coater or etching using plasma. The charging voltage (a) extends over the entire surface of the wafer, and is referred to as a global charging voltage ΔV _g . The global charging potential near the observation point is V _g = V _r + ΔV _g . The optical magnification M _obj at the global charging potential V _g can be expressed by the following equation similar to equation (1).

M _obj = M (V _o , V _g , Z) (10)
This function M can also be derived by electron optical simulation or actual measurement. On the other hand, the charging voltage ΔV _s by electron beam irradiation is local as shown in (b) and will be referred to as a local charging voltage. When both charges are superimposed, the local potential in (c) is V _s = V _g + ΔV _s .

FIG. 11 illustrates a mechanism by which the global charging potential V _g and the local charging voltage ΔV _s of the sample change the optical magnification M _obj of the objective lens. For wide area electrostatic voltage V _g is changing the potential of the objective lens 106a, the electrostatic lens is formed on a sample, focus will be shifts, combined with its focus appear as significant changes in the exciting current I ₆ . Since I ₆ changes and the incident energy to the sample also fluctuates, it converges like the trajectory 1a and the optical magnification M _obj fluctuates. On the other hand, V _g can be estimated from the fluctuation amount of I ₆ .

On the other hand, since the charging voltage ΔV _s due to the electron beam irradiation is local, there is almost no influence on the excitation current I ₆ . Nevertheless, since the local charging voltage ΔV _s forms a minute electrostatic lens 108b, the primary electrons 101 are focused like the trajectory 101b and the optical magnification M _obj is greatly changed. From the above facts, it has been confirmed that global charging has a large effect on the focus, and local charging has a large effect on the magnification.

  As described above, two charging phenomena with completely different characteristics are included, and the degree of influence of each charging on the magnification and focus is different. Therefore, even if correction is attempted without separating them, high-precision correction is possible. Could not be realized.

In order to solve such a problem, means for separately measuring and estimating the global charging voltage ΔV _g and the local charging voltage ΔV _s is provided, and means for calculating the correct optical magnification M _obj based on these data Should just be realized.

  By correcting the deflection intensity of the scanning deflector based on this magnification correction amount, it is possible to accurately display a two-dimensional scanned image at the designated observation magnification. However, in the dimension measurement of a semiconductor process, the magnification is added to the length measurement value itself. Simplification of correction is practical.

  The operation of the present invention described above will be described with reference to FIGS.

FIG. 12 shows the magnification fluctuation sensitivity coefficient T _g when the global charging potential V _g fluctuates in the range of −0.6 kV to −1.5 kV at the retarding potential V _r = −1.2 kV. From T _g and ΔV _g , the magnification fluctuation amount ΔM _g = (M _obj ′ −M _obj ) can be calculated by the following equation.

ΔM _g / M _obj = T _g * ΔV _g (11)
Here, since T _g varies depending on the observation condition before charging and the global charging potential V _g , it is necessary to store the graph of FIG. 8 in advance by experiment or calculation for each observation condition. Alternatively, there is no problem even if the magnification M _obj or M _obj ′ is obtained directly from the global charging potential V _g without using Equation (11).

On the other hand, FIG. 13 shows the magnification variation sensitivity coefficient T _s when the beam irradiation area is varied at the retarding potential V _r = −1.2 kV. T _s and [Delta] V _s ratio variation amount from .DELTA.M _s =
(M _obj ′ −M _obj ) can be calculated by the following equation.

ΔM _s / M _obj = T _s * ΔV _s / V _acc (12)
Here, T _s varies depending on the observation conditions before charging and the beam irradiation area, but as shown in the equation (12), the magnification correction amount ΔM _s and the local charging potential ΔV _s are in a good proportional relationship. Also,
T _s can be divided into four sections according to the beam irradiation area (ie, irradiation magnification).
It can be regarded as global charging at a magnification lower than 50 times. In the section from 50 times to 500 times, there is a transition region from global charging to local charging. It can be regarded as a substantially constant value in the section from 500 times to 5000 times. When the magnification is higher than 5000 times, T _s tends to gradually decrease. Therefore, if one side of the irradiation area is changed from 10 μm to 300 μm so that the magnification fluctuation sensitivity coefficient T _s includes a section of 500 times to 5000 times that can be regarded as a substantially constant value, it is stored in advance while maintaining the estimation accuracy of the correction value. This is advantageous because the number of data to be saved can be reduced.

  FIG. 5 shows a first specific example of the SEM in this embodiment. The primary electron beam 101 extracted from the cathode 104 is narrowed by the condenser lens 105 and further scanned two-dimensionally on the wafer 108 by the scanning deflector 107. The primary electron beam 101 is decelerated by a decelerating electric field between the objective lens 106 and the wafer 108 due to the negative retarding voltage applied to the wafer 108 via the sample stage 109, and on the wafer 108 by the lens action of the objective lens 106. It is squeezed thinly.

When the primary electron beam 101 irradiates the wafer 108, secondary electrons 102 are generated. The electric field created between the objective lens 106 and the wafer 108 acts as an accelerating electric field on the generated secondary electrons 102, so that it is attracted into the electron beam passage hole of the objective lens 106, and the magnetic field of the objective lens 106. Ascends while receiving the lens action. The raised secondary electrons 102 collide with the conversion electrode 110 with high energy and generate new secondary electrons 103. The secondary electrons 103 are attracted to the scintillator 111 to which a positive high voltage of about 10 kV is applied, and generate light when colliding with the scintillator 111. Although not shown, this light is guided to a photomultiplier tube with a light guide, converted into an electric signal, amplified, and then the luminance of the cathode ray tube is modulated with this output.

  In the description of FIG. 5, the control processor unit is described as being integrated with or equivalent to the scanning electron microscope. However, the control processor unit is not limited to this, and a control processor provided separately from the scanning electron microscope body is described below. Processing as described may be performed. In this case, a detection signal detected by the secondary electron detector is transmitted to the control processor, a transmission medium for transmitting the signal from the control processor to a lens or a deflector of the scanning electron microscope, and the transmission medium. An input / output terminal for inputting / outputting a received signal is required. Also, a program for performing the processing described below is registered in a storage medium, and the program is executed by a control processor having an image memory and means for supplying necessary signals to the scanning electron microscope. Also good.

In this embodiment apparatus, the global charging voltage [Delta] V _g of the measuring means (the potential distribution measuring apparatus), comprising a static electrometer as described for example in Example 1. Since the global charging voltage of the wafer is concentric, if the potential distribution is measured linearly including the center position on the wafer surface, the potential distribution of the entire sample can be roughly grasped. Therefore, specifically, as described in the first embodiment, a method in which the electrostatic electrometer probe 114 is fixed on the transfer path of the wafer 180 and the measurement is performed linearly using the movement of the transfer arm 181 is appropriate. It is. Using the measurement data, the global charging voltage ΔV _g is expressed as a function of the distance r from the wafer center, the retarding potential V _r is fed back each time the measurement point moves, and the primary electron beam 101 is incident on the wafer 108. The voltage is always set to a constant value V _acc = V _o + V _g . Here, V _o corresponds to the voltage of the cathode 104.

Furthermore, in this embodiment, an energy filter for secondary electrons is provided as a means for measuring the local charging voltage ΔV _s (potential measuring device). For example, the mesh electrode 112 is placed under the conversion electrode 110, and swept the voltage applied to the mesh electrode 112 a wide area electrostatic voltage V _g as a base point, measures a secondary electron signal amount changes (the so-called S curve). The S curve measured with the surface conductive sample is compared with the S curve at the observation point of the actual sample, and the shift voltage is defined as the local charging voltage ΔV _s .

The charging correction control unit 120 executes the above-described measurement of the global charging potential V _g and the S-curve measurement sequence until the local charging voltage ΔV _s is obtained. Furthermore, a magnification correction amount is calculated based on the obtained V _g , ΔV _s , excitation current of the objective lens 106, and the like, and the deflection intensity of the scanning deflector 107 is corrected.

  In this embodiment, in view of the fact that local charging has a greater influence on the magnification than global charging, based on the value obtained by subtracting the value corresponding to the global charging amount from the charging amount at a specific location (local charging amount). , Perform magnification correction. In the charge amount measurement using a simple energy filter, local charging and global charging (at least charging in a region larger than the scanning region, for example, a region larger than the observation region having a magnification of 50 times) are detected in a combined state. Then, by subtracting the charge amount at the scanning position of the electron beam measured by the electrostatic electrometer 114 from the charge amount measured by the energy filter, the local charge amount based on the true electron beam irradiation not depending on the global charge is measured. It becomes possible to do.

In the apparatus of the present embodiment, the deflection range of the scanning deflector can be adjusted based on the magnification fluctuation amount ΔM _s obtained by the above arithmetic expression, and the measured length measurement value can be corrected. It is said.

When adjusting the deflection range of the scanning deflector, in the case of an electromagnetic scanning deflector, it is preferable to add or subtract the current required to correct the magnification fluctuation amount ΔM _s to the original deflection current. As for feedback to the measurement value, an accurate measurement value can be obtained by multiplying the measurement value obtained by the measurement method used in known scanning electron microscopes, etc., by the multiplication factor. Can be calculated. The gist of the present embodiment is to separate and measure global charging and local charging, and the method for adjusting the scanning deflector and the method for correcting the measured value are not particularly limited to those described above.

FIG. 6 shows a second specific example of this embodiment. In this specific example, means for measuring the sample height was added in place of the electrostatic electrometer of the previous specific example. For example, the height detection laser emitter 115 irradiates the wafer 108 with laser light 116 from the time when the sample stage 109 approaches a predetermined observation point, and the position sensor 117 receives the reflected light. A so-called Z sensor that detects the height of the wafer from the position is provided. Since the global charging potential V _g is uniquely determined from the excitation current of the objective lens that is just focused and the sample height data, if the relationship between these three physical quantities is calculated in advance by experiment or electron optical simulation, The global charging potential V _g can be estimated without measuring the potential.

In this embodiment, the charging correction control unit 120 executes the sample height measurement by the Z sensor until the above-described global charging potential V _g is estimated, and the S curve measurement sequence until the local charging voltage ΔV _s is obtained. Further, as in the previous specific example, the magnification correction amount is calculated based on the obtained V _g , ΔV _s , the excitation current of the objective lens 106, etc., and the deflection intensity of the scanning deflector 107 or the obtained length measurement is obtained. Correct the value.

Next, further specific examples of the embodiment will be described. This is an SEM provided with both the electrostatic electrometer described in the above two specific examples and means for measuring the sample height. By providing both means, the global charging potential V _g and the local charging voltage ΔV _s can be stably measured with higher accuracy.

That is, the first approximate value V _g (1) is obtained from the measurement data of the electrostatic electrometer probe 114 or the fitted function, and excitation of the objective lens that is just focused by taking the sample height data from the Z sensor into consideration. If the current is estimated, the just focus search (so-called autofocus) can be completed in a short time. An accurate global charging potential V _g can be calculated from the difference between the excitation current of the objective lens calculated from V _g (1) and the excitation current obtained by autofocus. If V _g is accurate, ΔV _s = V _s −V _g, which is a difference from the local surface potential V _s , can be accurately calculated, and the magnification correction accuracy is improved.

FIG. 7 illustrates the energy filter provided in the above-described embodiment in more detail. A mesh electrode 112 sandwiched between a grounded mesh electrode 113 is installed under the secondary electron conversion electrode 110, and the mesh electrode is set with the global charging potential V _g or the first approximate value V _g (1) as an initial value. The applied voltage of 112 is run, and the S curve (secondary electron amount distribution when the applied voltage to the energy filter is changed) is measured. The grounded mesh electrode 113 prevents the electric field of the mesh electrode 112 from spreading unnecessarily to the conversion electrode 110 or the like. In addition, a certain amount of secondary electrons 102 collide with the lower mesh electrode 113 regardless of the voltage of the mesh electrode 112, and a certain amount of new secondary electrons 130 are generated. The secondary electrons 130 are attracted to the scintillator 131 to which a positive high voltage of about 10 kV is applied. By normalizing the current I ₁₁ from the scintillator 111 with the current I ₃₁ from the scintillator 131, the S curve can be measured with high accuracy. As with the scintillator 111, an image can be displayed on the cathode ray tube.

  FIG. 8 particularly illustrates the charge correction control unit 120 of the third specific example described above in detail. The charging correction control unit 120 includes an electrostatic voltage measurement data table 201, an autofocus control unit 202, a global charging potential calculation unit 203, an energy filter voltage control unit 204 that automatically measures an S curve, a local charging voltage calculation unit 205, and a magnification correction. The calculation unit 206 is included.

First, the electrostatic voltage measurement data table 201 stores the data of the voltage V ₁₄ with respect to the coordinates of the sample measured by the electrostatic voltmeter, or the fitted function form. Each time the observation point moves, the corresponding global charging voltage ΔV _g is calculated, and the desired acceleration voltage V _acc = V _o + ΔV _g +
The retarding voltage V ₉ (= V _r ) to the sample stage 109 is adjusted so as to satisfy V _r . The autofocus control unit 202 calculates the excitation current I ₆ (1) with respect to the sample height data Z _i from the Z sensor and the set acceleration voltage V _acc, and runs in the vicinity of this current to perform just focus excitation. Search for current I ₆ . Subsequently, the global charging potential calculation unit 203 considers that there is an error in the acceleration voltage V _acc when there is a difference between I ₆ (1) and I ₆ , corrects ΔV _g , and obtains an accurate global charging potential V _g . .

On the other hand, the energy filter voltage control unit 204 measures the S curve without charging and stores it in the local charging voltage calculation unit 205. As described above, the measurement sequence of the S curve is based on the global charging potential V _g or its estimated value V _g (1), and the applied voltage V ₁₂ of the mesh electrode 112 is run to change the secondary electron current I ₁₁ . Measure. It can also be normalized by the current I ₃₁ from the scintillator 31. The data to be stored may be processed data such as the S curve itself, a filter voltage exceeding a threshold, or a filter voltage at which the slope of the S curve is maximized. Since the S curve is slightly different depending on the material of the sample, the subsequent calculation accuracy can be improved by storing data for each material. The local charging voltage calculation unit 205 selects an S curve as a reference, and calculates the local charging voltage ΔV _s from the voltage shift amount. Finally, the magnification correction calculation unit 206 calculates the magnification correction amounts ΔM _g and ΔM _s from the global charging potential V _g and the local charging voltage ΔV _s by using the respective equations (1) and (2). By correcting the current I ₇ of the scanning deflector with the reciprocal of the magnification M + ΔM _g + ΔM _s , an image can always be observed at a desired magnification regardless of the charging voltage.

When a large number of wafers are automatically processed in the semiconductor element production line, it is effective to increase the processing speed by reducing the number of times of S curve measurement by the energy filter. Since the local charging voltage ΔV _s is the same in the same circuit pattern and the same material, the measured ΔV _s can be used. In some cases, it is possible to perform one S-curve measurement for each wafer. When the S curve is newly measured, it is automatically added to the database of the local charging voltage calculation unit 205.

  According to this embodiment, in the image observation and dimension measurement of the insulator sample, the fluctuation amount of the observation magnification can be calculated with high accuracy, and the measurement value that is fixed to the desired observation magnification or changes due to the magnification fluctuation can be corrected. Is also possible. As a result, high-accuracy dimensional management becomes possible in the process of manufacturing an ultrafine semiconductor device.

As an additional effect, stable image quality can be obtained because the energy of the primary electron beam incident on the sample can be controlled with high accuracy. Furthermore, by monitoring the local charging voltage ΔV _s , it can be used as an indicator of whether a charging voltage capable of preventing dielectric breakdown due to excessive charging and observing the bottom surface of a contact hole having a large aspect is obtained.

(Example 3)
As described with reference to FIG. 11, the local charging voltage ΔV _s changes the optical magnification M _obj of the objective lens. Since the charging voltage ΔV _s by the electron beam irradiation is local, there is almost no influence on the excitation current I ₆ . Nevertheless, since the local charging voltage ΔV _s forms a minute electrostatic lens 108b, the trajectory 101a of the primary electrons deflected by global charging (global charging) is focused like the trajectory 101b, and the optical magnification M _obj is set. The large variation is as described in the previous embodiment.

  As described above, a further method for performing accurate inspection and measurement which is difficult due to the combination of different charging phenomena will be described below.

In the present embodiment, a method is proposed in which the correct optical magnification M _obj is calculated by correcting the magnification fluctuation due to the local charging voltage ΔV _s .

Magnification variations caused by local charging, depending on the localized electrostatic voltage [Delta] V _s, localized electrostatic voltage [Delta] V _s is the irradiation magnification (in this embodiment of the electron beam inspection, prior to electron beam irradiation for the measurement, the In order to mainly explain that the electron beam irradiation as described in the embodiment is performed, it depends on M _pre , the electric field near the sample surface, and the kind of the sample.

FIG. 14 shows the local charging voltage ΔV _s when the _pre- dose magnification M _pre is varied at boosting voltages of 0.5 kV and 5 kV. The boosting here is a method of arranging a cylindrical electrode to which a positive voltage is applied at least in the objective lens so that the electron beam can pass through the objective lens with high acceleration. FIG. 14 shows the results of measuring the surface potential when the pre-dose magnification is changed when the voltage applied to the cylindrical electrode is 0.5 kV and 5 kV. The boosting technique is described in detail, for example, in JP-A-9-171791 (USP 5,872,358).

When the parameters for changing the local charging voltage ΔV _s are the sample surface electric field and the pre-dose magnification, the local charging voltage ΔV is calculated from the boosting voltage V _b , the retarding voltage V _r , the fitting coefficients A ₁ and a ₁ , and the _pre- dose magnification M _{pre. s} can be calculated by the following fitting function.

ΔV _s = A ₁ (V _b −V _r ) / M _pre + a ₁ (13)
Further, by using the magnification sensitivity coefficient T _s , the magnification fluctuation amount ΔM / M _obj can be calculated from ΔV _s by the following equation.

ΔM / M _obj = T _s * ΔV _s (14)
FIG. 15 shows the magnification variation sensitivity coefficient T _s when the irradiation area of the beam (∝1 / predose magnification = 1 / M _pre ) is varied at the retarding potential V _r = −1.2 kV. T _s can be divided into four sections according to the beam irradiation area. It can be regarded as global charging at magnifications lower than 50 times.

The section from 50 times to 500 times is in the transition region from global charging to local charging.
It can be regarded as almost constant in the section from 500 times to 5000 times. When the magnification is higher than 5000 times, T _s tends to gradually decrease. Therefore, if the irradiation area is set to an irradiation area (one side is 10 μm to 300 μm) corresponding to a magnification of 500 to 5000 times that the magnification variation sensitivity coefficient T _s can be regarded as a constant value, the estimation accuracy of the correction value is maintained, The number of data stored in advance can be reduced.

L true value of the dimension of the pattern and the actual measured value respectively, when the L _ex, magnification fluctuation amount B = .DELTA.M /
M _obj can be calculated by the following equation.

L / L _ex = 1 + B (15)
When the true length measurement value is estimated using the equations (13), (14), and (15), the unknown coefficients are A ₁ and a ₁ . Therefore, the true length measurement value L can be estimated by using the result of measuring twice or more by changing the _pre- dose magnification M _pre or (V _b −V _r ) proportional to the electric field on the sample surface.

In the case of observing an unknown insulator sample, the above-described method can be used for true length measurement by changing the _pre- dose magnification M _pre or the electric field on the sample surface and observing at two or more different local charging voltages ΔV _s. It also has the advantage that the value can be estimated. Further, when using the above method, the local charging voltage ΔV _s is not a fitting function having a pre-dose magnification or a surface electric field as a charging variable parameter as shown in the equation (13), but the incident beam energy, irradiation time, and in-sample. The same effect can be obtained even if the mobility of electrons and holes is expressed by another fitting function having a variable charging parameter.

Further, by storing the fitting coefficients a ₁ and A ₁ used in Equation (13) in the storage unit, the true dimension value can be estimated using one pre-dose magnification and the length measurement value in the surface electric field. I can do it. On the other hand, the fitting coefficient a ₁ used in Equation (13) is a term that does not depend on the pre-dose magnification or the surface electric field. For this reason, the fitting parameter used for correction is substituted by substituting the variation of a ₁ used when correcting the length measurement value of the same type of sample into the equations (13), (14), and (15). Reliability can be evaluated as a deviation of the measured value.

  FIG. 16 is a graph showing the relationship between the length measurement value before correction and the length measurement value after correction with respect to the pre-dose magnification. By storing the magnification variation amount B for each pre-dose magnification calculated from the relationship between the true dimension value and the length measurement value before correction according to the equation (15), the true dimension is obtained from the length measurement value under one observation condition. You can guess the value.

  Further, when pre-dose is performed, a high-contrast image can be obtained by using the optimum acceleration voltage disclosed in Japanese Patent Application Laid-Open No. 2000-200579, and a more accurate observation result can be obtained.

  In the first specific example of the present embodiment, the true dimension value is estimated by using a plurality of measured length values measured by changing the charging variable parameter for changing the local charging voltage as the true dimension value estimating means. An example in which the charging correction control unit 120 shown in FIGS.

  FIG. 17 is a schematic diagram of the charging correction control unit 120. The charging correction control unit 120 includes a global charging correction unit 302 and a local charging correction unit 303. The local charge correction unit 303 sets measurement conditions (charging variable parameters, acceleration voltage, primary electron irradiation time during pre-dose) via 313a.

A length measurement value measured under the set conditions is input from an input device (not shown) to the local charge correction unit 303 via 313b. Based on the input length measurement value and the set measurement condition, a magnification fluctuation amount B for performing local charge correction is input to the charge correction control unit 304 through 313d. Further, the magnification fluctuation amount calculated by the global charge correction unit 302 is also input to the charge correction control unit 304 through 313e.

The length measurement value changed by the influence of global charging and local charging is corrected from the magnification fluctuation amount derived by the respective correction units of the local charging correction unit 303 and the global charging correction unit 302 inputted by the charging correction control unit 304. The measured dimension value is output.

FIG. 18 shows a flowchart for correcting the length measurement value. First, in step s101, variable charging parameters and measurement conditions are set. In step s102, the sample is irradiated with an electron beam according to the conditions determined in step s101 to charge the sample. In step s103, a length measurement value L _ex measured under the charge variable parameter determined in step s101 or s109 is obtained. In step s104, the measurement value L _ex acquired in step s103, it is determined whether or length measurement performed with sufficient accuracy. If it is determined that the length measurement cannot be performed with sufficient accuracy, the observation conditions are reset in step s101.

  In step s106, it is determined whether a plurality of times of data necessary for the correction in step s107 are available. If the data is insufficient, a different charging variable parameter is set in step s109 and the length is measured again. In step s107, the length measurement value is corrected using the length measurement value measured in step s105 and the charging variable parameter determined in steps s102 and s109. The length measurement value corrected in step 108 is output to the monitor.

  If local charge correction is performed using this example, the true dimensional value can be estimated with high accuracy by measuring the shape and material of a sample that has not been subjected to preliminary measurement twice or more at different local charge voltages. I can do it. In addition, since preliminary measurement using an energy filter for each magnification is not necessary, the measurement speed can be improved.

  FIG. 19 shows a schematic diagram of sample charging when the pre-dose magnification is changed and the sample is charged. When two or more different pre-dose magnifications are used and the length is measured at each magnification, a stable local charge can be created quickly and stably by using the following procedure.

The sample 108 is charged in two types, a global charging potential V _g spreading over the entire surface and a local charging voltage ΔV _s caused by electron irradiation. In FIG. 19A, when observation is performed with the pre-dose magnification increased after observation with a small pre-dose magnification, a residual charged region 108d is formed. If the relaxation time is not taken until the charge in the residual charged region is sufficiently relaxed, the local charge correction is adversely affected. However, as shown in FIG. 19 (2), if observation is performed with a large pre-dose magnification and then the pre-dose magnification is lowered, there is no residual charged region, so there is no need for a relaxation time, and measurement is performed immediately after completion of the pre-dose. Put in. By using the above procedure, observation can be performed quickly under a highly accurate charged region.

Next, a second specific example of the embodiment of the present invention will be described with reference to FIGS. In this embodiment, the storage unit 301 incorporated in the charge correction unit 120 is provided with a database of fitting constants of functions representing the magnification fluctuation amount B or the local charging voltage ΔV _s . The measurement conditions (charging variable parameters, acceleration voltage, primary electron irradiation time during pre-dose) are set via the local charge correction unit 303 to 313a.

The length measurement value measured under the set conditions is input to the local charge correction calculation unit 303 via 313b. The charging variable parameter or the magnification fluctuation amount B is input to the storage unit 301 via 313g. A fitting coefficient or a magnification fluctuation amount B for associating ΔV _s adapted to the charging variable parameter input in the storage unit 301 with the charging variable parameter is input to the local charging correction unit 303 via 313h. The corrected length measurement value calculated by performing correction using the input data is output via 313d.

FIG. 21 shows a flowchart of the measurement procedure when correction data is stored. In step 201, charging variable parameters and measurement conditions are set. In step 202, the sample is charged by irradiating the sample with an electron beam according to the conditions set in step s201. In step s203, a length measurement value L _ex measured under the variable charging parameter determined in step s201 or s209 is obtained. In step s204, it is determined whether the length measurement can be performed with sufficient accuracy. If it is determined that length measurement with sufficient accuracy cannot be performed, the observation conditions are reset in step s209.

In step 206, it reads the fitting coefficients of the fitting function representing the same charge variable magnification change rate B or V _s derived in the past under the parameter as the correction data from the storage unit 301. In step s 207, the obtained measurement value L _ex, and from charge variable parameters established in step s201 or step s208, it is determined whether correction is possible or not. If it is determined that correction is impossible, the observation conditions are reset in step s209.

Inputting a measurement value L _ex measured in step s205 in localized electrostatic correction unit 303 at step 208. At the same time, correction is performed by the local charge correction unit 303 using this data to obtain a corrected length measurement value.

  By performing local charge correction using this embodiment, it is possible to perform local charge correction for at least the same pattern and the same condition portion of a sample once measured from a measured value measured under one charge variable parameter. It is possible to measure the shortening of the measurement time. If an optimal pre-dose condition as disclosed in Japanese Patent Laid-Open No. 2000-200579 is set in step s201, more accurate and stable observation can be performed.

In the third specific example of this embodiment, in order to improve the reliability of the corrected length measurement value, the storage unit 301 is fitted with measurement conditions and magnification fluctuation amounts B and V _s of the same kind of sample measured in the past. A configuration having function fitting coefficients as a database will be described.

  In the present embodiment, the storage unit 301 includes the above-described database, so that the accuracy of the length measurement value after the local charging correction is quantitatively evaluated as a deviation of the length measurement value after the correction. A threshold is set from the deviation of this measured value. When a length measurement value exceeding the set threshold is measured, it is determined that an abnormal measurement has been performed, and it is determined whether or not the cause of the abnormal measurement is due to the influence of abnormal charging due to the influence of impurities on the sample surface.

The procedure for using this embodiment is as follows. First, the database construction procedure is shown. A fitting coefficient of a plurality of local charging voltages is derived from a measured value between the same points measured by changing the local charging voltage ΔV _s of the sample. Similarly, the fitting coefficient of the fitting function of the local charging voltage ΔV _s is derived from the measured value between two different points of the same type of sample.

By repeating this, a variation in the fitting coefficient a ₁ of a term that does not depend on the charging variable parameter is extracted from the plurality of fitting coefficients obtained. Stores that fix the variation of the variation and the fitting coefficient a ₁ of the fitting coefficient a ₁ in the deviation of the measurement value in the storage unit 301.

  Next, a procedure for determining the presence or absence of abnormal charging using the constructed database will be described below. Each time the sample is replaced or measured, the measured value derived from the measured value obtained by changing the variable charging parameter exceeds the threshold value obtained from the deviation of the measured value stored in the storage unit 301. At that time, display examples of the screens shown in FIGS. 22A and 22B are displayed to inform the user of the presence or absence of charging abnormality.

When the fitting coefficient a ₁ used for the current correction is within the threshold value obtained from the variation of the fitting coefficient a ₁ measured in the past stored in the storage unit 301, the local charging is normally performed. It is estimated that there is no abnormal charging.

If the value of the fitting coefficient a ₁ used for the current correction exceeds the threshold value obtained from the variation of the fitting coefficient a ₁ stored in the storage unit 301, it indicates that abnormal charging has occurred. By using this embodiment, since it is possible to know the presence or absence of abnormal charging of local charging, a length measurement value after correction with high reliability can be obtained.

  The fourth specific example of this embodiment has a function that combines all of the above embodiments. A flowchart of this embodiment is shown in FIG. In step s1, the storage unit 301 determines whether or not there is correction data for the current observation sample. If there is no correction data necessary for the current length measurement in the storage unit 301, the process enters the flow of loop 1 and the process shown in the first specific example is performed in step s100 to derive the corrected length measurement value L.

  In step s120 and step s160, the correction result is displayed on the screen, and it is determined whether or not the correction data used this time will be used from the next time. When using the correction data from the next time, the correction data is stored in the storage unit 301 in step s170. If there is correction data in step s1, loop 2 is entered and the process shown in the second specific example is performed to derive a corrected length measurement value L. In step s120, the correction result is displayed on the screen shown in FIGS.

In step 210, when the evaluation shown in the third specific example is performed and a charging abnormality is detected, a warning is displayed and loop 3 is entered, and the procedure shown in the first specific example is repeated a plurality of times in step s100. The fitting coefficient a ₁ under a plurality of conditions is output. In step s300, a fitting function is created using the average value of variations of the plurality of fitting coefficients obtained in step s100.

The reliability of the fitting function created this time is evaluated from the deviation of the measured value due to the variation of the plurality of fitting coefficients a ₁ . If the fitting function cannot be determined to be reliable, the process returns to step s100 again. If it is determined in step s300 that a sufficiently reliable fitting function can be obtained for the abnormally charged portion, the corrected length measurement value and the deviation of the length measurement value using the fitting function created this time in step s120 are calculated. Display the results on the screen.

  By performing local charge correction using the present embodiment, when measuring a similar pattern carved in the same insulator sample, it is possible to measure the length at a higher speed and with a stable accuracy.

  In the fifth specific example of this embodiment, an example of a scanning electron microscope provided with an ultraviolet irradiation device 314 will be described in order to reduce the influence of the previously measured charging. FIG. 24 shows a schematic diagram of a scanning electron microscope provided with an ultraviolet irradiation device.

  Reference numeral 113 denotes an input device for inputting measurement conditions and the like. By irradiating the sample with ultraviolet rays for each observation from the ultraviolet irradiation device 314, the charge of the sample accumulated since the previous measurement can be initialized, and more stable measurement can be performed.

  According to the present embodiment, the variation amount of the observation magnification can be calculated with high accuracy in the image observation and dimension measurement of the insulator sample. As a result, in the process of manufacturing a miniaturized semiconductor device, high-accuracy dimension management can be performed in a short time.

It is a figure showing the whole structure of a scanning electron microscope. It is a figure which shows the positional relationship of the wafer in a conveyance path | route, a conveyance arm, and an electrostatic electrometer. It is a figure which shows the flow about the determination method of the charge distribution function on a wafer surface. It is a figure which shows the method of measuring the charging voltage of a wafer with a some probe. It is a whole block diagram of SEM provided with the electrostatic electrometer and the energy filter. It is a whole block diagram of SEM provided with the sample height measurement means and the energy filter. It is a detailed block diagram of an energy filter. It is a detailed block diagram of a charging correction control unit. It is a figure explaining the optical magnification of an objective lens. It is a figure explaining the charging mechanism of a sample. It is a figure explaining the mechanism by which a charging voltage is fluctuated. It is a graph which relates a global charging voltage and a magnification fluctuation sensitivity coefficient. It is a graph which relates a local charging voltage and a magnification fluctuation sensitivity coefficient. It is a graph which shows the irradiation area | region of an irradiation electron beam, and the relationship of local charging. It is a graph which shows the relationship between the magnification of an irradiation electron beam, and a magnification fluctuation sensitivity coefficient. It is a graph which shows the relationship between pre-dose magnification, measured length measurement value, and estimated length measurement value. It is a block diagram of a local charging correction unit. It is the flowchart which showed the procedure which performs the local electrification correction based on 2 or more electrification measurements. It is a figure explaining the charging condition at the time of performing electron beam irradiation by pre-dose magnification. It is a block diagram of another local electrification amendment part. It is a flowchart which shows the procedure of local charge correction | amendment when there exists correction data in a memory | storage part. It is a figure which shows the error confirmation screen of the length measurement value estimated by local charging correction. It is a flowchart which shows the procedure which combined the some specific example. It is a figure explaining the whole block diagram of SEM provided with the ultraviolet irradiation device.

Claims (18)

In the electron beam adjustment method for controlling the focus of the electron beam applied to the sample by adjusting the negative voltage applied to the semiconductor wafer,
Before introducing into the sample chamber for irradiating the electron beam to the semiconductor wafer, the potential distribution information of the potential distributed concentrically over the entire surface of the semiconductor wafer ,
A method of adjusting an electron beam, comprising: adjusting a negative voltage applied to a semiconductor wafer using the potential distribution information and coordinate information of the semiconductor wafer irradiated with the electron beam. In claim 1,
An electron beam adjustment method comprising: measuring potentials at a plurality of points on the semiconductor wafer; and creating a distribution function approximating the potential distribution based on the measurement data. In claim 2,
An electron beam adjustment method, comprising: measuring a potential on the semiconductor wafer; and re-creating the distribution function based on measurement data of the potential. In claim 2,
Calculating a difference between the measurement data and an approximate value of the potential obtained by the distribution function, and excluding the measurement data from creation of the distribution function when the difference is larger than a set threshold value. The electron beam adjustment method. In claim 1,
The electron beam adjustment method, wherein the distribution function approximating the potential distribution is an even function. In a charged particle optical system control device for transmitting a signal for controlling application of the negative voltage to a scanning electron microscope that controls the focus of an electron beam by applying a negative voltage to a semiconductor wafer,
The charged particle optical system control device is configured to obtain a potential distribution of potentials concentrically distributed on the entire surface of the semiconductor wafer obtained in advance before introducing the semiconductor wafer into the sample chamber for irradiating the electron beam. information and, wherein the electron beam is irradiated based on the coordinate information on the semiconductor wafer, the charged particle optical system control apparatus characterized by calculating a negative voltage applied to the semiconductor wafer. In claim 6,
The charged particle optical system control device creates a distribution function that approximates the potential distribution based on measured measurement data of the potential of the semiconductor wafer. In claim 7,
The charged particle optical system controller re-creates the distribution function based on the measured potential measurement data on the semiconductor wafer. In claim 7,
The charged particle optical system control device calculates a difference between the measurement data and an approximate value of a potential obtained by the distribution function, and when the difference is larger than a set threshold value, the measurement data is distributed to the distribution A charged particle beam optical system control apparatus characterized in that it is excluded from the creation of a function. In claim 6,
A charged particle beam optical system control apparatus, wherein a distribution function approximating the potential distribution is an even function. In a scanning electron microscope comprising an electron source and a control device that adjusts the focus of an electron beam emitted from the electron source by changing a negative voltage applied to a semiconductor wafer,
The control apparatus includes a potential distribution information on the potential concentrically distributed over the entire surface of the semiconductor wafer obtained in advance before introducing the semiconductor wafer into the sample chamber for irradiating the electron beam , A scanning electron microscope characterized by calculating a negative voltage applied to the semiconductor wafer based on coordinate information on the semiconductor wafer irradiated with an electron beam. In claim 11,
The control device creates a distribution function approximating the potential distribution based on measured measurement data of the potential of the semiconductor wafer. In claim 12,
The control device re-creates the distribution function based on the measured potential measurement data on the semiconductor wafer. In claim 12,
The charged particle optical system control device calculates a difference between the measurement data and an approximate value of a potential obtained by the distribution function, and when the difference is larger than a set threshold value, the measurement data is distributed to the distribution A scanning electron microscope characterized in that it is excluded from the creation of a function. In claim 11,
A scanning electron microscope characterized in that a distribution function approximating the potential distribution is an even function. In the electron beam adjustment method for controlling the focus of the electron beam applied to the sample by adjusting the negative voltage applied to the semiconductor wafer,
A potential distribution generated on the semiconductor wafer, with the center of the semiconductor wafer being the highest point, going down to the edge of the wafer and lowering across the entire surface of the semiconductor wafer, and for irradiating the electron beam Potential distribution information acquired in advance before introducing the semiconductor wafer into the sample chamber ;
A method for adjusting an electron beam, comprising: adjusting a negative voltage applied to a semiconductor wafer using coordinate information on the semiconductor wafer irradiated with the electron beam. In a charged particle optical system control device for transmitting a signal for controlling application of the negative voltage to a scanning electron microscope that controls the focus of an electron beam by applying a negative voltage to a semiconductor wafer,
The charged particle optical system control device has a potential distribution that occurs on the semiconductor wafer, has a center point of the semiconductor wafer as the highest point, and decreases toward the edge of the wafer and extends across the entire surface of the semiconductor wafer. Using the potential distribution information acquired in advance before introducing the semiconductor wafer into the sample chamber for irradiating the electron beam and the coordinate information on the semiconductor wafer irradiated with the electron beam, a semiconductor wafer A charged particle optical system control device, characterized in that a negative voltage applied to the device is adjusted. In a scanning electron microscope comprising an electron source and a control device that adjusts the focus of an electron beam emitted from the electron source by changing a negative voltage applied to a semiconductor wafer,
The control device, the results on a semiconductor wafer, the center of the semiconductor wafer as the highest point, toward the edge of the wafer, a potential distribution span the entire surface of the semiconductor wafer together with lower, the electron beam Negative potential applied to the semiconductor wafer using potential distribution information acquired in advance before introducing the semiconductor wafer into the sample chamber for irradiating the semiconductor wafer and coordinate information on the semiconductor wafer irradiated with the electron beam. A scanning electron microscope characterized by adjusting a voltage.

Images