JP2020056677A - Electron beam measuring device and electron beam measuring method - Google Patents

Abstract

PURPOSE: To cut off an electron beam in real time a prescribed interval after an edge peak is detected in real time and suppress irradiation of a pattern to suppress the damage to or the shrinkage of the pattern pertaining to an electron beam measuring device and an electron beam measuring method.CONSTITUTION: The electron beam measuring device comprises setting means for setting a partial scan region, detection means for detecting a pattern edge in real time, electron beam OFF signal generation means for detecting an edge in real time and generating an electron beam OFF signal in real time after the elapse of a prescribed time, and electron beam cutoff means for cutting off an electron beam in real time when an electron beam OFF signal is generated and cutting off irradiation of the pattern, with the electron beam measuring device constituted so as to automatically acquire a minimum required image for measurement.SELECTED DRAWING: Figure 1

Description

  The present invention relates to an electron beam measuring apparatus and an electron beam measuring method for measuring the size and area of a pattern of a distance measurement target formed on a sample or extracting the contour of the pattern.

  Semiconductor devices have been shrinking in minimum feature size over the last half century, and as new processes are developed, devices with higher transistor densities are being created. This makes it possible to dramatically reduce the unit transistor cost. At present, devices packed with billions of transistors in a 1 cm square chip can be purchased for tens of thousands of yen, and a computer with performance far exceeding the performance of a conventional large computer can be realized in a smartphone. Have been.

  The semiconductor industry has been prospering for more than half a century due to the principle of economic expansion by reducing the feature size, generally known by Moore's Law.

  At present, devices having a minimum dimension of the order of 10 nm are being developed and used, and it is expected that the trend toward smaller transistors of 7 nm, 5 nm and 3 nm will continue in the future.

  In order to manufacture a small transistor on a silicon wafer on which a semiconductor device is formed, an exposure technique called lithography for printing a wiring pattern is important. Since the limit of fine processing depends on the wavelength of light used for exposure, the wavelength has been shortened.

  At present, double or triple exposure using an ArF wavelength of 193 nm or an exposure technique using light having an EUV wavelength of 13.5 nm is mainly used. Nanoimprinting and DSA (Directed Self-Assembly) are also used as a nano-order miniaturization process using a resist having a similar chemical composition without using an exposure technique.

  Since the generation efficiency of a short wavelength laser light source such as ArF or EUV used in an exposure apparatus is extremely low, the amount of laser light finally obtained is also small. For example, in an EUV light source, only 1% of the input energy is converted into EUV light, and about 1% of the energy reaches the wafer surface and is used for actual exposure. More specifically, several hundred kilowatts of pumping laser is used to generate several hundred watts of EUV light, and finally several watts of light reach the wafer. Therefore, when a normal low-sensitivity resist is used, the exposure time is too long, and the throughput is significantly reduced. Therefore, a high-sensitivity resist called a chemically amplified resist capable of exposing even weak light with the same exposure time as the conventional one and improving the throughput has been developed and used.

  A typical chemically amplified resist is based on an esterified methacrylic resin and contains a photoacid generator (PAG) for generating an acid serving as a reaction catalyst in order to increase sensitivity. Accordingly, when light is applied to the resist, an acid is generated to accelerate the decomposition reaction of the resist, thereby increasing the sensitivity. In addition to onium salts, nitrobenzyl esters, diazomethane, triazine and the like are known as photoacid generators.

  Subsequent heating of the post-exposure bake (PEB) causes the acid-labile protecting group that protects the alkali-soluble group of the alkali-soluble resin in the positive type to undergo a deprotection reaction using an acid as a catalyst, and the resin to be deprotected. Is changed from alkali-insoluble to alkali-soluble by developing the polarity, and a positive pattern is obtained after developing with an alkaline solution. Examples of the alkali-soluble group include a phenolic hydroxyl group and a carboxyl group, and examples of the acid-labile protecting group include tert-butyl ter-butoxycarbonyl (t-Boc) and a tetrahydropyranyl group. In the negative type, by heating the post-exposure bake, the alkali-soluble resin and the cross-linking agent cause a cross-linking reaction using an acid as a catalyst, change from alkali-soluble to alkali-insoluble, and obtain a negative-type pattern after developing with an alkali solution. As the crosslinking agent, hexamethoxymethylmelamine, tetramethoxyglycouryl (TMCU) and the like are used. In either case, the reaction proceeds with the use of an acid as a catalyst, so that the resist is extremely sensitive compared to conventional resists.

  On the other hand, such a chemically amplified resist has a sensitivity amplification mechanism based on the generation of acid as described above, so that even when electron beam irradiation unrelated to light exposure generates acid in the resist, the resist is decomposed and the resist shrinks. Resulting in. For example, when a resist structure containing the above-mentioned material is irradiated with an electron beam converged to several nm in order to perform length measurement by CDSEM, the resist at the irradiated position shrinks and is different from the original length or structure. Is known to be. This problem has been widely recognized for many years as a hindrance to precise length measurement of resist structures by CDSEM, but there has been a major problem that a method for measuring CD without shrinking has not been found yet.

  At present, in the CDSEM for wafers of major manufacturers, etc., it is assumed that resist shrink of several nm to several tens of nm naturally occurs when measuring length by electron beam. A method for estimating or predicting a dimension when the electron beam is not irradiated is widely used.

  However, as described above, as the semiconductor device is miniaturized, the minimum width of the resist structure is on the order of 10 nm, and when shrinkage of several nm occurs, a random dimensional change of 10% or more of the main body size occurs, and In addition to the lack of measurement accuracy, it has become a major problem that adversely affects device characteristics. Further, there has been a serious problem that a defect may be generated by SEM observation.

  Therefore, the method of estimating the dimensions in the state where the electron beam is not irradiated after the resist is actually shrunk as in the prior art has not only made sense of the measurement but also originally performed to improve the yield. CDSEM measurements are even causing device manufacturing yields to drop. This has become a major problem in that even if a TEG (test element group) can be measured, it is impossible to measure a device or an element that actually operates.

  Therefore, in view of the problems and defects of the conventional length measurement method using an electron beam, the present invention introduces a method completely opposite to the conventional CDSEM measurement concept to essentially reduce the resist shrink itself, or The present invention proposes an SEM observation method in which a resist size is measured accurately and highly reproducibly for a device actually operating or a length measurement method which does not reduce the yield.

  The present invention is an electron beam measuring apparatus for extracting a dimension, an area measurement or a contour of a pattern of a distance measurement target formed on a sample, and setting means for setting a partial scanning area including an edge of the pattern of the distance measurement target, A scanning means for scanning a partial scanning area with an electron beam to detect generated electrons to acquire a partial scanning image, and at the time of acquiring the image, detecting means for detecting an edge of the pattern in real time; And an electron beam OFF signal generating means for generating an electron beam OFF signal in real time after a predetermined interval has passed, and an electron beam OFF signal in real time when the electron beam OFF signal generating means generates an electron beam OFF signal in real time. Electron beam cut-off means for cutting off the pattern and irradiating the pattern And acquire the small limit required image.

  At this time, edge extraction is performed on each of the partial scan images acquired facing the pattern to be measured, and the distance between edges of the pattern facing each other is obtained from each edge position information.

  In addition, an edge is extracted from the partial scan image acquired from the pattern to be measured, and the contour of the pattern to be measured is extracted.

  When the electron beam is scanned in a direction parallel to an edge to be measured of the pattern, the predetermined interval is an interval corresponding to N (integer) scanning lines, or the electron beam is scanned in a direction perpendicular to the edge to be measured of the pattern. In this case, the interval is set to correspond to M pixels (integer) of the scanning line.

  Further, the pattern to be measured is a pattern that is reduced by electron beam irradiation or a resist pattern.

  A feature of the present invention is that the width and length of the resist structure are measured without irradiating an electron beam to the resist structure (pattern) portion, which is completely opposite to the conventional method. In particular, the present invention is greatly characterized in that an electron beam scanning area can be automatically set in a very narrow range used for measurement using signal electrons obtained from a structure.

  In a conventional length measurement or inspection method using an electron beam, an image is acquired by scanning an electron beam, which is converged in the order of nanometers over the entire region from several microns to several mm called a FOV including a resist structure, and obtains an image. A portion where the luminance appearing in the obtained image greatly changes is defined as an edge of the structure, and a distance between the edges is obtained to perform length measurement and shape recognition.

  The present invention utilizes a technique for shortening a measurement time by irradiating an electron beam only to a necessary portion of a measurement target. By utilizing this, it is possible to irradiate only the edge portion of the resist with an electron beam, and it is possible to reduce shrinkage. For example, a line pattern having a width of about 100 nm can be measured by irradiating an electron beam only to a region having a width of about 10 nm constituting an edge. In this way, the shrinkage of the resist can be considerably reduced as compared with the case where the entire line is irradiated with the electron beam.

  However, since there is an error in the line formation position, the electron beam is accurately irradiated at a high speed only on the edge portion of the line pattern having the line width of the order of 10 nm, relying only on the position determined in advance by the design (CAD data). It is quite difficult to measure.

  Therefore, the present invention utilizes a caliper-like measuring method for an electron beam. That is, electron beam scanning is mainly performed in a region outside a boundary formed by a resist structure to be measured, and is characterized in that the length is measured so as to sandwich the structure. In particular, the present invention has a great feature in that a range for irradiating an electron beam using signal electrons obtained from a structure is automatically set (self-aligned ESR) in accordance with an edge position of a line actually formed using CAD data as a guide. is there.

  In other words, a signal electron generated from a structure irradiated with an electron beam is used in real time during measurement to immediately modulate the electron beam scanning method currently in progress or to change the measurement method. There is a great feature in having a scanning method (real-time measurement). Thus, the scanning method can be automatically changed by a change in a signal (electron, electromagnetic wave, heat, etc.) generated as a result of irradiating the measurement object.

  One more specific example is to use signal electrons from the object structure to recognize the boundaries of the pattern on the object and blank or stop the irradiation of the electron beam irradiating the object. There is a great feature in that the measurement object is reliably protected.

  The present invention sets each partial scan area including the opposing edge of the pattern of the distance measurement target, scans the partial scan area with an electron beam, detects electrons generated, and acquires the partial scan image. Then, after detecting a peak of an edge present at a boundary which enters from outside the pattern in real time, an electron beam OFF signal is generated in real time after a predetermined interval, and irradiation of the pattern with the electron beam is performed in real time. By cutting off the pattern and automatically acquiring only the minimum image necessary for the measurement, the irradiation on the pattern is suppressed, and the damage and shrink of the pattern are suppressed.

Embodiments of the present invention will be sequentially described in detail below.
(1) The electron beam scanning is started from a position where there is no resist structure derived from CAD data, and the scanning position is gradually approached toward the boundary with the resist structure. Secondary electrons or reflected electron signal electrons generated from the object to be measured as a function of the scanning position are detected and converted into digital values, and then recorded in a semiconductor memory or a storage device.
(2) Since the intensity of secondary electrons or reflected electrons at a place where there is no resist structure is different from the intensity of secondary electrons or reflected electrons or the direction of emission at a place where a resist is present, the difference between the intensity signal and the direction of emission is used. The region and the region other than the resist can be automatically distinguished by setting an appropriate threshold value, and the electron beam irradiation can be stopped as soon as the electron beam enters the resist region or stopped after scanning N pixels. The signal waveform obtained until the scanning is stopped is stored in a storage device such as a memory as one block.
(3) Since the measurement object is not affected by the electron beam scanning in the region where there is no resist, the above process can be repeated enough to obtain the required image SNR, and the electron beam irradiation can be performed sufficiently. The key point is that an image with a high SNR can be obtained. As a result of this characteristic, it is possible to suppress the noise floor to a very low level, and it is possible to detect a small signal indicating a boundary that slightly jumps out of the noise floor, so that the resist area and the other area can be easily distinguished. Is the basic principle of the present invention. Note that the electron beam scanning interval may be controlled so as to approach the resist structure with the shortest and maximum efficiency by learning the relationship between the interval and the waveform change using a learning function such as AI.

  With the above-described mechanism, it is possible to automatically acquire a minimum image necessary for measurement in a narrow area that cannot be realized only by setting using CAD data.

  FIG. 1 shows a system configuration diagram of the present invention.

  In FIG. 1, an electron gun 1 generates and reduces an electron beam. The electron generation source is a thermionic source using a W filament, a Schottky type using ZrO, and a cold-emitter of electrolytically polished W. Alternatively, a cold field emitter, a CNT cathode, LaB6, a photo-excited emitter (photocathode), etc. made using a micro-machining technique can be used. It is a known device including an accelerating electrode for accelerating generated electrons to a desired energy.

  The aperture 2 is a circular stop that passes an electron beam at a central portion (within a predetermined solid angle) of the electron beam emitted from the electron gun 1 and blocks the other portion. It can be switched to one with a diameter. As a moving means for changing the aperture diameter, a stepping motor, a servo motor, an ultrasonic motor using a piezo element, a shape memory alloy, or the like can be used. The illustrated aperture 2 is arranged above the CL 3, but is not limited to this, and may be arranged on the main surface (center of the lens) or below the CL 3.

  The CL (condenser lens) 3 is a known one that converges the electron beam emitted from the electron gun 1 and passing through the central portion of the aperture 2. The illustrated CL3 is schematically shown, and is actually constituted by a magnetic lens or an electrostatic lens.

  The blanking electrode 13 is used together with the blanking aperture 14 to turn on / off the electron beam in accordance with a signal from the automatic blanking control device 11 to determine whether the sample 8 is irradiated with the electron beam. . In order to be able to operate at ultra-high speed, design is made so that the capacitance between electrodes is as small as possible. Further, it is preferable to arrange the blanking aperture 14 at a place where the electron beam converges so that complete blanking can be performed by applying a small voltage.

  The automatic blanking control means 11 receives a secondary electron or reflected electron signal generated in the sample 8 due to the electron beam scanning, and determines whether the sample structure has changed significantly or the constituent material has changed (for example, an edge). This is a device that detects at high speed and automatically generates a blanking signal to prevent the resist structure from being irradiated with an electron beam. As described above, the place where the electron beam irradiation stop is actually started can be set after an arbitrary pixel (or after a line) after receiving the signal. The CAD data that determines the design dimensions and arrangement of the structure can be used for the determination of the ON / OFF determination. For example, when a certain amount of electron beam scanning enters the area of the resist structure specified by the CAD data, blanking is automatically or forcibly performed to completely protect the resist structure. In the above example, blanking is performed using signal electrons generated by the sample.However, a signal from the object to be measured is received, and the scanning interval, current amount, or duty of the irradiated electron beam is changed and the blanking is substantially performed. May be controlled so that the resist does not shrink.

  The blanking device 12 applies a voltage to the blanking electrode 13 based on a signal from the automatic blanking control means 11 to block an electron beam.

  The blanking aperture 14 is an aperture (aperture) for blocking the electron beam deflected by the blanking electrode 13.

  DEF (1) 4 and DEF (2) 5 represent two-stage deflectors, each of which is an electrostatic deflector having opposed electrodes, and irradiates the sample 8 with an electron beam 111 narrowed down to the order of nm. While scanning (scanning in the X and Y directions). It is desirable that this deflector be designed with a small electrode capacity so that the electron beam can be scanned only at a necessary place and can respond at high speed. It should be noted that a two-stage magnetic deflector having a particularly small hysteresis or the like may be used.

  The MCP 6 represents an example of a secondary electron detector, which detects emitted secondary electrons when an electron beam is scanned on a plane while irradiating the sample 8, and applies a positive voltage to the front surface. It was done. As the detector, any detector that can be used at present, such as a semiconductor detector such as an MCP, a scintillator, a PIN diode, an APD, and an MPPC can be used. In order to perform boundary detection in real time, it is desirable to accelerate secondary electrons generated on the sample surface so as to minimize the delay. For this purpose, it is preferable to apply an accelerating electric field between the sample and the electron detector. As a matter of course, it is desirable that the amplifier to be used performs amplification using an ampoule capable of frequency response of ns.

The objective lens 7 is for narrowly irradiating the sample 8 with the electron beam 111 in the order of nm, and is usually a magnetic field type objective lens. Of course, it can also be constituted by an electrostatic lens. In order to reduce the objective lens aberration, a method in which highly accelerated electrons are incident on the objective lens, a reverse bias is applied to the sample, and the electron beam is braked to control the landing energy of the electron beam can be used. (So-called retarding method and intermediate acceleration method)
The sample 8 is a sample with a resist pattern fixed on the stage 9 (for example, a photomask or a semiconductor wafer).

  The stage 9 fixes the sample 8 and moves it to a predetermined position while measuring the coordinate position in the X direction and the Y direction precisely by the length measuring device 27.

  The electron beam 111 is a primary electron beam emitted from the electron gun 1.

  The secondary electron beam 121 is an aggregate of secondary electrons emitted when the electron beam 111 is narrowly narrowed and irradiated onto the sample 8 while being scanned in a plane. The secondary electrons 121 emitted from the sample 8 travel toward the MCP 6 to which a positive voltage is applied while spiraling on the axis by the magnetic field of the objective lens 7, collide with the detection surface of the MCP 6, and are amplified. The signal current is appropriately amplified, converted into an analog signal, taken into a PC, and a secondary electron image having a gray scale of the signal current is generated on a display. The secondary electron beam includes reflected electrons and the like.

  The high-voltage power supply 21 applies a high positive voltage to the electron gun 1 and supplies a heating voltage to the cathode in accordance with an instruction from the PC 31.

  The CL power supply 22 supplies a predetermined current to the CL 3 according to an instruction from the PC 31.

  The deflection power supply 23 applies a predetermined deflection voltage to DEF (1) 4 and DEF (2) 5 in accordance with an instruction from the PC 31.

  The secondary electron detector 24 sends a secondary electron signal (secondary electron image) to the PC 31 by applying a predetermined voltage to the MCP 6 and further amplifying the signal amplified by the MCP 6.

  The objective lens power supply 25 supplies a predetermined current to the objective lens 7 in accordance with an instruction from the PC 31.

  The stage control power supply 26 supplies a predetermined control power to a motor for driving the stage 9 to move the sample 8 to a predetermined place.

  The length measuring device 27 is a laser interferometer or the like that precisely measures the position of the stage 9 or the sample 8. The position coordinates at which the electron beam irradiates the sample can be obtained by combining the measurement position of the laser interferometer and the scanning position of the electron beam. The irradiation start position is determined using this value and coordinates such as CAD data.

  The PC 31 is a known personal computer that performs various controls and processes according to a program. In this example, the scanning area setting unit 32, the scanning method designating unit 33, the global alignment unit 34, the image acquiring unit 35, the length measuring unit 36, It is composed of a judgment means 37, a DB 41 and the like.

  The scanning area setting means 32 sets the scanning area (scanning area, partial scanning area) of the pattern on the sample 8.

  The scanning method designating means 33 designates a scanning method for scanning the scanning area of the pattern on the sample 8.

  The global alignment means 34 generates a coordinate conversion equation between the pattern on the sample 8 and the pattern on the design data. Here, the precise position of the measurement target to be irradiated with an electron beam existing on a sample such as a semiconductor wafer or a photomask is precisely specified by using CAD data used for designing a semiconductor device after performing an appropriate alignment procedure. You can do it. Of course, if there is no CAD data, by specifying the range of the measurement target while observing the SEM image with the electron beam irradiation amount as small as possible at a low magnification that does not affect the resist, the resist and other And the position where the electron beam scanning is performed can be automatically determined. Of course, a high magnification optical microscope or AFM can also be used. When using something other than the electron beam, alignment is performed so that the electron beam irradiation position and the center position of the optical microscope coincide.

  The image acquisition means 35 scans the scanning area (scanning area, partial scanning area) of the pattern on the sample 8 with an electron beam that is narrowed down, and uses the amount of signal electrons generated on the sample as luminance information as an image (scanning area image). , Partial scanning area image).

  The length measuring means 36 measures a pattern by extracting a boundary portion from the image (scanning area image, partial scanning area image) acquired by the image acquiring means 35.

The reliability determining unit 37 determines the reliability of a value obtained by extracting a resist boundary region portion from the acquired image (scanning region image or partial scanning region image) and measuring the length of the pattern. The data is stored so that it can be easily searched. In this case, the data includes CAD data for determining a wiring pattern and a measurement position, measurement result data, reliability data, and the like.

  FIG. 2 shows an example of a photocathode type electron source capable of high-speed ON / OFF according to the present invention.

  FIG. 2A shows a configuration diagram, and FIG. 2B shows an example of a beam OFF signal.

  In FIG. 2A, a laser 51 oscillates a laser.

  The condenser lens 52 converges the laser generated from the laser 51 on the photoelectric film 53 as shown.

  The electron beam 54 is obtained by accelerating electrons emitted when the photoelectric film 53 is irradiated with a laser by an electrode (not shown).

  The automatic blanking control means 11 is the automatic blanking control means shown in FIG. 1, and generates a blanking signal when a predetermined interval (for example, an interval between N lines and M pixels) elapses from the edge of the pattern. The emission is stopped (blocked), and the irradiation of the sample with the electron beam is stopped (described later with reference to FIGS. 3 and 4).

  Here, a semiconductor laser 51 that can be turned on / off at high speed is realized by combining a photocathode type electron gun that irradiates a photoelectric film 53 such as GaAs to generate electrons. The semiconductor laser can be turned on / off much faster than ns, and can stop the electron beam in a time shorter than the time required for the electron beam to move in the X and Y directions by one pixel deflection scanning. Therefore, by inputting a signal for determining the resist boundary to the laser control signal, it is possible to reliably stop the electron beam irradiation within an error of one pixel or less and prevent the resist from shrinking. As a matter of course, if two pixels, three pixels, and N pixels are designated after receiving the resist boundary signal, the electron beam irradiation stops at that position. Similarly, by using a micro-emitter with a gate formed by using MEMS or semiconductor manufacturing technology, a blanking signal can be given to the gate of the electron gun itself, and the electron beam can be turned on / off at a high speed. That is, by using the blanking signal as the ONF / OFF signal of the electron source, it is possible to control on / off of the irradiation of the electron beam. The signal point should be corrected so that the blanking signal correctly corresponds to the boundary.

Next, the operation of the configuration of FIG. 2 will be described.
(1) When the electronic detector of FIG. 2A scans, for example, from the left to the right of a line and detects an edge on the left side of the line in real time, the automatic blanking control means 11 sets a predetermined interval from the edge. After (an interval of N lines or an interval of M pixels), a beam OFF signal is generated.
(2) The laser is turned off by the laser 51 based on the beam OFF signal of (1), and the emission of the laser is stopped. Then, the electron beam 54 emitted from the photoelectric film 53 is turned off, and the electron beam can be cut off at an extremely high speed.

  As described above, after detecting the edge of the line, a beam OFF signal is output after a predetermined interval, the laser is turned off, and the electron beam is turned off. As a result, after a predetermined interval after the electron beam passes through the edge of the line This makes it possible to cut off the electron beam at an extremely high speed, eliminate unnecessary electron beam irradiation, and reduce damage and shrink to a line (made of resist) to a necessary minimum.

  FIG. 3 shows an operation explanatory diagram (part 1) of the present invention.

  FIG. 3A shows a scanning example along the line longitudinal direction. The vertical direction represents the main scanning direction (Y axis), and the horizontal direction represents the sub scanning direction (X axis).

  In FIG. 3A, scanning (main scanning) of an electron beam is started from a space portion in a vertical direction parallel to a line, and scanning is performed sequentially (main scanning) toward a rightward line. , The intensity of the secondary electrons detected reaches a strong peak, and an edge (a boundary between a space and a line) can be detected. After the edge is detected, when an interval of N lines (for example, 3 lines) is reached, an electron beam OFF signal is generated, the electron beam is cut off, and the irradiation of the inner portion thereafter is cut off. . Thereby, the extra electron beam is not irradiated to the inside of the line, and the damage and shrink of the line can be reduced.

  Hereinafter, FIG. 3A will be described in detail.

  The present invention is characterized in that measurement is performed without irradiating an electron beam to a line as a resist structure as much as possible. Therefore, using the design CAD data, the design position of the line of the structure to be measured and the space around the line can be accurately known in advance. The electron beam scanning starts from a space region where there is no resist and scans for a required length in a direction along the line longitudinal direction. For example, a scan is started from a left side of the line for a preset distance (number of pixels) along the longitudinal direction of the line (scan 1). As in the conventional SEM, information (amount, emission direction, energy, etc.) of signal electrons generated during scanning is detected in real time as a function of the scanning position.

  When scanning the space portion, secondary electrons determined by the secondary electron emission capability of the base (quartz, metal film, etc.) are emitted. Since there is no resist in the space, even if a large amount of electron beam irradiation is performed, it does not affect resist shrink, so that electron beam scanning is performed by irradiating a sufficient amount of electrons so that an image signal with a necessary and sufficient SNR can be obtained. Do. When the scanning is gradually performed from the space area to the right side and approaches the line, a very thin remaining film of the resist structure forming the line structure is approached to the area where the remaining film remains. Scanning further to the right enters a thicker resist residual area. Although depending on the energy of the primary electrons, the resist has the property of absorbing electrons, so that the signal intensity temporarily decreases. The secondary electron detection ability is so sensitive that a change in the resist film thickness of several nm or less on the surface of the measurement target can be detected. After that, when it hits the side wall of the resist structure, the secondary electron emission rapidly increases due to the cosine effect. As described above, two opposing partial scan images acquired across the structure to be measured are used as a pair. In the present invention, since the SNR of the signal is very high, it is possible to accurately measure these signal changes.

  The relationship between the electron beam scanning position and the signal electron intensity can be replaced with a simple approximation or the like. By defining a first threshold value for determining the resist boundary in advance with respect to the obtained approximate expression, it becomes possible to determine the boundary between the space and the resist structure.

  The signal used for the determination may be used in a state where the secondary electron signal and the reflected electron signal are mixed, or the energy of the secondary electron signal and the reflected electron signal may be used so that the signal change corresponding to the boundary change appears strongly. Filtering may be performed using only electrons having desired energy, or a plurality of signals may be combined.

  FIG. 3B shows an example of a scanned image along the line longitudinal direction. FIG. 3A shows an example of an image when scanning is performed in the same manner as in FIG. 3A. The vertical direction represents the main scanning direction (Y-axis), the horizontal direction represents the sub-scanning direction (X-axis), and the image shown in FIG. This is the next electronic image.

  In FIG. 3 (b), the electron beam is cut off when it enters the left and right ends of the image of the vertical line (made of resist) by N lines, so that the electron beam is not irradiated into the inside. (The scanning image shown in the drawing shows all of them for explanation).

  As described above, the electron beam is irradiated from both ends of the line to N lines, and the electron beam is cut off inside the line, so that the irradiation of the electron beam inside the line is restricted in real time, and the damage and shrinkage of the line are reduced. It became possible to reduce.

  FIG. 4 shows an operation explanatory diagram (part 2) of the present invention. FIG. 4 shows a cross section of a resist structure (line) and a signal electron waveform generated by electron beam scanning. Unlike a conventional signal electronic waveform, there is a waveform corresponding to an edge portion, but no waveform corresponding to a structure.

  FIG. 4C shows an explanatory view in which scanning is stopped when an edge (+ N line) is reached, and FIG. 4D shows an example of electron beam irradiation.

Next, the operation will be described.
(1) As shown in FIG. 4C, for example, a first threshold value is set as a threshold value for measuring the length reflecting the bottom structure of the resist, and the length reflecting the top structure with the resist is measured. A second threshold is determined as a threshold for performing the determination (where the bottom or the top is defined by the user through experimentation in consideration of measurement reproducibility and the like). Usually, it is often determined by comparing with a peak position, a waveform height%, a differential waveform, or a value obtained by other measuring means such as a cross-sectional SEM or TEM.

When the amount of the electronic signal generated from the sample exceeds or falls below the respective threshold value, the automatic blanking control means judges that the scanning electron beam has reached the boundary between the space and the line (resist) and stops the electron beam irradiation or stops the electron beam irradiation. The electron beam irradiation is stopped by performing beam blanking to prevent the resist structure (line) from being irradiated with the electron beam. The position of the boundary obtained in this manner is recorded as boundary position 1 (X1, Y1).
(2) Then, scanning is started in the space on the right side of the line in reverse to the above (scanning 2). The two scans are in one straight line across the resist structure. It gradually scans to the left and goes closer to the line. As it approaches the line, it approaches a region with a thin resist film residue, and further scanning to the left enters a region with a thicker resist residue. As described above, the change in the resist thickness is changed so that the amount of the secondary electron signal can be sufficiently detected. By expressing this change by an approximate expression and distinguishing it by a second threshold value that defines the resist boundary, the boundary between the space on the right side of the line and the resist region (line) can be accurately known. By utilizing this signal, the automatic blanking control means operates to turn off the electron beam. The position of the boundary (edge) thus obtained is recorded as boundary position 2 (X2, Y2). Since the resolution of the CD measurement is proportional to each line interval, it is preferable to set the line scanning interval to a sub-nm or a small value when a high resolution is required, and to set it to 3 nm when a high resolution is not required.
(3) The position difference (X2-X1) between the two right and left resist boundaries (line edges) obtained as described above is used as the width of the resist structure, so that the measurement can be performed without shrinking the resist structure. You can do it.
(4) Where the threshold should be set for the signal waveform, as in the case of a normal CDSEM, learning such as AI is performed in correlation with the measurement reproducibility to optimize the threshold with the highest measurement reproducibility. You can also.

  FIG. 5 is an explanatory diagram of the line profile of the present invention. When associating the threshold values (first threshold value, second threshold value) shown in FIG. 5, the amount of secondary electrons or reflected electrons is represented by an approximate expression of an X1-dimensional or XY two-dimensional plane with respect to the electron beam scanning position, and its approximation is performed. The boundary may be obtained by applying a threshold to the expression. Alternatively, an approximate curve (boundary) connecting the respective coordinates of the boundary points obtained as the left and right boundaries of the resist structure may be considered, and the distance may be measured using the approximate curve.

  At this time, since the distances obtained by each electron beam scanning vary in distance due to edge roughness, the results of electron beam scanning of parts that are assumed to have the same distance by design are integrated and statistically averaged. When the processing is performed to determine the length, a measurement result with higher accuracy and good reproducibility can be obtained.

  FIG. 6 is a diagram (part 3) for explaining the operation of the present invention. This shows an example of scanning in a direction perpendicular to the line.

  In FIG. 6, this embodiment is characterized in that the direction in which electron beam scanning is performed is not parallel to the resist structure (line), but rather scans at an angle close to a right angle (may be oblique). If the scanning is performed obliquely, the line width becomes apparently longer, and the number of pixels of the image forming the boundary portion can be increased. Therefore, the measurement accuracy can be improved. When scanning is performed diagonally, the obtained distance is subjected to scanning angle correction and converted to a correct distance.

  FIG. 7 is a diagram (part 4) illustrating the operation of the present invention. This explains the details of FIG. This will be described in detail below.

  (1) The position on the sample at which electron beam irradiation is started is precisely determined by CAD guidelines 1 and 2 using CAD data or a manual. The scanning is started from a part of the space where no resist exists. Since the material of the space portion and the material of the resist portion (line) are different, different secondary electron emission amounts are shown for the same electron beam irradiation. For example, an electron beam is scanned from the left space portion toward the resist region (line), and secondary electrons or reflected electrons are continuously detected as a function of the electron beam scanning position, and the signal electron amount is detected as a function of the scanning position. From the space portion having the starting point specified by the CAD data to the resist region (line), the thin film of the resist approaches the residual region. At this time, the secondary electrons or the reflected electrons change the signal amount according to the thickness of the resist. When scanning proceeds to a resist area (line) where a thicker residual remains, a larger change in signal amount occurs. The boundary (edge) between the space region and the resist region can be detected by detecting the change in the signal amount (signal intensity of the edge) at this time (the boundary by the signal electrons).

  (2) Using the detected signal, the automatic blanking control means scans an extra N or M pixels from the boundary point (edge) and then cuts off the electron beam. Thereby, one piece of partial image data is obtained.

  (3) After stopping one electron beam scan, the second electron beam scan is started using the guideline indicated by each CAD data. By repeating this as many times as necessary, a set of partial images near the edge where the scan end point is self-aligned is obtained. The pixel positions of adjacent scanning lines can be arbitrarily determined as needed. The next scan after one line has been scanned is freely programmable in the present invention and can be varied depending on the signal electrons from the structure. For example, if the required SNR is not reached, the same location may be scanned again, or the horizontal pixel start position and end position may be changed. It is also possible to perform scanning from the scanning start point to the scanning end point, and to perform scanning so as to return to the scanning start point again (variable ESR).

  (4) In order to increase the blanking speed, an electrostatic blanking method is used, and blanking electrodes are designed so that the capacitance is minimized. Similarly, an extremely high-speed transistor having a response speed of several ns or less and a large output current is used as an output on the circuit side, so that a sharp rise and fall can be realized. It is desirable to place a blanking aperture at a position where the primary electron beam width becomes narrow. Since blanking has a finite speed, a tail occurs in ON / OFF of the primary electron beam. Therefore, if necessary, a method of reducing the influence of the tail by lowering the clock speed can be used so that blanking can be performed more accurately.

  (5) As described above with reference to FIG. 2, instead of performing blanking, the electron beam itself output from the electron source can be stopped using the automatic blanking control means together with the boundary detection (edge detection). In particular, when a photo-excitation type electron gun is used, a similar effect can be obtained by turning off a high-speed light source such as a laser for irradiating the electron source in accordance with blanking timing.

  (6) In the embodiment, line / space has been described for simplicity of description. However, since an arbitrary figure can be defined as a connection or aggregate of small lines, even an arbitrary shape other than line / space can be used. Needless to say, it can be measured similarly. In particular, when performing contour extraction of a pattern having an arbitrary shape, the approximate contour is specified using CAD data, electron beam scanning is started from that position, and the electron beam scanning range is determined using the boundary of the actual structure. By using the determined method, it becomes possible to scan only the contour portion of the pattern that needs to be inspected, which is present on the semiconductor wafer or the photomask, and the contour of the structure to be measured can be extracted at a very high speed.

  (7) Specifically, when electron beam scanning is started from the electron beam scanning start point specified by the CAD data and an edge is detected, the electron beam is stopped when N pixels have passed therefrom. Next, scanning is started with the CAD guideline corresponding to the pixel immediately next to the starting point as a starting point. The scanning direction is desirably scanning in a direction perpendicular to a partial line forming an arbitrary shape. By repeating this, it becomes possible to perform electron beam scanning only in the vicinity of the contour of the measurement object, and the contour can be extracted only by measuring a very small area.

  (8) Usually, an inspection apparatus scans a wafer or a photomask two-dimensionally, so it is necessary to acquire many pixels as an image. The number of acquired pixels is proportional to the inspection time, and the greater the number, the longer the inspection time. Inspections often have meaning in the inspection. According to the present invention, a measurement object can be expressed by performing a scan close to one dimension, so that inspection can be performed by acquiring a very small amount of pixels. Can be performed.

  FIG. 8 shows another explanatory diagram of the present invention.

  In FIG. 8, the container is a container for storing the ionic liquid.

  The ionic liquid is a liquid having ionic conductivity.

  The substrate is a substrate on which a pattern is formed, and is, for example, a wafer, a mask, or the like.

  The resist is a resist that forms a pattern on a substrate, and is immersed in an ionic liquid here.

Next, the configuration of FIG. 8 will be described.
(1) An ionic liquid is a salt (a substance that forms an ionic crystal) having a melting point of 100 ° C. or less and has a strong ionic bond. Have Therefore, it is used as an antistatic agent and a lubricant because it remains without evaporating even when it is left as it is in a vacuum chamber of an electron microscope in a vacuum state and exhibits conductivity. On the other hand, since it is the same as a salt and a lump of pure ions, each constituent substance has a positive ion minus an ion and is well mixed with a polar solvent.
(2) The following types of main ionic liquids are known. Since the ionic liquid itself is new, it will come out in the future and can be used.

 Basically, there are three types of cations: pyridines, alicyclic amines, and aliphatic amines. By selecting the type of anion to be combined with this, various structures can be synthesized. Examples of cations used include ammonium ions such as imidazolium salts and pyridinium salts, phosphonium ions, pyrrolidinium, and non-ion ions.Examples of anions used include halogen ions such as bromide ions and triflate, and tetraphenylborate. And phosphorous such as hexafluorophosphate.

ArF resists and the like used in the state-of-the-art exposure processes contain a large amount of the following solvents for viscosity adjustment. As the solvent, propylene glycol monomethyl ether has a melting point of -96 ° C, a boiling point of 120 ° C, a butyl acetate melting point of -74 ° C, a boiling point of 126 ° C, or an organic solvent is separated and generated as a resist decomposition reaction product.
(3) When exposure for forming a pattern on the resist is performed, and development, washing, and baking are performed, a large amount of the solvent remains inside the resist. Therefore, when the resist structure after development or baking is put in a vacuum chamber of an electron microscope, the solvent evaporates more and more and the loss occurs. In particular, where the electron beam irradiation converged to several nm was performed, the temperature of the resist locally increased, the acid generated by the acid generator was generated, the resist was cleaved, and the solvent jumped out. Loss or shrinkage occurs.

On the other hand, since the above-mentioned ionic liquid is conductive and has almost no vapor pressure, by replacing the solvent constituting the resist after development, the organic solvent evaporates from the resist to the outside even in a vacuum, and is reduced. Can be prevented. Substitution acceleration may be performed by applying ultrasonic waves or the like in order to quickly cause the ionic liquid to penetrate the resist structure. In the present invention, since the electron beam irradiation is automatically stopped at the edge portion, the edge portion only needs to be replaced with the ionic liquid, so that the replacement can be performed relatively easily.
(4) Conventionally, a technique of applying an ionic liquid to the resist surface for the purpose of preventing the resist from being charged after baking has been known. However, as in the present invention, the ionic liquid is developed after or during the development of the organic solvent constituting the resist. For the first time, the present invention has disclosed a method of preventing the organic solvent contained in the resist from evaporating in vacuum by suppressing the shrinkage.

Irradiation with an electron beam causes a reaction with an acid generator contained in the resist, thereby promoting a resist cleavage reaction. Although the organic solvent generated by this reaction flies in a vacuum, loss and shrinkage are promoted. However, when the present invention is used, the solvent is replaced by an ionic liquid having a low vapor pressure, so that almost no evaporation occurs. In other words, even if the resist is irradiated with an electron beam, it is possible to completely prevent the shrinkage and loss of the resist from occurring.
(5) By combining these techniques with the electron beam scanning method described above, it is possible to avoid a change in the resist that occurs when the electron beam is slightly scanned on the resist, and further to prevent the resist from being damaged by the electron beam irradiation. In other words, it goes without saying that the state before the electron beam irradiation can be maintained.

  FIG. 9 shows another explanatory view (part 2) of the present invention. FIG. 9 is characterized in that a resist or a sample is cooled by a cooling device in a vacuum chamber during measurement.

  In FIG. 9, an electron beam column is a well-known electron beam column for generating, converging, and narrowing an electron beam to scan a sample on a plane.

  The sample is a pattern or the like formed on the substrate.

  The cooling device is a stage for cooling to a low temperature, and mounts a sample and cools the sample.

  The XY stage moves the sample to a target position.

  The vacuum chamber is a container in which a sample, a cooling device, an XY stage, and the like are placed in a vacuum.

  The cooling helium gas is N gas for cooling the cooling device to a low temperature.

Next, the operation will be described.
(1) In FIG. 9, the shrink phenomenon that occurs when the resist is irradiated with an electron beam is a kind of chemical reaction. In general, a chemical reaction has a reaction rate determined by chemical energy, so that the progress of a chemical reaction can be generally controlled by temperature. Generally, if the temperature is lowered by 10 degrees, the reaction rate is halved. Therefore, if the temperature is lowered by 100 degrees, the reaction rate is reduced to 1/1000. To reduce the amount of shrink, which is considered to show substantially no change in the resist, to about 1/100, the temperature should be lowered by about 60 degrees from room temperature. That is, the temperature should be lowered from about minus 30 to about 40 degrees. The sample temperature can be reduced.
(2) Therefore, as shown in FIG. 9, a cooling device is provided on a metal or glass ceramic pallet, an XY stage or an electron beam column for carrying the sample, and the sample is cooled during the measurement. The whole sample may be cooled, or only the measurement point may be cooled. Generally, the melting point of the organic solvent generated or contained in the resist is higher than -100 degrees, so that the organic solvent can be solidified using a semiconductor cooling device using a Peltier element or the like. Of course, it may be cooled with liquid nitrogen, its cooling gas, or helium gas using a pipe so as not to leak in a vacuum. When the measurement is performed in a low vacuum state, a cooling gas may be sprayed on the measurement target portion.

If the cooling is performed as described above, the shrinkage becomes 1/100 or less of the room temperature, and the shrinkage can be substantially realized.
(3) In the case where the ionic liquid is replaced with the above-described one, the shrinklessness of the resist can be substantially realized only by cooling to around 0 degrees. When this method is used in combination with the above-described method, accurate and highly reproducible measurement with less resist damage can be performed.

  FIG. 10 shows another embodiment of the present invention. This is a configuration diagram of another embodiment in the case where the length of a sample is measured at atmospheric pressure and further increased pressure.

This will be described in detail below.
(1) As described above, the shrink of the resist is mainly caused by the local warming of the organic solvent contained in the resist or the organic solvent generated when the resist is irradiated with the electron beam and volatilization in a vacuum. Therefore, if the measurement environment is set at atmospheric pressure or under pressure, the reaction is suppressed by chemical equilibrium and the evaporation of the generated organic solvent can be reduced.
(2) There are various methods for measuring a sample in the atmosphere. A series of electron microscopes called atmospheric pressure SEMs can achieve that purpose. The device disclosed by the present inventor is one of them (for details, refer to JP-A-2014-157772). In these electron microscopes, a light element thin film such as carbon or silicon, which can pass an electron beam but cannot pass a gas such as air, is disposed at an exit of an electron beam to shut off the atmosphere. This is a device that can be irradiated with an electron beam through a film. However, simply irradiating the electron beam at atmospheric pressure cannot completely stop the shrinkage of the resist. Therefore, the self-aligned ESR electron beam scanning method, the ionic liquid replacement method, or the sample cooling method described above is used in combination. Needless to say, doing this will make you more effective.

  FIG. 11 shows an explanatory diagram of the present invention.

  FIG. 11A shows a scan for each pixel. This shows a state of scanning when the spot size of the electron beam and the scanning step interval are equal. In this case, when scanning (main scanning) from the left side to the line as shown in the drawing, the scanning is sequentially performed at the spot size of the electron beam as shown in the figure, so that the brightness of the secondary electrons detected at that time is (Amplitude) is a step-like waveform as shown in the figure.

  On the other hand, FIG. 11B shows a sub-pixel scan. This shows a state of scanning when the spot size of the electron beam is significantly larger (longer) than the scanning step interval. In this case, when scanning (main scanning) from the left side to the line as shown in the figure, the electron beam is sequentially scanned (shifted) rightward with a large spot size of the electron beam as shown in the figure. The brightness (amplitude) of the secondary electrons is a waveform having a peak at the edge portion (boundary portion between the space and the line) as shown in the figure.

  As described above, when the spot size of the electron beam and the scanning step are changed, the waveform of the detected secondary electrons is different. Therefore, in order to detect the boundary (edge) from the space to the line, the spot at that time is detected. It is necessary to obtain an approximate curve in advance by experiment based on the size and the scanning step, and to accurately detect the edge position using the approximate curve.

  FIG. 12 is an explanatory view (part 2) of the present invention. This explains the operation of generating a cutoff signal N pixels after detecting the edge detection signal, cutting off the electron beam, and cutting off the beam.

  In FIG. 12, a scanning signal is a signal for scanning (main scanning) an electron beam.

  The edge detection signal is a signal that detects a boundary (edge) between a space and a line.

  The cutoff signal is a signal for cutting off the electron beam, and is a signal for cutting off the electron beam generated N pixels (M lines) after the edge detection signal (edge) is detected.

  Secondary electrons are the intensity of secondary electrons generated from spaces, lines, and the like.

  Next, the operation will be described.

  (1) As shown in FIG. 7, for example, when an electron beam is scanned from a space to a line (resist) to detect a secondary electron peak that is a boundary (edge) between a space and a line, Output an edge signal.

  (2) After the edge signal is output, the cutoff signal is output when the scanning signal advances by N pixels.

  (3) When the cutoff signal is output, the electron beam is cut off (beam cutoff). Since the inside of the line after this is not irradiated with the electron beam, damage to the line (resist) and shrinkage can be prevented.

It is a system configuration diagram of the present invention. 5 is an example of a photocathode type electron source capable of high-speed ON / OFF according to the present invention. FIG. 3 is an operation explanatory diagram (part 1) of the present invention. FIG. 8 is an operation explanatory diagram (part 2) of the present invention. FIG. 3 is an explanatory diagram of a line profile of the present invention. FIG. 8 is an operation explanatory view (3) of the present invention. FIG. 4 is an operation explanatory diagram (part 4) of the present invention. It is another explanatory drawing of this invention. It is another explanatory drawing (the 2) of this invention. FIG. 6 is a configuration diagram of another embodiment of the present invention. It is explanatory drawing of this invention. It is explanatory drawing (the 2) of this invention.

1: electron gun 2: aperture 3: CL (condenser lens)
4: DEF (1)
5: DEF (2)
6: MCP
7: Objective lens 8: Sample 9: Stage 11: Automatic blanking control means 12: Blanking device 13: Blanking electrode 14: Blanking aperture 111: Electron beam 121: Secondary electron 21: High voltage power supply 22: CL power supply 23 : Deflection power supply 24: Secondary electron detector 25: Objective power supply 26: Stage control power supply 27: Length measuring device 31: PC (PC)
32: scanning area setting means 33: scanning method designation means 34: global alignment means 35: image acquisition means 36: length measurement means 37: reliability determination means 41: DB

Claims (6)

In an electron beam measuring apparatus for extracting the dimension, area measurement or contour of a pattern of a distance measurement target formed on a sample,
Setting means for setting a partial scan area including an edge of the pattern of the distance measurement target,
Detecting an electron generated by scanning the electron beam in the partial scanning area to acquire a partial scanning image, and, when acquiring the partial scanning image, detecting means for detecting an edge of the pattern in real time,
An electron beam OFF signal generating means for detecting the edge in real time by the detection means and generating an electron beam OFF signal in real time after a predetermined interval has passed;
When the electron beam OFF signal generation means generates an electron beam OFF signal in real time, the electron beam cutoff means for cutting off the electron beam in real time to cut off the irradiation on the pattern,
An electron beam measuring apparatus characterized by automatically acquiring a minimum image required for measurement.   Means for extracting an edge in each of the partial scan images acquired in opposition to the pattern to be measured, and obtaining a distance between opposing edges of the pattern from each edge position information. Item 7. An electron beam measurement device according to Item 1.   2. The electron beam measuring apparatus according to claim 1, wherein an edge is extracted from the partial scan image acquired from the measurement target pattern to extract a contour of the measurement target pattern.   The predetermined interval is an interval corresponding to N (integer) scan lines when the electron beam is scanned in a direction parallel to the measurement target edge of the pattern, or the electron beam is scanned in a direction perpendicular to the measurement target edge of the pattern. 4. The electron beam measuring apparatus according to claim 1, wherein an interval corresponding to M pixels (integer) of a scanning line is set in the case.   5. The electron beam measuring apparatus according to claim 1, wherein the pattern to be measured is a pattern that is reduced by electron beam irradiation or a resist pattern. In the electron beam measurement method for extracting the dimension, area measurement, or contour of the pattern of the distance measurement target formed on the sample,
A setting step of setting a partial scan area including an edge of the pattern of the distance measurement target,
A detection step of detecting an electron generated by scanning the electron beam in the partial scanning region to acquire a partial scanning image, and detecting the edge of the pattern in real time when acquiring the partial scanning image,
The detecting step detects the edge in real time, an electron beam OFF signal generating step of generating an electron beam OFF signal in real time after a predetermined interval,
An electron beam cutoff step of cutting off the irradiation of the pattern by cutting off the electron beam in real time when the electron beam OFF signal generation step generates an electron beam OFF signal in real time,
An electron beam measurement method characterized by automatically acquiring a minimum image required for measurement.