Drift reduction device and drift reduction method
The drift reduction device and method address the challenges of thermal drift in photomask repair by using a reference pattern and temperature control to achieve precise, automatic photomask repair.
Patent Information
- Application Number
- JP2023075338
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-04-29
- Publication Date
- 2024-11-11
AI Technical Summary
Conventional photomask repair methods suffer from drift due to thermal expansion caused by high-temperature gas injection, making high-precision repair difficult, and real-time confirmation of drift reduction is challenging.
A drift reduction device and method that uses an electron beam irradiation system with a reference pattern as a base point to correct and reduce drift, incorporating a radiant heat shield and temperature control of the photomask to minimize thermal effects, and a repair area setting mechanism to ensure precise alignment and repair.
Enables highly accurate, fully automatic photomask pattern repair by correcting drift and eliminating the need for separate measurement devices, ensuring precise alignment and real-time confirmation of repair quality.
Smart Images

Figure 2024160038000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to a drift reduction apparatus and a drift reduction method for reducing drift of a pattern to be repaired on a photomask that occurs when the photomask is repaired. [Background technology]
[0002] Conventionally, the repair of a pattern on a photomask has been performed by manually creating, selecting, or copying an existing pattern to determine the target pattern, and then repairing the pattern by deposition or etching using an electron beam while injecting gas.
[0003] In this case, the photomask is sprayed with high-temperature gas, and the repair target (location) drifts due to thermal expansion, making it extremely difficult to perform repairs with high precision (on the order of a dozen nm or nanometers).
[0004] Furthermore, even if the palette holding the photomask was kept at a constant temperature and temperature changes were reduced, it was not possible to check in real time whether the drift had actually been reduced, making it extremely difficult and time-consuming to confirm whether high-precision repair had been performed. DISCLOSURE OF THEINVENTION [Problem to be solved by the invention]
[0005] As mentioned above, conventional photomask repair involves manually creating, selecting, or copying an existing shape to be repaired, and then injecting gas into it to repair (remove or deposit). However, there is a problem in that drift occurs when gas is injected, making it impossible to repair with high precision.
[0006] Furthermore, even if the photomask is controlled to a constant temperature, it is extremely difficult to check in real time whether the drift of the photomask to be repaired is being reduced with high precision. [Means for solving the problem]
[0007] In order to solve the above-mentioned problems, the present invention irradiates an electron beam while injecting gas (etching gas and deposition gas) onto the portion of the repair pattern to be repaired, using a reference pattern selected or newly created from near the repair target on the photomask as a base point. Even if the photomask drifts, the drift can be corrected and reduced by expressing and controlling the irradiation position of the electron beam used for repair as a function of the reference pattern on the photomask as a base point.
[0008] For this reason, the present invention provides a drift reduction device for reducing drift of a repair target on a photomask, comprising: a reference pattern selected or newly created from the vicinity of the repair target on the photomask; a repair area setting means for setting a repair area to be repaired by spraying an etching or deposition gas from a nozzle while irradiating an electron beam onto a portion of the repair pattern of the repair target based on the selected or created reference pattern; and a repair means for repairing the repair area by spraying an etching or deposition gas from a nozzle while irradiating an electron beam onto the repair area.
[0009] At this time, the repair area setting means measures the amount of drift of the reference pattern, corrects the repair area by the measured amount of drift, and reduces the influence of the drift.
[0010] The restoration means performs one or more scans starting from either side of the restoration area to restore the entire area.
[0011] In addition, the repair means acquires a secondary electron image of a reference pattern, and compares the acquired secondary electron image of the pattern to be repaired with the repair pattern to be repaired based on the acquired secondary electron image of the reference pattern, and terminates the repair if it is within a predetermined tolerance, and outputs a message to that effect or causes the repair to be retried if it is outside the predetermined tolerance.
[0012] Moreover, the reference pattern is set within a range that can be acquired within one screen together with the repair pattern.
[0013] In addition, a flat radiation heat shielding plate is provided adjacent to and facing the area to be repaired on the photomask irradiated by the electron beam, and has holes through which the electron beam passes when irradiated and through which gas is sprayed from a nozzle onto the photomask, thereby reducing temperature drift of the photomask due to radiant heat.
[0014] In addition, the temperature of the bottom surface of the flat pallet that holds the photomask is controlled to a predetermined temperature to reduce temperature drift of the photomask.
[0015] In addition, the temperature of the photomask is controlled by injecting gas at a predetermined temperature through holes in the photomask, thereby reducing temperature drift of the photomask. Effect of the Invention
[0016] The present invention uses a reference pattern selected or newly created near the repair target on the photomask as a base point, and irradiates the repair pattern portion of the repair target while spraying gas (etching gas and deposition gas) onto it, making it possible to correct and reduce the drift even if the photomask drifts by using the reference pattern on the photomask as a base point.
[0017] This has made the following possible: 1 During fully automated mask repair within a single device, photomask drift can be corrected and high-precision photomask pattern repair can be achieved. 2. No separate equipment is required to measure or obtain precise repair pattern information. 3. There is no need to evaluate the repair results of the repair pattern using a separate optical simulation microscope. Example 1
[0018] Fig. 1 shows a configuration diagram of one embodiment of the present invention. This Fig. 1 shows the configuration diagram of an apparatus for repairing defects in a pattern formed on a photomask 12, in which the photomask is irradiated with an electron beam while gas is being injected thereon, and the defect is repaired by etching or deposition within the reactive area of the electron beam. A detailed description is given below.
[0019] In FIG. 1, a column 1 is used to generate a primary electron beam, focus it, narrow it down, and irradiate it onto a photomask 12, and is made up of an electron gun 2 and other components.
[0020] The electron gun 2 generates electrons.
[0021] The blanking electrode 3 passes or blocks the primary electron beam generated by the electron gun 2 at high speed (on the order of nanoseconds), and is essentially a parallel plate electrode to which a high voltage is applied. It is designed to prevent the blanked primary electron beam from becoming charged and drifting.
[0022] The U-MCP4 is an MCP (electron detector) located on the upper side (closer to the electron gun 2) that multiplies electrons and detects them with high sensitivity, and mainly detects and multiplies backscattered electrons here. The detected signal is converted to digital data by an AD converter, then sent to a PC for imaging or used for elemental analysis required for endpoint determination.
[0023] The D-MCP5 is an MCP (electron detector) located on the lower side (closer to the objective lens 7) that multiplies electrons and detects them with high sensitivity, mainly detecting and multiplying secondary electrons here. The detected signal is converted into digital data by an AD converter, and then sent to a PC to be imaged.
[0024] The deflector 6 deflects the primary electron beam in two stages (X and Y directions) and scans a desired position on the photomask 12 while irradiating it with a finely focused electron beam.
[0025] The objective lens 7 narrows the primary electron beam and irradiates it onto the photomask 12. Here, an aberration corrector (spherical and chromatic aberration corrector) (not shown) is used in combination as necessary to reduce spherical and chromatic aberration to a minimum.
[0026] The differential exhaust device 8 exhausts gas from the middle part of the differential exhaust device consisting of a two-stage disk with a hole in the center through which primary and secondary electrons pass, and exhausts gas from the middle part of the differential exhaust device so that the gas exhausted into the sample chamber 21 does not flow toward the column 1 side.
[0027] The radiant heat shield 9 is a thin plate that blocks heat from piping etc. connected to the nozzle 10 from reaching the photomask 12 as radiant heat, and has a hole in the center through which the primary electron beam, the nozzle 10 or the gas from the nozzle 10 can pass.
[0028] Nozzle 10 injects gas onto a predetermined area of photomask 12, and normally injects gas heated to a predetermined temperature (for example, about 80 to 100° C.).
[0029] The Z sensor 11 precisely measures the distance (WD) of the photomask 12 in the Z direction.
[0030] The photomask 12 is a photomask on which defects are to be repaired, and is a photomask on which a predetermined pattern is formed based on design data (CAD data) (not shown). The photomask 12 refers to the design data, irradiates the pattern to be repaired easily and with high precision with a primary electron beam, while ejecting gas from the nozzle 10, and performs etching and deposition to repair the defect pattern so that it matches a repair target pattern generated from the design data or the like (described later).
[0031] The heating / cooling controller 13 is disposed on the rear side of the photomask 12, and heats and cools the photomask 12 to control it to a predetermined temperature, thereby reducing drift of the photomask.
[0032] The support table 14 holds the photomask 12, the heating / cooling controller 13, and the like.
[0033] The temperature sensor 15 detects the temperature of the support table 14 or the photomask 12 .
[0034] The XYZ stage 16 is equipped with a temperature sensor 15, a support stand 14, a heating / cooling controller 13, and a photomask 12, and moves precisely to a predetermined position in the X, Y, and Z directions. The stage moves while irradiating a laser onto a bar mirror on the stage, etc., to perform precise measurements in real time (measurements using a laser interferometer).
[0035] The TMP17 is an oil-less vacuum pump.
[0036] The dry pump 18 is an oil-less vacuum pump.
[0037] The controller 19 automatically controls the temperatures of the radiant heat shield 9, the support stand 14, etc. to predetermined temperatures.
[0038] The vibration isolation device 20 is a device that has the sample chamber 21, the column 1, etc. mounted thereon, and isolates external vibrations.
[0039] The sample chamber 21 is a container that is evacuated and in which the photomask 12 and the like are placed.
[0040] The gas temperature adjustment device 22 adjusts the temperature of the gas sprayed from the nozzle 10 onto the photomask 12 by heating the gas to a predetermined temperature and generating gas by sublimation or the like, and further adjusting the pressure of the gas to a predetermined level before supplying it.
[0041] The gas controller 23 controls the gas temperature regulator 22 to generate gas at a predetermined temperature, adjusts the flow rate and pressure of the generated gas using a mass flow controller, and sprays the gas from the nozzle 10 onto the photomask 12.
[0042] The high voltage power supply 31 generates a high DC voltage to be applied to the electron gun 2 .
[0043] The BL controller 32 applies a predetermined high voltage to the blanking electrode 3, and controls the passage and blocking of the primary electron beam at high speed.
[0044] The reflected signal processing device 33 further amplifies the reflected signal detected and amplified by the U-MCP 4, and generates a reflected electron image of the repair area of the photomask.
[0045] The secondary signal processing device 34 further amplifies the secondary signal detected and amplified by the D-MCP 5, and generates a secondary electron image of the repair area of the photomask, the reference pattern, and the like.
[0046] The deflection control device 35 applies a high deflection voltage to the deflector 6 and controls the primary electron beam, which is narrowed and irradiated onto the photomask 12, to be scanned at high speed in the X and Y directions.
[0047] The stage position control device 36 receives real-time measurement signals from the laser interferometer and refers to design data (CAD data) to position the photomask 12 mounted on the XYZ stage 16 so that the primary electron beam is irradiated onto a predetermined position on the photomask 12 (as is well known).
[0048] PC37 is a personal computer that controls various processes and controls according to a program. This computer is equipped with a database and selection software for automatically selecting a reference pattern, design data (CAD data) of the photomask to be corrected, a calculation device for calculating drift vector functions, a device for learning drift, and a database and inference engine for inference calculations. In addition, it is equipped with databases and software required for image calculation and inference for element estimation used for endpoint judgment.
[0049] The display device 38 is a display that displays a secondary electron image of the region to be repaired on the photomask 12, a reflected electron image, a secondary electron image of the reference pattern, the photomask temperature, and the like.
[0050] Next, the operation of the configuration in FIG. 1 will be described.
[0051] The mask repair device in Figure 1 uses a primary electron beam with an energy of about 0.5 to 1 KV. Typical reactive gases are xenon fluoride for etching, trimethyl platinum for deposition, and water vapor or nitrogen as auxiliary gases. The reactive gas is heated to about 100°C and injected from a nozzle 10 toward a photomask 12. The injection amount is a very small amount, ranging from 0 to several tens of SCCM. Different types of gases may be injected simultaneously if necessary.
[0052] Since the activation energy of the electron beam excited gas reaction is several electron volts, secondary electrons with an energy of several electron volts dominate the reaction. The reaction region is not the small spot size of the primary electron beam, which is several nm or less, but the range of the secondary electrons generated when they collide with the photomask 12 and are scattered. The minimum repair resolution of an electron beam mask repair device depends heavily on the acceleration voltage. The reaction region tends to become smaller as the energy of the primary electron beam decreases; for example, the diameter is about 20 nm at 1 KV and 6 nm at 0.5 KV.
[0053] An electron beam irradiation current of about 0 to 100 pA is used. The reaction speed between the electron beam and gas is proportional to the irradiation current, so the higher the irradiation current, the faster the reaction proceeds. The primary electron beam can be irradiated to any position within the deflection area by the deflector 6. It can also be irradiated for any time. For example, by using raster scanning to scan the electron beam in a certain direction at a constant speed while turning it on and off at the desired timing, it is possible to irradiate the electron beam only to a desired position on the photomask 12. A pixel clock of about 1 KHz to 100 MHz is used. The repair process can be completed in one scan by scanning slowly, or the repair process can be completed by scanning multiple times at high speed.
[0054] The reaction speed of the electron beam and the gas depends on the concentration of the gas adsorbed on the surface of the photomask 12 when the electron beam irradiation current is constant. Maximum efficiency is obtained when the gas supply speed and the amount of the electron beam are well balanced. The surface temperature of the photomask 12 is an important control parameter because the gas surface concentration is proportional to the amount of gas adsorption. Since the adsorption of the reactive gas is physical adsorption, generally, the higher the temperature, the smaller the gas adsorption capacity, and the lower the temperature, the larger the adsorption capacity. To increase the speed, a lower temperature is desirable, but since most of the reactive gases are materials gasified by heating a solid source, if the surface temperature of the photomask 12 is made too low below the vaporization temperature, the gas will accumulate on the surface of the photomask 12 with cavities like frost or snow, and the deposited film may become porous.
[0055] As described above, since the surface temperature of the photomask 12 affects the reaction rate and the process results, it is desirable to control it to a desired temperature. It is desirable to use a non-contact temperature sensor using infrared rays to measure the temperature of the photomask 12 so as not to contaminate the photomask 12.
[0056] The pixel clock for electron beam gas reaction and the pixel clock for SEM observation can be the same, or they can be set independently. For example, you can set it to high speed (10 ns per pixel) for SEM observation and slower (1 ms per pixel) for gas reaction. The electron beam irradiation range can also be determined as needed. Similarly, the electron beam spot size for electron beam gas reaction and the spot size for SEM observation can be set to the same or different values.
[0057] The sample chamber 21 of the electron beam mask repair equipment is made of pure iron or Invar to prevent electromagnetic noise from the environment in which the equipment is installed, and its inner surface is coated with corrosion-resistant electroless nickel plating. The sample chamber 21 is evacuated by the TMP 17 and dry pump 18, and is maintained at a high vacuum of about 10-3 Pascals. The inside of column 1 is maintained at a high vacuum of 10-5 Pascals, and the electron gun chamber is maintained at a high vacuum of 10-7 Pascals or higher.
[0058] Since the column (electron beam column) 1 is exposed to corrosive gases, there is a risk that the lifespan of a normal column made of permalloy or pure iron will be shortened due to corrosion. In this embodiment, the gas supply nozzle 10, which is directly exposed to the corrosive gases, and the opening of the objective lens 7 at the tip of the column 1 are covered with a corrosion-resistant material such as metal such as Hastelloy C, nickel phosphorus plating, or antistatic Teflon (registered trademark). A differential exhaust device 8 is installed on the side of the objective lens 7 where the electron beam column 1 is connected to the sample chamber 21. Corrosive gases such as xenon fluoride, chlorine, and iodine injected from the nozzle 10 onto the surface of the photomask 12 are exhausted to the outside before entering the inside of the column 1 by a powerful exhaust system using two differential exhaust apertures and a TMP (not shown) connected between the apertures. This system can prevent the parts inside the column from corroding.
[0059] The repair of the photomask 12 using an electron beam is roughly performed as follows. A photomask is placed on the photomask port. When a load command is sent, a transport device (loader) linked to the photomask transport port automatically transports the photomask to the subchamber and then to the main vacuum chamber (sample chamber 21). Defect position information is obtained from a separately installed photomask defect inspection device. Alignment is performed to match the stage coordinates with the coordinates on the mask using OM or SEM image pattern matching. The XYZ stage 16 is used to move the position of the defect to be repaired on the photomask 12 so that it is on the axis of the primary electron beam. The primary electron beam is scanned to perform SEM observation (secondary electron image observation) of the defect to be repaired. The position of the repair target is measured using the SEM image, and it is moved to a more accurate position, that is, to the center of the image. At this time, it is also possible to automatically recognize the defect using deep learning AI and automatically position it at the center.
[0060] Next, a reactive gas is injected onto the surface of the repair area on the photomask 12 using a gas injection nozzle 10, while a primary electron beam is irradiated onto the repair area. A reaction occurs between the primary electron beam and gas at the repair area. The repair process is observed in real time by superimposing an SEM image and a reflected image obtained in real time during repair on the SEM image before repair. Finally, the repair area is observed with an SEM after the reactive gas is removed to confirm whether the desired repair has been performed. Alternatively, reflected electrons are used to detect material changes as the process progresses. If the repair is insufficient, gas is injected again and the electron beam is irradiated to continue the repair.
[0061] If the repair shape matches the target shape, the repair process is determined to be complete. After all of the parts to be repaired on the photomask have been repaired, the photomask is removed from the repair device by the transport device.
[0062] In this manner, the mask repair work using the primary electron beam is carried out.
[0063] However, in reality, the temperature of the photomask 12 fluctuates due to continuous localized irradiation of a large amount of primary electron beam for several minutes or by blowing gas onto the photomask, causing thermal expansion or contraction, resulting in the so-called drift phenomenon in which the observed image appears in a position shifted from the original position. When this drift occurs, not only is it impossible to properly repair the intended defect, but there is also the risk of repairing the wrong position and creating a new defect.
[0064] An apparatus and method for reducing the effect of photomask drift will be described in detail below.
[0065] Fig. 2 shows an explanatory diagram of the present invention (drift explanatory diagram). Fig. 2 shows how a defective repair pattern (repair area) 43 extracted from design data is scanned from above and below with a primary electron beam while ejecting gas from the nozzle 10 in Fig. 1 to repair the defective area (no drift, drift in the scanning direction, or drift in the opposite scanning direction) with a reference pattern 41 (or a reference pattern may be selected from existing patterns) formed on a photomask 12 as a base point.
[0066] 2(a) shows a case where there is no drift of the reference pattern 41 of the photomask 12, with the upper part showing an image without correction (secondary electron image) and the lower part showing an image with correction of the present invention (secondary electron image). In the case of Fig. 2(a), there is no drift of the reference pattern 41 fixed to the photomask 12 in the vicinity of the repair pattern 43 to be repaired, so the image without correction (upper repair pattern 43) and the image with correction (lower repair pattern 43) are naturally the same, and this is an example where repair is possible without the drift correction of the present invention.
[0067] On the other hand, Fig. 2(b) shows a case where the reference pattern 41 of the photomask 12 drifts in the scanning direction, with the upper image (secondary electron image) showing an uncorrected image and the lower image (secondary electron image) showing an image with correction of the present invention. In the case of Fig. 2(b), the reference pattern 41 fixed to the photomask 12 near the repair pattern 43 to be repaired drifts in the scanning direction (downward in the figure), so the uncorrected image (upper repair pattern 43) becomes a rectangle that is long in the downward direction (corresponding to the repair area scanned by the electron beam). Therefore, when scanning the upper repair pattern 43 in FIG. 2B from top to bottom with the primary electron beam, the drift of the reference pattern 41 (here, drift in the scanning direction (downward)) is tracked, and the repair pattern 43 is scanned in the scanning direction from top to bottom by adding it as an offset, so to speak, and scanning the repair pattern 43 with the primary electron beam. This cancels (subtracts) the drift of the reference pattern 43, and makes it possible to scan and repair (etch, deposit) the repair pattern in the corrected image in FIG. 2B.
[0068] In other words, the present invention repairs the repair pattern 43 by sequentially scanning the primary electron beam from top to bottom (scanning direction Y) using the reference pattern 41 on the photomask 12 as a base point, so that even if drift occurs in the photomask 12, for example in the scanning direction, it is not affected by this.
[0069] Similarly, Fig. 2(c) shows a case where the reference pattern 41 of the photomask 12 drifts in the counter-scanning direction, with the upper image (secondary electron image) showing no correction and the lower image (secondary electron image) showing the correction of the present invention. In the case of Fig. 2(c), the reference pattern 41 fixed to the photomask 12 near the repair pattern 43 to be repaired drifts in the counter-scanning direction (upward in the figure), so the image without correction (upper repair pattern 43) becomes a short rectangle in the downward direction (corresponding to the repair area scanned by the electron beam). Therefore, when scanning the upper repair pattern 43 in FIG. 2C from top to bottom with the primary electron beam, the drift (here, drift in the opposite scanning direction (upward)) of the reference pattern 41 is tracked, and the repair pattern 43 is scanned in the top to bottom scanning direction by subtracting it as an offset, so to speak, and scanning the repair pattern 43 with the primary electron beam. This cancels (subtracts) the drift of the reference pattern 43, and makes it possible to scan and repair (etch, deposit) the repair pattern in the corrected image in FIG. 2C.
[0070] In other words, the present invention repairs the repair pattern 43 by sequentially scanning the primary electron beam from top to bottom (scanning direction Y) on the photomask 12, always using the reference pattern 41 on the photomask 12 as the base point, so that even if drift occurs in the anti-scanning direction on the photomask 12, it is not affected. Drift in the scanning direction has been explained above, but even if drift occurs in another direction, it can be divided into drift in the scanning direction and drift in a direction perpendicular to the scanning direction, and each can be corrected in the same way.
[0071] Fig. 3 shows the drift correction method 1 of the present invention. In Fig. 3, a unique pattern formed on a photomask 12 is adopted as a reference pattern 41, and a repair pattern 43 to be repaired is scanned with a primary electron beam using this reference pattern 41 as a base point while gas is sprayed from a nozzle 10 to perform repair.
[0072] FIG. 3(a) shows the selection of a unique reference pattern 41 formed on the photomask 12. The reference pattern 41 is selected from an existing pattern that is within a range where the repair pattern 43 to be repaired and the electron beam scanning are possible. The selection may be performed automatically or may be manually instructed by an operator. In the figure, among the large patterns, a pattern within a range enclosed by a square in the figure is automatically selected as the reference pattern 41. A method for automatically selecting a reference pattern can be to register the reference pattern figure in a database in advance and select from patterns present within the scanning range of the electron beam using pattern matching or the like. In addition to a specific pattern figure, a method can be used in which a numerical value representing the size or shape or a function expressing the figure is specified as the reference pattern. Alternatively, a figure similar to the registered figure can be searched within a specified range and automatically set. In semiconductor devices, it is highly likely that similar pattern shapes will appear. It is also possible to use machine learning in advance for design data to learn frequently selected reference figures and automatically select a reference pattern based on that inference. Then, as the base point of the reference pattern 41, the intersection point of the maximum and minimum in the X and Y directions is determined as, for example, base point (Xb, Yb) (determined on design data not shown).
[0073] FIG. 3B shows a state in which scanning of a pattern 43 to be repaired with a primary electron beam is started from a position distant from the reference pattern 41 by L (vector).
[0074] FIG. 3C shows a state in which the pattern 43 to be repaired is scanned with the primary electron beam, starting from a position L (vector) away from the reference pattern 41, and the scanning has been completed about halfway.
[0075] 3(d) shows the state where the pattern 43 to be repaired is scanned with the primary electron beam, starting from a position L (vector) away from the reference pattern 41, and the entire scan is completed. The end of the scan is determined when the secondary electron image of the pattern 43 to be repaired matches the target repair pattern determined from the design data within an allowable range. Alternatively, in the case of etching, the end is determined when it is determined that the base (quartz) of the photomask 12 is exposed and becomes a reflected image of the brightness of the base.
[0076] As described above, the reference pattern 41 is automatically selected near the repair pattern (nearby within the range of one screen) that is fixed on the photomask 12, and the position of L (vector) is used as the base point to sequentially scan the specified repair pattern 43 with the primary electron beam while injecting gas to repair the repair pattern 43. Even if the photomask 12 drifts, the primary electron beam is scanned in sequence in the scanning direction with the reference pattern 41 on it as the base point, so that a remarkable effect is obtained in that the influence of the drift is nullified and it is possible to confirm that the repaired repair pattern 43 is within a predetermined tolerance range.
[0077] The details are explained below.
[0078] (1) In Fig. 3, the amount of drift is obtained from an SEM image of the reference pattern 41, and the deposition and etching end points are determined by detecting material changes at the electron beam irradiation point using reflected electrons. Alternatively, a secondary electron image may be obtained, and the end determined when it is within a specified tolerance range with respect to the target repair pattern of the design data to be repaired. In addition, while the process is in progress, the progress of the repair process can be shown in real time on the display by overlaying an SEM image and a reflected electron image obtained from the repaired area during repair on the SEM image before repair. The difference from when there is no drift can also be displayed to show the amount of drift. At this time, the length measurement function of the device is used to accurately measure and display how many nm the movement has occurred.
[0079] (2) In the case of deposition, for example, to achieve the deposition of Zn, conditions are set in advance, such as irradiating Y coulombs of electrons at X kV per pixel, so that the desired height is deposited. In the case of etching, the repair process can be stopped automatically or manually when the underlying material is detected by a backscattered electron signal or image with material contrast. In particular, in the case of etching using xenon fluoride, if the underlying material is quartz, etching stops automatically, so there is no need to worry about excessive removal of the underlying material. Rather, a preliminary experiment is conducted to set the end point so that etching is sufficient so that no etching residue is left. The parameter is set to an overetching amount of +50% or +100% relative to the reference etching amount.
[0080] 4 shows a flow chart of the drift correction of the present invention, which is a flow chart of the drift correction of FIG.
[0081] In Fig. 4, S1 acquires an SEM image and performs correlation calculations such as pattern matching with a registered reference pattern to calculate (measure) the amount of drift. This involves selecting (selecting on design data not shown) a reference pattern 41 in Fig. 3(a) from the vicinity of the position where the repair pattern 43 is to be created (within a range within one screen that can be scanned with an electron beam) and acquiring an SEM image (secondary electron image) of the selected reference pattern 41 and measuring the amount of drift (nm / sec, nm / min) of the reference pattern 41. In the case of Fig. 8 described later, instead of selecting the reference pattern 41, a new reference pattern 511 is created and the amount of drift is measured.
[0082] S2 introduces reactive gas after S1 selects reference pattern 41 (a) in FIG. 3, measures the drift of reference pattern 41, and calculates the repair area taking into account the offset of the drift. The reactive gas is sprayed from nozzle 10 (b) in FIG. 3 on photomask 12.
[0083] S3 irradiates the electron beam to a position at a distance L (vector quantity) from the center position of the reference pattern. This irradiates the electron beam to a position at a distance L (vector quantity) from the center of the reference pattern 41 in FIG. 3B, that is, irradiates the electron beam to the starting point of the upper left corner of the repair pattern 43 in FIG. 3B. This vector means that the coordinates of the electron beam irradiation point are expressed as a function of the reference position coordinate. Therefore, the vector means a function expressing the drift. The functional form expressing the vector can be given as a linear function, quadratic function, Nth order function, periodic function, or a function of irradiation time. Similarly, in addition to a simple mathematical function, an inference function obtained by machine learning the drift transition in the past can also be adopted as a function of the vector.
[0084] In S4, the electron beam is scanned from the start point to the end point of the repair pattern 43, and further scanned from the start point to the end point of the Nth stage of the repair pattern 43. This is repeated. In this manner, one or more scans are sequentially performed in the scanning direction on the repair pattern 43 from (b) to (c) and then (d) in FIG. 3, the entire repair pattern 43 is scanned with the electron beam, and gas is sprayed to repair the repair target pattern 43. This repair is performed until it is almost completed, or until it is slightly less than completed. Then, in the procedures from S5 to S7 described later, the repair pattern (secondary electron image) is compared with the target repair pattern in the design data, and the procedure is repeated until it is determined that the drift correction (repair) is completed when it is within the allowable range.
[0085] In step S5, the reactive gas is removed and an inactivation process is performed by stopping the injection of the active gas from the nozzle 10, evacuating the reactive gas, and injecting an inactivation gas to perform an inactivation process, followed by evacuation.
[0086] In S6, an SEM image and a BS image are acquired. This is the state after the processing of S5 in Fig. 3(d) is performed, and an SEM image (secondary electron image) and a BS image (backscattered electron image) of the repair pattern 43 in Fig. 3(d) are acquired. Note that in this state of S6 (after the processing of S5), even if the electron beam is irradiated onto either the repair pattern 43 or the reference pattern 41 to acquire the secondary electron image and the backscattered electron image, no etching or deposition is performed.
[0087] In the case of FIG. 8 described later (when creating a new reference pattern 511), if it does not matter whether the reference pattern 511 is present on the photomask 12 or not, it is possible to scan the reference pattern 511 with an electron beam and obtain the status of the reference pattern 511 (information on drift, etc.) in real time even during the repair process by injecting gas, without performing the process of S5 (removal of reactive gas, inactivation process) (the reference pattern 511 is also etched or deposited, as a matter of course).
[0088] In S7, it is determined whether the drift correction is completed. This is done by comparing the secondary electron image of the repair pattern 43 in FIG. 3(d) acquired in S6 with the secondary electron image of the repair pattern 43 targeted by the design data, and determining whether the former is within a predetermined allowable range of the latter and the drift correction has been performed normally. Alternatively, in the BS image of the repair pattern 43, a BS image with a brightness equivalent to that of the quartz underlying the photomask is obtained, and this is within a predetermined allowable range, determining whether the drift correction has been performed normally. If the answer is YES, it is determined that the drift correction has been performed and the repair pattern 43 has been repaired within the allowable range, and therefore the repair with drift correction of the present invention is terminated. On the other hand, if the answer is NO, the process returns to S2 and is repeated.
[0089] 3(a) is selected from the photomask 12, and the repair pattern 43 is sequentially scanned with an electron beam while gas is injected from the reference pattern 41 to the distance L, and the repair is completed when the secondary electron image of the repair pattern 43 is within a predetermined allowable range, or when the reflected electron image of the repair pattern 43 is the luminance of the background of the photomask 12 and within a predetermined allowable range. As a result, even if the photomask 12 drifts, the repair pattern 43 is repaired by scanning it in the scanning direction in sequence from the reference pattern 41 above it, so that the drift of the photomask 12 can be canceled and the drift of the repair pattern 43 can be reduced to zero.
[0090] 5 shows an example of a table for drift correction according to the present invention, which is necessary when performing the drift correction shown in FIG. 3 according to the procedure shown in the flow chart of FIG.
[0091] Fig. 5(a) shows an example of the reference pattern table 51. For example, the reference pattern 41 in Fig. 3 has its center coordinates (X, Y) registered in association with the pattern No. By referring to this reference pattern table 51, the center coordinates (X, Y) can be known.
[0092] 5(b) shows an example of the pattern table 52. This is data representing patterns (standard patterns, repair patterns), and here, is expressed (described) using one or more of a bitmap, opposing coordinate data, and contour function.
[0093] FIG. 5(c) shows an example of the repair pattern table 53. For example, the repair pattern 43 in FIG. 3 represents a figure to be repaired, such as a contour function, in association with a pattern No. By referring to this repair pattern table 53, it is possible to extract the repair pattern 53 to be repaired, and repair the repair pattern 53 by sequentially scanning the inside of the repair pattern 53 with an electron beam and injecting gas. At this time, if there is drift in the reference pattern 41 in FIG. 3 or the like, the drift is cancelled, that is, the repair pattern 53 is cancelled by enlarging or reducing it in the direction opposite to the drift direction, and the inside of the repair pattern 53 after enlarging or reducing (referred to as a repair area) is sequentially scanned with an electron beam while injecting gas to repair it, so that the target repair pattern 53 can be accurately repaired even if there is drift in the reference pattern 41.
[0094] FIG. 6 shows an example of a representation of the repair pattern of the present invention.
[0095] Fig. 6(a) shows a schematic diagram of a bitmap (binary image). This shows an example of a bitmap to be registered in the pattern table 52 of Fig. 5(b) described above. A bitmap is a representation of a figure as shown in the figure, in which the inside is 1 and the outside is 0, for example.
[0096] Fig. 6(b) shows a schematic representation of opposing coordinate data. This shows an example of opposing coordinate data to be registered in the pattern table 52 of Fig. 5(b) described above. The opposing coordinate data is expressed as a pair of a start point and an end point in the length direction for the inside of a figure as shown in the figure.
[0097] Fig. 6(c) shows a schematic diagram of a contour function. This shows an example of the contour function to be registered in the pattern table 52 of Fig. 5(b) described above. The contour function is a vector representation of the contour of a figure as shown in the diagram.
[0098] FIG. 7 shows an example of the brightness of the BS image of the present invention. This is a schematic representation of the brightness of the BS image (backscattered electron image) acquired in the flow chart of FIG. 4 described above. Here, the horizontal axis represents the atomic number, and the vertical axis represents the brightness of the backscattered electrons. The brightness (brightness) of the backscattered electrons emitted by irradiating an electron beam has the property of increasing as the atomic number increases. As a result, in the case of the photomask 12, when the base is quartz (Si) and, for example, a Cr pattern is formed thereon, when the Cr pattern is removed by etching and the base Si appears, the brightness (brightness) decreases, so that it becomes possible to observe the BS image and determine whether etching has been performed sufficiently (see S6 and S7 in FIG. 4). Automatic determination of the end point can be achieved by setting an appropriate threshold value for the brightness.
[0099] Fig. 8 shows an explanation of the present invention (drift correction method 2). In Fig. 8, a reference pattern 511 is newly created in a patternless portion on the photomask 12, and a repair pattern 531 to be repaired is scanned with a primary electron beam using this reference pattern 511 as a base point while gas is sprayed from the nozzle 10 to perform repair.
[0100] 8(a) shows a state where a new unique reference pattern 511 is created (a convex reference pattern 511 is created by deposition, or a deep hole-like reference pattern 511 is created by etching) in an empty portion (a portion without a pattern) on the photomask 12. As the base point of the reference pattern 511, the intersection point of the maximum and minimum in the X and Y directions is determined as the base point (Xb, Yb), for example (determined on design data not shown).
[0101] FIG. 8B shows a state in which scanning of the repair pattern 531 is started with the primary electron beam, starting from a position distant from the reference pattern 511 by L (vector).
[0102] FIG. 8C shows a state in which the primary electron beam is scanned over the repair pattern 531 starting from a position L (vector) away from the reference pattern 511, and the scanning has been completed about halfway.
[0103] 8(d) shows the state where the primary electron beam has been used to scan the repair pattern 531, starting from a position L (vector) away from the reference pattern 511. The end of the scan is determined when the secondary electron image of the repair pattern 531 matches the target repair pattern determined from the design data within an allowable range. Alternatively, in the case of etching, the end is determined when it is determined that the base (quartz) of the photomask 12 is exposed and becomes a reflected image of the luminance of the base.
[0104] As described above, the reference pattern 511 is generated in an empty space on the photomask 12 and in the vicinity of the repair pattern to be repaired (within the range of one screen), and then, starting from a position of the distance L (vector) from this reference pattern 511, a specified repair pattern 531 is scanned in sequence with the primary electron beam while injecting gas to repair the repair pattern 531. Even if the photomask 12 drifts, the primary electron beam is scanned in sequence in the scanning direction from the reference pattern 511 formed thereon, resulting in a remarkable effect that the influence of the drift is nullified and it is possible to confirm that the repaired repair pattern 531 is within a specified tolerance range.
[0105] The details are explained below.
[0106] (1) In the drift correction method of FIG. 8, a reference pattern 511 that can be deformed by the repair process is prepared in advance near a pattern 531 to be repaired by deposition and etching. A tall or very deep reference pattern 511 is desirable, and furthermore, a unique shape (star, ellipse, etc.) that is advantageous for pattern matching is desirable, with an area as small as possible so that it does not adversely affect the performance of the photomask 12 even if it remains on the photomask 12. The data is recorded in a reference pattern shape database in advance. The reference pattern 511 is observed as the process progresses. Since the area observed with the SEM is also processed, the minimum area that includes the reference pattern 511 for drift correction is observed with the SEM. At the same time, the repair pattern 531 is also irradiated with the electron beam. An image in which the SEM image and the backscattered electron image obtained from the repaired area during repair are superimposed on the repair SEM image is displayed in real time. The time for irradiating the electron beam, the pixel speed, etc. are independently performed for each area. In FIG. 8, the reference pattern 511 is unnecessary for the photomask 12 and can be omitted, so there is no need to stop or purge the reaction gas in order to measure the drift amount, and therefore repair can be achieved more quickly.
[0107] FIG. 9 shows an example of a photomask arrangement and a nozzle arrangement according to the present invention.
[0108] 9(a) shows an example of the arrangement of a photomask. As shown in the figure, the photomask 12 is fixed by being pressed against the fixed end 62 at the upper right end of the pallet 61. For this reason, drift occurs mainly due to thermal expansion or thermal contraction of the members including the photomask 12. For example, if the photomask 12 is a plate made of quartz with one side measuring 15 cm, it has a thermal expansion coefficient of about 0.6 ppm, so that a temperature change of ±1 degree causes the length of the photomask 12 to expand or contract by a maximum of about ±90 nm, which is the amount of drift.
[0109] The size of the pattern on the photomask 12 for manufacturing the most advanced semiconductor devices is smaller than 100 nm. If the photomask drifts by as much as 90 nm during repair, there is a risk that the photomask may go outside the range during repair even if drift correction is performed using Figures 1 to 8.
[0110] Therefore, it is necessary to reduce the drift of the photomask 12 to within a range in which the drift correction of Figures 1 to 8 works well. Since the repair of the photomask takes place over several minutes to several tens of minutes, it is necessary to strictly control the photomask temperature so that the drift is less than several nm per minute of photomask repair time, for example, so that the drift correction of Figures 1 to 8 works effectively.
[0111] 9B shows an example of the arrangement of nozzles for supplying reactive gases. As shown in the figure, the nozzle 10 has a temperature of 100° C. in this example and is arranged above the repair site on the photomask 12.
[0112] Here, there are various factors that change the temperature of the photomask 63, but the heat source that causes a drift large enough to affect the drift correction in Figs. 1 to 8 is relatively clear in the case of the electron beam repair device in Fig. 1. The electron beam repair device has a gas supply nozzle 10 at a position several mm above the photomask 63. The reactive gas is introduced to the surface of the photomask 63 using the nozzle 10 while being heated to nearly 100 degrees. Since the periphery of the nozzle 10 is a vacuum and insulated, the nozzle 10 is also heated to a temperature of nearly 100 degrees. The amount of gas injected from the nozzle 10 is minute and the total amount of heat is negligibly small. In addition, since there is no gas in a vacuum, there is no heat conduction by convection, and since the nozzle 10 does not directly contact the photomask 63, there is no heating by conduction, but the photomask 63 is heated by radiant heat from the heated nozzle 10. The photomask 63 is supported by point contact from the stage so as not to distort the photomask 63, and since the periphery is in a state close to vacuum insulation, heat conduction is very poor. If the photomask 63 and other components continue to be heated due to radiant heat or a sudden temperature change that occurs when the photomask 63 is transferred to the sample chamber 21, thermal expansion occurs, causing drift on the order of microns in the worst case.
[0113] Radiant heat is an electromagnetic wave, so it can be blocked by materials that do not transmit electromagnetic waves. For example, the thin gold film seen on artificial satellites has an emissivity of 0.01, which means it can block 99% of infrared rays. For example, since radiant heat is proportional to the area, if a gold film shielding plate is inserted between the nozzle 10 and the photomask 63, except for the vicinity of the nozzle's ejection hole, the exposed radiation area of the nozzle 10 facing the photomask 63 will be 1 square centimeter or less, making it possible to block most of the radiant heat that falls on the entire photomask 63.
[0114] 9(b), if the thermal radiation radiated from the nozzle 10 to the photomask 63 is designed to be symmetrical with respect to the electron beam irradiation axis, the temperature distribution of the photomask 63 near the electron beam irradiation point can be made symmetrical (concentric). The symmetrically heated photomask 63 does not expand or contract in one direction, and only internal stress is generated, resulting in extremely small drift.
[0115] When the above is done, drift can in principle be reduced to zero (small) in the portion of the photomask 63 facing the nozzle 10. However, as time passes, the radiant heat received by the entire photomask 63 is not zero, so the average temperature of the photomask 63 rises slowly. The rate of thermal expansion of the photomask 63 becomes extremely small but not zero, so drift occurs in the direction shown in Figure 9(a) in proportion to the overall average temperature rise of the photomask 63. To completely eliminate the cause of drift, it is desirable to keep the temperature of the entire photomask 63 constant.
[0116] 10 shows an example (part 1) of photomask temperature control according to the present invention. This shows an example in which a water-cooling head / radiant heat shielding plate 74 is provided above the pattern of a photomask 77 to block radiant heat from above, and a water-cooling head is used to control the temperature to a constant level from below a pallet 78 on which the photomask 77 is fixed.
[0117] In FIG. 10, a gas supply system 73 supplies gas (usually at about 100° C.) to be sprayed onto a photomask 77 from a nozzle.
[0118] The water-cooled head / radiant heat shielding plate 74 is a water-cooled plate that blocks the radiant heat radiated from the piping of the gas supply system 73 toward the photomask 77, and is a flat plate with holes at the center for injecting gas and passing the electron beam, and also for allowing the secondary electrons emitted from the photomask 77 to travel upward.
[0119] The water-cooling tube 75 is a thin tube for water cooling attached to the water-cooling head / radiation heat shield plate 74 .
[0120] The gold plating film 76 is a gold plating film that blocks radiant heat emitted from the gas supply system 73.
[0121] The water cooling head 71 is attached to the underside of the pallet 78 and controls the temperature of the pallet 78 to a constant value. The set temperature and current temperature of the pallet or photomask, as well as control parameters, are shown on the display.
[0122] The water-cooled pipe (or heat pipe) 72 supplies cooling water to the water-cooled head 71 to adjust the temperature of the moving body on the stage. These water-cooled pipes must be flexible and have low gas permeability, and can be made of bellows-shaped metal pipes, flexible Teflon (registered trademark) resin pipes, composite pipes mainly made of polyimide resin, etc. Heat pipes, for example, have a structure in which a bellows-shaped metal pipe that emits little gas in a vacuum, flexible Teflon resin, or composite pipes mainly made of polyimide resin are sealed with water, alcohol, naphthalene, etc. as a working fluid in a linear or mesh-shaped tube wall material made of an inorganic or organic material that causes a capillary phenomenon called a wick in the bellows-shaped metal pipe, which emits little gas in a vacuum, flexible Teflon resin, or composite pipe mainly made of polyimide resin.
[0123] The non-contact temperature sensor 83 measures the temperature of the pallet 78 (photomask 77) in real time without contacting it, and mainly uses a highly sensitive MEMS thermoelectric element capable of receiving infrared rays.
[0124] Next, the temperature control of the photomask 77 will be described. (1) A water cooling head / heat radiation shielding plate 74 is provided opposite the upper surface of the photomask 77 having the pattern, to block heat radiation from the gas supply system 73. (2) The bottom surface of the photomask 77 is fixed to a pallet 78, and the pallet 78 is controlled to a constant temperature by a water-cooling head 71. The underside of the pallet is made black so that it radiates more heat toward the underside of the chamber. For example, it is desirable to cover it with carbon, ceramics, or platinum black to make the radiant heat to infrared rays close to 100% so that the heat accumulated on the pallet is naturally released to the outside. In this way, the photomask is naturally cooled, so the temperature of the photomask is stabilized.
[0125] As described above, the water-cooled head / heat radiation shield 74 is used on the upper surface of the photomask 77 where the pattern is located to block radiant heat from above, and the water-cooled head 71 is used to control the temperature at a constant level from the underside of the pallet 78 on which the photomask 77 is placed, thereby minimizing temperature changes in the photomask 77 and reducing thermal drift. The temperature is usually kept close to room temperature, but it can also be cooled or heated to a negative temperature. If an abnormality is detected in the temperature control, a warning signal is generated, a warning is displayed on the display, and interlock measures can be taken to automatically stop the photomask repair operation, etc.
[0126] 11 shows a second example of temperature control according to the present invention. This shows an example in which a radiant heat shielding plate 741 is provided above the pattern of a photomask 77 to block radiant heat from above, and a Peltier element 831 is used to control the temperature to a constant level from the underside of a pallet 77 to which the photomask 77 is fixed.
[0127] In FIG. 11, a gas supply system 73 supplies gas (usually at about 100° C.) to be sprayed onto a photomask 77 from a nozzle.
[0128] The radiant heat shielding plate 741 is a plate that blocks the radiant heat radiated from the piping of the gas supply system 73 toward the photomask 77, and is a flat plate with a hole at the center for injecting gas and passing the electron beam, and further for allowing the secondary electrons emitted from the photomask 77 to travel upward.
[0129] The Peltier element 831 is a known device that is placed between the pallet 78 and, in this embodiment, the Z stage 82, and sends heat from the pallet 78 to the Z stage 82 by passing an electric current through it. Since it is sufficient to dissipate heat generated by the photomask to a location other than the photomask, it is sufficient that one end is connected to the photomask side and the other end is connected to the outside, so the Peltier element may be placed between the XY stage and the Z stage. When using water cooling at the same time as in the previous embodiment, if a water cooling head is connected to the heat dissipation section of the Peltier element, heat from the photomask can be dissipated to the outside of the sample chamber.
[0130] The non-contact temperature sensor 83 measures the temperature of the pallet 78 (photomask 77) in real time without contact.
[0131] Next, the temperature control of the photomask 77 will be described. (1) A heat radiation shielding plate 741 is provided opposite the upper surface of the photomask 77 having the pattern thereon to block heat radiation from the gas supply system 73 . (2) The underside of the photomask 77 is fixed to a pallet 78, and the pallet 78 is controlled to a constant temperature by passing a current through a Peltier element 831. The current temperature, target temperature, or control value of the Peltier element can be displayed as necessary. If an abnormality occurs, the repair operation can be automatically stopped and a warning can be displayed on the display.
[0132] As described above, by blocking heat radiation from the upper surface on which the pattern of photomask 77 is formed by the radiant heat shielding plate 74, and controlling the temperature to a constant value by passing a current through the Peltier element 781 from the lower surface of the pallet 78 on which the photomask 77 is placed, it is possible to minimize temperature change in the photomask 77 and reduce thermal drift.
[0133] 12 shows a third example of temperature control according to the present invention. This shows an example in which a radiant heat shielding plate 741 is provided above the pattern of a photomask 77 to block radiant heat from above, and a local temperature control gas 731 is blown onto the photomask 77 from above to control the temperature of the photomask 77 to a constant temperature.
[0134] In FIG. 12, local temperature control gas 731 is a gas that is sprayed onto photomask 77 to control the temperature. The temperature control gas is supplied using a gas temperature control device (not shown) that is separately installed outside the correction device. This device is equipped with a heat exchanger and can control the temperature of the temperature control gas to the desired temperature. The control accuracy is about ±0.1°C. The temperature control gas can be irradiated continuously or intermittently. Through these controls, the temperature of the photomask is accurately controlled to about ±0.01°C.
[0135] Next, the temperature control of the photomask 77 will be described. (1) Figure 12 shows a method of directly blowing temperature-controlled gas onto a photomask 77 in order to control the temperature of the photomask 77 without contact. The gas can be blown onto either the front or back side of the photomask. However, when blowing gas onto the front side, it is necessary to take measures to ensure that the gas does not mix directly with or affect the reactive gas. (2) For example, the position where the reactive gas is sprayed onto the photomask 77 is on the electron beam axis, but the temperature control gas 731 is sprayed onto the photomask 77 at a position slightly away from the electron beam axis using another nozzle. Alternatively, it is possible to detect that the repair process is in an idling state, temporarily stop the reactive gas, and spray the temperature control gas 731 onto the photomask 77 during that time. In this case, the gas 731 can be sprayed onto the electron beam axis. As a nozzle for irradiating the temperature control gas, a separate nozzle dedicated to temperature control can be used, or one nozzle may be switched to spray the reactive gas and the temperature control gas 731. It is desirable that the nozzle for the local temperature adjustment gas 731 is arranged symmetrically with respect to the electron beam axis so that it can be cooled symmetrically with respect to the electron beam axis.
[0136] Fig. 13 shows a fourth example of temperature control according to the present invention. In Fig. 13, instead of blowing local temperature control gas 731 from above the pattern of photomask 77 as in Fig. 12, local temperature control gas 732 is blown from the underside of pallet 78 holding photomask 77.
[0137] Next, the temperature control of the photomask 77 will be described. (1) In FIG. 13, local temperature adjustment gas 732 is sprayed from the back side of photomask 77. Specifically, a nozzle for local temperature adjustment gas 732 is provided on the back side of photomask 77, facing the reactive gas nozzle on the front side of photomask 77. This is a method of controlling the temperature by spraying temperature-controlled gas from this nozzle directly onto the back side of photomask 77. In this arrangement, the heat-generating portion does not move even if XY stage 84 moves. As a result, if the heat-generating portion of photomask 77 to which reactive gas is sprayed generates a certain amount of heat, the temperature of photomask 77 can be maintained at a constant temperature by setting local temperature adjustment gas 732 to just take away that heat (setting local temperature adjustment gas 73 to take out heat from the back side of the heat-generating portion heated by reactive gas from the front side). (2) Here, the sample chamber 21 is evacuated to a pressure of 10-3 Pascals or less by a powerful pump. The photomask 77 acts as a partition because the gas is sprayed on the back side of the photomask 77. The local temperature control gas 732 is exhausted to the outside before it reaches the surface of the photomask 77, so it does not mix with the reaction gas used on the surface of the photomask 77. Unlike the reaction gas, there is no limit to the flow rate, so while only a very small amount of reaction gas is used, such as a few SCCM, the local temperature control gas 732 can be injected in a large amount of gas, such as 100 SCCM or more. Hydrogen and helium, which have excellent thermal conductivity, are excellent in principle as the local temperature control gas 732, but they are rare and expensive. The photomask 77 does not generate as much heat as is used in semiconductor wafer processing, so gases such as purified dry air, oxygen, and nitrogen, which do not have as good thermal conductivity but are inexpensive and can be flowed in large quantities, can be used. 1, the gas temperature is precisely controlled to within 0.1 degrees and then sprayed onto the photomask 77 to control the temperature. The temperature of the photomask 77 is measured using a non-contact temperature sensor 83 that uses an infrared sensor, and is controlled so that the surface of the photomask 77 reaches the desired temperature.
[0138] If photomask drift is caused by thermal expansion, in addition to keeping the photomask temperature constant, it is also possible to predict and correct the thermal expansion by measuring the photomask temperature. It is difficult to keep the photomask temperature constant precisely to the 0.01 degree level, but it is easy to measure the temperature precisely to the 0.01 degree level. 14 is a diagram illustrating the relationship between the photomask temperature (T) and the drift amount (D=kT) of the present invention. The horizontal axis of Fig. 14 represents the photomask temperature (T / °C), and the vertical axis represents the drift amount (D / mm).
[0139] In FIG. 14, the line before correction represents the drift amount D when no thermal expansion correction of the photomask 12 is performed.
[0140] The post-correction line segment diagrammatically represents the state in which the drift amount D becomes nearly zero when the thermal expansion correction of the photomask 12 is performed.
[0141] Here, the drift of the position of the photomask 77 mounted on the XY stage 84 is caused by the thermal expansion or thermal contraction of the photomask itself and the member supporting the photomask 77. A laser interferometer is used to measure the position of the photomask, and in order to align the position measurement of the photomask with the XY bar mirror position, the photomask is often fixed at one point on the end face close to the XY bar mirror. The amount of thermal expansion of the photomask 77 is expressed as a function of the average temperature of the photomask when the supporting end of the photomask is fixed at one point (see FIG. 15). As shown by the line before correction in FIG. 14, the drift amount D of a specific position increases in proportion to the photomask temperature T multiplied by the coefficient K. The drift occurs when the center position of the electron beam irradiation is offset by a certain amount from the reference position (fixed ends 79, 62 in FIG. 15 and FIG. 9). Therefore, if the data shown by the line before correction in FIG. 13 is obtained, the offset amount of the repair target position existing at an arbitrary temperature and at an arbitrary location on the photomask can be calculated using the average temperature of the photomask 77 and the coordinates of the repair target position. Then, by sending a signal to the electron beam deflection control device (or moving the XYZ stage 12) so that the calculated offset becomes zero, it is possible to correct the thermal drift as shown by the post-correction line in Figure 14. This correction can be used in conjunction with the drift correction using the reference pattern described above to achieve even greater effectiveness. In this embodiment, for simplicity, drift correction is performed using the average value of the photomask temperature, but correction may also be performed by accurately measuring the temperature distribution of the photomask and calculating the respective thermal expansions to calculate the amount of drift at each repair point.
[0142] FIG. 15 shows an example of temperature measurement according to the present invention.
[0143] FIG. 15(a) shows an example of measuring the temperature by inserting a temperature sensor 832 into the contact portion between the pallet 78 and the Z stage 82. In this example, the average temperature of the photomask 77 is measured by the temperature sensor 832 provided on the photomask support member, and one end of the Z stage 82 is in direct contact with a transport jig called the pallet 78 for transporting the photomask 77. The temperature of the photomask 77 can be measured by providing the temperature sensor 832 at the place where the Z stage 82 contacts the pallet 78. It is desirable to measure the temperature at all the installation points of the Z stage 82 and obtain the average value so that the average temperature of the photomask 77 can be measured. The contact point between the Z stage 82 and the pallet 78 is the only place where the heat accumulated in the photomask 77 flows to the outside by heat transfer. The amount of temperature change of the photomask 77 can be determined by increasing or decreasing the thermal conductivity of the contact point to control the inflow or outflow of heat from the outside.
[0144] 15(b) shows an example in which a non-contact temperature sensor 83 is used to measure the temperature of the underside of the pallet 78. In this example, the average temperature of the photomask 77 is measured in a non-contact manner using a non-contact temperature sensor 83 such as an infrared sensor. Temperatures at multiple locations are measured and averaged so that the average temperature of the photomask 77 can be measured. Measurement from the backside of the photomask 77 can be achieved by arranging multiple measurement devices on the XY stage 84. [Brief description of the drawings]
[0145] [Figure 1] FIG. 1 is a configuration diagram of one embodiment of the present invention. [Diagram 2] FIG. 2 is an explanatory diagram (drift explanatory diagram) of the present invention. [Diagram 3] FIG. 1 is an explanatory diagram of the present invention (drift correction method 1). [Figure 4] 4 is a drift correction flowchart of the present invention. [Diagram 5] 4 is an example of a table for drift correction according to the present invention. [Figure 6] 1 is an example of a representation of a repair pattern of the present invention. [Figure 7]13 is a diagram showing an example of brightness of a BS image according to the present invention. [Figure 8] FIG. 2 is an explanatory diagram of the present invention (drift correction method 2). [Figure 9] 2 shows an example of a photomask arrangement and a nozzle arrangement according to the present invention. [Figure 10] This is the temperature control property (part 1) of the present invention. [Figure 11] 2 is a second example of temperature control according to the present invention. [Figure 12] This is a third example of temperature control according to the present invention. [Figure 13] This is the fourth temperature control example of the present invention. [Figure 14] FIG. 2 is a graph showing the relationship between the photomask temperature (T) and the amount of drift (D=kT) according to the present invention. [Figure 15] 1 is a temperature measurement example of the present invention. [Explanation of symbols]
[0146] 1: Column 2: Electron gun 3: Blanking electrode 4: U-MCP 5:D-MCP 6: Deflector 7: Objective lens 8: Differential pumping device 9: Radiation heat shield 10: Nozzle 11: Z sensor 12: Photomask 13: Heating and cooling controller 14: Support stand 15:Temperature sensor 16: XYZ Stage 17:TMP 18: Dry pump 19: Controller 20: Vibration isolator 21: Sample chamber 22: Gas temperature regulator 23: Gas controller 31: High voltage power supply 32:BL Controller 33: Reflection signal processing device 34: Secondary signal processor 35: Deflection control device 36: Stage position control device 37: PC 38:Display device 41: Reference pattern 42, 43: Repair pattern 421: SEM image acquisition range 50: Photomask 51: Reference pattern table 511: Reference pattern 52: Pattern table 53: Repair pattern table 531: Repair pattern 61: Palette 62: Fixed end 63: Photomask 631: Repaired area 632: Nozzle temperature distribution 64: Bar mirror 71: Water-cooled head 72: Water cooling pipe (or heat pipe) 73: Gas supply system 731: Gas for local temperature control 732: Gas required for local temperature control 74: Water-cooled head and radiation heat shield 741:Radiant heat shielding board 75:Water cooled pipe 76: Gold plating 77: Photomask 78: Palette 79: Fixed end 80: Platinum Black 81: Heat flow 82: Z Stage 83: Non-contact temperature sensor 831: Peltier element 832:Temperature sensor 84: XY Stage
Claims
1. A drift reduction apparatus for reducing drift of a repair target on a photomask, comprising: A reference pattern selected or newly created from the vicinity of the repair target on the photomask; a repair area setting means for setting a repair area to be repaired by injecting an etching or deposition gas from a nozzle while irradiating an electron beam onto a portion of the repair pattern to be repaired, based on the selected or created reference pattern; a repair means for repairing the repair area by injecting an etching or deposition gas from a nozzle while irradiating the repair area with an electron beam; A drift reduction device comprising:
2. 2. The drift reduction device according to claim 1, wherein said restoration area setting means measures an amount of drift of said reference pattern, corrects said restoration area by the measured amount of drift, and reduces the influence of said drift.
3. 3. The drift reduction device according to claim 1, wherein the restoration means performs one or more scans starting from one of the restoration areas to restore the entire restoration area.
4. 4. The drift reduction device according to claim 1, wherein the repair means acquires a secondary electron image of the reference pattern, compares the acquired secondary electron image of the pattern to be repaired with the repair pattern to be repaired based on the acquired secondary electron image of the reference pattern, and terminates the repair when the difference is within a predetermined tolerance, and outputs a message to that effect or causes the process to be retried when the difference is outside the predetermined tolerance.
5. 5. The drift reduction device according to claim 1, wherein the reference pattern is within a range that can be acquired within one screen together with the repair pattern.
6. 6. The drift reduction device according to claim 1, further comprising a flat plate-like radiation heat shielding plate disposed adjacent to and facing the region to be repaired on the photomask irradiated by the electron beam, the plate having holes through which the electron beam passes when irradiated and through which the gas is sprayed from a nozzle onto the photomask, thereby reducing temperature drift of the photomask due to radiation heat.
7. 7. A drift reduction device for reducing temperature drift of a photomask by controlling the temperature of a lower surface of a flat pallet for holding the photomask according to claim 6 to a predetermined temperature.
8. 8. A drift reduction device for reducing temperature drift of a photomask according to claim 6, wherein a gas having a predetermined temperature is injected through said holes to control the temperature of the photomask, thereby reducing temperature drift of the photomask.
9. 1. A drift reduction method for reducing drift of a repair target on a photomask, comprising: A reference pattern is provided that is selected or newly created from the vicinity of the repair target on the photomask; a repair area setting step of setting a repair area to be repaired by irradiating an electron beam onto a portion of the repair pattern to be repaired while spraying an etching or deposition gas from a nozzle, based on the selected or created reference pattern; a repair step of repairing the repaired area by injecting an etching or deposition gas from a nozzle while irradiating the repaired area with an electron beam; A drift reduction method comprising the steps of:
10. 10. The drift reduction method according to claim 9, wherein the restoration area setting step measures an amount of drift of the reference pattern, corrects the restoration area by the measured amount of drift, and reduces an effect of the drift.
