JP2005189491A - Photomask defect correction method using transfer or light intensity simulation - Google Patents

Abstract

PROBLEM TO BE SOLVED: To determine the necessity of more accurate defect correction by light intensity (transfer) simulation, and to determine the necessary scanning area by light intensity simulation considering the limitations of defect correction method even when defect correction is necessary By obtaining this, the time required for fine adjustment of the beam irradiation area of the FIB device and the removal area by AFM scratch processing is shortened.
An AFM image of a defect is acquired, a light intensity simulation is performed based on this three-dimensional shape, transfer characteristics are estimated, and whether or not defect correction is necessary is determined. In the case of black defect correction, it is required for each correction part on the premise that the processed surface sagging and transmittance decrease during FIB etching, and in the case of white defect correction, the actual shape, halo, and shading rate are preconditions. An irradiation area that satisfies the optical characteristics is obtained from a light intensity simulation, and a black defect or a white defect is corrected by removing the FIB or forming a light shielding film.
[Selection] Figure 1

Description

  The present invention relates to a photomask defect correcting method.

  Due to the progress of miniaturization of semiconductor integrated circuits, demands required for the performance of a photomask serving as a transfer original plate are becoming stricter than ever. With regard to defect correction, it is more strongly demanded that the corrected portion satisfies the required optical characteristics than ever before.

  With improved computer performance, it is now possible to perform high-precision light intensity simulations and transfer simulations of photomasks, and along with this, the optical effects of defects can also be quantitatively analyzed. (For example, Non-Patent Document 1).

  When correcting a black defect with a focused ion beam (FIB), light is less than that without a defect due to a slight decrease in the transmittance of the quartz substrate due to the edge of the correction area and the implantation of Ga used as the ion beam. Strength decreases. For this reason, in order to increase the light intensity, the light intensity is increased by cutting slightly from the line of the normal pattern. The optimum value of the amount of shaving is obtained experimentally using a transfer simulation microscope. The transfer simulation microscope can obtain an image and light intensity distribution after transmitting light through the mask under the same conditions as when exposing an actual mask. That is, the black defect is cut out while observing with a transfer microscope, and the cutting is stopped when the optimum light intensity distribution is obtained.

  However, this takes time, and a desired optical performance may not be obtained with a real defect having a complicated shape.

  In addition, when correcting a white defect by FIB, the light intensity is reduced as compared with a defect-free area because of the edge of the FIB-CVD light shielding film and the halo component at the correction site. For this reason, in order to increase the light intensity, the light shielding film is formed by being slightly retracted from the normal pattern line. The optimum value of the amount of retraction is also experimentally obtained using the above-described transfer simulation microscope, but it takes time to determine.

  A proposal has been made to perform light intensity simulation using an image observed with an FIB apparatus to guarantee the presence / absence of defect correction and the quality of a defect correction portion (for example, Patent Document 2). This is because the FIB apparatus is provided with light intensity simulation means, and image data of a mask pattern including a target defect correction portion is input to the light intensity simulation means, and simulation calculation is performed based on the image data and exposure conditions. The exposure light intensity distribution at the time of exposure transfer is output.

Recently, black defects in photomasks and convex phase defects in Levenson masks have been removed by scratching with an atomic force microscope (AFM) probe that is harder than the workpiece.
CR Musil, DK Stewart, and RF Clark, Proceedings of SPIE 4889 1048-1055 (2002) JP 2000-347384 (Page 3-4)

  In the method using the light intensity simulation means, the image data taken by the FIB device input to the light intensity simulation means has no three-dimensional information on defects, and therefore no information on the edge or halo. . Further, the reduction in transmittance due to Ga implantation and the optical characteristics of the light shielding film component are not taken into consideration. For this reason, the exposure light intensity distribution obtained as a result of the light intensity simulation differs from the exposure light intensity distribution obtained with the transfer simulation microscope.

  In addition, in the method of scratching a defective portion with the above AFM, the processing shape depends on the shape of the AFM probe and the scanning conditions, and is not an ideal shape, and the transmittance of the removal region is slightly descend. For this reason, in order to obtain an appropriate light intensity, fine adjustment of the removal region was experimentally performed using a transfer simulation microscope, and it took time for the condition.

  The present invention solves the above-mentioned problems, makes it possible to determine the presence or absence of defect correction with higher accuracy by transfer or light intensity simulation, and the beam irradiation area at the time of defect correction using FIB and the removal area by AFM scratch processing. It is intended to shorten the time of fine adjustment and guarantee a higher quality of the corrected part.

In order to solve the above problems, in the photomask defect correction method of the present invention,
First, a three-dimensional shape of a photomask defect is obtained with an atomic force microscope (AFM), and based on this three-dimensional shape, transfer or light intensity simulation (hereinafter abbreviated as light intensity simulation) is performed to obtain the defect. It is characterized in that the exposure light intensity in a region including the above is estimated and the necessity of defect correction is determined by comparing the exposure light intensity with that of a normal mask pattern. That is, transfer or light intensity simulation is performed based on the above three-dimensional shape to estimate transfer characteristics to the wafer, and when the expected line width of the defective portion does not fall within the allowable value, the defect is corrected and enters the allowable value. If it is determined that the defect is not corrected.

  Secondly, when it is determined that the defect needs to be corrected in the above case, and this defect is a black defect, the three-dimensional shape at the time of correcting the focused ion beam defect at this black defect correction location, or damage caused by FIB irradiation Calculate the light intensity simulation in anticipation of the transmittance decrease based on the above, find the amount of cutting from the normal pattern edge line to the pattern side, correct the black defect with the focused ion beam with the amount of cutting described above, normal A light intensity distribution similar to that of a simple pattern portion is obtained.

  Third, when the defect is a white defect, light simulation calculation is performed by looking at the three-dimensional shape and light shielding ratio of the white defect correction film by the focused ion beam, and the white defect correction film from the edge line of the normal pattern The amount of retraction is obtained, white defects are corrected with a focused ion beam in consideration of the amount of retraction, and a light intensity distribution similar to that of a normal pattern portion is obtained.

  Fourthly, the defect is a black defect, and when this black defect is corrected with an atomic force microscope, the actual processed shape and transmittance after scratch processing caused by the probe shape of the atomic force microscope probe Calculate the light intensity simulation to see the decrease, find the amount of cutting from the edge line of the normal pattern to the pattern side, and correct the black defect by scratching the atomic force microscope probe with the amount of cutting described above It is characterized by performing.

  Light intensity simulation is performed by inputting an accurate 3D shape obtained by AFM, so light intensity simulation is performed using secondary electron or secondary ion images lacking height information of the conventional FIB. It is possible to perform a light intensity simulation with higher accuracy than that performed, and to determine whether or not defect correction with high accuracy is necessary.

  The time required for fine adjustment of the beam irradiation area, which was performed to obtain a high transmittance at the correction point with the conventional FIB, is shortened, and the necessary correction amount of the correction part is accurately determined and transferred only by light intensity simulation. The optical characteristics can be guaranteed.

  The time required for fine adjustment of the removal area of the scratch processing of the AFM probe can also be shortened, and the required correction amount of the correction portion can be obtained more accurately by only light intensity simulation, and the optical characteristics when transferred can be guaranteed.

  Defect correction is performed according to the procedure shown in FIG. Hard defects that can be corrected with a normal defect correction device are not a problem on the glass substrate, and the interface between the glass substrate and the light-shielding film is flat and only the light-shielding film pattern has excess or lack. Since the area can be determined, first acquire a faithful AFM image of the defect, perform transfer or light intensity simulation (hereinafter abbreviated as light intensity simulation) based on this three-dimensional shape to estimate the transfer characteristics to the wafer, and simulate The exposure light intensity distribution obtained in step 1 is compared with that of a normal pattern to determine whether or not defect correction is necessary.

  When it is determined that the defect correction is necessary, the defect correction is optimized in consideration of the deterioration of the optical characteristics of the defect correction portion. In other words, light intensity simulation calculation is performed by looking at the actual three-dimensional shape at the time of FIB defect correction at the black defect correction point and the decrease in transmittance based on damage due to FIB irradiation, and from the pattern edge line of the normal pattern to the pattern side The optimum cutting amount is obtained, and the black defect is corrected by the focused ion beam with the cutting amount obtained by the simulation.

  When the defect is a white defect, the normal pattern of the correction film is used to obtain the optimal light intensity distribution by calculating the light simulation considering the three-dimensional shape of the white defect correction film by the focused ion beam and the light blocking ratio. The amount of retraction from the edge line is obtained, and white defects are corrected with the focused ion beam in consideration of the amount of retraction.

  When the defect is a black defect and this black defect is corrected with an atomic force microscope, the actual machined shape and transmittance decrease after scratching due to the probe shape of the atomic force microscope probe are not observed. Then, the light intensity simulation calculation is performed to obtain the amount of cutting from the edge line of the normal pattern to the pattern side, and the black defect is corrected by scratching the atomic force microscope probe with this amount of cutting included.

  The operation procedure shown in FIG. 1 will be specifically described.

  First, a photomask having a defect is introduced into the AFM, and the stage is moved to a position where the defect is found by the defect inspection apparatus. An AFM image of a region containing defects is obtained with a probe having a small diameter and a high aspect ratio. A standard sample is measured in advance with the same probe, the influence of the probe shape is determined, and the obtained AFM image is deconvolved to obtain an AFM image that is more faithful to the defect shape. Based on this three-dimensional shape, light intensity simulation is performed. That is, using the acquired three-dimensional shape image data, the light intensity distribution when the mask pattern including the defective area is exposed is obtained by simulation calculation. This light intensity distribution is compared with that of a normal pattern to determine whether or not a defect needs to be corrected. In other words, light intensity simulation is performed based on the above three-dimensional shape to estimate the transfer characteristic of the defect to the wafer, and when the expected line width of the defective portion does not fall within the allowable value, the defect is corrected and enters the allowable value. When it is, it is determined that the defect is not corrected.

  When it is determined that defect correction is necessary, defect correction is performed as follows.

  The shape and optical characteristics associated with FIB defect correction, which are prerequisites for light intensity or transfer simulation, are obtained in advance. For this purpose, first, as shown in FIG. 2, the region 3 of 3 to 5 μm square in which the transmittance can be correctly obtained with a transfer simulation microscope is etched by FIB. Next, a three-dimensional shape image of the processing area is acquired by AFM with a probe having a diameter as small as possible and a high aspect ratio. In this case, in order to remove the influence of the shape of the probe, the difference between the shape measured by the probe and the actual shape is obtained in advance, the machining area is measured using the probe, and then the difference is obtained. Is corrected to obtain an accurate shape. Find exactly the slope of the edge of the edge corrected by etching with FIB in the above AFM. At the same time, the amount of decrease in the transmittance based on the damage caused by the FIB of the transmission part in the processing region 4 is obtained with a transfer simulation microscope. Similarly, as shown in FIG. 3, the light shielding film 6 is formed in the deposition region 5 of 3 to 5 μm □, which can be obtained with the transfer simulation microscope with FIB while supplying the light shielding film source gas, and etched. As in, the position and inclination of the shape and the thickness and width of the halo are determined accurately. At the same time, the relationship between the thickness and transmittance of the FIB-CVD light shielding film is obtained with a transfer simulation microscope.

  A photomask having a defect is introduced into the FIB apparatus, and the stage is moved to a position where the defect is found by the defect inspection apparatus. In the case of a black defect, the ion beam irradiation area was changed as a precondition for lowering the transmittance of the transmissive part of the actual corrected part obtained by the transfer simulation microscope and the edge of the actual processed part obtained by the AFM. In each case, the 3D finished shape of the etching is predicted, and the light intensity simulation of the area including the defect is performed with the predicted 3D shape, and the line width (the light intensity at which the resist is exposed) (Line width exceeding the threshold value) is calculated to obtain the correction amount (pattern cutting amount) 8 that will be the same line width as the normal pattern, and the black defect recognized by FIB as shown in Fig. 4 The area obtained by adding the correction amount (pattern cutting amount) 8 obtained by simulation to area 7 is the ion beam irradiation area necessary for black defect correction, and FIB-GAE (gas assist energy) is supplied while supplying FIB or assist gas. Remove and correct with (chipping). Thereby, the same light intensity distribution as that of a normal pattern portion can be obtained. In the case of white defect 9, the position and inclination of the shape of the white defect corrected portion obtained by AFM, the thickness and width of the halo, the film thickness and transmittance of the actual FIB-CVD light shielding film obtained by a transfer simulation microscope As a precondition, the final shape of the three-dimensional light-shielding film is predicted when the ion beam irradiation region is changed, and the light intensity simulation of the region including the defect is performed using the predicted three-dimensional shape. The amount of correction (recession from the normal pattern) is calculated by calculating the line width that will be transferred (the line width that exceeds the threshold value of the light intensity that the resist is exposed to) to the same line width as the normal pattern. As shown in Fig. 5, the area that takes into account the correction amount of the light-shielding film (retraction amount from the normal pattern) 10 calculated in the simulation to the white defect area recognized by the FIB as shown in FIG. Beam irradiation area Depositing a light shielding film 6 by FIB while supplying to the light-shielding film material gas.

  Even when the black defect or the Levenson mask convex phase defect is corrected by scratching the AFM probe, the area 3 of 3 to 5 μm □ is cut by scratching with the AFM probe harder than the material to be processed in advance. A three-dimensional shape image of the processed region is acquired by AFM with a probe having a diameter as small as possible and a high aspect ratio. In this case, in order to remove the influence of the shape of the probe, the difference between the shape measured by the probe and the actual shape is obtained in advance, the machining area is measured using the probe, and then the difference is obtained. Is corrected to obtain an accurate shape. Using the AFM, the inclination of the edge of the edge corrected by scratching with the AFM is accurately obtained. At the same time, the amount of decrease in the transmittance of the processed region is obtained with a transfer simulation microscope. After that, a photomask having a defect is introduced into the AFM, and the stage is moved to a position where the defect is found by the defect inspection apparatus. In the case of a black defect, obtain the amount of light transmission required for the defect correction part on the assumption that the edge of the actual edge obtained by AFM and the transmittance reduction of the transmission part of the actual correction part obtained by the transfer simulation microscope are preconditions. The amount of correction for the probe scanning range (pattern cutting amount) is calculated from the light intensity simulation using the correction amount as a parameter, and the correction amount (pattern cutting) calculated for the black defect area 7 recognized by the AFM is calculated. The area to which 8) is added is made the scanning range necessary for black defect correction, and it is removed and corrected by scratching with an AFM probe harder than the material to be processed. 6A and 6B are schematic cross-sectional views of the shape before and after the correction of the convex phase defect of the Levenson mask. In FIG. 6A, the light shielding pattern 1 is etched with an undercut, and phase defects 11 (portions that can be confirmed by AFM) and 12 (a convex phase under the undercut collar) are formed in the etched portion. Defects). As shown in FIG. 6B, it is assumed that the edge defect of the actual machining location obtained by AFM and the transmittance reduction of the transmission portion of the actual correction location obtained by the transfer simulation microscope are applied to the phase defect portion. As a condition, the 3D finished shape when the scanning area of scratch processing is changed is predicted, and the light intensity simulation of the area including the defect is performed with the predicted 3D shape, and the defective part is transferred. The line width (line width exceeding the threshold value of the light intensity to which the resist is exposed) is calculated, and a correction amount (Cr wrinkle cut amount) 13 is obtained so as to obtain the same line width as a normal pattern. Scratching with an AFM probe that is harder than the material to be processed, with the area 14 added to the defect area 12 under the Cr arm obtained in the simulation added to the defect area 11 recognized in step 1 as the necessary range for defect correction Remove and fix.

5 is a correction flowchart that best represents the features of the present invention. It is a figure explaining the shape of a black defect correction area | region, and the transmittance | permeability fall part by FIB. It is a figure explaining the shape and halo part of a white defect correction area | region by FIB. It is explanatory drawing in the case of correcting a black defect by FIB. It is explanatory drawing in the case of correcting a white defect by FIB. It is explanatory drawing in the case of correcting the black defect or the convex phase defect of the Levenson mask by AFM scratch processing.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Light-shielding pattern 2 Glass (quartz) board | substrate 3 FIB etching area | region 4 Ion beam damage area | region 5 FIB deposition area | region 6 FIB-CVD light shielding film 7 Black defect area | region 8 Correction area | region 9 at the time of black defect etching obtained from simulation 9 White defect area | region 10 Correction area 11 of white defect FIB-CVD film obtained from simulation Convex phase defect 12 confirmed by AFM Convex phase defect 13 under the undercut collar 13 Undercut structure Cr defect obtained from the simulation Cutting area 14 Area for removing convex phase defects

Claims (5)

  Obtain the three-dimensional shape of the mask pattern defect with an atomic force microscope, perform transfer or light intensity simulation based on this three-dimensional shape to estimate the exposure light intensity of the mask pattern region including the defect, A photomask defect correction method using transfer or light intensity simulation, wherein the exposure light intensity is compared with that of a normal mask pattern to determine whether or not there is a transfer required for defect correction.   When it is determined that the defect needs to be corrected in determining whether the defect needs to be corrected, when the defect is a black defect, the three-dimensional shape at the time of correcting the focused ion beam defect at this black defect correction location, Perform transfer simulation or light intensity simulation calculation in view of the decrease in transmittance due to damage caused by irradiation, and calculate the amount of cutting from the normal pattern edge line to the pattern side. 2. The photomask defect correction method using transfer or light intensity simulation according to claim 1, wherein black defect correction is performed to obtain light intensity similar to that of a normal pattern portion.   When it is determined that the defect needs to be corrected in the determination of whether or not the defect needs to be corrected, if the defect is a white defect, the three-dimensional shape of the white defect correction film and the light shielding rate by the focused ion beam are not considered. Calculate the amount of retraction of the white defect correction film from the edge line of the normal pattern, and correct the white defect with the focused ion beam in consideration of the amount of retraction. 2. The photomask defect correction method using transfer or light intensity simulation according to claim 1, wherein the light intensity is obtained.   When it is determined that the defect needs to be corrected in determining whether the defect needs to be corrected, the defect is a black defect, and when the black defect is corrected with an atomic force microscope, the atomic force microscope is searched. Perform transfer simulation or light intensity simulation calculation in consideration of the actual processing shape after scratch processing and the transmittance decrease due to the probe shape of the needle, and calculate the amount of cutting from the edge line of the normal pattern to the pattern side 2. The photomask defect correction method using transfer or light intensity simulation according to claim 1, wherein the black defect is corrected by scratching the atomic force microscope probe with the amount of cutting.   5. The photomask defect correction method according to claim 4, wherein the defect to be corrected is a convex phase defect of a Levenson type phase shift mask.

Images