JP2007299017A - Method for correcting mask pattern - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To correct a mask pattern highly precisely in consideration of an exposure margin without significantly increasing the load of proximity effect correction on the mask pattern. <P>SOLUTION: The method for correcting the mask pattern aims, to the mask pattern formed in an exposure mask to be served for pattern exposure, to impart correction to the pattern form so as to suppress influences of an optical proximity effect upon transferring the pattern onto an exposure substrate through an exposure device. The method includes: a step S2 of extracting a pattern within a range affected by the optical proximity effect; a step S3 of classifying the extracted pattern into a correction object pattern where a pattern edge is actually shifted and a reference object pattern where an edge is not shifted upon calculating correction; and a step of correcting the shape of the correction object pattern so as to finish the object pattern within an allowable edge shift amount ΔPos. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a mask pattern correction method for correcting an optical proximity effect generated when a pattern of an exposure mask is transferred onto a substrate to be exposed, and in particular, various variations (exposure amount, exposure, The present invention relates to a correction method in consideration of focus and the like.

  The present invention also relates to an exposure mask created using the above correction method, and further to a recording medium storing a program for realizing the above correction method by a computer.

  In recent years, with the miniaturization of LSIs, the influence of so-called optical proximity effects (OPE), in which dimensional variations and shape changes occur between the pattern of the exposure mask and the pattern obtained on the wafer, has become apparent. Have been doing. Therefore, in order to obtain a desired pattern on the wafer, an optical proximity correction (OPC) technique of correcting the mask pattern in consideration of the influence of the optical proximity effect in advance is required. ing.

  The conventional OPC technique predicts OPE under a certain exposure condition (exposure amount, focus position) by an experimental result or calculation, and corrects the mask pattern. However, as the allowable dimension variation during transfer pattern formation becomes smaller with miniaturization, an optical proximity effect correction technique that takes into account various variations during transfer (such as variations in the focus position of the exposure apparatus and variations in exposure amount) is required. It is coming.

  As the OPC technique considering the above-described variation, a method mainly considering the focus variation has been shown so far (see Non-Patent Document 1). However, with this method, it is not always possible to obtain an optimal correction pattern when exposure amount variation is assumed.

Further, the present inventors have proposed a correction method for maximizing an exposure range (exposure tolerance) and a focus range (focus depth) that can form a pattern within a predetermined dimensional variation amount with respect to the memory cell portion. However, in this document, since correction in a very limited area of a cell base is targeted and a correction calculation is performed using a kind of random search method by simulated annealing, When trying to apply to a random pattern, it takes too much calculation time, and it is considered impossible to process in a practical time.
“Fast proximity correction with zone sampling” by John P. Stirniman et al. (Proceeding of SPIE, Vol.2197, pp.294-300) “Automatic optical proximity correction with optimization of stepper condition” by S. Inoue et.al (Proceeding of SPIE, Vol.2440, pp.240-251)

  As described above, conventionally, when the optical proximity effect correction for the mask pattern is performed in consideration of the exposure margin, the calculation time required for the correction processing is greatly increased, high-speed processing is impossible, and sufficient correction accuracy is obtained. It was difficult to get.

  The present invention has been made in consideration of the above-described circumstances, and the object of the present invention is to provide a high accuracy in consideration of an exposure margin without significantly increasing the load of the optical proximity effect correction processing on the mask pattern. An object of the present invention is to provide a mask pattern correction method capable of performing correction.

  Another object of the present invention is to provide an exposure mask that is produced using the above correction method and can contribute to improvement of pattern transfer accuracy. Still another object of the present invention is to provide a recording medium storing a program for executing the above correction method by a computer.

  In order to solve the above problems, the present invention adopts the following configuration.

  That is, one embodiment of the present invention suppresses the influence of an optical proximity effect when a pattern is transferred to an exposed substrate via an exposure apparatus with respect to a mask pattern formed on an exposure mask used for pattern exposure. In a mask pattern correction method for correcting a pattern shape for the purpose, a step of extracting a plurality of patterns including the pattern within a range that exerts an optical proximity effect on a predetermined pattern, and the extracted pattern is actually The correction target pattern for moving the pattern edge to the reference pattern and the reference target pattern for which the edge movement is not performed during the correction calculation, and the correction target pattern so as to finish both within the allowable edge position deviation amount ΔPos. And a step of correcting the shape of the target pattern.

  According to the present invention, when correcting the optical proximity effect of the mask pattern to be formed on the exposure mask, high-accuracy correction considering the exposure margin without significantly increasing the load of the optical proximity effect correction processing on the mask pattern. And its usefulness is great.

  The details of the present invention will be described below with reference to the illustrated embodiments.

  Here, a method of calculating a correction amount in consideration of a lithography margin at an edge of interest of a pattern obtained by one-dimensionally extracting a mask pattern will be described.

  FIG. 1 shows an algorithm for correction processing according to this embodiment. First, in step S1, a pattern layout including a correction target pattern edge is input. Then, a pattern is extracted in the range covered by the OPE with the pattern edge of interest as the center (step S2). OPE here refers to a phenomenon in which the transfer position of the pattern edge of interest varies depending on the adjacent pattern layout, and is not limited only by optical factors, but can also include development and etching processes. It is.

  Next, the extracted pattern is classified into two types, a correction target pattern and a reference target pattern (step S3). The correction target here refers to a location where the pattern edge is actually moved by performing pattern correction together with the target pattern edge correction, while the reference target is considered as a pattern existing at the time of pattern correction. A point where the pattern edge is not moved.

  Next, only the correction target pattern is collectively corrected under the set exposure conditions so that the transfer dimension is as designed (step S4). The exposure conditions set here include exposure parameters (exposure wavelength, numerical aperture NA of the projection optical system of the exposure apparatus, light source size (coherence factor) σ, light source shape, etc.) and the type of mask used (COG ( Chrome on glass) mask, halftone phase shift mask, etc.), exposure amount, and focus position. Normally, the just focus position is set as the focus position, and the exposure amount is set so that the representative pattern is obtained as desired.

  The correction calculation is executed so that the edge position of the correction target pattern in the obtained transfer pattern is simultaneously within a certain variation. That is, under the set exposure conditions, the transfer pattern edge position at the just focus position is within the deviation amount from the desired pattern edge position determined by the positional deviation amount ΔPos1 smaller than the preset allowable edge positional deviation amount ΔPos. It correct | amends so that it may be formed.

  Next, in step S5, an exposure amount for finishing (desired edge position ± allowable variation) is obtained at a focus position (DOF / 2) necessary to secure a depth of focus (DOF) (Einf, Esup). . FIG. 2 shows an EDtree at this time (showing an exposure amount and a focus position necessary for obtaining a certain line width). In the figure, the horizontal axis indicates the exposure amount, the vertical axis indicates the focus, the solid line indicates the EDtree obtained in step S4, and the broken line indicates the EDtree corrected in step S6 described later.

  Based on FIG. 2, the exposure amount (E0inf, E0sup) at the exposure amount (E0) ± allowable exposure amount variation used in step S4 is compared with the exposure amount. Then, the correction amount is changed as follows (step S6).

i) If E0inf <Einf: Change mask pattern edge position correction amount by (Einf-E0inf) / M
ii) When Esup <E0sup: Change mask pattern edge position correction amount by (E0sup-Esup) / M
iii) In other cases: The correction amount is not changed. The coefficient M shown here is a coefficient indicating how much the exposure amount to be finished at the desired edge position varies due to the edge position variation of the mask pattern.

  FIG. 3A shows the relationship between the exposure mask and the image intensity, and FIG. 3B shows the relationship between the edge position variation of the mask pattern and the finished exposure amount. In the figure, 11 is a transparent substrate such as quartz, and 12 is a light shielding film such as Cr. As shown in FIG. 3B, when the edge position variation of the mask pattern is very small, it is considered that the edge position variation and the exposure amount variation are in a linear relationship. As shown in FIG. 2, the exposure amount can be shifted in the exposure amount direction while substantially maintaining the EDtree shape on the ED diagram (exposure amount-defocus diagram).

In FIG. 3B, M is expressed as a ratio of exposure difference due to mask edge position fluctuation, but the following equation M ′ = (logE−logE0) / Δ = {log (E / E0)} / Δ… (1)
As shown in FIG. 5, the ratio of the difference between the logarithms of the exposure doses is not limited.

  The result of correcting the mask pattern shown in FIG. 4 according to the algorithm of the present embodiment shown in FIG. 1 is shown below. The attention edge of the correction pattern shown in FIG. 4 is a space portion edge at the center of the pattern, and the evaluation uses the center line width of a pattern in which 12 types of S (distance to an adjacent pattern) are changed from 0.18 to 3 μm. The conditions used for the correction are shown below.

Exposure wavelength 248 nm, NA = 0.6, annular zone shielding rate = 2/3
COG (chrome on glass) mask,
The pattern edge position is determined by the aerial image + threshold value (Ith = 0.292).
Ith corresponds to an exposure amount for finishing a 0.18 μm line and space (L & S) pattern to a desired dimension with just focus.

  The range covered by OPE was not set in this calculation. In addition, for the classification of the correction target pattern and the reference target pattern, a total of five patterns were calculated with two patterns adjacent to the left and right of the target pattern as correction target patterns. Regarding the target exposure tolerance and DOF, in this calculation, when the allowable dimensional variation is ± 10%, the exposure tolerance = 10% and DOF = 0.6 μm. That is, the calculation is performed with the allowable edge position deviation amount (ΔPos) = 9 nm.

  FIG. 5 and subsequent figures show results corrected by the method of the present embodiment. 5A shows an EDtree evaluation result before OPC (without correction), FIG. 5B shows an EDtree evaluation result after transfer of a pattern corrected by a conventional method, and FIG. 6C shows an EDtree evaluation result by just focus correction. FIG. 6D shows the EDtree evaluation result after transfer of the pattern corrected according to this embodiment. Further, FIG. 7 shows a result of comparison regarding the common margin. The conventional method described here refers to a method of performing correction so that the exposure amount is fixed and the DOF is 0.6 μm.

  5 and 6, in the conventional method, the exposure amount is fixed in advance, and when correcting to make the DOF desired, in order to suppress the line width variation at the just focus and defocus positions as much as possible, This amount of variation in line width cannot be set small, and as a result, the common margin is reduced. The allowable line width variation amount in this correction calculation was set to ± 14 nm. That is, a value of 7 nm is used as ΔPos1. If it tries to suppress below this, the convergence of the correction calculation will be remarkably deteriorated. What is described as just focus correction is correction only at the just focus position, but in this case, the amount of variation in line width can be suppressed as compared with the conventional method. The allowable line width variation amount in this calculation could be ± 2 nm. That is, a value of 1 nm is used as ΔPos1. However, since no consideration is given to DOF, it is conceivable that a margin can not be obtained depending on variations in pattern arrangement.

  On the other hand, in the present embodiment, after just focus correction, further correction is added so as to ensure a given exposure latitude and DOF. In this embodiment, M described above is obtained for each pattern at a defocus position (0.3 μm) that determines DOF, and the correction amount of the target pattern is further changed. As described above, in the present embodiment, when the correction is performed so as to secure the exposure margin, the correction check is not performed on all the boundaries of the EDtree, but after the correction at the optimum focus position, the correction target pattern is changed with respect to the exposure amount fluctuation. It is only necessary to add an offset to the dimension according to the coefficient of changing the finished dimension, and high-speed processing is possible.

  Further, as is clear from FIG. 7 regarding the correction accuracy, the exposure tolerance at the desired depth of focus (DOF) = 0.6 μm is increased as compared with the conventional correction and the just focus correction method. It is clear that the accuracy obtained is improved compared to other methods.

  For comparison, FIG. 8 shows an algorithm for calculating a correction amount in consideration of a lithography margin by a conventional method. The process up to pattern input and pattern extraction is the same as in the present embodiment. After that, a check is performed under exposure condition 1 (just focus), and correction is performed under exposure condition 1 if NG. After correction to satisfy the exposure condition 1, the exposure condition 2 (defocus) is checked. If it is NG, the correction is performed under the exposure condition 2. After the correction is performed so that the exposure condition 2 is satisfied, the check and the correction under the exposure condition 1 are resumed.

  Since one set of checking and correction under a certain exposure condition corresponds to the batch correction in step S4 in the present embodiment, the batch correction in the present embodiment is required twice in the example of FIG. Further, in the example of FIG. 8, after the correction is performed under the exposure condition 2, the check and the correction under the exposure condition 1 are necessary, and the check and the correction are performed again.

  On the other hand, in the present embodiment, the batch correction in step S4 only needs to be performed once, and it is not necessary to redo the batch correction, so that the time required for the correction process can be significantly shortened compared to the example of FIG. . In addition, in this embodiment, the extraction pattern is separated into the correction target pattern and the reference target pattern, and the correction target pattern is narrowed down. From this point, the time required for the correction process can be shortened.

  As described above, according to the present embodiment, the transfer pattern edge position at a predetermined focus position at a predetermined exposure amount is determined by a positional deviation amount ΔPos1 smaller than a preset allowable edge positional deviation amount ΔPos. In order to correct it so that it is formed within the deviation from the pattern edge position, and then to consider the exposure tolerance, the mask pattern is obtained so as to obtain a desired depth of focus at each exposure tolerance boundary position. By adding the additional correction, the time required for the correction calculation can be greatly shortened. As a result, it is possible to perform highly accurate correction in consideration of the exposure margin without significantly increasing the load of the correction process.

  Further, by classifying the extracted pattern into a correction target pattern and a reference target pattern and correcting only the correction target pattern, it is possible to perform high-precision correction on the pattern edge of interest while suppressing the load of correction processing. It becomes possible.

  In addition, this invention is not limited to embodiment mentioned above. The processing presented in the embodiment is not limited to a one-dimensional pattern, and does not limit the extension to a two-dimensional pattern. Further, step S3 in FIG. 1 and subsequent steps 4 to 6 all contribute to shortening the processing time, and it is most preferable to perform both of these, but only one of them is performed. However, the processing time can be shortened.

  The methods described in the embodiments are various programs written on a recording medium such as a magnetic disk (FD, hard disk, etc.), optical disk (CD-ROM, DVD, etc.), semiconductor memory, etc. as programs that can be executed by a computer. The present invention can be applied to an apparatus or can be applied to various apparatuses by being transmitted through a communication medium. A computer that implements the present invention may be any computer that reads the program recorded on the recording medium and executes the above-described processing by controlling the operation by this program.

  In addition, various modifications can be made without departing from the scope of the present invention.

The figure which shows the correction | amendment algorithm concerning one Embodiment of this invention. The figure which shows typically the exposure amount evaluation position on an ED diagram. The figure which shows the relationship between a mask and image intensity, and the definition of the mask correction coefficient M. The figure which shows the example of the mask pattern used for this embodiment. The figure which shows evaluation by the ED figure of the correction result of a mask pattern, and is a figure by no correction and conventional correction. The figure which shows evaluation by the ED figure of the correction result of a mask pattern, and is a figure which shows the correction by just focus correction | amendment and embodiment. The figure which shows exposure amount tolerance-DOF created from the ED figure obtained from FIG. The figure which shows the algorithm of the correction amount calculation in a conventional method.

Explanation of symbols

11 ... Transparent substrate 12 ... Light shielding film

Claims (3)

A mask that corrects the pattern shape to suppress the effect of the optical proximity effect when transferring the pattern to the substrate to be exposed via the exposure device with respect to the mask pattern formed on the exposure mask used for pattern exposure In the pattern correction method,
Extracting a plurality of patterns including the pattern within a range that exerts an optical proximity effect on a predetermined pattern; and
Classifying the extracted pattern into a correction target pattern that actually moves a pattern edge and a reference target pattern that does not perform edge movement during correction calculation;
Correcting the shape of the correction target pattern so that both the correction target patterns are finished within the allowable edge position deviation amount ΔPos;
A mask pattern correction method comprising:   An exposure mask, wherein a mask pattern correction amount is determined using the mask pattern correction method according to claim 1, and a pattern is formed based on the correction amount. To correct the pattern shape in order to suppress the influence of the optical proximity effect when transferring the pattern to the substrate to be exposed via the exposure device with respect to the mask pattern formed on the exposure mask used for pattern exposure The program of
A plurality of patterns including the above pattern are extracted within a range in which the optical proximity effect is exerted on a predetermined pattern, and the pattern to be corrected is moved in the extracted pattern and the edge is moved during the correction calculation. A computer-readable program storing a program for controlling the computer to correct the shape of the correction target pattern so that the correction target pattern is finished within the allowable edge position deviation amount ΔPos. recoding media.

Images