JPS6110236A - Electron beam exposure - Google Patents

Abstract

PURPOSE:To enable a pattern to be printed with high precision, by actually exposing a substrate to be exposed while correcting the magnitude of beam deflection in the original data in accordance with the magnitude of the corrected deflection obtained by detecting the position of a mask for drift calibration provided in the course of the model exposure. CONSTITUTION:A substrate to be model exposed has a chip region 4, as shown in the figure (a), which is provided with four chip-positioning marks M1, M2, M3 and M4 and two drift calibrating (displacement correcting) marks M5 and M6. When the model exposure is carried out, the positions of the marks M1, M2, M3 and M4 are first detected by beam scanning similarly to the actual exposure, in order to position the chip. Subsequently, the region a1 terminating at the mark M5 for drift calibration is pattern exposed at the completely same rhythm with that of the actual exposure, according to the original pattern data of the chip. The pattern exposure operation is interrupted when it reaches said drift calibrating mark M5, and then the position of the drift calibrating mark M5 is detected by beam scanning. Thus, the chip is actually exposed to light while any deflection of the original pattern data is corrected properly.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to an electron beam exposure method, and in particular to a variable shaping method that has a function of pre-correcting deviations in exposure position that occur over time during direct exposure of a chip. This invention relates to a beam type electron beam exposure method.

As semiconductor integrated circuits become more highly integrated, pattern widths are being reduced to the submicron range.

The light exposure technology that was mainly used in the past in such situations has become impractical due to its wavelength (approximately 0.5 μm), and recently it has been
Direct exposure technology using an electron beam, which can provide pattern accuracy on the order of m, has come into widespread use.

Among direct exposure techniques using electron beams, the variable shaped beam type electron beam exposure method is currently in practical use as a method that can sufficiently compete with optical exposure in terms of throughput.

Even with the electron beam exposure method, which ideally has the characteristics of obtaining extremely high precision as described above, it is difficult to ensure the above precision due to various usage conditions of the equipment, and it is necessary to approach the above pattern precision as much as possible. That is MILSI
This is strongly required when manufacturing extremely miniaturized semiconductor integrated circuits such as.

[Conventional technology]

FIG. 3 shows an example of a device used in the most widely used variable shaped beam type electron beam exposure method as described above, that is, an example of a variable shaped electron beam exposure device 2 shows a beam blanking electrode for blocking the electron beam; 3 is an illumination lens that converges the beam, 4 is a first aperture (beam limiting aperture), 5 is a variable shaping deflector that deflects the limited beam, and 6 projects the deflected limited beam onto a second aperture. A projection lens, 7 is a second aperture (beam shaping aperture), 8 is a reduction lens, 9 is a blur correction coil, 10 is shaped into a predetermined rectangular shape by the first aperture and the second aperture, and is reduced by the reduction lens. A projection lens projects the blur-corrected electron beam onto the substrate to be exposed, 11 is a main deflection coil that deflects the irradiation position of the electron beam to a predetermined position, and 12 is a sub-coil that precisely adjusts the deflection position.・Deflector, 13 is a backscattered electron detector that detects the amount of change in backscattered electrons (including secondary electrons) at the mark position during chip alignment, 14 is a semiconductor substrate to be exposed, and 15 is for moving the exposure field. X-Y stage, EB indicates electron beam.

Conventionally, when actually exposing a chip using the above-mentioned electron beam exposure apparatus, the semiconductor substrate 14 to be exposed is set on the X-Y stage 15, and with the stage fixed at a predetermined position, for example, at least the four corners of the chip are exposed. An electron beam scans the positioning mark formed on the surface, the mark position is detected by the backscattered electron detector, and a predetermined rhythm (time and deflection speed) is set using the original pattern data using the detected mark position as a reference. The actual exposure was done by

Figure 4 shows a time chart of the conventional exposure method described above, in which eV is the electron beam intensity, [ is time, 48
0 represents a chip positioning operation (time) area, ExP represents an actual exposure operation area according to a predetermined rhythm, and m-3L represents a stage movement operation area.

The normal deflection range (field) is about 10 dragon mouths (the size that fits 1 to 4 chips), and the size of the shaping beam is 2 to 4
It is on the order of μm.

[Problem that the invention seeks to solve]

However, as mentioned above, in the conventional method of only aligning the chip before the actual exposure and using the original pattern data to actually expose the chip, when the actual exposure is performed, there is no difference in the progress of the chip exposure. As a result, there is a problem in that a large deviation (displacement) in the exposure position gradually occurs.

This misalignment is (1) Resist attached to column parts of exposure equipment.

(2) Charge-up of the resist deposited on the exposed substrate, (3) Thermal drift due to temperature rise of the main deflector/amplifier, (4) Main This is caused by the combined effects of hysteresis of the magnetic material near the deflector, etc., and the positional deviation is as shown in FIG.

In FIG. 5, (al is a plan view showing the magnitude of the deviation of each point P in the field in the Y direction by an arrow d, and (bl is the elapsed exposure time t and the deviation amount ΔY of the above point P in the Y direction). FIG.

Since such a positional shift also occurs in the X direction, the exposed tip shape is deformed and enlarged as shown by the chain line C' in the plan view of FIG.

Therefore, in order to suppress the above-mentioned positional deviation to a small level, various measures have been taken to eliminate the causes (1) to (4), etc., but it is extremely difficult to completely eliminate the above-mentioned causes with these measures. , a positional shift of about 0.3 to 0.5 μm was unavoidable.

The present invention solves the above-mentioned conventional problems and uses a variable shaped beam electron beam exposure method that can suppress positional deviation during actual chip exposure to about 0.1 μm, which is necessary for submicron pattern formation. This was done for the purpose of providing a method.

(Means for solving the problem) The above problem is solved by performing model exposure of the chip using the same original pattern as in the actual exposure, at the same timing and route as in the actual exposure, and at a predetermined point in the middle of the model exposure. At the position, the model exposure is temporarily interrupted to detect the mark position, and the amount of deflection to be corrected is calculated by adding it to the beam deflection amount from the difference between the detected mark position data and the original mark position data, and the amount of beam deflection is calculated. This problem is solved by the variable shaped beam type electron beam exposure method according to the present invention, which corrects the problem and performs actual exposure of the chip using the corrected beam deflection data.

[Operation] That is, the method of the present invention uses a model substrate that is similar to an actual exposure substrate and has a chip area in which a predetermined drift calibration mark is formed at one or more accurate positions, and Using the same original pattern data as the actual exposure, model exposure of the chip was performed on the chip area with an exposure rhythm having the same timing and deflection path as the actual exposure, and the electron beam reached the first drift calibration mark. At this point, the model exposure is temporarily interrupted, the first drift calibration mark position is detected by scanning the electron beam, and then the model exposure is continued according to the rhythm of the actual exposure, and the electron beam is scanned at the second drift calibration mark position. When reaching the drift calibration mark, model exposure is again interrupted and the position of the second drift calibration mark is detected by scanning the electron beam. And the third. Fourth, the above operation is repeated and continued for the number of marks for drift calibration, and the amount of deflection to be corrected is added to the amount of beam deflection based on the difference between the position data of the drift calibration mark obtained thereby and the original mark position data. is calculated, and the chip is actually exposed to the actual exposure substrate while correcting the amount of beam deflection using the corrected deflection data.

By doing this, the beam position deviation (
Since both the beam drift (beam drift) and the beam drift caused by the correlation with the exposure position are corrected, a highly accurate pattern can be drawn. .

FIG. 1 is a schematic plan view ta+ of a chip and a time chart fb) showing one example of model exposure operation, and FIG. 2 is a time chart showing another example of model exposure. Identical objects are indicated by the same reference numerals throughout the figures.

In the method of the present invention, the original pattern data of the chip is used as described above on a model substrate similar to the exposed substrate on which actual exposure is performed, for example, a model substrate on which a resist film is deposited on a silicon substrate. Model exposure is performed with the same exposure rhythm as exposure, and the beam deflection amount of the original pattern data is corrected using the deflection correction amount obtained by detecting the position of the drift calibration mark provided in the middle of the model exposure. While performing this, actual exposure is performed on the actual substrate to be exposed.

Next, a model exposure method, which is a feature of the present invention, will be described with reference to examples.

For example, as shown in FIG. 1 (al), the exposed substrate for model exposure has four chip positioning marks Ml, M2, M3, and M4 near the corners of chip area C, and further inside chip area C. Two drift calibration (displacement correction) marks!'15.M6 are provided.

Then, during model exposure, as in the case of actual exposure, first mark old, M2. M3. The position of M4 is detected and the chip is positioned.

Next, the drift calibration mark M is created using the chip's original pattern data in exactly the same rhythm as the actual exposure.
A pattern exposure operation is performed for the area a1 up to this side of 5.

The pattern exposure operation is interrupted before the drift calibration mark gate 5, and then the beam is scanned over the mark to detect the position of the drift calibration mark M5.

After the detection of mark i5 is completed, the pattern exposure operation according to the same rhythm as the actual exposure after mark M5 is continued until just before the drift calibration mark M6. This area is represented by a2.

Next, the position of the drift calibration mark M6 is detected in the same manner as described above, and the remaining areas a,
Perform pattern exposure operations using the actual exposure rhythm. Note that the pattern exposure operation for this asSM area (remaining area) may be omitted.

Figure 1(b) shows the time chart of the above model exposure, where eV is the intensity of the electron beam, t is the time, and m
st is a stage movement operation area, 4WD is a chip positioning operation area for detecting marks M1 to i4, EχP,
EXP! , EχP, respectively, are the above aI+''2.
The pattern exposure operation area according to the actual exposure rhythm corresponding to a3, MDS and MD6 represent the detection operation area of the drift calibration mark i5 and the belly.

Although the above embodiment shows an example in which the number of detection marks is small, a large number of tryst calibration marks may be provided in order to further improve exposure accuracy.

In such a case, the rhythm of pattern exposure corresponding to the actual exposure will be disrupted, so the detection of the drift calibration mark is divided into multiple model exposures to make the rhythm of the model exposure as close to the actual exposure rhythm as possible. It will be done.

FIG. 2 shows a time chart of one embodiment of the model exposure method described above, which is performed in multiple steps.

In this method, for example, in the first model exposure ME, the detection operation MDI of the first drift calibration mark is performed by performing the pattern exposure operation EXP++ according to the actual exposure rhythm up to the first drift calibration mark. Then, in the second model exposure ME, the pattern exposure operation EXh+ was performed continuously up to the second drift calibration mark.The detection operation MDZ□ of the second drift calibration mark was performed, and the third model exposure ME, pattern exposure operation EXP3 continues up to the third drift calibration mark.
Detection operation MDss of the acid third drift calibration mark for which + has been performed is performed. In addition, EXPI2.
EXP2z and EXP3z represent the remaining pattern exposure operation areas. This EXP+z, [1Xhz, E
You don't have to do this for XP3z.

The beam deflection data at the mark position obtained by detecting the position of the drift calibration mark using the model exposure method shown in the example above is the beam deflection data when irradiating the beam to the position corresponding to the mark during actual exposure. The data will be approximately equal to the data.

Therefore, a computer included in the electron beam exposure apparatus calculates correction deflection data to be added and corrected from the beam deflection data and the original pattern data of the drift calibration mark, and uses the correction deflection data to calculate the deflection data of the original pattern data. By performing the actual exposure of the chip while making the correction, the exposure accuracy of the pattern can be greatly improved, and an exposure accuracy of about 0.1 μm can be obtained.

〔Effect of the invention〕

As explained above, according to the electron beam exposure method of the present invention, the amount of beam drift that occurs during actual chip exposure is determined in advance by detecting marks for drift calibration during model exposure performed at the same exposure rhythm as the actual exposure. Since the actual exposure is made by correcting the amount of drift,
Exposure accuracy during actual exposure is significantly improved.

Therefore, the present invention provides a VLSI with a submicron pattern.
It is extremely effective for the production of products such as

[Brief explanation of drawings]

FIG. 1 is a schematic plan view of the Sennobu (al) and time chart diagram (bl) showing an example of model exposure operation in the electron beam exposure method of the present invention, and FIG. 2 shows another example of model exposure. Time chart diagram, Figure 3 is a schematic cross-sectional view showing an example of a variable shaping electron beam exposure system, Figure 4 is a time chart diagram of a conventional exposure method, and Figure 5 is a diagram of each point in the field in the Y direction in the conventional method. (a) is a plan view showing the magnitude of the deviation; (a) is a plan view showing the relationship between the elapse of exposure time and the amount of deviation ΔY in the Y direction at each point (bl); and (fcl) is a plan view showing the deformed state of the chip shape. In the figure, C is the chip area, Ml, M2.M3.M4 are chip positioning marks, M
5. M6 is a mark for drift calibration, eV
is the intensity of the electron beam, t is the time, mSt is the stage movement operating area, 11MD is the chip positioning operation tilting position to detect marks from old to -4, EXP, , EXP, , l1XP are the pattern exposure operating areas according to the actual exposure rhythm , MDS and MDa indicate detection operation areas of the drift calibration marks M5 and M6. 1st Inne 2 @ 1--------2 Grass 3 Picture grass 4 Picture grass S Picture 111-2

Claims (1)

[Claims] Model exposure of the chip is performed using the same original pattern data as the actual exposure at the same timing and route as the actual exposure, and the model exposure is temporarily interrupted at a predetermined position during the model exposure to detect the position of the mark. The amount of deflection to be corrected is calculated by adding it to the amount of beam deflection from the difference between the mark position detection data and the original mark position data, the beam deflection amount is corrected, and the actual exposure of the chip is performed using the corrected beam deflection data. An electron beam exposure method characterized by performing the following steps.