JPH0434813B2 - - Google Patents

Description

[Detailed Description of the Invention] [Technical Field of the Invention] The present invention relates to charged beam exposure technology that exposes fine circuit patterns such as semiconductor integrated circuits, and particularly to charged beam exposure technology that performs pattern exposure by varying beam dimensions. Concerning improvements in methods.

[Technical background of the invention and its problems]

In recent years, various electron beam exposure apparatuses have been developed to form fine patterns for LSIs, VLSIs, and the like. In this type of device, generally 1 [μm]
In order to form a circuit pattern with a size of about 1 [μm] or less, a method is adopted in which a pattern is drawn using an electron beam with a circular cross section having a diameter of about 0.25 [μm]. It takes a long time to expose a single wafer, making it impractical. Recently, an electron beam exposure system using a variable beam type using a beam with a rectangular cross section has been developed, which has increased the exposure area per unit time by nearly 10 times, resulting in high productivity. On the other hand, as is well known, VLSI circuit patterns require finer patterns year by year, and accordingly, variable-dimension beam writing apparatuses are also required to be capable of higher precision exposure.

An overview of the variable dimension beam system will be explained with reference to FIG. In the figure, 1 is an electron gun that is a light source, 2 is a first molded aperture mask having a rectangular aperture, and 3
is a second molded aperture mask, and 4 is a deflector. An image of the aperture 2a of the mask 2 is projected onto the mask 3 by a lens (not shown). the relative position between the aperture image of mask 2 and mask 3;
That is, by controlling the optical overlap of the apertures 2a and 3a by the deflector 4, a beam with a rectangular cross section of variable dimensions can be obtained. FIG. 2 is a circuit diagram of a control system for driving the deflector 4. As shown in FIG.
The beam dimension data supplied from CPU5 is
The signals are converted into analog values by D/A converters (DACs) 6a and 6b, and voltage or current signals for driving the deflector 4 are output from amplifiers (deflection amplifiers) 7a and 7b. Therefore, the accuracy of setting the beam dimensions depends on the accuracy of setting the gain and offset of the deflection amplifier. Normally, the gain value and offset value of the amplifier are set by comparing the beam dimension data value and the measured beam dimension value so that the measured value becomes the data value.

However, this type of method has the following problems. That is, due to the imaging aberration of the electron optical system for projecting and imaging the beam onto the sample, beam dimension measurement errors, etc., the relationship between the beam dimension data value and the dimension measurement value is as shown in Figure 3. In the region of small beam dimensions, the relationship deviates from the linear relationship. Normally, for gain and offset values, the maximum dimension data value is H, the dimension data value about 1/2 of that value is L,
or the dimensional measurements obtained for these, H′,
Assuming L', the points (H, H'), as shown in Figure 4,
Set so that the straight line connecting (L, L') passes through the origin. However, due to the reasons mentioned above, L' is estimated to be large, so the relationship between the beam size data value and the beam current value becomes as shown in Figure 5. The current value becomes zero, and this has the disadvantage that pattern defects occur in portions of the drawn pattern that should be exposed with a beam of minute size.

As a method to eliminate this drawback, the beam currents I H and I L corresponding to the above H and L are measured, and the beam currents I _H and I _L corresponding to the points (H,
L so that the straight line connecting I _H ) and (L, I _L ) passes through the origin.
A possible method is to set the gain and offset values by combining the voltage or current signal values corresponding to the .
This method is effective when the assumption that the relationship between the beam dimension data value and the beam current value is linear as shown in FIG. 5 has recently been established.
For example, as shown in Fig. 6, when dimension a of a rectangle in one direction (Y direction) is fixed to a large value (>1 μm) and dimension b in the other direction (X direction) is changed, This assumption holds true. However, this assumption does not hold if the dimension a in the Y direction is 1 [μm] or less. The reason for this is that the direction of each side of the apertures 2a, 3a of the molded aperture masks 2, 3 must be perfectly matched with the deflection direction of the deflector 4 due to mechanical assembly adjustment errors, deflection distortion of the electron optical system, and This is nearly impossible due to correction errors, and a certain finite value of direction deviation must be tolerated. Therefore, when the dimension of the short side is reduced to, for example, about 0.4 μm, the relationship between the dimension data value of the long side and the beam current value becomes a quadratic curve as shown in FIG. Whether this curve is upwardly convex or downwardly convex depends on the relative angle difference between each aperture side and the deflection direction. In FIG. 7, the beam current values corresponding to the dimension data values H and L are measured according to the method described above, and the points (H, I _H ), (L,
If the gain and offset of the deflection amplifier are set so that the straight line connecting I _L The beam current for the data deviates significantly from the designed value, causing a local dose error during exposure, resulting in pattern dimensional errors or pattern defects.

Note that the above-mentioned problems are not limited to electron beam exposure methods, but also apply to ion beam exposure methods using ion beams.

[Purpose of the invention]

An object of the present invention is to provide a charged beam exposure method that can set the gain and offset of a deflection amplifier so as not to cause dimensional errors in exposure patterns or pattern defects due to dose errors, and can perform highly accurate pattern exposure. Our goal is to provide the following.

[Summary of the invention]

The dose error due to the gain and offset setting error of the deflection amplifier described above occurs at a portion of the pattern exposed with a beam size sufficiently smaller than the maximum beam size. Therefore, the gain of the deflection amplifier is adjusted so that the relationship between the beam size data
By setting the offset, it is possible to effectively eliminate dose errors.

The present invention focuses on such points, and deflects a charged beam emitted from a charged beam source using a beam dimension variable deflector, thereby varying the optical overlap of each aperture of at least two beam shaping aperture masks. In the charged beam exposure method, the dimension-controlled beam is scanned on a sample by a beam scanning deflector to expose a desired pattern on the sample. The gain and offset of the deflection amplifier that drives the deflector are set so that the maximum dimension data value of the beam and the beam dimension measurement value corresponding to the data value are approximately equal, and the minimum dimension data value of the beam and the data value. In this method, the corresponding beam current measurement values are set to be approximately equal.

〔Effect of the invention〕

According to the present invention, it is possible to suppress the dose error in a portion of an exposure pattern exposed with a beam having a size sufficiently smaller than the maximum size of the beam to a few [%] or less, making it possible to perform highly accurate patterning. Therefore, it will be able to fully handle the formation of fine circuit patterns for future VLSIs, making it extremely useful.

[Embodiments of the invention]

FIG. 8 is a schematic configuration diagram showing an electron beam exposure apparatus used in a method according to an embodiment of the present invention. 1 in the diagram
1 is a sample chamber, and within this sample chamber 11 is arranged a sample stage 13 on which a sample 12 such as a semiconductor wafer is placed. The sample stage 13 is moved by a stage drive system 14 in the X direction (left and right directions in the paper) and Y direction (front and back directions in the paper). The moving position of the stage 13 is measured by a laser length measuring system 15.

On the other hand, above the sample chamber 11 are an electron gun 21, deflectors 22, 23, 24, first and second beam shaping aperture masks 31, 32, and various lenses 3.
Electron optical lens barrel 4 consisting of 3, 34, 35, 36, etc.
0 is set. Here, the deflector 22 is a blanking deflector that is given a deflection voltage by the blanking control circuit 41 to blank the beam. The deflector 23 is a beam size variable deflector that deflects the beam by applying a deflection voltage by the beam size control circuit 42 and changes the overlapping state of each aperture of the aperture masks 31 and 32. The dimensions of the beam can be varied by the deflector 23. Here, the beam size control circuit 42 consists of a DAC, a deflection amplifier, etc. as shown in FIG.
The gain and offset of the deflection amplifier are set by the method described later. Further, the deflector 24 is a beam scanning deflector that is applied with a deflection voltage by the beam deflection control circuit 43 and scans the beam on the sample 12. The size of the electron beam emitted from the electron gun 21 is controlled to a predetermined size by the beam size variable deflector 23 and the beam shaping aperture masks 31 and 32, and the size-controlled beam is used for beam scanning. sample 1 by the deflector 24.
2 to expose the sample 12 to a desired pattern.

In the figure, 44 is a control computer for controlling the drive system 14 and various control systems 41, 42, 43, and 45 is a control computer for controlling the drive system 14, the drive system 14, the length measurement system 15, and the control circuits 41, 42, 45. 43.

Next, the gain of the deflection amplifier in the above device,
The offset setting method will be explained. The maximum dimensions of the beam in this example are 3.2 [μm] x 3.2 [μm]
m]. Further, for the sake of simplicity, only the dimensional change in the X direction will be described, but the same applies to the Y direction.

First, the output V _H of the deflection amplifier whose maximum dimension in the X direction is L _H =3.2 [μm] is determined. That is, the beam dimension L is measured while changing the deflection amplifier output V, and the amplifier output V=V ₁ where L ₁ =L _H is obtained.
Similarly, obtain amplifier output V=V ₂ such that L ₂ =L _H /2. Here, if the relationship between beam length and output is linear, L=α(V-V ₁ )+L ₁ . However, α is the sensitivity of the deflector, and α=
(L ₂ −L ₁ )/(V ₂ −V ₁ ). On the other hand, the relationship between the beam current I and the amplifier output V is approximately expressed as I=β(V-V ₁ )+I ₁ . Here, β is determined by the sensitivity α of the deflector, the current density of the beam, and the dimension in the Y direction, and there is a relationship β=γα . However, γ is a coefficient and I ₁ is the beam current at V=V ₁ . Let G' and O' be the temporary gain and offset values of the deflection amplifier, respectively, and X be the beam size data, and control the amplifier so that V=G'X+O' . However, G′,
O' is defined from α, X ₁ , X ₂ , L ₁ , L _{2 G} ' = (L ₂ − L ₁ )/α ( _{X 2 −} ... is. The above formula is I=β[G'X+O'-V ₁ ]+I ₁ =βG'(X-X ₁ )+I ₁ . As described in the conventional example, the above equation does not hold strictly, and an error occurs in the beam current value near X=O. Therefore, X ₃ = L ₁ /10, X ₄ = L ₁ /
The beam currents I ₃ and I ₄ at 30 are measured, and the value G of G' such that I ₄ =I ₃ /3 is determined. Also, the offset is V ₁ ,
From G and X ₁ , set so that O = V ₁ - GX ₁ ....

By using the above setting method, it was possible to substantially eliminate the dose error caused by a slight angular error between each aperture side of the beam shaping aperture mask and the deflection direction by the deflector. Here, the relationship between the beam size data X and the beam current I after optimizing the gain and offset is shown by a solid line in FIG. If the deviation from linearity is ΔI, then ΔI/I ₁ is a value determined by the beam length in the Y direction, the rotation mechanism of the beam shaping aperture mask, and the orientation accuracy of the deflector, and is usually ΔI/I ₁ It is adjustable from =0.01 to 0.05. Therefore, in the case of FIG. 9, the dose error is 1 to 5%. In contrast,
The relationship between X and I when using the conventional method is the first
The result will be as shown in Figure 0. That is, when X=0, the beam current I=ΔI. At this time, the beam current at X=X ₁ /30 is I ₄ '=I ₁ /30 + (14/15)ΔI, and when ΔI=0.01I ₁ , the dose error is (0.01×14/15)/ 30=0.28, which means an error of nearly 30%.

As described above, by setting the gain and offset of the deflection amplifier as in the method of this embodiment,
We were able to reduce the dose error to 1/30 of the conventional level.

Note that the present invention is not limited to the embodiments described above. For example, when setting the deflection amplifier corresponding to the minute beam dimensions, the gain and offset are not necessarily set so that the ratio of the two types of beam dimension data and the corresponding beam current measurement value approximately match. This is not necessary, and the beam size data value that is sufficiently smaller than the maximum beam size may be set to be approximately equal to the beam current measurement value that corresponds to the data value. Furthermore, the structure of the electron optical lens barrel is not limited to that shown in FIG. 8, and can be used in various types of electron beam exposure apparatuses of various beam dimension variable methods equipped with a beam shaping aperture mask and a beam dimension variable deflector. Of course, it can be applied to Furthermore, the invention is not limited to electron beam exposure apparatuses, but can also be applied to ion beam exposure apparatuses. In addition, various modifications can be made without departing from the gist of the present invention.

[Brief explanation of drawings]

Figure 1 is a schematic diagram for explaining the principle of changing beam dimensions in the variable dimension beam method, Figure 2 is a circuit configuration diagram showing the deflector drive control system, and Figures 3 to 7 are problems with the conventional method. The third purpose is to explain
The figure is a characteristic diagram showing the relationship between beam dimension data and beam dimension measurements, Figure 4 is a characteristic diagram showing the relationship between beam dimension data and beam dimension measurements obtained by the conventional setting method, and Figure 5 is a characteristic diagram showing the relationship between beam dimension data and beam dimension measurements. Figure 6 is a schematic diagram showing an example of a variable dimension beam, and Figure 7 is a characteristic diagram showing the relationship between beam size data and beam current measurements using the conventional setting method. 8 is a schematic configuration diagram showing an electron beam exposure apparatus used in an embodiment method of the present invention, and FIG. 9 shows beam dimension data when the gain and offset of the deflection amplifier are set by the above embodiment method. FIG. 10, a characteristic diagram showing the relationship with the beam current measurement value, is a characteristic diagram corresponding to the above-mentioned FIG. 9 using a conventional setting method as a comparative example. 1, 21... Electron gun, 2, 31... First beam shaping aperture mask, 3, 32... Second beam shaping aperture mask, 2a, 3a...
Aperture, 4, 23... Beam dimension variable deflector, 5, 44... Control computer (CPU), 6a,
6b...D/A converter, 7a, 7b...amplifier (deflection amplifier), 11...sample chamber, 12...sample,
13... Sample stage, 14... Stage drive system, 15... Laser length measurement system, 22... Blanking deflector, 24... Beam scanning deflector, 3
3, -, 36... Lens, 40... Electron optical lens barrel, 41... Blanking control circuit, 42... Beam size control circuit, 43... Beam deflection control circuit, 45... Interface.

Claims (1)

[Claims] 1. A charged beam emitted from a charged beam source is deflected by a beam size variable deflector, and the optical overlap of each aperture of at least two beam shaping aperture masks is varied to change the beam size. In a charged beam exposure method in which dimensions are variably controlled and the dimension-controlled beam is scanned over a sample by a beam scanning deflector to expose a desired pattern on the sample, the beam dimension variable deflector is driven. The gain and offset of the deflection amplifier are set so that the maximum dimension data value of the beam and the beam dimension measurement value corresponding to the data value are approximately equal, and the beam micro dimension data value and the beam current corresponding to the data value are adjusted. A charged beam exposure method characterized by setting the measured values to be substantially equal. 2. Select two types of beam dimension data values that are sufficiently smaller than 1/2 of the maximum dimension of the beam as the settings of the deflection amplifier corresponding to the minute dimensions of the beam, and set the ratio of these data values to correspond to the data value. 2. A charged beam exposure apparatus according to claim 1, wherein the gain and offset of said deflection amplifier are set so that the ratio of measured beam current values substantially matches.