Multi-charged particle beam lithography method and multi-charged particle beam lithography device

Abstract

To improve lithography accuracy by suppressing the influence of coulomb effects by reducing an average total current of on-beams during a shot cycle.SOLUTION: A multi-charged particle beam lithography method includes: a step of grouping a multi-beam into a plurality of beam groups; a step of determining an irradiation time on a beam basis for each shot; and a shot division step of dividing each shot into a plurality of irradiation steps including irradiation steps of different irradiation times, grouping a plurality of irradiation steps of a predetermined first irradiation time as a first group at a part of the plurality of divided irradiation steps, and selecting a set of irradiation steps in which the total of the irradiation times of the irradiation steps becomes the irradiation time on the beam basis. In the shot division step, timing of the irradiation step of the first group of irradiation steps belonging to each of the plurality of beam groups is determined based on an allocation order which is predetermined to be different for each beam group, and the number of irradiation steps to be turned on.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a multi-charged particle beam writing method and a multi-charged particle beam writing apparatus. [Background technology]

[0002] As LSIs become more highly integrated, the circuit line widths used in semiconductor devices are becoming finer year by year. In order to form a desired circuit pattern for a semiconductor device, a method is adopted in which a high-precision original pattern (mask, or particularly one used in steppers and scanners is called a reticle) formed on a light-shielding film on a glass substrate is reduced and transferred onto a wafer using a reduction projection exposure apparatus. To create a high-precision original pattern, a so-called electron beam lithography technique is used in which a resist pattern is formed using an electron beam drawing apparatus. In some cases, a method called direct wafer writing is used in which a pattern is formed on a resist coated on a wafer using an electron beam.

[0003] In a multi-beam lithography device using a blanking aperture array substrate, which is one form of a multi-beam lithography device, for example, an electron beam emitted from one electron gun is passed through a shaping aperture array member having multiple openings to form a multi-beam (multiple electron beams). Downstream of the shaping aperture array member is a blanking aperture array substrate. The blanking aperture array substrate has, for each beam of the multi-beam, an electrode pair (blanker) for individually deflecting the beam and an opening for passing the beam formed between the electrode pair, which are arranged in an array on the blanking aperture array substrate. The blanking aperture array substrate switches the blanking deflection of the passing beam on and off by controlling the electrode pairs corresponding to each beam of the multi-beam to the same potential or to different potentials. The multi-beam formed by the shaping aperture array substrate passes through the passage holes between the corresponding blankers of the blanking aperture array substrate. The optical lens barrel of the multi-beam lithography device is configured so that the electron beam deflected by the blanker is shielded and the electron beam not deflected is irradiated onto the substrate.

[0004] A lithography system using multiple beams can irradiate many beams at once compared to lithography using a single electron beam, and thus can greatly improve throughput. On the other hand, because the total beam current can be set large, the lithography accuracy can be degraded due to the Coulomb effect. Specifically, the repulsive force between electrons can cause degradation of the beam resolution, and the beam position and focus on the sample surface. The Coulomb effect is most prominent in places where the electron beam density is high in the optical system, such as crossovers where multiple beams converge to a single point. However, the Coulomb effect at crossovers downstream of the blanking aperture array substrate is not constant, but varies depending on the total current of the on-beams in the multiple beams.

[0005] As a measure against the deterioration of drawing accuracy due to the Coulomb effect, for example, a method of lowering the total current of the on-beam has been proposed (see Patent Document 1). This method is effective when the waiting time during which the beam cannot be turned on during a shot cycle is longer than the time the beam is turned on due to reasons such as not being able to control the beam at high speed, but the effect becomes smaller when the drawing system is optimized and the waiting time during which the beam cannot be turned on becomes shorter. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] JP 2007-329220 A [Patent Document 2] US Patent Application Publication No. 2010 / 0124722 [Patent Document 3] Japanese Patent Application Publication No. 10-308341 [Patent Document 4] JP 2012-243802 A [Patent Document 5] Japanese Patent Application Publication No. 6-302506 Summary of the Invention [Problem to be solved by the invention]

[0007] An object of the present invention is to provide a multi-charged particle beam drawing method and a multi-charged particle beam drawing apparatus that suppresses the influence of the Coulomb effect by lowering the average total current of the on-beam during a shot cycle without significantly degrading the drawing throughput, thereby improving the drawing accuracy. [Means for solving the problem]

[0008] A multi-charged particle beam writing method according to one embodiment of the present invention includes a step of emitting a multi-beam of charged particles, a step of grouping a plurality of beams constituting the multi-beam into a plurality of beam groups, a step of determining an irradiation time for each beam of each shot of the multi-beam from writing pattern data, a shot division step of dividing each of the shots into a plurality of irradiation steps including irradiation steps having different irradiation times, and selecting a set of irradiation steps in which a sum of the irradiation times of each irradiation step is the predetermined irradiation time for each beam, the plurality of irradiation steps being grouped as a first group among the divided plurality of irradiation steps, and a step of performing a multi-beam shot by switching each beam of the multi-beam on and off to execute the set of the plurality of irradiation steps, wherein in the shot division step, the timing of the irradiation step of the first group of the plurality of irradiation steps belonging to each of the plurality of beam groups is determined based on a predetermined allocation order different for each of the beam groups and the number of the irradiation steps to be turned on.

[0009] In a multi-charged particle beam writing method according to an aspect of the present invention, each of the plurality of irradiation steps includes the first group and a second group of irradiation steps each having an irradiation time shorter than the first irradiation time.

[0010] In a multi-charged particle beam writing method according to one aspect of the present invention, the first irradiation time is the longest irradiation time among the irradiation times of the plurality of irradiation steps.

[0011] In a multi-charged particle beam writing method according to one aspect of the present invention, the irradiation time of each of the multiple irradiation steps is determined to be proportional to a power of two.

[0012] In a multi-charged particle beam writing method according to an aspect of the present invention, in the second group, the timings at which the beams are turned on in the plurality of irradiation steps of the second irradiation time differ for each of the beam groups.

[0013] In a multi-charged particle beam writing method according to one aspect of the present invention, in the multiple beams, adjacent beams in the writing progress direction or in a direction linearly independent of the writing progress direction are classified into different groups.

[0014] A multi-charged particle beam drawing apparatus according to one embodiment of the present invention includes an emission unit that emits a multi-beam of charged particles, and a drawing control unit that switches on and off each beam of the multi-beam to control the irradiation time of each beam of the multi-beam calculated from drawing pattern data, wherein the drawing control unit groups a plurality of beams that constitute the multi-beam into a plurality of beam groups, divides each of the shots into a plurality of irradiation steps including irradiation steps of different irradiation times, groups a plurality of irradiation steps of a predetermined first irradiation time among a portion of the divided irradiation steps into a first group, selects a set of irradiation steps whose total irradiation time of each irradiation step is the predetermined irradiation time for each of the beams, and determines the timing of the irradiation steps of the first group of the plurality of irradiation steps belonging to each of the plurality of beam groups based on an allocation order that is predetermined to be different for each beam group and the number of the irradiation steps that are turned on. Effect of the Invention

[0015] According to the present invention, by lowering the average total on-beam current during a shot cycle, it is possible to suppress the influence of the Coulomb effect and improve the writing accuracy. [Brief description of the drawings]

[0016] [Figure 1] 1 is a schematic diagram of a multi-charged particle beam writing apparatus according to an embodiment of the present invention; [Diagram 2] FIG. 2 is a plan view of a shaping aperture array member. [Diagram 3] FIG. 2 is a cross-sectional view showing a configuration of a blanking aperture array substrate. [Figure 4]FIG. 2 is a schematic diagram of a control circuit in a blanking aperture array substrate. [Diagram 5] FIG. 2 is a configuration diagram of an input / output circuit and a cell array circuit. [Figure 6] FIG. 2 is a schematic diagram of an individual blanking mechanism. [Figure 7] FIG. 1 is a diagram showing an example of irradiation steps within one shot cycle. [Figure 8] FIG. 13 is a diagram showing an example of beam-on timing. [Figure 9] FIG. 13 is a diagram showing the order of allocation of irradiation steps. [Figure 10] FIG. 13 is a diagram showing an example of beam-on timing. [Figure 11] FIG. 13 is a diagram showing an example of beam-on timing. [Figure 12] 13(a) to 13(c) are diagrams showing examples of grouping of beam arrays. [Figure 13] 11 is a flowchart illustrating a drawing method. [Figure 14] FIG. 11 is a diagram illustrating a drawing operation. [Figure 15] FIG. 13 is a diagram showing the order of allocation of irradiation steps. [Figure 16] FIG. 13 is a diagram showing an example of beam-on timing. [Figure 17] FIG. 13 is a diagram showing the order of allocation of irradiation steps. [Figure 18] FIG. 13 is a diagram showing an example of beam-on timing. [Figure 19] 11 is a graph showing an example of beam current when the beam-on timing is shifted and when it is not shifted. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0017] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In the embodiment, a configuration using an electron beam as an example of a charged particle beam will be described. However, the charged particle beam is not limited to an electron beam, and an ion beam may be used.

[0018] Fig. 1 is a schematic diagram of a drawing apparatus according to an embodiment. As shown in Fig. 1, the drawing apparatus 100 includes a drawing unit 150 and a control unit 160. The drawing apparatus 100 is an example of a multi-charged particle beam drawing apparatus. The drawing unit 150 includes an electron lens barrel 102 and a drawing chamber 103. In the electron lens barrel 102, an electron gun 201, an illumination lens 202, a shaping aperture array member 203, a blanking aperture array substrate 204, a reduction lens 205, a limiting aperture member 206, an objective lens 207, and a deflector 208 are arranged.

[0019] An XY stage 105 is disposed in the writing chamber 103. A substrate 101 to be written is disposed on the XY stage 105. A resist to be exposed to an electron beam is applied to the upper surface of the substrate 101. The substrate 101 is, for example, a mask blank or a semiconductor substrate (silicon wafer) to be processed into a semiconductor device. In addition, a mirror 210 for measuring the stage position is disposed on the XY stage 105.

[0020] The control unit 160 has a control computer 110, a deflection control circuit 130, a stage position detector 139, and a storage unit 140. Drawing data is input from the outside and stored in the storage unit 140. The drawing data defines information on a plurality of graphic patterns to be drawn. Specifically, a graphic code, coordinates, size, etc. are defined for each graphic pattern. Other information, for example, control information for the dose of irradiation, may be additionally defined in the drawing data.

[0021] The control computer 110 has an area density calculation unit 111, an irradiation time calculation unit 112, a data processing unit 113, and a drawing control unit 114. Each unit of the control computer 110 may be configured with hardware such as an electric circuit, or may be configured with software such as a program that executes these functions. Alternatively, each unit may be configured with a combination of hardware and software.

[0022] The stage position detector 139 irradiates a laser onto a mirror 210, receives the light reflected from the mirror 210, and detects the position of the XY stage 105 by laser interference.

[0023] Fig. 2 is a conceptual diagram showing the configuration of the shaping aperture array member 203. The shaping aperture array member 203 is a plate-like member, and as shown in Fig. 2, a plurality of openings 203a are formed in the plane of the shaping aperture array member 203 along the vertical direction (y direction) and horizontal direction (x direction). It is preferable that each opening 203a is formed, for example, in a rectangular shape of the same or approximately the same dimensions. The shape of each opening 203a may be circular.

[0024] Electron beam 200 emitted from electron gun 201 (emitter) illuminates shaping aperture array member 203 via illumination lens 202. Electron beam 200 illuminates an area of ​​shaping aperture array member 203 that includes all of the apertures 203a. A portion of electron beam 200 passes through multiple apertures 203a in shaping aperture array member 203, and the remaining beam is stopped by shaping aperture array member 203, thereby forming multiple electron beams, i.e., multi-beams 20a to 20e. The shape of each multi-beam follows the shape of the apertures 203a in shaping aperture array member 203, and is, for example, rectangular.

[0025] As shown in Fig. 3, the blanking aperture array substrate 204 has a support base 204a and a semiconductor substrate 204b made of silicon or the like provided on the support base 204a. The central portion of the semiconductor substrate 204b is thinly scraped from the back side and processed into a thin membrane region 204c. The membrane region 204c is surrounded by a thick outer peripheral region, and the semiconductor substrate 204b is held on the support base 204a at the back side of the outer peripheral region. The central portion of the support base 204a is open, and the membrane region 204c is located in the open region of the support base 204a.

[0026] In the membrane region 204c, a plurality of beam passage holes H are formed in accordance with the respective positions of the plurality of openings 203a of the shaping aperture array member 203. A blanker 50 consisting of a pair of electrodes 51, 52 is disposed in each passage hole H, and one of the multi-beams passes between the pair of electrodes and through the passage hole H. The blanker 50 switches the deflection of the beam passing through the passage hole H on and off by grounding one electrode 52 and keeping it at ground potential and switching the other electrode 51 to ground potential or a potential other than ground potential. In this way, the blanker 50 performs blanking control to set each of the multi-beams to either a beam-on or beam-off state. The principle of blanking control will be described below.

[0027] When one of the multiple beams is controlled to a beam-on state, the opposing electrodes 51, 52 of the blanker 50 are controlled to the same potential, and the blanker 50 does not deflect the beam passing through the passage hole H. When the one of the multiple beams is controlled to a beam-off state, the opposing electrodes 51, 52 of the blanker 50 are controlled to different potentials, and the blanker 50 deflects the beam passing through the passage hole H.

[0028] The multiple beams 20 a to 20 e that have passed through the blanking aperture array substrate 204 are reduced in size by the reduction lens 205 and travel toward a central opening formed in the limiting aperture member 206 .

[0029] Here, the beam controlled to the beam-off state is deflected by the blanker 50 and passes through a trajectory that passes outside the opening of the limiting aperture member 206, and is therefore blocked by the limiting aperture member 206. On the other hand, the beam controlled to the beam-on state is not deflected by the blanker 50 and passes through the opening of the limiting aperture member 206. The trajectory of the beam is adjusted by an alignment coil (not shown) so that the beam controlled to the beam-on state is located within the opening of the limiting aperture member 206. In FIG. 1, the trajectory of the multi-beam in the beam-on state is adjusted so that the multi-beam converges to one point at the position of the limiting aperture member 206, but it is preferable to adjust the alignment coil so that this one point is located at the center of the opening of the limiting aperture member 206. In this way, the on / off state of each beam of the multi-beam is controlled by a combination of the on / off operation of the deflection of the blanker 50 and the blocking of the beam by the limiting aperture. In other words, blanking control is performed.

[0030] As described above, the blanking aperture array substrate 204 uses the multiple blankers 50 to individually control blanking of each beam of the multi-beam 20. In other words, the beam-on / beam-off state can be independently switched for each beam of the multi-beam 20. In the irradiation process described below, blanking control is performed to change from a state in which all beams of the multi-beam are controlled to a beam-off state to a state in which only certain beams are blanked to a beam-on state, and after a certain time has elapsed, all beams are controlled to a beam-off state. By such blanking control, it is possible to control only selected beams of the multi-beam to a beam-on state for a certain time.

[0031] The multi-beams that have passed through the limiting aperture member 206 are focused by the objective lens 207, and are formed as a pattern image with a desired reduction ratio on the substrate 101. Ideally, the multi-beams are arranged on the substrate 101 at a pitch obtained by multiplying the arrangement pitch of the multiple openings 203a of the shaping aperture array member 203 by the desired reduction ratio described above. The beams (all of the multi-beams that are in a beam-on state) that have passed through the limiting aperture member 206 are deflected collectively in the same direction by the deflector 208, and are irradiated at a desired position on the substrate 101 with the focus on the surface of the substrate 101.

[0032] Irradiation of the substrate 101 with the multibeams is possible whether the XY stage 105 is stationary or moving continuously. When the XY stage 105 is moving continuously, the stage position detector 139 measures the amount of change in stage position, and the result is used to continuously change the position of the multibeams using the deflector 208 so as to follow the movement of the XY stage 105. This is called stage tracking deflection. Stage tracking deflection makes it possible to fix the position of the multibeams on the substrate 101. At least while the substrate 101 is being irradiated with the beams, stage tracking deflection is performed to control the position of each beam of the multibeams on the substrate 101 to be fixed.

[0033] The blanking aperture array substrate 204 includes the blanker 50 and the passage hole H described above, as well as a control circuit for applying a desired voltage to the blanker 50. This control circuit includes an input / output circuit 31 and a cell array circuit 34, as shown in FIG.

[0034] As shown in FIG. 5, the cell array circuit 34 is provided with a plurality of cells that constitute the individual blanking mechanism 40 that drives the blanker 50. FIG. 5 shows an example of a blanking aperture array substrate having 262,144 cell array circuits and blankers, each consisting of 512 rows and 512 columns. One individual blanking mechanism 40 drives one blanker 50. The input / output circuit 31 outputs data received from the deflection control circuit 130 to the cell array circuit 34. For example, the input / output circuit 31 has an input / output circuit 31a that outputs data to the individual blanking mechanism 40 arranged on one half side of the cell array circuit 34, and an input / output circuit 31b that outputs data to the individual blanking mechanism 40 arranged on the other half side.

[0035] The input / output circuit 31 is provided with a plurality of selectors 320 (demultiplexers). The selectors 320 receive irradiation time control data that defines the on / off state of each beam via the amplifiers 310, and output the data from the corresponding output lines. A plurality of individual blanking mechanisms 40 are connected in series to each output line.

[0036] The selector 320 has, for example, eight output lines row1 to row8, and each output line is connected to 256 individual blanking mechanisms 40. By arranging 64 selectors 320 in each of the input / output circuits 31a and 31b, irradiation time control data can be transferred to 512×512 individual blanking mechanisms 40 constituting the cell array circuit 34.

[0037] 6, the individual blanking mechanism 40 includes a shift register 41, a pre-buffer 42, a buffer 43, a data register 44, a NAND circuit 45, and an amplifier 46. The shift register 41 transfers data output from the shift register of the preceding cell to the shift register of the succeeding cell in accordance with a clock signal (SHIFT).

[0038] The pre-buffer 42 stores the irradiation time control data for the cell output from the shift register 41 in accordance with the clock signal (LOAD1).

[0039] The buffer 43 captures and holds the output value of the pre-buffer 42 in accordance with the clock signal (LOAD2).

[0040] The data register 44 captures and holds the output value of the buffer 43 in accordance with the clock signal (LOAD3).

[0041] An output signal of the data register 44 and a shot enable signal (SHOT_ENABLE) are input to the NAND circuit 45. An output signal of the NAND circuit 45 is provided to an electrode 51 of a blanker 50 via an amplifier 46 (driver amplifier).

[0042] When the output signal of the data register 44 and the shot enable signal are both high, the output of the NAND circuit 45 goes low, electrodes 51 and 52 have the same potential, the blanker 50 does not deflect the beam, and the beam is turned on. When at least one of the output signal of the data register 44 and the shot enable signal is low, the output of the NAND circuit 45 goes high, electrodes 51 and 52 have different potentials, the blanker 50 deflects the beam, and the beam is turned off.

[0043] The shot enable signal is input to the NAND circuits 45 of all the individual blanking mechanisms 40 , and by setting the shot enable signal to Low, all the beams can be turned off regardless of the output signal of the data register 44 .

[0044] When the shot enable signal is maintained at High, the beam is switched on / off by the output of the data register 44. That is, when the irradiation time control data is 1 (High), the beam is turned on, and when the irradiation time control data is 0 (Low), the beam is turned off.

[0045] In order to form a resist pattern with the desired dimensions by exposing the resist, it is necessary to control the amount of irradiation to an appropriate value. In other words, blanking control requires not only switching the beam on / off but also controlling the time when the beam is on. Since blanking control is performed by a logic circuit, i.e., a digital circuit, the irradiation time is expressed as an integer value. For example, the irradiation time T is divided by the quantization unit Δ to calculate the gradation irradiation time N. The quantization unit Δ can be set in various ways, but it is set so that the maximum irradiation time obtained by multiplying the maximum value of the gradation irradiation time N by the quantization unit Δ is greater than the maximum value of the irradiation time required for exposure. The quantization unit Δ is set to, for example, 1 ns.

[0046] In multi-beam writing, some of the beams are positioned inside the pattern, and other beams are positioned at the edge of the pattern or in an area without a pattern, so the irradiation time control data assigned to each beam of the multi-beam is not the same for all beams. Therefore, the irradiation time of each beam of the multi-beam needs to be controlled independently. As one embodiment of such control, a circuit of the blanking plate 204 shown in FIG. 5 is used to divide one shot into multiple irradiation steps of different lengths, and control is performed to turn on each beam for the irradiation steps required to obtain the desired exposure time for each beam. For this purpose, the gradation value N is converted into a binary number of n digits, and the length of the irradiation step corresponding to each digit of the binary number of n digits, that is, the time for turning on the beam, is set to the time obtained by multiplying the decimal number corresponding to each digit of the binary number by Δ.

[0047] For example, if the gradation exposure time N=50 and the number of digits n=8, the relationship expressed as the sum of decimal numbers of powers of 2 is 50=2 5 +2 4 +2 1 Therefore, the irradiation time control data is uniquely determined to be "00110010." Similarly, if the gradation irradiation time N=100 and the number of digits n=8, the irradiation time control data is "01100100."

[0048] The first lowest digit of the irradiation time control data corresponds to an irradiation step with an irradiation time of 1Δ. The second lowest digit of the irradiation time control data corresponds to an irradiation step with an irradiation time of 2Δ. The third lowest digit of the irradiation time control data corresponds to an irradiation step with an irradiation time of 4Δ. The fourth lowest digit of the irradiation time control data corresponds to an irradiation step with an irradiation time of 8Δ. The fifth lowest digit of the irradiation time control data corresponds to an irradiation step with an irradiation time of 16Δ. The sixth lowest digit of the irradiation time control data corresponds to an irradiation step with an irradiation time of 32Δ. The seventh lowest digit of the irradiation time control data corresponds to an irradiation step with an irradiation time of 64Δ. The eighth lowest digit of the irradiation time control data corresponds to an irradiation step with an irradiation time of 128Δ. In other words, one shot is divided into the same number of irradiation steps as the number of digits (number of bits) n of the irradiation time control data, and each irradiation step is Δ×2 k-1 The irradiation time is (k=1, 2, . . . , n).

[0049] Figure 7 shows an example of one shot irradiation step when the number of digits n=8 and the quantization unit Δ=1ns. In this example, each irradiation step is performed in order from the longest irradiation step. The first irradiation step has an irradiation time of 128ns. The second irradiation step has an irradiation time of 64ns. The third irradiation step has an irradiation time of 32ns. The fourth irradiation step has an irradiation time of 16ns. The fifth irradiation step has an irradiation time of 8ns. The sixth irradiation step has an irradiation time of 4ns. The seventh irradiation step has an irradiation time of 2ns. The eighth irradiation step has an irradiation time of 1ns. Note that the total irradiation time does not change even if the irradiation steps are swapped, so they may be performed in a different order than that shown in Figure 7.

[0050] When N=100, the irradiation time control data is “01100100”, and as shown in FIG. 8, the beam is controlled to be on in the second (64 ns), third (32 ns), and sixth (4 ns) irradiation steps, and the beam is controlled to be off in the first, fourth, fifth, seventh, and eighth irradiation steps.

[0051] As mentioned above, the required irradiation time is set for each beam of the multi-beam. Even if a certain gradation irradiation time, for example N=100, is set for one beam in a certain shot, a different gradation irradiation time, for example N=50, is set for another beam. Multiple irradiation steps in one shot cycle are performed simultaneously for all multi-beams, but each beam is independently controlled to be beam-on or beam-off in each irradiation step. In other words, the beam with a gradation irradiation time of N=100 and the beam with N=50 are controlled with different irradiation time data for beam-on and beam-off in each irradiation step, that is, with different combinations of beam-on and beam-off corresponding to each irradiation step. In this way, the irradiation time of each beam of the multi-beam in one shot cycle is independently controlled for each beam.

[0052] When electron beam irradiation is used for drawing, if a pattern is irradiated with a uniform dose, the pattern dimensions become thicker where the pattern density is high, which is a problem known as the proximity effect. This occurs because electrons that pass through the resist applied to the top surface of the substrate to be drawn are backscattered by the substrate and re-enter the resist, causing secondary resist exposure. To correct this proximity effect, a method is used in which the dose is corrected based on the pattern density around the beam irradiation position.

[0053] In this method, the dose is reduced as the pattern density increases so that the amount of exposure of the resist, that is, the sum of the primary exposure due to irradiation and the secondary exposure due to backscattering, remains constant regardless of the pattern density. As a result, the pattern dimensions can be made constant regardless of the pattern density. The secondary exposure due to backscattering is about half the amount of the primary exposure due to irradiation, so in this case, for a pattern with 100% density, half the dose is used as for a pattern with 0% density. The proximity effect correction dose D is given, for example, by the following formula:

[0054]

number

[0055] Here, D baseis the reference dose, η is the backscattering coefficient, and U is the average pattern density at the irradiation position.

[0056] In multi-beam lithography, especially in multi-beam lithography using many multi-beams, there may be cases where both high-density and low-density regions exist in the region where the multi-beams are lithographed. Therefore, in order to be able to lithograph even if there exists a region where the density is always zero, that is, a region where the irradiation time after the proximity effect correction is the longest, it is necessary to set the maximum shot time and the shot cycle time to a time longer than the irradiation time after the proximity effect correction in the region where the density is zero. In addition, to simplify the control, the stage speed that allows lithography in the shot cycle time in the region where the density is zero is calculated, and the stage speed is fixed at that speed, so-called constant stage speed running is often used.

[0057] As a result, in the case of a multi-beam exposing an area with low pattern density, the beam is on for most of the shot cycle, but in the case of a multi-beam exposing an area with high pattern density, the beam is on for only part of the shot cycle, and in particular in the case of a multi-beam exposing an area with 100% pattern density, the beam is controlled to be on for only about half the shot cycle and off for the rest of the time.

[0058] The feature of the multi-beam writing device is that the total current of the multi-beams is increased by using a large number of beams, and low-sensitivity resist can be written at high speed. On the other hand, there is a fundamental problem that the beam resolution and the resolution of the written pattern deteriorate due to the Coulomb effect when the total current of the multi-beams increases. In this embodiment, in order to alleviate the deterioration of the writing accuracy due to the Coulomb effect without reducing the writing speed, control is performed to shift the timing at which each beam of the multi-beams is turned on during the shot cycle. In addition, by utilizing the fact that the beam-on time during the shot cycle is short in the region with high proximity effect density as described above and that one shot is performed in multiple irradiation steps, blanking control of the multi-beams is performed so as to reduce the total current of the multi-beams in the on state without significantly extending the shot cycle or significantly reducing the writing speed.

[0059] In this embodiment, multiple (m) irradiation steps with an irradiation time of T1 and multiple irradiation steps with an irradiation time less than T1 are provided in one shot cycle. If the irradiation time of each irradiation step is proportional to a power of 2, irradiation time control becomes easier and is preferable, so an example of this case will be described. The irradiation steps with an irradiation time less than T1 have different irradiation times. Then, the multi-beams are divided into m groups, and the allocation order of the irradiation steps with an irradiation time of T1 is changed in each group, and irradiation is performed according to the allocation order. Group information indicating which group each beam belongs to is stored in the memory unit 140.

[0060] Such irradiation steps can be generated by dividing some of the irradiation steps from a set of multiple irradiation steps with irradiation times of powers of 2. For example, from the multiple irradiation steps shown in FIG. 7, the first irradiation step with an irradiation time of 128 ns is divided into two irradiation steps with an irradiation time of 64 ns. Although this division increases the number of irradiation steps, in order not to reduce the drawing speed, it is preferable to delete the irradiation step with the shortest irradiation time (irradiation time of 1 ns) so as not to increase the number of digits (number of bits) of the irradiation time control data. As a result, as shown in FIG. 9, the first to third irradiation steps have an irradiation time of 64 ns. The fourth irradiation step has an irradiation time of 32 ns. The fifth irradiation step has an irradiation time of 16 ns. The sixth irradiation step has an irradiation time of 8 ns. The seventh irradiation step has an irradiation time of 4 ns. The eighth irradiation step has an irradiation time of 2 ns. Since the irradiation step with the shortest irradiation time (irradiation time of 1 ns) has been deleted, the quantization unit Δ can be set to 2 ns, which is twice 1 ns.

[0061] Furthermore, in order to shift the timing of turning on the multi-beams, the multi-beams are classified into a plurality of groups, and blanking control is performed by a different method for each group. For example, in the example of Fig. 9, the multi-beams are classified into three groups A to C, and the beams of group A are assigned the first irradiation step as number 1, the second irradiation step as number 2, and the third irradiation step as number 3. The beams of group B are assigned the third irradiation step as number 1, the first irradiation step as number 2, and the second irradiation step as number 3. The beams of group C are assigned the second irradiation step as number 1, the third irradiation step as number 2, and the first irradiation step as number 3.

[0062] For example, if the exposure time for one shot is 80ns, the quantization unit is 2ns, so the gradation exposure time is 40. The relationship expressed as the sum of decimal numbers that are powers of 2, 40=2 5 +2 3 From the relationship, in FIG. 9, if exposure is performed in one of the three irradiation steps with an irradiation time of 64 ns and in an irradiation step with an irradiation time of 16 ns, the desired exposure time of 80 ns can be obtained. Here, when the allocation order is set as described above, the beams of group A are turned on in the first and fifth irradiation steps as shown in FIG. 10. The beams of group B are turned on in the third and fifth irradiation steps. The beams of group C are turned on in the second and fifth irradiation steps. Since the beams of groups A to C are turned on at different timings for the irradiation steps with an irradiation time of 64 ns, the current amount of the on beam can be reduced.

[0063] For example, if the exposure time for one shot is 180 ns, the quantization unit is 2 ns, so the gradation exposure time is 90. 90=2 6 +2 4 +2 3 +2 1From the relationship above, in one shot cycle, the beam is turned on in two of the three irradiation steps with an irradiation time of 64 ns in FIG. 9, and in the irradiation steps with irradiation times of 32 ns, 16 ns, and 4 ns. When the allocation order is set as described above, the beam of group A is turned on in the first, second, fourth, fifth, and seventh irradiation steps, as shown in FIG. 11. The beam of group B is turned on in the first, third, fourth, fifth, and seventh irradiation steps. The beam of group C is turned on in the second, third, fourth, fifth, and seventh irradiation steps. For the irradiation step with an irradiation time of 64 ns, only two groups are turned on at the same time, and all three groups are not turned on at the same time, so the amount of current of the on beam can be reduced.

[0064] Specifically, when all the beams of the multi-beam are irradiated for an irradiation time of 180 ns, the average total current of the on-beam during the shot cycle can be reduced to a level of (64×2+64×2+64×2+32×3+16×3+4×3) / (64+64+64+32+16+4) / 3=73%, which is the average value obtained by weighting the number of beams turned on in each irradiation step by the irradiation time of each irradiation step. In the example of FIG. 11, when the irradiation time after the proximity effect correction in the area with the pattern density of 0 is set to 250 ns, the irradiation time in the area with the pattern density of 40% is about 180 ns. In other words, in the area with the pattern density of 40% or more, the number of the multi-beams controlled by the beam on and the current amount of the multi-beams controlled by the beam on are large because the pattern density is high, but in this embodiment, the average value of the total current of the on-beam during the shot cycle can be reduced by 30%.

[0065] In this way, by setting the allocation order of the irradiation steps to be different for each group, the timing at which the multi-beam is turned on can be shifted between groups, lowering the average value of the total current of the on-beam of the multi-beam during a shot cycle, reducing the influence of the Coulomb effect, and improving the writing accuracy. Also, as shown in Figures 7 and 9, the number of irradiation steps constituting one shot and the total irradiation time of each irradiation step do not change, so the shot cycle and writing time do not extend. In other words, the average value of the total current of the on-beam of the multi-beam during a shot cycle can be lowered without extending the writing time.

[0066] An example of grouping multiple beams into groups A to C is shown in Figs. 12(a) to 12(c). For convenience of explanation, Figs. 12(a) to 12(c) show an 8x8 beam array. It is preferable to group the beams so that there is no bias in the beam current of the on-beam. Therefore, rather than the grouping shown in Fig. 12(c), it is more preferable to group the beams in groups A to C so that they are arranged in order, as shown in Figs. 12(a) and 12(b). In Fig. 12(a), beams adjacent in the x and y directions belong to different groups. In Fig. 12(b), beams adjacent in the x direction belong to different groups.

[0067] Next, the pattern writing method according to this embodiment will be described with reference to the flowchart shown in Fig. 13. In the pattern area density calculation step (step S1), the area density calculation unit 111 virtually divides the writing area of ​​the substrate 101 into a plurality of mesh areas. The size of the mesh area is, for example, approximately the same as the size of each beam constituting the multi-beam, and each mesh area becomes a pixel (unit irradiation area). The area density calculation unit 111 reads out the writing data from the storage unit 140, and calculates the pattern area density ρ of each pixel using a pattern defined in the writing data.

[0068] In the irradiation time calculation step (step S2), the irradiation time calculation unit 112 multiplies the pattern area density ρ by the reference irradiation amount D0 and a correction coefficient for correcting the proximity effect, etc. to calculate the irradiation amount of the beam irradiated to each pixel. The irradiation time calculation unit 112 divides the irradiation amount by the current density to calculate the irradiation time.

[0069] In the irradiation time control data generation step (step S3), the data processing unit 113 distributes the irradiation time to a plurality of irradiation steps while taking into consideration the allocation order of the irradiation steps, and generates irradiation time control data. For example, the data processing unit 113 divides the irradiation time by the quantization unit to calculate a gradation value t (irradiation time converted into an integer). In the example shown in FIG. 9, the data processing unit 113 calculates the irradiation time control data from a sequence T k (2 5 , 2 5 , 2 5 , 2 4 , 2 3 , 2 2 , 2 1 , 2 0 ) for each T k ON / OFF flag column b corresponding to k Find (k=1,2,···,7).

[0070] The upper bits b1, b2, and b3 correspond to the time T M =T1=T2=T3=2 5 m=FLOOR(t / T M ) and the order of allocation of irradiation steps for each group. However, m is a number not exceeding 3. The lower bits b4, b5, b6, b7, and b8 are integer values ​​tm × T M is converted to binary. More generally, the sequence T k Even if some of the data are not powers of two, the dose control data b can be calculated from t using the following formula. k It is possible to obtain

[0071]

number

[0072] If the quantization unit Δ of the irradiation time is determined so that the irradiation time is shorter than the sum of the irradiation times of each irradiation step, the bit string b k is decided.

[0073] In the data transfer process (step S4), the drawing control unit 114 outputs the irradiation time control data to the deflection control circuit 130. The deflection control circuit 130 outputs the irradiation time control data to the blanking aperture array substrate 204. The input / output circuit 31 of the blanking aperture array substrate 204 transfers the irradiation time control data to the corresponding individual blanking mechanism 40.

[0074] In the individual blanking mechanism 40, the irradiation time control data is transferred one bit at a time to the subsequent buffer in accordance with the clock signal. Data corresponding to the irradiation time of each irradiation step is transferred, and the beam in each irradiation step is switched on / off in accordance with the irradiation time control data stored in the final stage buffer 44. When one irradiation step is performed, the irradiation time control data for the next irradiation step is transferred to each buffer and the final stage buffer 44. In this way, each beam is switched on / off for each irradiation step in accordance with the irradiation time control data for each beam in each irradiation step.

[0075] In the drawing step (step S5), the drawing control unit 114 controls the drawing unit 150 to execute drawing processing. The drawing control unit 114 controls the deflector 208 using the deflection control circuit 130 to position the multi-beam so that each beam of the multi-beam irradiates each pixel corresponding to the transferred data. After the positioning is completed, the drawing control unit performs an irradiation step using the deflection control circuit 130, and performs blanking control so that each beam of the multi-beam irradiates the corresponding pixel with a predetermined irradiation amount. When the irradiation step constituting one shot is completed, a data transfer step and a drawing step are performed to irradiate the next set of pixels.

[0076] In this manner, by repeating each step in FIG. 13, the drawing unit 150 performs a drawing operation for drawing on a drawing area by a raster scan method using multiple beams.

[0077] Fig. 14 is a conceptual diagram for explaining the drawing operation. As shown in Fig. 14, a drawing region 80 on a substrate 101 is virtually divided into a plurality of rectangular stripe regions 82 with a predetermined width in the y direction (first direction). First, the XY stage 105 is moved and adjusted so that an irradiation region 84 that can be irradiated with one multi-beam irradiation is positioned at the left end of the first stripe region 82, and drawing is started.

[0078] When drawing the first stripe region 82, the XY stage 105 is moved in the -x direction to relatively advance drawing in the +x direction. During drawing, the multibeam is deflected by the deflector at least in the Y direction, thereby switching the pixels to be exposed by the multibeam. By repeating the switching of the exposure pixels and the exposure operation, the multibeam exposes all the pixels defined in the stripe region 82. The XY stage 105 is moved continuously at a predetermined speed. At this time, the stage speed is set within a range in which the above-mentioned drawing operation is possible. After the drawing of the first stripe region 82 is completed, the stage position is moved in the -y direction and adjusted so that the beam array 84 is located at the right end of the second stripe region 82. Next, drawing is performed in the -x direction by moving the XY stage 105 in the +x direction.

[0079] In the third stripe region 82, writing is performed in the +x direction, and in the fourth stripe region 82, writing is performed in the -x direction. Writing time can be reduced by alternately changing the direction. Each stripe region 82 may be written in the same direction.

[0080] In this embodiment, multiple irradiation steps with the same irradiation time are provided in one shot cycle without changing the number of irradiation steps, and the multi-beams are classified into multiple groups, and the order of allocation of the irradiation steps is changed between the groups. As a result, the timing at which the beam is controlled to be on is shifted between the multiple groups of the multi-beams in at least some of the irradiation steps in one shot cycle, so that the average total current of the on-beam during the shot cycle can be reduced, the influence of the Coulomb effect can be reduced, and the writing accuracy can be improved.

[0081] In this embodiment, the number of groups into which the multi-beams are grouped may be set to 2, instead of 3 in Fig. 9. In this case, groups A and B are set, and the beams of group A are assigned the order of the first irradiation step as 1, the order of the second irradiation step as 2, and the order of the third irradiation step as 3, while the beams of group B are assigned the order of the third irradiation step as 1, the order of the second irradiation step as 2, and the order of the third irradiation step as 3. In the case of an irradiation time in which one or two of the first, second, and third irradiation steps are used, the timing of part of the irradiation steps of group A and group B is controlled to be shifted.

[0082] The number of bits of the irradiation time control data may be increased to increase the number of irradiation steps for the same irradiation time, thereby efficiently shifting the timing at which the beam is turned on.

[0083] In this embodiment, for example, as shown in FIG. 15, the number of digits (number of bits) of the irradiation time control data is increased from 8 digits in FIG. 7 to 10 digits, the quantization unit Δ is set to 2 ns, the first irradiation step is irradiated for 64 ns, and the second to sixth irradiation steps are irradiated for 32 ns. The seventh irradiation step is irradiated for 16 ns. The eighth irradiation step is irradiated for 8 ns. The ninth irradiation step is irradiated for 4 ns. The tenth irradiation step is irradiated for 2 ns. These ten irradiation steps correspond to the irradiation steps obtained by decomposing the first irradiation step with an irradiation time of 128 ns in the eight irradiation steps in FIG. 7 into four irradiation steps with an irradiation time of 32 ns. Therefore, the total value of the irradiation time of each irradiation step in FIG. 15 is 255 ns, which is the same as FIG. 7. However, since the number of drawing steps increases and the amount of irradiation time data transferred to the blanking aperture during drawing increases, the overhead time of drawing, such as the data transfer time, increases, and the drawing time becomes longer than when the irradiation time control data is eight digits.

[0084] In this embodiment, the multi-beams are classified into two groups A and B, and the allocation order of the second to sixth irradiation steps is different between group A and group B. For example, for the beams of group A, the allocation order of the second irradiation step is number 1, the allocation order of the third irradiation step is number 2, the allocation order of the fourth irradiation step is number 3, the allocation order of the fifth irradiation step is number 4, and the allocation order of the sixth irradiation step is number 5. For the beams of group B, the allocation order of the sixth irradiation step is number 1, the allocation order of the fifth irradiation step is number 2, the allocation order of the fourth irradiation step is number 3, the allocation order of the third irradiation step is number 4, and the allocation order of the second irradiation step is number 5.

[0085] For example, if the irradiation time of one shot is 180 ns, within one shot cycle, the beam is on for an irradiation step with an irradiation time of 64 ns, three irradiation steps with an irradiation time of 32 ns, an irradiation step with an irradiation time of 16 ns, and an irradiation step with an irradiation time of 4 ns.

[0086] When the irradiation step allocation order is set as described above, the beam of group A is turned on in the first to fourth irradiation steps, the seventh irradiation step, and the ninth irradiation step, as shown in FIG. 16. The beam of group B is turned on in the first irradiation step, the fourth to seventh irradiation steps, and the ninth irradiation step. The beams of groups A and B are turned on simultaneously only once for the irradiation step with an irradiation time of 32 ns, and are turned on at different timings otherwise, so that the current amount of the on-beam can be reduced. In this example, the number of groups into which the multi-beams are grouped is set to 2, but it may be set to 3. For example, by adding group C to groups A and B in FIG. 15, the beam of group C may be assigned as 4th in the second irradiation step, 5th in the third irradiation step, 1st in the fourth irradiation step, 2nd in the fifth irradiation step, and 3rd in the sixth irradiation step.

[0087] The number of bits of the irradiation time control data may be increased and the shot cycle may be extended to shift the timing at which the beam is turned on more efficiently.

[0088] For example, as shown in FIG. 17, the number of digits (number of bits) of the irradiation time control data is set to 14, the quantization unit Δ is set to 2 ns, and the first to sixth irradiation steps are set to irradiation of 32 ns. The seventh and eleventh irradiation steps are set to irradiation of 16 ns. The eighth and twelfth irradiation steps are set to irradiation of 8 ns. The ninth and thirteenth irradiation steps are set to irradiation of 4 ns. The tenth and fourteenth irradiation steps are set to irradiation of 2 ns. That is, two sets of four irradiation steps with irradiation times of 16 ns, 8 ns, 4 ns, and 2 ns are provided. Therefore, the total value of the irradiation time of each irradiation step is increased by 14 ns from 255 ns in FIG. 7. That is, the shot cycle becomes 14 ns longer, and the writing time is extended.

[0089] The multi-beams are divided into two groups, A and B, and the allocation order of the 1st to 6th irradiation steps is different between group A and group B. For example, for the beams in group A, the allocation order of the 1st irradiation step is number 1, the allocation order of the 2nd irradiation step is number 2, the allocation order of the 3rd irradiation step is number 3, the allocation order of the 4th irradiation step is number 4, the allocation order of the 5th irradiation step is number 5, and the allocation order of the 6th irradiation step is number 6. For the beams in group B, the allocation order of the 6th irradiation step is number 1, the allocation order of the 5th irradiation step is number 2, the priority allocation order of the 4th irradiation step is number 3, the allocation order of the 3rd irradiation step is number 4, the allocation order of the 2nd irradiation step is number 5, and the allocation order of the 1st irradiation step is number 6.

[0090] In addition, the set of the 7th to 10th irradiation steps is used only by group A, and the set of the 11th to 14th irradiation steps is used only by group B.

[0091] For example, if the irradiation time of one shot is 140 ns, within one shot cycle, the beam is turned on for four irradiation steps with an irradiation time of 32 ns, one irradiation step with an irradiation time of 8 ns, and one irradiation step with an irradiation time of 4 ns.

[0092] When the irradiation steps are assigned in the order described above, the beams of group A are turned on in the first to fourth irradiation steps, the eighth irradiation step, and the ninth irradiation step, as shown in Fig. 18. The beams of group B are turned on in the third to sixth irradiation steps, the twelfth irradiation step, and the thirteenth irradiation step.

[0093] The beams of groups A and B are turned on simultaneously only twice for the irradiation step with an irradiation time of 32 ns, and are turned on at different times otherwise, so that the average value of the total current of the on-beam during a shot cycle can be reduced. As can be seen from Figures 16 and 18, the irradiation step in Figure 18 has a set of irradiation steps with short irradiation times for each of groups A and B, so that the time when the beams of group A and group B are turned on simultaneously can be shortened more efficiently than the irradiation step in Figure 16.

[0094] Figure 19 shows the average value of the total on-beam current during a shot cycle when the beam on-timing is shifted and when it is not shifted as shown in Figures 9, 15, and 17. It was confirmed that the average value of the total on-beam current during a shot cycle is reduced by shifting the beam on-timing.

[0095] It was also confirmed that the higher the pattern areal density, the greater the reduction in the average value of the total on-beam current during a shot cycle. This is because the higher the pattern areal density, the smaller the dose due to proximity effect correction, and the more irradiation steps in which the beam is turned off within one shot cycle are increased, allowing the beam on timing to be shifted more efficiently between irradiation steps.

[0096] In the above embodiment, the beam shaping and blanking control may be performed by the apertures of the blanking aperture array substrate, and the shaping aperture array member may be eliminated.

[0097] The present invention is not limited to the above-described embodiment, and the components can be modified and embodied in the implementation stage without departing from the gist of the invention. In addition, various inventions can be formed by appropriately combining the multiple components disclosed in the above-described embodiment. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, components from different embodiments may be appropriately combined. [Explanation of symbols]

[0098] 40 Individual Blanking Mechanism 50 Blanka 100 drawing device 110 Control computer 111 Area density calculation section 112 Irradiation time calculation unit 113 Data Processing Unit 114 Drawing control section

Claims

1. emitting multiple beams of charged particles; grouping the beams constituting the multi-beam into a plurality of beam groups; determining an irradiation time for each beam of each shot of the multi-beam from drawing pattern data; a shot division process for dividing each of the shots into a plurality of irradiation steps including irradiation steps with different irradiation times, and selecting a set of irradiation steps in which a plurality of irradiation steps having a predetermined first irradiation time are grouped as a first group among a part of the divided irradiation steps, and a total of the irradiation times of the irradiation steps is the predetermined irradiation time for each of the beams; performing a multi-beam shot by switching each beam of the multi-beam on and off to execute the set of the plurality of irradiation steps; Equipped with a multi-charged particle beam writing method, characterized in that, in the shot division process, timing of the irradiation steps of the first group of the multiple irradiation steps belonging to each of the multiple beam groups is determined based on an allocation order that is predetermined to be different for each beam group and the number of the irradiation steps that are turned on.

2. 2. The multi-charged particle beam writing method according to claim 1, wherein each of the plurality of irradiation steps includes the first group and a second group of irradiation steps each having an irradiation time shorter than the first irradiation time.

3. 3. The multi-charged particle beam writing method according to claim 2, wherein the first irradiation time is the longest irradiation time among the irradiation times of the plurality of irradiation steps.

4. 2. The multi-charged particle beam writing method according to claim 1, wherein the irradiation times of the respective irradiation steps are determined to be proportional to a power of two.

5. 3. The multi-charged particle beam writing method according to claim 2, wherein, in the second group, beam-on timings of a plurality of irradiation steps in the second irradiation time differ for each of the beam groups.

6. 3. The multi-charged particle beam writing method according to claim 1, wherein in the multiple beams, beams adjacent to each other in the writing progress direction or in a direction linearly independent of the writing progress direction are classified into different groups.

7. an emission section for emitting multiple beams of charged particles; a writing control unit that switches on and off each of the multi-beams to control an irradiation time of each of the multi-beams in each shot determined from writing pattern data; Equipped with the drawing control unit groups the multiple beams constituting the multi-beam into multiple beam groups, divides each of the shots into a multiple irradiation steps including irradiation steps with different irradiation times, groups the multiple irradiation steps having a predetermined first irradiation time among some of the divided multiple irradiation steps into a first group, selects a set of irradiation steps whose total irradiation time of each irradiation step is the predetermined irradiation time for each of the beams, and determines the timing of the irradiation step of the first group of the multiple irradiation steps belonging to each of the multiple beam groups based on an allocation order that is predetermined to be different for each beam group and the number of the irradiation steps that are turned on.

Images