Method and device for producing curved lines on an irradiation sensitive resist

[0001] [Technical field to which the invention belongs] The present invention relates to the field of electron beam lithography, and in this case, in particular, a method and an apparatus for forming a line drawing (line stroke) by an electron beam on a substrate coated with a photosensitive resist (drawing a line). An electron beam is directed toward the surface of the coated substrate in the direction of the Z coordinate, and the substrate is gradually slid in the direction of the X coordinate and / or the Y coordinate of the Cartesian coordinate system (raster). Regarding the form that acts on the resist. [0002] [Conventional technology] The electron beam lithography apparatus includes a light source that emits a focused electron beam directed at a substrate coated with a photosensitive resist. The substrate is the surface of the substrate that is oriented at right angles to the electron beam and supports the resist layer, and is placed and held on a table that is slidable along the Cartesian coordinate system. By a predetermined table sliding, electromagnetic deflection of the electron beam, or simultaneous operation of both, at this time, the collision point of the electron beam defined with respect to the substrate surface is changed, and the resist has a geometric structure or pattern, particularly. , Draw a line. [0003] This principle of causing the substrate and the electron beam to move relative to each other by sliding the substrate table with respect to at least an almost immobile electron beam is known from, for example, Microelectronic Engineering 27 (1995) P135-138. In the case of another method of appropriately changing the collision point on the substrate surface, the electron beam is deflected by, for example, an electron beam deflection system. [0004] Since this type of device for positioning a substrate table or electron beam generally utilizes a Cartesian coordinate system having X and Y coordinates, the geometric data to be constructed on the substrate is also converted to Cartesian coordinates. It is preferable to describe. This is suitable for describing structures that are primarily rectangular or trapezoidal in shape, as is typically used in microscopes. The straight line drawing that generally appears in this case can be transferred (transferred) to the substrate while maintaining a slight edge roughness by using a positioning device that is controlled stepwise. [0005] However, there is great interest in drawing curved lines with high precision so that, for example, elliptical grids, circular grids and curved transmitters can be manufactured, especially in the case of substrate exposure considered for optical applications. Be held. This type of delineation can be approximated with the accuracy of unit increments of the X, Y coordinates of the appropriate positioning device anyway, using the Cartesian raster. If the unit increment value is chosen small enough for the desired width of the line, the edge roughness due to the stepwise approximation or the shape deviation of the line that is nearly approximate to the theoretically ideal line can be ignored. [0006] The construction of a certain line segment on the substrate is performed by closely juxtaposing a large number of irradiation points based on the table displacement of the incremental stroke. In the case of irradiation at one point, the electron beam is directed to that point for a short residence time until the desired energy input is achieved. Then, move to the next point and operate in the same manner. In this case, it is not always necessary to block the electron beam between the following adjacent points. Depending on the magnitude of the energy input at the collision point of the electron beam on the resist, small or large exposed "points" are generated on the resist. [0007] Therefore, when constructing a line drawing that is extremely thin but as uniform as possible, the input energy applied to the photosensitive resist along the line drawing, that is, the effective dose (charge per area) or the effective line dose (per length). It is extremely important to keep the charge) sufficiently constant. The width variation that sometimes appears in the drawing is different from the edge roughness by shortening the unit stroke when positioning the substrate with respect to the electron beam.<u style="single">Is</u>Exclusion<u style="single">Can not</u>.. [0008] [Problems to be Solved by the Invention] A positioning device that is operated stepwise (stepwise) at the X and Y coordinates feeds only one coordinate of the X and Y coordinates by a unit increment, and a unit stroke parallel to the axis and feeds to both the X and Y coordinates. It is possible to carry out a unit process. In the latter case, it will be clear that the distance between the start and end points of the unit stroke is larger than in the case of the unit stroke parallel to the axis. If both coordinates, i.e. the unit strokes of the X and Y coordinates, are equal, the length increases by 2 times. [0009] That is, when the same energy input is performed after each unit stroke, the dose in the X coordinate direction or the Y coordinate direction is 45 ° with respect to the X coordinate or the Y coordinate for the diagonal line drawing for the unit stroke parallel to the axis. Greater than the dose in the direction of a straight line tilted at an angle. This different energy input results in a line width deviation. That is, the diagonal line has a smaller width than the line parallel to the axis. [0010] On the other hand, on the contrary, when the input energy is set to be 2 times larger, an excessive dose is generated with respect to the straight line formed parallel to the axis. This is because the path length is shorter than that of a sloping straight line. From the quantitative calculation, it is known that in this example of consideration, the deviation is maximum at an angle of 22.5 ° and is about 8%. Theoretically, the deviation can be reduced by adjusting the input energy with respect to the straight line depending on the angle. Of course, this would lead to a significant increase in computing work. [0011] In the case of curved lines, especially in the case of circles, the slope of the path changes steadily. The separate calculation of input energy after each unit stroke would result in a huge increase in computational work, thus drastically reducing the speed of work when drawing curved lines. Therefore, this mode of operation, which determines the exact input energy depending on the actual slope of the delineation after each unit stroke, is largely unsatisfactory with respect to efficient manufacturing. On the other hand, in the case of a circular lattice, the dose fluctuations resulting from the approximation are particularly remarkable, and therefore, in this case, the creation of auxiliary means is of interest. [0012] Therefore, an object of the present invention is to improve a known electron beam lithography method so that a curved line with a uniform line width can be constructed with a small amount of calculation work. [0013] [Means for solving problems]<u style="single">Book</u>Of invention<u style="single">First</u>Based on the viewpoint<u style="single">The following method is provided in which a curved line is formed by an electron beam on a substrate coated with a photosensitive resist. In this method, an electron beam is directed toward the surface of the coated substrate in the direction of the Z coordinate of the Cartesian coordinate system, and the substrate is slid stepwise in the direction of the X coordinate and / or the Y coordinate of the Cartesian coordinate system. During this time, a predetermined effective dose of electron beam is allowed to act on the resist.</u>After each unit stroke<u style="single">、</u>Depends on the line drawing transition obtained from multiple preceding unit strokes<u style="single">、</u>Configuration<u style="single">Using the arithmetic circuit connected to the intermediate storage device, the effective dose to be determined is (a) determined according to the stroke width of the unit increment, and the category for the unit stroke is determined, and (b) the category corresponds to the category. By assigning a correction coefficient proportional to the stroke width of the unit increment to the unit stroke, and (c) forming an average correction coefficient from the correction coefficients for multiple preceding unit strokes stored in the intermediate storage device.</u>Determine. [0014] Thus, an energy input that is well adapted to the draw transition can be achieved for the resist and thus an arbitrarily curved draw with a very uniform line width can be constructed. By considering a plurality of preceding unit strokes (steps), grasped or calculated information regarding the delineation transition is used to determine the input energy after the last positioning unit stroke. The computational work required for this is far less than the case where the theoretically accurate input energy is calculated separately from the description of the transition of the delineation in relation to the execution of the unit stroke. The history of each process can be taken into consideration with almost no reduction in work speed. [0015]<u style="single">line</u>The basic principle of the pulling method is to define a group of categories of unit strokes with specific correction coefficients to obtain the input energy in the subsequent unit strokes. In this case, in the categories of various unit processes<u style="single">To the associated correction factor</u>On the other hand, the shortest distance between the starting point and the final point<u style="single">Sutra</u>It is suitable for the purpose if the weights are weighted according to the road length and the unit processes with the same weight are associated with one category. [0016] BEST MODE FOR CARRYING OUT THE INVENTION Of this method<u style="single">Good</u>Better embodiment<u style="single">In</u>, The first category A of the axis-parallel unit strokes in the form of unit increments in the direction of the X and Y coordinates, and the form of the unit increments in the X and Y coordinates, respectively. Use with the second category B of the diagonal unit stroke of. [0017] A correction factor of 1 is assigned to the unit stroke of category A, and 2 is assigned to the unit stroke of category B. The input energy of the next unit stroke is determined based on the unit stroke already performed in categories A and B based on the present embodiment. By orienting the unit increments to the X and Y coordinates, and also by reducing the number of different types of unit strokes, the control and arithmetic work when using this method is minimal. Moreover, an extremely uniform transition of the effective dose and an extremely uniform transition of the line width along the arbitrarily curved path curve are achieved regardless of the inclination angle of the path curve. [0018] For example, claim 3 takes into account the eight immediately preceding unit strokes and depends on the number of unit strokes in categories A and B.<u style="single">To</u>When the average correction coefficient Kn is determined in response to the disclosure relevance provision, the calculation work is particularly small. [0019] However, if necessary, it can be easily realized that the number of unit steps (steps) considered retroactively is reduced or increased in order to further improve the accuracy. Of course, as the number of unit strokes to consider increases, so does the hardware and arithmetic work. Further, by returning the unit stroke, the error of the average correction coefficient can be obtained. To take this issue into account, it has proved particularly advantageous to consider the eight preceding unit strokes. It is preferable to input energy in proportion to the calculated average correction coefficient Kn. [0020] If the electron beam output is kept constant, the input energy can be adjusted in proportion to the correction coefficient Kn via the action time of the electron beam for each unit stroke. This operation method has an advantage that a field emission cathode characterized by extremely uniform electron beam output can be used as an electron beam source in steady operation. The action time is set by the holding time of the table positioning device after the unit stroke. Therefore, the time interval for fixing the position of the table after the unit stroke performed is proportional to the average correction coefficient Kn. [0021] [0021] In the case of other embodiments with respect to the above embodiment, the action time or the holding time is kept constant, and energy is input by changing the electron beam output. In this case, for the unit stroke, the electron beam output is changed or newly set in proportion to the correction coefficient Kn. In this case, there is an advantage that the table can be slid in a constant work cycle. When a field emission cathode is used as the electron beam source, of course, the purpose is to not change the electron beam output during operation. This is because the operation of adapting the electron beam output to the change of the applied voltage must be performed extremely slowly. When controlling the input energy, the electron beam radiated from the electron beam source is sequentially weakened by using, for example, an air core coil. To this extent, the first embodiment in which the holding time is changed is preferable. [0022] The first of the present invention<u style="single">2</u>From this point of view, the subject of the present invention is further<u style="single">Cartesian coordinate system</u>Z coordinate<u style="single">Oriented to</u>Electron beam<u style="single">For</u>Light source and<u style="single">Cartesian coordinate system</u>In the XY plane<u style="single">For each unit stroke</u>A feeder that can slide and forms a relative motion between the substrate and the electron beam,<u style="single">An intermediate storage device formed to store a predetermined number of correction factors assigned to a unit stroke category, in which the correction factors are proportional to the stroke width of the unit increment and are in the stroke width. The categories are different accordingly, and are the effective doses of electron beams acting on the resist to determine the average correction factor from the correction factors stored in the intermediate storage for the preceding unit process, respectively. For the unit process of</u>Depends on the line drawing transition obtained from multiple preceding unit strokes<u style="single">, Connected to an intermediate storage device and a light source trigger circuit to determine the effective dose to be set in proportion to the average correction factor.</u>Including arithmetic circuit<u style="single">I'm</u>On a substrate coated with a photosensitive resist<u style="single">Curved</u>Wire drawing (<u style="single">Curved</u>Line stroke)<u style="single">Electron beam lithography</u>Solved by the device. [0023] According to the apparatus according to the present invention, the above method can be applied to apply energy to the resist in accordance with the transition of the drawing, thus forming an arbitrarily curved drawing having a uniform line width. [0024] [Example] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.<u style="single">It should be noted that the drawing reference reference numerals added to the scope of claims of the present application are solely for the purpose of facilitating the understanding of the present invention, and do not limit the present invention to the following examples. ..</u>[0025] FIG. 1 shows an embodiment of the electron beam lithography apparatus 1. The device includes a table 2 that holds a flat substrate 4 that can move along the X, Y coordinates and is fixed in the XY plane. Substrate 4 comprises a photosensitive material or at least one coating layer made of such a material. [0026] Table 2 includes a feeder 3. The feeder 3 allows the table to travel stepwise (step-by-step) to the X and Y coordinates. The smallest feed unit in the direction of both X and Y coordinates, that is, the unit increment of the feeder 3, is approximately in the range of 1 nm to 10 nm for each of the X and Y coordinates, in which case the feed rate is approximately. It is 1 mm / s. For example, a 2.5 nm unit increment can be used for both coordinates. [0027] The electron beam lithography apparatus further includes an electron beam source 5 that directs a constant density focused electron beam 6 toward the table 2 or the substrate 4. The electron beam 6 acts in the direction of the Z coordinate perpendicular to the XY plane with respect to the substrate 4. In this case, it is preferable to use a thermal field emission cathode characterized by constant electron beam density in the electron beam as the electron beam source 5. [0028] An electromagnetic deflection device 7 capable of positioning the collision point of the electron beam 6 on the surface of the substrate 4 is provided in the subsequent stage (downstream) of the electron radiation source 5. In this case, the deflector 7 can deflect the electron beam 6 in the direction of the X coordinate and the direction of the Y coordinate, in which case the orientation can be performed digitally, eg, with 16-bit resolution. FIG. 1 shows an electron beam 6 in a non-biased orientation state perpendicular to the surface of the substrate 4 or the table 2. [0029] When drawing one line on the surface of the substrate 4, the table 2 is passed under the position-invariant electron beam 6 in the first approximation corresponding to a predetermined path curve. The deviation of the actual position of Table 2 with respect to a predetermined position is measured, for example, with a laser interferometer and dynamically compensated by the corresponding deflection of the electron beam 6. In this case, the deflection amount of the electron beam 6 is on the order of a maximum of about 20 μm. [0030] Further, first, a line (curve) to be configured on the substrate 4 is determined, and then an arithmetic circuit for converting the feed command of the feeder 3 and the setting of the input energy to the substrate 4 is provided. The desired delineation can be defined by a number of small unit strokes (steps), or from the origin of the curve in Cartesian coordinates. The high resolution of the nanometer range of delineation, of course, results in an extremely large amount of data that requires a lot of processing work. In this case, the speed at which the line is drawn on the substrate 4 is relatively low in the case of the curved line. [0031] Therefore, in the case of this case, it is preferable that the delineation is approximated by the curved portion of Bezier. In this case, the target position set in the table 2 can be calculated from a small number of characteristic quantities of the curved portion of the Bezier, corresponding to the stroke width for each stroke of the feeder 3. Thus, the data range can be significantly reduced and the speed at which the line is drawn can be increased, especially for curved lines. A suitable algorithm for calculating the unit stroke from the Bezier curve is disclosed in German Patent Publication No. 4244462 (which is incorporated herein by reference if necessary) and is therefore not detailed here. .. [0032] It is preferable to use the smallest unit stroke of the feeder 3 unit stroke for particularly high resolution and close approximation in the line drawn defined by the Bezier characteristic quantity. As mentioned above, the feeder can move stepwise along the X and Y coordinates. Thus, as the smallest unit stroke, table sliding is obtained in which the unit increment in the coordinate direction is performed. Thus, there are four possibilities for the two X and Y coordinates. [0033] In addition, unit increments of sliding at both X and Y coordinates can be performed simultaneously, thus resulting in four more diagonal unit strokes. Figure 2 shows the directions of a total of eight possible unit strokes in this operating method. Under the precondition that the unit increments of the X and Y coordinates are equal, the diagonal unit stroke has a path length 2 times that of the axially parallel unit stroke. [0034] The direction of the unit stroke shown in FIG. 2 forms the basic element used for the line drawing approximation in this case. The above basic elements can be classified into two categories according to the path length of the unit stroke. In this case, one category A is associated with a shorter unit stroke, that is, a unit stroke parallel to the axis, and the category B is associated with the unit stroke. , Diagonal unit strokes are associated. [0035] However, basically, other types of unit strokes can also be defined, for example, by defining (defining) such strokes by multiple unit increments in one or both coordinate directions. In this case, the process is also categorized according to its theoretical path length. In addition, unit increments of different lengths can be used at both X and Y coordinates. [0036] When drawing a line on the substrate 4, the unit stroke to be performed by the table 2 is continuously calculated online from the Bezier characteristic quantity (value) and transmitted via the positioning device, thus being almost continuous. Table movement occurs. For the unit stroke calculated in this way, it is confirmed whether this unit stroke belongs to category A or category B. According to this attribution (relevance), the correction coefficient is associated with the relevant unit stroke. In this case, each correction factor is proportional to the shortest path length of the unit stroke. In the case of the embodiment described here, the correction factor 1 is associated with the unit stroke of category A, and the correction factor 2 is associated with the unit stroke in the diagonal direction of category B based on the standardization. [0037] The obtained correction values are filed in an intermediate storage device in the form of a shift register capable of intermediate storage of a predetermined number of correction values. With the storage of new correction values, each old value stored is removed from the intermediate storage. Thus, for example, eight new correction factors are always retained. [0038] The above correction factors allow the slope of the actual path of the line to be approximated. In this application, therefore, from the correction factors filed in the intermediate storage, an average value is formed as the mean correction factor Kn, which gives more accurate information about the slope of the route. To find the average correction factor Kn, find the number distributed to category A and the number distributed to category B out of the eight new unit strokes. The mean correction factor Kn is associated with each of the nine possible combinations in this case, so the following relationship arises: n Number of unit strokes in category A Number of unit strokes in category B Average correction coefficient Kn 1 8 0 1.000 2 7 1 1.008 3 6 2 1.031 4 5 3 1.068 5 4 4 1.118 6 3 5 1.179 7 2 6 1.250 8 1 7 1.329 9 0 8 2 [0039] The average correction factor Kn calculated for one unit stroke is used to determine the input energy to substrate 4 after that unit stroke, thus the energy input to substrate 4 is proportional to the correction factor Kn. That is, it is performed corresponding to the inclination of the actual route of drawing. [0040] When a thermal field emission cathode is used as the electron source 5, it is not convenient to change the electron beam output of the cathode for the reasons described above. However, the input energy, that is, the dose introduced into the substrate 4 at the collision point P of the electron beam 6, can be controlled via the action time of the electron beam 6 at the collision point P. In this case, the residence time of Table 2 after one unit stroke is easily adjusted in proportion to the correction factor Kn. [0041] In parallel with the above, the calculation of the next unit stroke can already be performed in the above-described embodiment. Figure 3 shows the corresponding flow chart. [0042] In the case of the modified example of the present invention, the input energy is adjusted by changing the electron beam output acting on the collision point P after one unit stroke. Basically, in this case, it is conceivable to directly adjust the electron beam output emitted from the electron beam source 5 at the electron beam source. However, this operation is not convenient when using a thermal field emission cathode. This is because the cathode only responds very slowly to changes in its regulatory parameters. [0043] In this case, the constant electron beam 6 emitted from the field emission cathode is reduced by a subsequent weakening (attenuation) device (for example, an air-core coil). After the weakening of the electron beam 6, the weakening device is adjusted depending on the average correction coefficient Kn so that it is proportional to the average correction coefficient Kn obtained for the unit stroke. [0044] By considering the "previous history" of the above unit process in both cases when determining the electron beam energy to be introduced into the substrate 4 after the unit process, the over-irradiation effect mentioned at the beginning is as small as about 0.5%, actually. It is reduced to a negligible residual error. [0045] At the beginning of the curve drawing operation, of course, no old information about the preceding unit stroke is available. In this case, support can be obtained by arbitrarily setting a correction value in the intermediate storage device for a unit process that does not exist. However, the initial error thus generated has no practical effect. In this case, for example, when considering eight preceding unit strokes, the transition of the delineation up to now can be completely considered after the ninth unit stroke. The unknown unit stroke is tentatively associated with category A to obtain a correction factor of 1. [0046] By the way, Fig. 4 contrasts the relationship between the dose and the angle with respect to the coordinate axes when drawing a circular curve. In this case, the angle of 0 ° corresponds to the path portion parallel to the axis, while the angle of 45 ° with respect to the path portion is parallel to the first (angle) bisector. [0047] If the annulus is approximated by the unit strokes of categories A and B described above and an equal amount of energy is applied after each unit stroke, the line is drawn with the correct dose only in the parallel portion of the line drawing. However, for all other sloped pathways, the dose is too low, as evidenced by curve L1. [0048] On the other hand, after the diagonal stroke of category B, unlike the unit stroke parallel to the axis of category A, if the input energy is multiplied by 2, the correct dose is applied to both the parallel part of the line and the 45 ° angle part. Is obtained. However, with respect to the intermediate slope of the path, the input energy is excessive, thus resulting in an undesired width change of delineation depending on the slope of the path. This is shown by the curve L2 in FIG. [0049] By considering the preceding unit strokes when determining the input energy to the substrate, the eight leading unit strokes are taken into account, and the very small and practically negligible dependence of the dose on the inclination angle of the delineation path. The curve L3 showing is obtained. [0050] [Effect of the invention] According to the basic configuration (first viewpoint) of the present invention, electron beam lithography capable of constructing and drawing a curved line with a uniform line width with a small amount of calculation work is realized. The specific principle can be obtained from the second viewpoint of the present invention. From the third viewpoint of the present invention, the correction coefficients are further categorized to simplify the calculation. From the fourth viewpoint of the present invention, an electron beam lithography apparatus can be obtained by a predetermined arithmetic circuit. [Simple explanation of drawings] FIG. 1 is a schematic diagram of an embodiment of an electron beam lithography apparatus. FIG. 2 is a graph showing possible unit strokes for sliding the substrate by the table of FIG. FIG. 3 is a flow chart for determining the input energy to the substrate, which is filed in the control / calculation circuit of the device of FIG. FIG. 4 is a graph showing the relationship between the relative dose when drawing a ring in the XY plane and the angle with respect to the X coordinate or the Y coordinate. [Explanation of symbols] 1 device 2 tables 3 Feeder 4 board 5 electron source 6 electron beam 7 Deflection device X, Y, Z coordinates Kn average correction factor