JP2005056788A - Deflector and its drive method as well as charged particle beam exposure system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a deflector with less noise in driving. <P>SOLUTION: As for a CY, two coils of an outside coil 2 and an inside coil 3 are formed at a vane 1 made of a plate insulator. And as for a PY, two coils of an outside coil 5 and an inside coil 6 are formed at a vane 4 made of a plate insulator. A length of the outside coils 2, 5 in a light axis 7 direction is set longer than that of the inside coils 3, 6 in a light axis 7 direction. Therefore, when the same current is passed in the outside coils 2, 5, and the inside coils 3, 6, a deflecting field formed by the outside coils 2, 5 is larger than that formed by the inside coils 3, 6. Further, it is so constructed that the outside coil 2 of the CY and the inside coil 6 of the PY are connected in series and driven by current I<SB>1</SB>of a first power supply 8, and the inside coil 3 of the CY and the outside coil 5 of the PY are connected in series and is driven by current I<SB>2</SB>of a second power supply 9. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a deflector that is used in a charged particle beam exposure apparatus or the like and has low noise during driving, a deflector having a function as an aligner, a driving method thereof, and these deflectors. The present invention relates to a charged particle beam exposure apparatus using the above.

  In recent years, research and development of electron beam exposure apparatuses capable of exposing finer semiconductor integrated circuits have been actively conducted. Here, the electron beam exposure apparatus is an apparatus for reducing and transferring a circuit pattern formed on a reticle circuit onto a wafer by a projection lens of an electron optical system. Among them, the one that is being developed for practical use is that the reticle surface is divided into exposure areas called subfields, the patterns in the subfields are exposed and transferred in one batch, and the exposure patterns are joined together to form a wafer. This is an electron beam exposure apparatus (USP 4,376,249, Variable axis electron beam projection system: Patent Document 1) of a so-called divided projection exposure system that exposes and transfers the entire circuit pattern.

  The size of the subfield, that is, the exposure area of this divided projection exposure type electron beam exposure apparatus is about 1 mm □ on the reticle and about 250 μm □ on the wafer. In contrast, a reticle is made from a wafer having a diameter of 200 mm. In order to connect exposure patterns, a technique for repeating exposure while deflecting an electron beam has been proposed. In this technique, a deflection field, which is a lateral magnetic field, is superimposed on a lens field, which is a longitudinal magnetic field, and the optical axis of the magnetic lens is effectively shifted laterally. In addition, the deflector deflects the electron beam so that the electron beam passes through the vicinity of the optical axis of the magnetic lens shifted in the lateral direction.

  The trajectory of the electron beam is shown in FIG. The electron beam 32 vertically passing through the pattern formed on the reticle 31 is deflected so as to cross the optical axis 33 by the deflector CY, passes through the optical axis 33 at the position of the scattering aperture 34, and then deflected by the deflector PY. It is deflected in the reverse direction and landed vertically on the wafer 35 to form an image of the reticle 31. The trajectory of this electron beam 32 is called Variable Axis. In the figure, for simplification, the lens is not shown, and only the deflectors CY and PY are shown. Thus, an excellent deflector has been proposed as a deflector for deflecting the electron beam 32 (USP 6,153,885: Troidal charged particle deflector with high mechanical stability and accuracy: Patent Document 2).

  The deflector generates a transverse magnetic field (deflection field) in two orthogonal directions, and deflects the electron beam with Lorentz force. In the following description of this patent, the beam is deflected in the X direction on the wafer surface by the magnetic field in one direction generated by the deflector, so that the beam is deflected in the Y direction on the wafer surface by the magnetic field in the other direction. It is assumed that the angle of the deflector is set so as to be deflected by the angle. The deflector can deflect the beam in an arbitrary direction by superimposing two orthogonal deflection fields. Therefore, it is not necessary to set the angle of the deflector parallel to the wafer coordinates. However, this time, in order to make the explanation easy to understand, the case where it is arranged in that way will be explained. Further, in the following description, a case where a deflector in the X direction out of two orthogonal directions of the deflector is excited will be described. Of course, the following description can also be applied to the Y direction. It should be noted that the direction of the magnetic field generated by the deflector itself and the direction in which the beam is actually deflected on the wafer surface are different because the beam rotates in the lens field. For this reason, the X deflector does not necessarily generate a magnetic field in the X direction, and the Y deflector does not necessarily generate a magnetic field in the Y direction.

  The upper deflector (abbreviated as “Collimator Vane Yoke” or “CY”) and the lower deflector (abbreviated as “Projector Vane Yoke” or “PY”) have the electron beam 32 substantially opposite to the opposite direction on the surface of the wafer 35 and the same distance. Only deflect. For example, when only CY and only PY are excited in the x-axis direction, the trajectories of the electron beam 32 are as shown in FIGS. 14B and 14C, respectively. The beam 32 cannot pass through the scattering aperture 34, but is shown as passing for illustrative purposes). When the landing position of the electron beam 32 is displayed as a vector on the surface of the wafer 35, as shown in FIG. 16A, the landing position vector of the electron beam 32 by CY and the landing position vector of the electron beam 32 by PY are vector lengths. Are the same vector in the reverse direction on the x axis.

  By the way, the driver that drives the deflector always generates electrical noise. Therefore, when CY and PY are driven by separate drivers 36 and 37, as shown in FIG. 16B, the currents flowing through CY and PY change, and the deflection field also changes accordingly. Along with this, the landing position of the electron beam 32 on the wafer 35 varies. However, if CY and PY are wired in series and driven by a single driver, the electrical noise of CY and PY will be the same, and if the deflection vectors of CY and PY have the same vector length on wafer 35, the reverse If facing the direction, all noise is canceled, and the landing position of the electron beam 32 on the wafer does not fluctuate as shown in FIG. Such means is disclosed in, for example, Japanese Patent Application Laid-Open No. 11-224845 and Japanese Patent Application Laid-Open No. 11-168059, and is called pairing.

Although the beam stability is directly related to the accuracy of exposure transfer, noise can be canceled and exposure transfer accuracy can be improved by such means. In this case, since CY and PY are connected in series, the load on the driver increases, and one driver cannot drive the two deflectors CY and PY. For this reason, as shown in FIG. 16C, the same number of drivers (38 and 39) as before pairing are used, and they are arranged in parallel and driven in cooperation.
USP4,376,249 USP6,153,885 Japanese Patent Laid-Open No. 11-224845 Japanese Patent Laid-Open No. 11-168059

  However, when CY and PY are connected in series as described above, if there is a manufacturing error in CY and PY, the deflection vector is set to each deflection vector, although it is desired to deflect electron beam 32 as shown in FIG. Due to the error, a predetermined deflection trajectory cannot be obtained.

  In this case, if CY and PY are driven independently, the respective excitation currents can be finely adjusted so that the directions of the deflection vectors are opposite and the vector lengths are the same. The trajectory shown in Fig. 14 (a) is obtained. However, if CY and PY are connected in series (when paired), the degree of freedom is lost. To supplement the degree of freedom, for example, when an auxiliary deflector is added, the driver is newly required, and the position of the electron beam becomes unstable due to noise of the driver.

  Furthermore, an aligner is often used in the charged particle beam apparatus. The aligner, like the deflector, basically generates a transverse magnetic field and deflects the electron beam, but it is excited only by direct current and statically aligns the electron beam passage position (axis alignment). used. Axis alignment refers to an operation for correcting, for example, when the mechanical structure of the illumination system and the mechanical structure of the projection system are misaligned or these axes are inclined.

  It is also possible to apply an alternating current to the aligner and scan the electron beam. This is done, for example, to observe the position and shape of the electron beam by scanning the electron beam over the pinhole and measuring the amount of electron beam passing through the pinhole. When an aligner is added, the driver is newly required, and the beam position becomes unstable due to the noise of the driver.

  The present invention has been made in view of such circumstances, and a deflector with low noise during driving, a deflector having a function as an aligner, a driving method thereof, and charging using the deflector. It is an object to provide a particle beam exposure apparatus.

  A first means for solving the problem is a deflector composed of a pair of unit deflectors, wherein the coil of each unit deflector is divided into a first coil and a second coil. And the second coils are connected in series, and power supply devices for supplying current to the coils connected in series are separately provided, and the coils connected in series are connected to the coils connected in series. The ratio of the deflection amount by the first coil of the first unit deflector and the deflection amount by the second coil of the first unit deflector when the same current flows is the first unit deflector's first ratio. A deflector having a ratio different from a deflection amount by the second coil of the second unit deflector and a deflection amount by the second coil of the second unit deflector (claim 1).

  Here, in the column of the claims and the means for solving the invention, the unit deflector means what is normally treated as one deflector, and a pair of these is simply referred to as a deflector. I'm calling.

By doing so, when the current supplied to the first coil of the first unit deflector is I ₁ and the current supplied to the second coil of the first unit deflector is I ₂ , deflection sensitivity by I ₁ in one unit deflector (deflection amount per unit current) becomes larger than the deflection sensitivity due to I _2, deflection sensitivity by I ₂ in the second unit deflectors, deflection by I ₁ Greater than sensitivity. Therefore, by utilizing this property, it is possible to implement a method of using a deflector with high flexibility, as will be described later, while reducing deflection fluctuation due to minute noise of the power source.

  The second means for solving the above-mentioned problem is the first means, wherein the pair of unit deflectors has a deflection that occurs when the same current flows when the deflection vectors on the image plane are approximately in opposite directions. The vector lengths are substantially equal (claim 2).

  In this means, as described later, it is possible to implement a method of using a deflector that is rich in flexibility while reducing fluctuation in deflection due to minute noise of the power source.

  The third means for solving the problem is the first means or the second means, wherein the sum of the lengths of the first coil in the optical axis direction is the light of the second coil. It is longer than the sum of the lengths in the axial direction (claim 3).

  A deflection magnetic field formed in the vicinity of the optical axis of the deflector is mainly given by a component parallel to the optical axis of the coil of the deflector. Therefore, a coil having a longer length in the optical axis direction can have a larger deflection sensitivity than a coil having a shorter length in the optical axis direction. Since the coil is formed to circulate, the portion of the coil near the optical axis gives a large deflection magnetic field, the portion of the coil far from the optical axis gives a small magnetic field, and the strength of the magnetic field is almost the same as that of the coil. It is inversely proportional to the distance between the part and the optical axis. Therefore, a coil with a long length in the optical axis direction has a larger magnitude of a reverse magnetic field than a coil with a short length in the optical axis direction. Since it is located away from the optical axis, the influence of the reverse magnetic field is small and does not affect much.

  A fourth means for solving the problem is any one of the first to third means, wherein a center position of the first coil is more than a center position of the second coil. It is arrange | positioned in the position close | similar to an optical axis (Claim 4).

  As described above, the deflection magnetic field formed in the vicinity of the optical axis of the deflector is mainly given by a component parallel to the optical axis of the deflector coil. Therefore, if the center position of the first coil is arranged closer to the optical axis than the center position of the second coil, the first coil is closer to the optical axis than the second coil. The deflection sensitivity can be increased by the amount.

  A fifth means for solving the problem is any one of the first means to the fourth means, wherein the deflection magnetic field in the vicinity of the optical axis by the first coil and the second coil is determined. The pattern of the distribution in the optical axis direction is made to be similar to each other (claim 5).

  In the first to fourth means, the distribution in the optical axis direction of the deflection magnetic field in the vicinity of the optical axis often does not have a similar shape due to the difference in shape between the first coil and the second coil. As a result, the charged particle beam meanders, which may cause aberrations. In this means, the pattern of the distribution in the optical axis direction of the deflection magnetic field in the vicinity of the optical axis by the first coil and the second coil is made similar. Therefore, meandering of the charged particle beam is reduced.

  A sixth means for solving the above-mentioned problem is the fifth means, wherein the first coil and the second coil are provided in a co-winding (claim). 6).

  According to this means, the pattern of the distribution in the optical axis direction of the deflection magnetic field in the vicinity of the optical axis by the first coil and the second coil can be made similar by an easy method.

  A seventh means for solving the problem is the fifth means, wherein at least one of the first coil and the second coil is meandering in a portion parallel to the optical axis. (Claim 7).

  In this means, a deflection magnetic field in the vicinity of the optical axis by the first coil and the second coil can be easily obtained by meandering a portion parallel to the optical axis of at least one of the first coil and the second coil. The pattern of the distribution in the optical axis direction can be made similar.

  The eighth means for solving the problem is the fifth means, wherein the sum of the lengths of the first coil in the optical axis direction is the length of the second coil in the optical axis direction. The center position of the first coil is arranged at a position closer to the optical axis than the center position of the second coil (Claim 8). is there.

  By adopting the configuration like this means, the pattern of the distribution in the optical axis direction of the deflection magnetic field in the vicinity of the optical axis by the first coil and the second coil can be made more similar. it can.

  A ninth means for solving the problem is a deflector having a first coil and a second coil, wherein the first coil and the second coil are connected in series. In the deflector in which the deflection amount by the first coil and the deflection amount by the second coil are substantially the same when the same current is supplied to the divided coils, the first coil and the second coil are substantially the same. It is a shape, and is provided separately on the front and back surfaces of the vane, respectively (claim 9).

  According to this means, since the deflection fields generated by the respective coils are almost the same size and have the same shape, the combined on-axis deflection field becomes almost zero, and extra aberrations even when the charged particle beam passes. Is less likely to occur.

  As a tenth means for solving the above problem, one of the unit deflectors constituting any one of the first to eighth means is a deflector which is the ninth means. It is characterized by this.

An eleventh means for solving the above problem is a driving method of a deflector of any one of the first to eighth means or the tenth means, wherein the first unit deflector The sensitivity (deflection amount per unit current) of the first coil is a, the sensitivity of the second coil of the first unit deflector is b, and the sensitivity of the first coil of the second unit deflector is The sensitivity is c, the sensitivity of the second coil of the second unit deflector is d, and the first coil of the first unit deflector and the first coil of the second unit deflector. I ₁ , the current flowing through the second coil of the first unit deflector and the second coil of the second unit deflector as I ₂ ,
| (Ac) I ₁ | >> | (db) I ₂ |
While maintaining the above state, excitation is performed with currents I ₁ and I ₂ in the same direction, or | (ac) I ₁ | <| (db) I ₂ |
While maintaining this state, excitation is performed by currents I ₁ and I ₂ in the same direction, respectively (claim 11).

As will be described later in the section of the embodiment, according to this means, the magnetic fields generated by the first unit deflector and the second unit deflector can be arbitrarily adjusted by adjusting I ₁ and I _2. Can do. In particular, when there is a difference in gain (deflection field formed when the same current flows) between the first unit deflector and the second unit deflector, this difference is expressed as I ₁ and I ₂ . It can be countered by adjustment. In addition, deflection deflection due to minute noises of I ₁ and I ₂ can be reduced.

A twelfth means for solving the problem is a driving method of a deflector of any one of the first to eighth means or the tenth means, wherein the first unit deflector The current that flows through the first coil and the first coil of the second unit deflector is I ₁ , the second coil of the first unit deflector and the second of the second unit deflector. When the current flowing through the coil is I ₂ and the minute changes are ΔI ₁ and ΔI ₂ ,
While maintaining the state of | ΔI ₁ |> | ΔI ₂ |, the currents I ₁ and I ₂ are changed in the opposite directions, respectively, or the current in the opposite direction is maintained while maintaining the state of | ΔI ₁ | <| ΔI ₂ |. I ₁ and I ₂ are changed (claim 12).

As will be described later in the section of the embodiment, according to the present means, a minute change in the deflection magnetic field of only one of the first unit deflector or the second unit deflector can be achieved by adjusting I ₁ and I _2. Can be adjusted to give In addition, deflection deflection due to minute noises of I ₁ and I ₂ can be reduced.

A thirteenth means for solving the above problem is a driving method of a deflector of any one of the first means to the eighth means or the tenth means, wherein the first unit deflector The sensitivity (deflection amount per unit current) of the first coil is a, the sensitivity of the second coil of the first unit deflector is b, and the sensitivity of the first coil of the second unit deflector is The sensitivity is c, the sensitivity of the second coil of the second unit deflector is d, and the first coil of the first unit deflector and the first coil of the second unit deflector. the current flowing as I _1, the current flowing in the second coil of the second coil and the second unit deflector of the first unit deflector I _2, with the deflector of the configuration,
ΔD = (a−c) I ₁ − (b−d) I ₂
The current is changed by I ₁ and I ₂ while keeping constant, and the deflector is driven while keeping the offset of the deflection amount of the first unit deflector and the second unit deflector constant. (Claim 13).

As will be described later in the section of the embodiment, according to this means, the deflector can serve as an aligner by adjusting I ₁ and I ₂ .

  A fourteenth means for solving the above problem is a charged particle beam exposure apparatus (Claim 14), comprising a deflector which is one of the first means to the tenth means.

  According to this means, since the shake of the charged particle beam due to the noise of the deflector is small, a charged particle beam exposure apparatus with high exposure accuracy can be obtained, and the flexibility is high in the use of the deflector. be able to.

  As described above, according to the present invention, a deflector with little noise during driving, a deflector having a function as an aligner, a driving method thereof, and a charged particle beam exposure apparatus using these deflectors. Can be provided.

  Hereinafter, an example of an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a schematic diagram showing a first example of an embodiment of the present invention, and FIG. 1A shows vane type deflectors (toroidal deflectors) CY and PY. The structure of such a vane type toroidal deflector is described in the aforementioned USP 6,153,885. In the figure, 1 indicates one vane of one toroidal deflector, and 4 indicates one vane of another toroidal deflector. Actually, the toroidal deflector is constructed by assembling a plurality of vanes radially around the optical axis, and connecting coils formed in each vane in series by wiring. In FIG. 1, one toroidal deflector is represented by one vane for convenience. These toroidal deflectors CY and PY are called “unit deflectors” in the claims and the means for solving the invention. Each unit deflector is divided into first and second coils on all vanes, and the first and second coil groups are all wired in series in order to operate as a toroidal deflector. . Furthermore, in the present invention, the first coil group of one unit deflector and the first coil group of another unit deflector are further connected in series. This is called pairing. Specifically, in FIG. 1, the outer coil 2 and the inner coil 6 and the inner coil 3 and the outer coil 5 are paired, respectively.

  In the claims and the means for solving the invention, the unit deflector thus paired is called a “deflector”. It should be noted here that, in FIG. 1, one vane of a unit deflector is paired with one vane of another unit deflector, so that each corresponding vane of each unit deflector Each pair seems to be paired. Actually, the first coil group of one unit deflector and the first coil group of another unit deflector make one pair, and one corresponding sheet A vane does not make many pairs.

  In CY, two coils of an outer coil 2 and an inner coil 3 are formed on a vane 1 made of a plate-like insulator. In this example, both the outer coil 2 and the inner coil 3 have one turn. Moreover, in PY, two coils of an outer coil 5 and an inner coil 6 are formed on a vane 4 made of a plate-like insulator. In this example, both the outer coil 5 and the inner coil 6 have one turn.

  In order to explain the deflection current and deflection distance of the paired deflector in an easy-to-understand manner, the operation principle will be described using a notation method similar to polar coordinates as shown in FIG. The horizontal axis in FIG. 1B literally represents the deflection distance in the X direction on the wafer surface, as indicated by X. The vertical axis is like the imaginary axis of complex coordinates and has no actual meaning.

The arrow labeled CY _{1 in the} figure has a fixed length, which is the deflection on the wafer when the outer coil of CY is excited with the maximum current I _1max of the deflector. It is a distance. This CY ₁ arrow is oriented in the same direction as the vertical axis when the current is zero, and rotates clockwise around the root of the arrow as the current increases. Then, when the maximum current I _1max of the deflector is passed through the coil outside the CY, it is defined so as to be directed in the direction of the X axis.

Similarly, the arrow displayed as CY ₂ in FIG. 1B represents the deflection distance on the wafer when the maximum current I _2max of the deflector is excited in the coil having the length inside CY. The arrow of CY ₂ is always connected to the tip of the arrow of CY _{1 at} the root, and when the current is zero, it is oriented in the same direction as the vertical axis. As the current increases, the arrow rotates clockwise around the root of the arrow. Rotate to. Then, when the maximum current I _2max of the deflector is passed through the coil inside the CY, it is defined so as to face the X axis. The arrow of PY ₁ is the deflection distance on the wafer when the maximum current of the deflector is excited in the coil inside PY, and the arrow of PY ₂ is the maximum current of the deflector in the coil outside PY. This is the deflection distance on the wafer.

The arrow of PY ₁ always rotates around the origin, whereas the root of the arrow of PY ₂ is always connected to the tip of the arrow of PY ₁ , and the rotation around that is the same as CY. is there. However, it is defined that the arrow of PY is directed downward when the current is zero and rotates clockwise around the root of the arrow as the current increases.

Here, since the arrows of CY ₁ and PY ₁ represent the paired coils, the same current flows through each of the coils. Therefore, the angle between each arrow and the coordinate system changes in the same way as the current increases or decreases. Therefore, since the arrows are directed in opposite directions in the initial direction (when the current is zero), the orientation relationship is always preserved. In other words, the arrows of CY ₁ and PY ₁ are always facing in opposite directions, and in this state, behaves to rotate around the origin. Similarly, the arrows of CY ₂ and PY ₂ represent the other paired coils, and naturally the same current flows in each coil. For this reason, the direction of each arrow changes in the same way as the current increases or decreases, and therefore, the arrows are directed in opposite directions at the initial time (when the current is zero), so that the direction relationship is always preserved. In other words, the arrows of CY ₂ and PY ₂ always behave in a state of rotating in accordance with increase / decrease of current while facing the opposite direction while maintaining a parallel relationship.

As can be seen from the above relationship, the angle α ₁ between CY ₁ and PY _{1 and the} vertical axis is α ₁ = sin ⁻¹ (πI ₁ / (2I) where I _{1 is} the current flowing through the coil outside CY. _1max )). Further, the angle α ₂ formed by CY2 and PY2 and the vertical axis is represented by α ₂ = sin ⁻¹ (πI ₂ / (2I _2max )), where I ₂ is the current flowing through the coil inside CY.

In FIG. 1B, the length of CY ₁ is longer than the length of CY ₂ , and the size of the deflection field formed by the outer coil 2 is the size of the deflection field formed by the inner coil 3 as described above. This is because the length of PY ₁ is shorter than the length of PY _{2 because} the deflection field formed by outer coil 5 is larger than the deflection field formed by inner coil 6. It is. Here, if I ₁ and I ₂ are adjusted independently, the deflection amount due to CY and the deflection amount due to PY can be made the same, or either can be increased, and the conventional CY shown in FIG. A deflection field that can be realized with a PY deflection vector can be created, and even if there is an imbalance between CY and PY, the deflection fields of CY and PY can be adjusted by adjusting I ₁ and I _2. Can be of the same size but opposite orientation.

Thus, in the prior art, as shown in FIG. 16A, the deflection vectors of CY and PY cancel each other. However, in the present invention, the lengths of CY ₁ and PY ₁ , CY ₂ and The lengths of PY ₂ are different from each other. Therefore, the conventional one is called pairing, and the one in this embodiment is called unbalanced pairing. Hereinafter, the deflection vector driven by I ₁ is indicated by a solid line, and the vector driven by I ₂ is indicated by a broken line. It will be apparent from the above description that the arrows indicated by solid lines and the arrows indicated by broken lines are parallel.

FIG. 2 shows an example of deflectors CY and PY that have undergone these unbalanced pairings. FIG. 2A shows a case where I ₁ = I ₂ = 0 (α ₁ = α ₂ = 0 °). In this case, the projection of any arrow on the X axis is 0, indicating that no deflection field is generated. On the other hand, FIG. 2B shows the relationship of arrows when I ₁ and I ₂ are both maximized (α ₁ = α ₂ = 90 °), and the arrows are on the X axis. If the sum of the projections of CY ₁ and CY ₂ onto the X axis is equal to the sum of the projections of PY ₁ and PY ₂ onto the X axis, then CY and PY are balanced.

For further understanding, a method of exciting only CY with the above configuration will be described. As shown in FIG. 3 (a), at a certain excitation current _{I 1,} CY, when excited PY, solid arrow is tilted by alpha _1. At this time, since the upper half (CY ₁ ) of the solid line arrow is longer than the lower half (PY ₁ ), the projection onto the horizontal axis is shorter for PY ₁ . Then only to cancel the deflection amount due to the PY ₁ passing an excitation current I _2.

In this case, I ₁ > I ₂ (α ₁ > α ₂ ) is satisfied because the rotated projection of the short PY ₁ is canceled by the rotated projection of the long PY ₂ . That is, the solid line arrow rotates more than the broken line arrow. At this time, in the CY system, since the solid arrow with a long arrow rotates more than the broken arrow with a short arrow, the projection of each other is not canceled and a certain value difference remains. This is the finally obtained CY deflection amount.

Similarly, FIG. 3B shows a method of exciting only PY. This time, under the condition of I ₁ <I ₂ (α ₁ <α ₂ ), it is possible to create a situation equivalent to canceling the generated magnetic field of the CY system and exciting only the PY system. From the above description, it can be seen that an arbitrary deflection field can be generated in each of CY and PY. This proves that with this configuration, the beam can be deflected along a variable axis.

  Basically, if the balance of the excitation currents is adjusted well, the function as an auxiliary yoke (AUX Yoke) can be reproduced in addition to the original deflector function of deflecting to Variable Axis. Although the function of the auxiliary yoke has been described above, the movement is represented by the notation here as shown in FIG.

For example, if a CY deflector is made slightly smaller than a predetermined size due to a manufacturing error, in order to deflect the beam to the original variable axis, the gain corresponding to the manufacturing error must be added to the CY system. In FIG. 4A, when excitation is 0 (arrow is vertical), CY ₁ and PY ₁ are excited with a certain current I ₁ , and CY ₂ and PY with current I ₂ having a smaller absolute value and the same direction. _{When 2} is excited, a solid line arrow with a long CY rotates more than a broken line arrow with a short CY length. On the other hand, the reverse occurs in the PY system, and as a result, the excitation of CY becomes larger. If the excitation is performed to the same sign while maintaining the state of | I ₁ |> | I ₂ |, the deflection amount of CY always increases. The relationship between I ₁ and I ₂ is determined by how much additional gain is desired to be applied to the CY system.

  Note that giving additional gain (putting gain) means adjusting the deflection amount of one deflector to be a predetermined magnification of the deflection amount of the other deflector. Even when there is a difference in the amount of deflection formed when the same current flows between the deflectors (ie, there is a difference in gain as hardware between the two deflectors), it is compensated. Thus, the deflection amount by the two deflectors can be made the same.

For easier understanding, when the arrow in FIG. 4A is compared to an insect spider, it can be said that the degree of opening of the front foot is always larger than that of the rear foot. Next, the case of maximum excitation under this condition will be described. As shown in FIG. 4B, when I ₁ is increased from 0 and the maximum current I ₁ = MAX is reached (solid arrow becomes horizontal), | I ₂ | is always smaller than | I ₁ |. The dashed arrow still does not turn 90 ° completely. Therefore, the total projection does not reach the maximum deflection amount shown in FIG. Similarly, as shown in FIG. 4C, when I ₁ is decreased from 0 and reaches the minimum current I ₁ = MIN (maximum negative value) (solid arrow becomes horizontal), | I ₂ Since | is always smaller than | I ₁ |, the dashed arrow still does not completely turn 90 °. Therefore, the total projection is also shorter than when the minimum current is supplied to each. That is, as the CY system gain is increased, the difference between I ₁ and I ₂ needs to be increased, and as a result, the maximum deflection amount decreases. When designing, it is necessary to estimate in advance how much gain is required and to select the maximum current required for the driver appropriately.

Next, using FIG. 5, an operation for applying a gain to the PY system will be described. Unlike FIG. 4A, when the excitation is performed to the same sign while maintaining the state of | I ₁ | <| I ₂ |, the deflection amount of PY always increases. To explain more clearly, if the arrow in FIG. 5 is compared to an insect emblem, it can be said that the degree of opening of the front foot is always smaller than that of the rear foot. FIG. 5 shows a state where the broken arrow is rotated until it becomes horizontal, and the projected image in this case is the maximum deflection amount of each of CY and PY. Similar to the case where a gain is placed on CY, the maximum deflection amount is limited to be smaller than that when a maximum current is supplied to each pair.

  Next, a method of performing a scanning operation with the configuration of the present invention will be described with reference to FIG. The scanning operation is to generate an alternating magnetic field in CY or PY in a state where the beam is deflected to some variable axis. For the sake of explanation, in this case, it is assumed that there is no additional gain in CY and PY (there is no problem even if there is gain).

As an example, in order to deflect the beam to Variable Axis, excitation of half the maximum current is performed with I ₁ = I ₂ (in a state where the arrow is rotated by 45 °), and an AC magnetic field is generated only from CY therefrom. As shown in FIG. 6A, the excitation of I ₁ is first decreased by ΔI ₁ and at the same time the excitation of I ₂ is increased by ΔI ₂ (Δα ₁ corresponds to ΔI ₁ and Δα ₂ corresponds to ΔI ₂ ). At this time, when attention is paid to the image of the PY system, the current changes in the opposite direction while maintaining the state of | ΔI ₁ |> | ΔI ₂ |. It can be canceled with a smaller current of a long arrow. On the other hand, in the CY system, since the change in the angle of the long solid arrow is large, as long as the broken arrow changes at a smaller angle, the size of the CY image changes. This means that a scanning magnetic field of an arbitrary magnitude can be generated only in CY in a certain predetermined beam deflection state.

When the scanning operation is performed in the PY system, the current may be changed in the opposite direction while maintaining the state of | I ₁ | <| I ₂ | in reverse to the case of the CY system (FIG. 6B). . In this state, it is possible to change only the PY image while canceling the change in the size of the CY image. Whether the scanning operation is performed in the CY system or the scanning operation in the PY system, no more alternating current can be applied when the maximum deflection current is applied. Therefore, before designing, it is necessary to consider how much larger scanning operation is required at the time of maximum deflection and to select the maximum current of the driver appropriately.

  Finally, a method for performing an aligner operation in the configuration of the present invention will be described with reference to FIG. The aligner operation is an operation in which the offset amount is always maintained even if CY or PY initially has an offset in a state where the beam is not deflected by Variable Axis and then the beam is started to be deflected by Variable Axis. For the sake of explanation, in this case, it is assumed that there is no additional gain in CY and PY.

First, the case where a positive offset is put on CY will be described with reference to FIG. The state in which only CY is excited is the same as that already described with reference to FIG. When the current is changed while maintaining the state of I ₁ > I ₂ from this state, a state in which a positive offset is always put on CY (a state in which the deflection amount by CY is a fixed amount larger than the deflection amount by PY) Both the deflection amount by CY and the deflection amount by PY can be changed. Similarly, the method of exciting only PY is described with reference to FIG. 3B. From that state, as shown in FIG. 7B, the current is changed while maintaining the state of I ₁ <I _2. Then, this time, in a state where a positive offset is always put on PY (a state in which the deflection amount by PY is a certain amount larger than the deflection amount by CY), both the deflection amount by CY and the deflection amount by PY can be changed.

The above is the description of the basic operation. Next, what kind of coil patterns are paired unbalanced will be described. First, as a basic idea, how to determine the degree of unbalance will be explained. When the degree of unbalance is increased, even if the current difference between I ₁ and I ₂ is small, the range of adjustment functions as an auxiliary yoke or aligner can be expanded, and the scan range can be increased. However, since the deflection vectors of the upper and lower yokes are unbalanced, the noise cancellation rate decreases.

On the other hand, if the pair is balanced, a sufficient auxiliary yoke, aligner, and scanner range cannot be secured unless the current difference between I ₁ and I ₂ is increased. Then, the maximum range as the original deflector is reduced. However, the noise cancellation rate increases. It is necessary to understand the above and design an appropriate imbalance rate.

  Next, a method for forming a mechanical coil shape for obtaining an optimal unbalance rate will be described. FIG. 8A shows an example in which a two-turn coil is provided as the inner coil 3 and a two-turn coil is provided as the outer coil 2 inside and outside. In this case, since the outer coil 2 has a longer length in the optical axis direction, the deflection amount by the outer coil 2 is larger than the deflection amount by the inner coil 3 when the same current flows.

  When it is desired to make the deflection amounts closer to each other, the inner and outer coils may be wound together as shown in FIG. By doing so, the difference between the length of the outer coil 2 in the optical axis direction and the length of the inner coil 3 in the optical axis direction becomes smaller than in FIG. 8A, and the deflection amount by both coils becomes closer. .

  On the other hand, when it is desired to increase the difference between the deflection amounts of the two, the number of turns of the outer coil 2 may be larger than the number of turns of the inner coil 3. However, when such a difference is not necessary, the outermost outermost coil 2 may be extended in the vertical direction as shown in FIGS. Similarly, the center of the outer coil 2 is made closer to the optical axis 7 as compared with the inner coil 3 by bringing the outermost peripheral pattern closer to the optical axis 7 as shown in FIGS. Also good.

  In the above description, the method for adjusting the amount of deflection when a unit current is passed through the coil has been described. In the following, the shape of the magnetic field generated by the deflector will be discussed. As shown in FIG. 9 (a), when coils are arranged in a co-winding manner, the magnetic field generated by applying a unit current to each coil is as shown in FIG. 9 (b). Here, the magnetic field generated on the optical axis 7 when the outer coil 2 in FIG. 9A is excited is indicated by a broken line in FIG. On the other hand, the magnetic field generated on the optical axis 7 when the inner coil 3 in FIG. 9A is excited is indicated by a one-dot chain line in FIG.

  Since the outer coil 2 is closer to the optical axis, the magnitude of the generated magnetic field is larger than the magnitude of the magnetic field generated by the inner coil 3. Further, since the outer coil 2 is longer in length, the magnetic field profile is broader than the magnetic field profile of the inner coil 3. Therefore, the magnetic field (dashed line) when the outer coil 2 is excited has a higher peak than the magnetic field (dashed line) when the inner coil 3 is excited, and is broader.

  However, it should be noted that the magnetic field tends to be broad because the inner coil 3 is farther from the optical axis than the outer coil 2. In this example, the case where the effect of broadening the magnetic field due to the outer coil 2 being larger than the inner coil 3 is greater than the broadening of the magnetic field due to the separation of the inner coil 3 from the optical axis. In an actual design, the relationship of the magnetic field spread may be reversed depending on the line width, spacing, and size of the coil.

  Next, while maintaining the deflection amount obtained in FIG. 9B, the current of the outer coil 2 is decreased and the current of the inner coil 3 is increased. The magnetic field generated on the optical axis 7 by each coil at that time is shown in FIG. In FIG. 9C, the magnetic field generated by the outer coil 2 that was broadened is weakened, and the magnetic field generated by the more localized inner coil 3 is increased. As a result, the magnetic field obtained by synthesizing each magnetic field is more localized than that in FIG. 9B. When the magnetic field is more localized, the electron beam is strongly bent at a shorter distance. Therefore, even if the deflection amount of the electron beam is the same in FIGS. 9B and 9C, the trajectory of the electron beam up to that point is different.

  In a charged particle beam exposure apparatus, the amount of deflection is a very important amount, but the deflection trajectory is also important. Therefore, even if the same deflection amount is obtained, it is not preferable that the deflection trajectory varies depending on conditions. This is because the lens system is optimized so as to reduce the aberration with respect to a certain deflection trajectory, and thus the aberration increases when the electron beam passes through the trajectory deviated therefrom. In actually configuring the apparatus, it is not necessary to secure an excessive degree of freedom in adjustment, and therefore the degree of unbalanced pairing does not need to be increased so much. Therefore, the amount of adjusting the current within the minute unbalanced range is small. Therefore, the deviation of the deflection trajectory due to this only generates a negligible additional aberration. However, an increase in aberrations is not desirable. Ideally, the magnetic field profiles generated by the inner and outer coils should be similar.

  FIG. 10 shows an embodiment in which both magnetic field profiles are more similar. In this embodiment, the coil pattern near the optical axis 7 is uneven with respect to the optical axis 7 in the inner coil 3 and the outer coil 2. Since the magnetic field is mainly generated in a direction perpendicular to the coil, the magnetic field of the outer coil 2 that originally generated the broad magnetic field becomes narrower, and the magnetic field of the coil that originally generated the localized magnetic field becomes wider. In this way, the magnetic field profile of each coil can be made similar.

  In this case, since the pattern can be formed on the same plane, the manufacturing becomes easier. You may adjust the intensity | strength of an unevenness | corrugation as needed, and may form an unevenness | corrugation only in the pattern of the inner coil 3 or the outer coil 2. FIG. Since the direction of the unevenness cannot be determined unconditionally depending on the distance from the optical axis 7 of the coil pattern, the number of turns, and the length, there are various combinations. In the example shown in FIG. 10, the unevenness is provided only in the pattern close to the optical axis 7 among the coil patterns, but the same processing may be applied to the coil pattern far from the optical axis 7. If necessary, it may be applied above and below the coil pattern. Further, in the example shown in FIG. 10, the sides are straight and bent, but may be curved and uneven.

  Moreover, you may employ | adopt the method as shown to Fig.11 (a). In this embodiment, first, the vertically long coil 14 is formed on both surfaces of the substrate 13 (in FIG. 11, the coil on the back side of the substrate 13 is hidden and cannot be seen, but has the same shape as the coil 14). ). On the other hand, a short coil 16 is formed on one surface of another substrate 15 or 15 '(the coil formed on the substrate 15' is hidden behind because it is formed on the back side in the figure, but it is the same as the coil 16) Have a shape). They are bonded together to make the coils 14 and 16 two-story. Since the coil pattern is usually provided on the back surface of the vane, the same treatment may be applied to the back surface.

  In this way, a magnetic field profile (indicated by a broken line in FIG. 11B) generated by the long coil 14 close to the optical axis 7 and a magnetic field profile generated by the short coil 16 far from the optical axis 7 (FIG. 11B). ) Can be made similar to.

  The reason can be explained as follows. When the coil is close to the optical axis 7, the magnetic field profile becomes more localized. Therefore, the coil close to the optical axis 7 has a long profile and a broad profile. On the other hand, when the coil is far from the optical axis 7, the magnetic field profile is broad. In order to suppress the broad spread, the coil far from the optical axis 7 is shortened. However, naturally, a coil that is far from the optical axis has a low deflection sensitivity, and therefore, unless a larger current is passed through the coil, a deflection field having the same strength as the magnetic field generated by the other coil is not generated. In this way, the magnetic field profiles generated by the coils 14 and 16 can be made similar.

  So far, the method of making the magnetic field profiles of the two coils similar to each other has been described, but from here on, the deflection amount when the same current is passed through the two coils and the method of making the magnetic field profiles exactly the same. Will be explained. In this case, if both the CY and PY unit deflectors are divided into exactly the same, unbalance will be lost. Therefore, either CY or PY must be divided exactly the same and the other unbalanced. However, if only one of them is divided into exactly the same, the magnetic field profile of each coil will naturally be the same, so there is no change in trajectory when adjusting the current ratio of each coil as described above, and aberration increases There is an advantage that there is no. As the method, the unit deflector is not divided in the plane of the vane but on the front and back of the vane.

  FIG. 12A is a top view of one unit deflector in which the vanes 1 are radially arranged on the cylinder 17. The vane 1 is usually provided with a coil pattern on the front and back, and a first coil (black) 11 is provided on one side thereof, and a second coil (white) 12 is provided on the back side thereof.

  If the unit deflector is divided in this way, the deflection amount and the magnetic field profile of the two coils are exactly the same. However, in this case, the coil 11 and the coil 12 have the same size and the same number of turns, but the deflection direction of each deflector obtained by dividing the unit deflector is rotated by the angle of the vane 1 thickness. Will end up. In FIG. 12A, the deflection vector by the first coil 11 is indicated by a real arrow A11, and the deflection vector by the second coil 12 is indicated by a broken line arrow A12. As shown in the figure, even if magnetic field vectors having the same magnitude are generated, the resultant vector A does not become zero because each vector rotates slightly. However, this error is very small and is not a problem in practice.

  However, in order to make the deflecting direction of the deflector exactly the same, a pattern as shown in FIGS. 8A and 8B is formed on the front and back surfaces of the vane 1 and the hatched hatch pattern on the front surface is shaded on the back surface. The hatched pattern and the hatched hatch pattern on the front surface may be electrically connected to the hatched hatch pattern on the back surface. A cross-sectional view of the unit deflector in this case is shown in FIG. Similarly to FIG. 12A, the cross section of the first coil 11 is displayed in black, and the cross section of the second coil 12 is displayed in white. Compared to the case of FIG. 12A, in this case, the first coil 11 and the second coil 12 do not have a relative angle, so that the respective deflection directions exist on the same straight line, and have the same size, the reverse Directional deflection vectors cancel out completely.

  In the above embodiments, the toroidal type deflector has been described as an example. However, a coil can also be formed by a saddle type deflector based on the above concept. In the above description, there is one CY deflector and one PY deflector. However, if the number is the same, unbalanced pairing can be performed.

  Even if CY and PY are not the same number, the deflection vector on the wafer when a predetermined current is passed through a combination of deflectors with CY causes the same current to flow through a deflector with a combination of PY. If the length is substantially equal to the deflection vector on the wafer and is directed in the opposite direction, pairing can be performed by connecting all these coils in series.

  When there are two or more deflectors, some of them may be unbalanced paired deflectors as described above, and the others may be deflectors as in this embodiment. it can.

Up to now, the function of unbalanced pairing has been explained using a special notation. In the following, a more general explanation is given using mathematical expressions. The deflection distance on the wafer of the beam by the yoke CY of the deflector is D _cy , and the deflection distance of the beam on the wafer by the yoke PY of the deflector is D _py . Assuming that the currents flowing through the paired pairs are I ₁ and I ₂ as described above, the deflection distances of CY and PY are expressed by the following equations.
D _cy = aI ₁ + bI ₂ (1)
−D _py = cI ₁ + dI ₂ (2)
Here, a, b, c, and d are the sensitivities of the coils to be paired, each taking a positive value and having a unit of [distance / current]. In (1), D _cy is unsigned (meaning a plus sign), whereas in (2) D _py has a minus sign because the beam is on the wafer surface by applying a positive current. This means that deflection is performed in the + X direction by the yoke CY of the deflector and in the −X direction by the yoke PY of the deflector.

  However, the deflection direction by the deflector yoke CY and the deflection direction by the deflector yoke PY do not have to be on the same straight line. Further, when the same current is passed through the coils connected in series, the deflection distances need not be equal. However, if the directions are opposite and the distances are equal, only noise cancellation can be obtained, and even if this is not the case, individual operation of each yoke CY, PY, gain adjustment, scan operation, and aligner operation are possible.

First, consider a case where only CY is operated alone. In this case, when the currents I ₁ and I ₂ are supplied, the deflection distance of PY must be zero, so the equation (2) is as follows: −D _py = cI ₁ + dI ₂ = 0 (3) )
This will be organized for the _{I 2,}
I ₂ = − (c / d) I ₁ (4)
It becomes. Substituting (4) into (1),
D _cy = b (a / _bc −d) I ₁ (5)
It becomes.

Here, since b> 0, in order to _satisfy D _cy ≠ 0 in the equation (5), it is necessary that (a / b−c / d) ≠ 0. This meaning will be described below. As shown in FIG. 15, when considering a coordinate system in which the horizontal axis (X axis) is I ₁ and the vertical axis (Y axis) is I ₂ , the equation (4) has zero X and Y intercepts (that is, the origin is Street) is a straight line having an inclination of −c / d (FIG. 15A). Zero X and Y intercepts mean zero deflection distance. That is, when applying the current values I ₁ and I ₂ on this straight line, the PY generates a magnetic field that cancels each pair, and the deflection distance becomes zero. I _{1, I 2} are maximum and minimum current range of each driver (respectively _{I 1min ~I 1max, I 2min ~I} 2max) for take a value between, the area inside the dashed square in FIG. 15 (a), This is the actual current region.

On the other hand, when equation (1) is solved for I ₂ ,
I ₂ = − (a / b) I ₁ + D _cy / b (6)
It becomes. This is a straight line with a Y intercept of D _cy / b and a slope of − (a / b). Here, (a / b−c / d) ≠ 0, which is a condition for _satisfying D _cy ≠ 0 in equation (5), means that the slopes of the straight lines in equations (4) and (6) are different. To do.

If the inclinations of the two are different, when I ₁ and I ₂ in the equation (4) are changed, the equation (6) can move in the hatched area. It is, for example (4) of the straight line of _I 1, the _{I 2} sets of equations, _{if I 1} takes the smallest _{I 1min,} I 1, _{I 2} is the minimum _{I 1} of (4) and the current range It is on the broken line shown. At this time, the equation (6) passes through the same point and becomes a straight line having an inclination of-(a / b). The Y intercept (D _cy / b) at this time is a small negative value, which means that the deflection distance is small (large in the negative direction).

From this point, I ₁ is gradually increased, and I ₂ is moved so as to decrease I ₂ , and when the value of I ₁ becomes I _1max , the values of I ₁ and I ₂ are now I _1max Is on the intersection of the broken line and (4). At this time, the equation (6) passes through the same point and becomes a straight line having an inclination of-(a / b). At this time, the Y intercept (D _cy / b) is a large value, which means that the deflection distance is large (large in the positive direction). That is, the length of the movement distance of the intercept in the equation (6) is proportional to the deflection distance on the wafer when only CY is excited without moving PY.

  As is clear from FIG. 15A, when the slopes of the equations (4) and (6) are equal, the two overlap completely, so that the Y intercept of the equation (6) can take only one value. This means that each pair has been perfectly balanced and paired and has lost all of its freedom.

  On the other hand, as the degree of unbalance is increased, the area of the shaded portion in FIG. 15A gradually increases, and the length of the Y-intercept that can be moved by equation (6) becomes longer. This means that the excitation range of CY alone is widened. However, this naturally reduces the noise cancellation rate.

To summarize the above, when CY alone is desired to be excited, I ₁ and I ₂ are changed under the conditions of I ₁ and I ₂ so that the PY deflection distance is kept at zero while the balance of the pairing is lost. You can see that

  The embodiment shown in FIG. 3A is an example in which a> b and c <d, but it can be understood that it is sufficient if the ratio of a: b and c: d is different. A method of exciting only PY without moving CY can be explained in exactly the same way.

A state similar to the single excitation of CY is the CY scanning operation shown in FIG. In this case, since the deflection distance of PY needs to be a constant value, when D _py in equation (2) is set to D _py0 (constant) and (2) is solved for I ₂ ,
_{I 2 = - (c / d} ) I 1 -D py0 / d ... (7)

It becomes. This means that the Y intercept shown in FIG. 15B is a straight line of -D _py0 / d. Substituting equation (7) into equation (1),
D _cy = b (a / bc / d) I ₁ − (b / d) D _py0 (8)
It becomes. Since the second term on the right side of equation (8) is a constant, the condition of (a / b−c / d) ≠ 0 is also necessary in order for the value of D _cy to change as I ₁ changes. That is, in the same way as in FIG. 15A, the slopes of Equations (7) and (6) are different, and the area of the shaded portion in FIG. 15 (b) can be changed according to the degree of imbalance. , (6), the length of the Y-intercept that can move is increased.

In summary, if you want CY to perform a scanning operation, change I ₁ and I _{2 under} the conditions of I ₁ and I ₂ to keep the PY deflection distance constant in a state where the balance of pairing is lost. You can see that The embodiment shown in FIG. 3A is an example in which a> b and c <d, but it can be understood that it is sufficient if the ratio of a: b and c: d is different. In the embodiment shown in FIG. 6A, since c <d, the inclination of the equation (7) is gentle. Therefore _I 1, [Delta] I a range for changing the _{I 2} respectively _1, when [Delta] I _2, the absolute value is more of [Delta] I ₁ as shown in FIG. 15 (b) is large _{(| ΔI 1 |> | ΔI} 2 |) . This is consistent with the embodiment shown in FIG. The same can be said for the case where the PY performs the scanning operation.

Next, the movement of the auxiliary yoke shown in FIG. 4 will be described. In the equations (1) and (2), when it is desired to make the absolute value of the deflection distance of CY larger than that of PY, that is, | D _cy |> | D _py | (9)
Consider the case. First, when the currents I ₁ and I ₂ are positive, since D _cy > 0 and D _py <0, | D _cy | = D _cy , | D _py | = −D _py . Therefore, the formula (9) is expressed as D _cy > −D _py
aI ₁ + bI ₂ > cI ₁ + dI ₂ (10)
It becomes.

Rearranging the equation (10) for I ₁ , if (ac)> 0,
I ₁ > ((db) / (ac)) I ₂ (11)
And if (ac) <0,
I ₁ <((db) / (ac)) I ₂ (12)
It becomes. I ₁ and I ₂ may be changed while maintaining this condition.

When the currents I ₁ and I ₂ are negative, since D _cy <0 and D _py > 0, | D _cy | = −D _cy and | D _py | = D _py . Therefore, equation (9) is
-D _cy > D _{py
-AI 1 -bI 2> -cI 1 -dI} 2 ... (13)
It becomes. When the equation (13) is rearranged for I ₁ , when (ac)> 0,
I ₁ <((db) / (ac)) I ₂ (14)
And if (ac) <0,
I ₁ > ((db) / (ac)) I ₂ (15)
Thus, the conditions are opposite to those in the expressions (11) and (12). I ₁ and I ₂ may be changed while maintaining this condition.

Conditions that satisfy the expressions (11), (12), (14), and (15) are difficult to understand because they include four unknowns. In the following, this condition will be explained in an easy-to-understand manner. In the embodiment shown in FIG. 4, as shown in FIGS. 2A and 2B, when a current of I ₁ = I ₂ is passed, the deflection distances of CY and PY are equal (the directions are opposite). Is). This condition is not necessarily required for the function of the auxiliary yoke, but it is easier to understand the condition if the discussion is advanced only in this case. When I ₁ = I ₂ = I, where the currents are equal to each other, the respective deflection distances are equal, and therefore, D _cy = −D _py from Equations (1) and (2)
aI + bI = cI + dI
It becomes. Solving the above equation for d, d = ac−b (16)
It becomes.

When the equation (16) is substituted into the equations (11) and (12), I ₁ > I ₂ and I ₁ <I ₂ respectively. In this case, since I ₁ and I _{2 are both} positive, I ₁ = | I ₁ |, I ₂ = | I ₂ | Therefore, | I ₁ |> | I ₂ | (17)
| I ₁ | <| I ₂ | (18)
It can be rewritten as

On the other hand, when the equation (16) is substituted into the equations (14) and (15), I ₁ <I ₂ (19)
I ₁ > I ₂ (20)
It becomes. In this case, since both I ₁ and I ₂ are negative, it can be written as −I ₁ = | I ₁ |, −I ₂ = | I ₂ |. Therefore, the expressions (19) and (20) are expressed as | I ₁ |> | I ₂ | (21), respectively.
| I ₁ | <| I ₂ | (22)
It can be rewritten as

To summarize the above, in order to satisfy the condition (| D _cy |> | D _py |) of the expression (9), it is desired to make CY larger than the absolute value of the PY deflection distance in the CY auxiliary yoke operation. If
If a> c: | I ₁ |> | I ₂ | (23)
When a <c: | I ₁ | <| I ₂ | (24)
It can be seen that the currents I ₁ and I ₂ may be changed while maintaining the above relationship. In the embodiment shown in FIG. 4, since a> c, it is only necessary to satisfy the condition (23), which is consistent with the description of FIG.

On the other hand, in the operation of the auxiliary yoke of PY, when it is desired to make PY larger than the absolute value of the deflection distance of CY (| D _cy | <| D _py |), this can be explained in the same manner as above, and the following conditions It becomes.
If a> c: | I ₁ | <| I ₂ | (25)
In the case of a <c: | I ₁ |> | I ₂ | (26)
In the embodiment shown in FIG. 5, since a> c, it is only necessary to satisfy the condition (25), which is consistent with the description of FIG.

4B and 4C illustrate the loss of the maximum deflection distance when a difference in gain between CY and PY is given, this can also be explained by mathematical expressions. A large gain means that the difference in deflection distance between CY and PY is large. If the difference is ΔD,
ΔD = D _cy + D _py (27)
It becomes. Substituting the equations (1) and (2) into the equation (27), ΔD = aI ₁ + bI ₂ −cI ₁ −dI ₂ (28)
It becomes. When d is deleted using the equation (16),
ΔD = (ac) (I ₁ −I ₂ ) (29)
It becomes.

  Here, since a large gain means that ΔD is large, one method of increasing the gain is a method of increasing the value of (ac) in equation (29). This means that the balance of pairing is further lost, and a trade-off that increases noise as the gain increases.

Another method is to excite the yoke while keeping the current difference between I ₁ and I ₂ large. In this way, when one excitation reaches the maximum value of the current, the other excitation is performed with a value smaller than the maximum value, and the total deflection distance is shortened. This indicates a decrease in the maximum deflection distance due to an increase in gain, which coincides with that shown in FIGS. 4 (b) and 4 (c).

Next, the aligner operation by CY shown in FIG. The aligner operation is to excite the CY with a certain offset (ΔD = constant) on the PY. (28)
ΔD = (a−c) I ₁ + (b−d) I ₂ (30)
It becomes. An offset operation can be obtained by changing I ₁ and I ₂ while satisfying this condition.

Equation (30) also contains four unknowns and is difficult to understand. Therefore, as described in the explanation of the auxiliary yoke operation, the case where the current of I ₁ = I ₂ is passed and the case where the deflection distances of CY and PY are equal will be further discussed. Similar to the auxiliary yoke operation, this condition is not necessarily required for the aligner operation, but it is easy to understand the condition if the talk is advanced only in this case.

Substituting Equation (16), which is the condition in this case, into Equation (30) yields Equation (29). In summary:
When a> c: I ₁ > I ₂ (31)
When a <c: I ₁ <I ₂ (32)
In the embodiment of FIG. 7A, since a> c, the condition of the expression (31) is satisfied, which coincides with the description of FIG. On the other hand, putting an offset on PY as shown in FIG. 7B is equivalent to putting a negative offset on CY, and ΔD becomes negative. In that case, it can be explained in the same way, and it is summarized as follows.
When a> c: I ₁ <I ₂ (33)
If a <c: I ₁ > I ₂ (34)
In the embodiment of FIG. 7B, since a> c, the condition of (33) is satisfied, which matches the description.

As described above, according to the embodiment of the present invention, the following advantages can be obtained as compared with the conventional pairing deflector.
1. There is no need to provide an additional auxiliary yoke.
2. A driver to drive the auxiliary yoke becomes unnecessary.
3. There is no need to provide an additional aligner.
4). A driver to drive the aligner is not required.

In the prior art, the aligner often has a scan function, but due to restrictions, the aligner may not have a scan function. Compared to that case, the following advantages can be mentioned.
5). There is no need to provide an additional scan coil.
6). A driver to drive the scan coil is not required.

  In the above description, a single deflector (a set of CY and PY) has been described. Usually, in a charged particle beam exposure apparatus, it is usual to use three or more deflectors. . Since each deflector requires x-axis and y-axis, a total of six coil groups and six drivers are unnecessary. The driver is an extremely high-precision electric circuit, is expensive and large in size, consumes high power, and is not accurate unless used with a high-precision temperature control device. Therefore, the advantage that the driver is unnecessary is very large, leading to cost reduction, space saving, power consumption reduction, and temperature control device reduction.

  On the contrary, as a demerit in the above embodiment, since the balance of pairing is slightly lost, the noise canceling effect is lowered. However, in many cases, the ratio of unbalance may be about 5%, and a noise canceling effect of 95% is still obtained. So this is not really a disadvantage. In addition, when the driver's maximum current is passed, the maximum deflection amount slightly decreases, but this is also about 5%, and the driver's maximum current may be increased slightly.

  When the maximum current of the driver is increased, the noise increases with it, but even if the noise increases by 5%, 95% of the noise is canceled due to the effect of unbalanced pairing. Then, the noise can be greatly reduced in total.

  Comparing the conventional pairing and the unbalanced pairing of the present invention in terms of noise, the conventional pairing was able to completely cancel the noise from the deflector, but additional auxiliary yokes and aligners had to be added. At that point, the coil generates noise. The noise and the noise of unbalanced pairing can be said to be almost equivalent. In that sense, the effect of the present invention that can realize the same function and noise characteristics with a small number of coils and drivers is very high.

  FIG. 13 is a schematic diagram of an optical system of a charged particle exposure apparatus which is an example of an embodiment of the present invention. In FIG. 13, 21 is a charged particle beam source, 22 is an illumination lens, 23 is a beam shaping aperture, 24 is an aperture stop, 25 is a mask, 26 is a projection lens, 27 is a scattering aperture, and 28 is a wafer.

  The charged particle beam emitted from the charged particle beam source 21 uniformly illuminates the mask 25 by the illumination lens 22. The image of the pattern formed on the mask 25 is formed on the wafer 28 by the projection lens 26, and the resist on the wafer 28 is exposed. An aperture stop 24 is provided to cut the scattered rays and limit the aperture angle, and a scattering aperture 27 is provided to cut the scattered rays.

  In this embodiment, the projection optical system is provided with six deflectors indicated by CYa, CYb, CYc, PYa, PYb, and PYc. Since the deflectors CYa and PYc, CYb and PYb, and CYc and PYa are unbalanced, these deflectors can be used in the above-described usage.

It is a schematic diagram which shows the 1st example of embodiment of this invention. It is a figure which shows the example of the relationship of a deflection vector. It is a figure explaining the method of exciting only deflector CY and only PY. It is a figure for demonstrating the method to give an effect | action as an auxiliary yoke to a deflector. It is a figure for demonstrating the method to give an effect | action as an auxiliary yoke to a deflector. This is for explaining a method of causing the deflector to perform a scanning operation. It is a figure for demonstrating the method of making an aligner operation | movement to a deflector. It is a figure for demonstrating the formation method of the mechanical coil shape for obtaining the optimal imbalance rate. It is a figure explaining the shape of the magnetic field generated by a deflector. It is a figure for demonstrating the example which makes the magnetic field shape of the two coils in CY or PY similar. It is a figure for demonstrating the example which makes the magnetic field shape of the two coils in CY or PY similar. It is a figure for demonstrating the example of the method of equalizing the deflection amount of two coils in CY or PY. It is a schematic diagram of the optical system of the charged particle exposure apparatus which is an example of embodiment of this invention. It is a figure which shows the track | orbit of the electron beam in the conventional electron beam exposure apparatus. It is a figure for demonstrating the relationship of the electric current value sent through two yokes. It is a figure which shows the excitation method of the deflector in the conventional electron beam exposure apparatus.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Vane, 2 ... Outer coil, 3 ... Inner coil, 4 ... Vane, 5 ... Outer coil, 6 ... Inner coil, 7 ... Optical axis, 8 ... 1st power supply device, 9 ... 2nd power supply device, 11 ... Coil , 12 ... coil, 13 ... substrate, 14 ... coil, 15 ... substrate, 15 '... substrate, 16 ... coil, 17 ... cylinder, 21 ... charged particle beam source, 22 ... illumination lens, 23 ... beam shaping aperture, 24 DESCRIPTION OF SYMBOLS Aperture stop, 25 ... Mask, 26 ... Projection lens, 27 ... Scattering aperture, 28 ... Hue, CY, PY, CYa, CYb, CYc, PYa, PYb, PYc ... Deflector

Claims (14)

A deflector comprising a pair of unit deflectors, wherein the coil of each unit deflector is divided into a first coil and a second coil, and the first coil and the second coil are each in series. The first unit deflector when the same current is supplied to each of the coils connected in series is separately provided with a power supply device that supplies current to each of the coils connected in series. The ratio of the deflection amount by the first coil and the deflection amount by the second coil of the first unit deflector is such that the deflection amount by the first coil of the second unit deflector and the second unit deflector The deflector is different from the deflection amount ratio of the second coil. The pair of unit deflectors is characterized in that the deflection vectors on the image plane are approximately opposite in direction and the lengths of the deflection vectors generated when the same current flows are substantially equal. The sum of the lengths of the first coil in the optical axis direction is longer than the sum of the lengths of the second coil in the optical axis direction. Deflector. 4. The center position of the first coil is disposed at a position closer to the optical axis than the center position of the second coil. Deflector. 5. The deflector according to claim 1, wherein the pattern of the distribution in the optical axis direction of the deflection magnetic field in the vicinity of the optical axis by the first coil and the second coil is similar. 6. A deflector characterized in that the deflector is adapted to be close to. The deflector according to claim 5, wherein the first coil and the second coil are provided together. 6. The deflector according to claim 5, wherein a portion parallel to the optical axis of at least one of the first coil and the second coil meanders. 6. The deflector according to claim 5, wherein a sum of lengths of the first coils in the optical axis direction is longer than a sum of lengths of the second coils in the optical axis direction. The deflector according to claim 1, wherein a center position of the first coil is disposed closer to the optical axis than a center position of the second coil. A deflector having a first coil and a second coil, wherein the first coil and the second coil are connected in series, and when the same current is passed through the divided coils, In the deflector in which the deflection amount by the first coil and the deflection amount by the second coil are substantially the same, the first coil and the second coil have substantially the same shape, and respectively on the front and back surfaces of the vane, A deflector that is provided separately. A deflector according to claim 9, wherein one of the unit deflectors constituting the deflector according to claim 1 is the deflector according to claim 9. The sensitivity (deflection amount per unit current) of the first coil of the first unit deflector is a, the sensitivity of the second coil of the first unit deflector is b, and the second unit deflection. The sensitivity of the first coil of the detector is c, the sensitivity of the second coil of the second unit deflector is d, and the first coil and the second unit of the first unit deflector. I ₁ is a current flowing through the first coil of the deflector, and I ₂ is a current flowing through the second coil of the first unit deflector and the second coil of the second unit deflector.
| (Ac) I ₁ | >> | (db) I ₂ |
While maintaining the above state, excitation is performed with currents I ₁ and I ₂ in the same direction, or | (ac) I ₁ | <| (db) I ₂ |
The method of driving a deflector according to any one of claims 1 to 8, wherein excitation is performed with currents I ₁ and I ₂ in the same direction while maintaining the state. A current flowing through the first coil of the first unit deflector and the first coil of the second unit deflector is I ₁ , the second coil of the first unit deflector and the second coil When the current flowing through the second coil of the unit deflector 2 is I ₂ and the minute changes are ΔI ₁ and ΔI ₂ ,
While maintaining the state of | ΔI ₁ |> | ΔI ₂ |, the currents I ₁ and I ₂ are changed in the opposite directions, respectively, or the current in the opposite direction is maintained while maintaining the state of | ΔI ₁ | <| ΔI ₂ |. The method for driving a deflector according to claim ₁ , wherein I ₁ and I ₂ are changed. The sensitivity (deflection amount per unit current) of the first coil of the first unit deflector is a, the sensitivity of the second coil of the first unit deflector is b, and the second unit deflection. The sensitivity of the first coil of the detector is c, the sensitivity of the second coil of the second unit deflector is d, and the first coil and the second unit of the first unit deflector. Assuming that the current passed through the first coil of the deflector is I ₁ , the current passed through the second coil of the first unit deflector and the second coil of the second unit deflector is I ₂ , With the deflector of the configuration,
ΔD = (a−c) I ₁ − (b−d) I ₂
The currents I ₁ and I ₂ are changed while keeping constant, and the deflector is driven while keeping the offset ΔD of the deflection amount of the first unit deflector and the second unit deflector constant. The method of driving a deflector according to any one of claims 1 to 8, or claim 10. A charged particle beam exposure apparatus comprising the deflector according to claim 1.

Images