JP7280237B2 - Automatic sample tilt correction device and sample tilt automatic correction method - Google Patents

Description

The present invention relates to an automatic sample tilt correction device and an automatic sample tilt correction method for automatically correcting the tilt of a sample.

Conventionally, electron microscopy technology has been essential in order to check whether a mask or the like formed by exposing a fine pattern on a wafer has the correct dimensions. Furthermore, autofocus technology is very important for obtaining high-definition images. However, since state-of-the-art electron microscopes are equipped with aberration correctors, the depth of focus is very shallow, only 0.00. The image to be obtained is blurred even if the height is different by only a few microns.

Under such circumstances, when autofocusing is performed by changing the focus of the objective lens, autofocusing may fail depending on the mask. If the autofocus fails, not only can measurements not be made, but erroneous measurements are provided to process control, causing additional yield losses.

A sample introduced into an electron beam observation device is placed on a mask holder or an XY stage, so it is tilted on the order of microns no matter how precisely it is manufactured. If an image is acquired by moving the XY stage while the mask is tilted, the distance between the sample and the objective lens, ie, WD, changes each time the mask moves, requiring time-consuming focus adjustment each time.

If the sample is greatly tilted, the WD varies greatly depending on the measurement position. Therefore, in order to achieve successful autofocus, it is necessary to search the wide height range corresponding to the maximum tilt amount to find the just-focus state. and it takes more time to focus. If the focus search is performed by roughly changing the focus value in order to shorten the time, there is a risk that the just focus may be missed with the recent low-contrast photomasks.

In addition, if the sample is greatly tilted, the angle of electron beam irradiation will change, causing problems such as deterioration in reproducibility of measurement.

In addition, since the WD change due to the tilt when the mask holder is installed (fixed) is larger than the tilt change due to the movement of the XY stage, if it can be reduced, the WD fluctuation range can be reduced. A mask is generally transported into the measuring apparatus using a transport device called a pallet. Since there are many masks to be measured, this transport device must be detachable and can be loaded with various masks. Although it is necessary to be easily attachable and detachable and it is difficult for displacement to occur, there has been no such device in the past.

SUMMARY OF THE INVENTION It is an object of the present invention to measure the height of a sample precisely and correct its tilt in a charged particle beam device, and to realize high-speed, precise and reliable autofocus.

Therefore, the present invention provides a sample tilt automatic corrector for automatically correcting the tilt of a sample, in which a sample or a sample holder containing the sample is held, and the tilt of the sample or the sample holder containing the sample in the height direction is corrected. A stage Z to be corrected, a height measuring device for measuring the height of three or more arbitrary points of the sample fixed on the stage Z, and a height direction of the sample fixed on the stage Z by the height measuring device height correction means for adjusting the height direction of the stage Z so that the tilt in the height direction becomes zero.

At this time, the height measuring device has a mirror that reflects the laser beam on the lower surface of the objective lens that scans the sample while irradiating the sample with a finely focused electron beam. The light is reflected downward in the direction of the light to irradiate the sample at three arbitrary points, and the height is precisely measured based on the light reflected from each of the three points.

In addition, the height measuring device uses an objective lens that scans a plane while irradiating the sample with a narrowly focused electron beam. I try to measure the height precisely.

Further, the stage Z has a mechanism that can be automatically adjusted at any three points in the height direction.

In addition, a piezoelectric element is used as the mechanism.

INDUSTRIAL APPLICABILITY The present invention makes it possible to precisely measure the height of a sample in a charged particle beam device, correct the tilt to 0.1 μm or less, and realize high-speed, precise, and reliable autofocus.

In addition, since the height of the sample is automatically adjusted each time the sample or the holder to which the sample is fixed is replaced, automatic focusing can be performed quickly and reliably.

In addition, since the sample is automatically placed horizontally (corrected), it is possible to reduce the autofocus search range, which contributes to an increase in throughput.

Also, since WD can be kept constant, image rotation can be prevented.

In addition, since the sample is placed horizontally, the angle of the electron beam irradiating the sample is constant at any position on the sample, improving the reproducibility of the measurement.

In addition, in the case of semiconductor wafers, deep contact holes are formed, but since the incident angle can be kept constant, the bottom of the hole can be irradiated with an electron beam at a constant angle, improving the reproducibility of measurement. .

In addition, the photomask on the pallet can be easily attached and detached, and masks of various sizes can be carried into the apparatus and measured simply by exchanging the pallet.

Utilizing the pallet of the present invention provides the following additional advantages.

Even if a transport error occurs and the photomask falls off from the transport robot, damage to the mask is less than in bare transport. It is easy to ground the mask to prevent charging up. The pallet functions as a plate-like electrode and stabilizes the electric field behind the mask. Since the positioning of the mask with respect to the pallet is easy with the mounting device, positioning within the apparatus can be performed with the pallet positioning accuracy.

FIG. 1 shows a block diagram of one embodiment of the present invention. This will be described in detail below using an electron beam as an example of a charged particle beam.

In FIG. 1, an electron gun 1 is a known electron beam generator that generates an electron beam 2 accelerated to several hundred volts to several tens of kilovolts.

The electron beam 2 is an electron beam generated and accelerated by the electron gun 1 . The electron beam 2 is converged by a converging lens (not shown).

The deflector 3 deflects the focused electron beam 2 in the X and Y directions (usually two-stage deflection), and the electron beam 2 narrowed by the objective lens 5 scans the sample 7 in a plane (X and Y directions). scanning in the Y direction).

The electron detector 4 detects and amplifies secondary electrons, reflected electrons, etc. emitted when the sample 7 is planarly scanned with the narrowly focused electron beam 2, and produces an image (secondary electron image, reflected electron image, etc.). is for generating

The objective lens 5 narrows the electron beam 2 and irradiates it onto the surface of the sample 7 .

The sample holder 6 is for fixing the sample 7 .

Piezoelectric elements Z1, Z2, Z3 (stage Z) 8 are installed between the XY stage 9 and the sample holder 6, and are contracted by applying a voltage to the sample 7 in the axial direction (high direction) is for automatic adjustment.

The XY stage 9 precisely moves the sample 7 in the XY directions, and moves the sample 7 while precisely measuring it in real time with a laser interferometer (not shown).

The general control 11 is software that controls the entire apparatus.

The just-focus determining means 12 shifts the position where the electron beam 2 is focused on the surface of the sample 7 by the objective lens 5, and detects the current flowing through the objective lens 5 at the time of just focusing (actually, an air-core dynamic current flowing in the coil). Just-focus values are detected at three arbitrary points on the sample 7, and voltages are applied to the piezoelectric elements Z1, Z2, and Z38 so that the inclination from the axial direction of the sample 7 is zero based on these three just-focus values. and the inclination of the sample 7 is automatically corrected.

The image processing means 13 generates an image (secondary electron image, reflected electron image, etc.) based on the signal detected by the electron detector 4 (a signal obtained by detecting and amplifying secondary electrons and reflected electrons). be.

The autofocus control means 14 supplies a current to the objective lens (actually a dynamic coil) 5 so that the electron beam is automatically focused on the surface of the sample 7 for autofocusing.

The sample height information processing means 15 uses height sensors 151 and 152 to radiate laser beams in the direction of the sample 7, and measures height information of any three points on the sample 7 based on the laser beams that have returned. It is.

The stage Z control means 16 applies a predetermined DC voltage to the piezoelectric elements Z1, Z2, Z3 (stage Z) 8 to control the tilt of the sample 7 fixed to the sample holder 6 mounted on the XY stage 9 in the X-axis direction. is automatically corrected to zero.

The XY stage control means 17 moves the sample 7 to a predetermined position while accurately measuring the length of the XY stage 9 based on a laser interferometer (not shown).

FIG. 2 shows an optical height measuring device (bottom view) of the present invention.

In FIG. 2, reflecting mirrors 153 and 154 are reflecting mirrors that are one of the light guides, and reflect the laser beams from the height sensors Z1 and Z2 toward the sample 7 and The reflected laser beams are reflected in the directions of the height sensors Z1 and Z2. The size of the laser beam emitted from the laser head has a diameter of about 0.1 to 1 mm, and a diameter of about 0.6 mm was used in the experiment. Therefore, the reflecting mirrors 153 and 154 for irradiating the sample 7 with the laser beam perpendicularly may be very small as long as they have a reflecting portion of about 0.1 to 2 mm. This can prevent mechanical interference when the XY stage 9 moves.

Height sensors (Z1, Z2) 151, 152 output the laser beams from the spectroscope to reflecting mirrors 153, 154 to irradiate the sample 7, and the laser beams reflected by the sample 7 pass through the reflecting mirrors 153, 154. , the light is taken in and sent to the spectroscope. The size of the laser beam emitted from the height sensors Z1, Z2 is between 0.1 and 1.0. It has a diameter of about several millimeters, and a diameter of about 0.6 mm was used in the experiment. The laser beams emitted from the height sensors Z1 and Z2 are reflected by the reflecting mirrors 153 and 154 to irradiate the sample 7, and the laser beams reflected by the sample 7 are reflected by the reflecting mirrors 153 and 154 to form the height sensor. In order not to change the optical path length when incident on Z1 and Z2, the portion where the height sensors Z1 and Z2 and the reflecting mirrors 153 and 154 are attached (usually the objective lens 5 or its peripheral members) has an error due to expansion due to thermal change. Therefore, try to keep the optical path length unchanged by shortening the optical path as much as possible, or by using materials that are not affected by thermal changes, or have a small thermal expansion coefficient, or materials with the same thermal expansion coefficient. do.

A spectroscope (not shown) is integrated with the height sensors Z1 and Z2 and the reflecting mirrors 153 and 154 to form a spectral interference laser displacement measuring device. Of the spectroscope, height sensors Z1 and Z2, and reflecting mirrors 153 and 154 that make up the spectral interference laser displacement measuring device, the height sensors Z1 and Z2 are typically several centimeters long and several millimeters wide. small size. Optical elements are built in the height sensors Z1 and Z2, and optical fibers are attached to the ends of the height sensors Z1 and Z2 for inputting laser beams from the outside (spectroscope). Since there is no electrical circuit inside the height sensors Z1, Z2, they do not become a source of electrical noise. Since the power of the laser beam used is as small as mW or less, it does not act as a heat source, and the object to be measured (for example, the sample 7) and members in the vicinity thereof are not affected by the temperature. Height sensors Z1 and Z2 are covered with a non-magnetic metal housing. The lens for laser beam emission is made of glass, and is normally placed inside the sample chamber in the case of the structure of FIG. 2, and can withstand plasma cleaning and the like. Since the height measurement speed of the sample 7 is as high as 200 μs, the height measurement can be performed almost in real time.

Further, the spectroscope, height sensors Z1 and Z2, and reflecting mirrors 153 and 154 (not shown) have a measurement resolution of 1 nm and a measurement reproducibility of 10 nm or more. This embodiment utilizes two height sensors. The two height sensors are arranged symmetrically with respect to the center of the objective lens 5 in this embodiment. Since they can be corrected by calculation, they do not necessarily have to be arranged symmetrically. One is fine.

Also, in FIG.
・The height of the sample 7 at the center of the objective lens 5 = Ze
・The height of the sample 7 on the reflector 153 of the height sensor Z1=Z1
・The height of the sample 7 on the reflector 154 of the height sensor Z2=Z2
is defined as
The height of the sample 7 at the center of the objective lens can be calculated as Ze=(Z1+Z2)/2 (when the height sensors Z1 and Z2 are arranged symmetrically with respect to the optical axis).

2, the operation and characteristics of the height measurement of the sample 7 by the spectrometer, height sensors Z1 and Z2, and reflecting mirrors 153 and 154 according to the present invention will be described.

(1) The laser beam emitted from the height sensor Z1 is emitted right beside the sample 7. As shown in FIG. The traveling direction of this laser beam is bent by 90 degrees by a 90-degree reflecting mirror 153, and the surface of the sample 7 is irradiated with the laser beam almost perpendicularly.

(2) The laser beam reflected by the surface of the sample 7 returns along the same path, passes through the height sensor Z1, and reaches the spectroscope via an optical fiber.

(3) The emitted laser beam (wide wavelength band laser beam) and the laser beam reflected by the sample 7 are caused to interfere with each other in the spectroscope to create distance information (displacement data). Absolute values can be measured with this method. Based on this, focus control is performed.

(4) In the method of measuring the height of the sample 7 of the present invention, the incident angle is almost perpendicular to the sample 7, so the variation (error) of the measured value due to the inclination of the sample 7 is extremely small. It is easy to use because it does not require tilt correction of the sample 7 . Moreover, it is less likely to be affected by the pattern on the surface of the sample 7 .

(5) Furthermore, the laser beam guided from the height sensor to the outside by an optical fiber is taken into an external spectroscope and converted into a distance (the height of the sample 7). Although this spectrometer generates heat, it does not cause errors in measurement because it is not normally inside a vacuum sample chamber. The data measured by the spectrometer is led to a personal computer by means of communication such as TCP/IP or USB, and is subjected to desired processing such as storage of the correspondence with the coordinates of the measurement points.

(6) Two height sensors Z1, Z2 and reflecting mirrors 153, 154 shown in FIG. Since it is flat, for example, if two height sensors Z1 and Z2 are arranged at the same distance from the center of the objective lens 5, the average value of the output values of both will be the corresponding sample of the irradiation point on the sample 7 of the electron beam 2. represents (estimate) the height of By performing such calculations on a personal computer, the height of the sample at the position of the irradiation point of the electron beam 2 on the sample 7 can be obtained precisely and in real time without directly measuring it.

(7) In addition, since the distance between the tip of the pole piece of the immersion type objective lens 5 and the sample 7 is usually as narrow as several millimeters, it is practically impossible to place a sample height sensor at that position. In the present invention, the reflecting mirrors 153 and 154 can be arranged at a position away from the tip of the pole piece of the objective lens 5, and a light guiding device (reflecting mirror, prism, optical fiber, etc.) having a size of about 0.1 to 2 mm can be used as desired. It can be easily placed in a location and the height of the sample can be measured with high accuracy and in real time.

The outline of the present invention shown in FIGS. 1 and 2 will be described below.

(1) When the present invention is used, tilt correction can be performed by the stages Z1, Z2, and Z3 before the electron beam autofocus is performed, and a state sufficiently close to the electron beam focused state can be created. However, the range for searching for the just focus value can be reduced to about 1/10 of the conventional range, and autofocus can be performed at high speed and with high accuracy.

(2) When the tilt of the sample 7 is corrected, the main factor of WD change due to stage movement disappears, so even if the XY stage 9 moves in all directions, the change in WD can be made very small. In this case, since there is no need to greatly change the strength of the objective lens 5 when electron beam focusing is performed, high-precision and high-reproducibility measurement can be realized without image rotation.

In particular, as shown in FIG. 2, if the horizontal installation of the photomask, which is sample 7, can be ensured, it is possible to measure with one height sensor or measure the distance with two height sensors and take the average value. , the difference is very small. Therefore, it is possible to use one height sensor instead of two, which is simple and reduces the cost.

(3) Unlike the conventional lever type, the optical height sensor is extremely compact, so it does not require optical axis adjustment and is easy to mount and maintain. In particular, opening the lid of the chamber during maintenance does not affect the optical system at all. A height sensor is installed in the narrow space between the tip of the pole piece of the objective lens 5 and the surface of the sample 79, because there is no need for mechanical tilt correction using a shim or the like for the sample 7, and there is no need for measurement at the electron beam irradiation point. No need to squeeze. In the conventional laser beam lever method, light hits the electron beam irradiation point, changes the conductivity and electron accumulation, and affects the measurement reproducibility. In this method, the electron beam irradiation point is not exposed to light, so such a measure is unnecessary.

(4) Since the sensor head is passive, it does not generate electromagnetic waves, heat, etc. during sensor operation, so the temperature at the measurement point does not change, and the measurement can be very stable. Contamination of the vacuum can be avoided because the material is glass and non-magnetic metal. No moving parts, long life. Since the distance between the sensor and the measurement point is short, the effect of disturbance is small.

(5) Since the distance between the pole piece of the objective lens 5 and the surface of the sample 9 can be measured in real time, even if the height of the sample 9 changes due to changes in the external environment, the change can be measured in real time. , the electron beam autofocus can be more successful.

(6) Even if the two optical height sensors cannot measure in real time at the same time, it is possible to perform height measurements almost in real time by using a pre-made focus map. I can.

(7) Even for the sample 9, which is difficult to autofocus only with the electron beam image, the height is first measured by the optical height sensor, and it can be brought to a state close to the focused state, so very fine electron beam images can be obtained. Beam focus parameters can be changed. This makes the just focus more accurate, and makes it easier to obtain an electron beam focused state for a three-dimensional object or a measurement object with low edge contrast.

(8) When the three-point support Z stage 22 is used, the sample 9 becomes horizontal and the height of the sample 9 can be kept constant, so that the electron beam irradiation can be made vertical. become. This makes it possible to acquire a more accurate electron beam image.

(9) When the electron beam irradiation angle becomes oblique, vertical irradiation can be achieved by controlling the three-axis stage.

(10) Within the movable range of the Z-axis stage, the electron beam irradiation angle can be controlled at an arbitrary angle by positively tilting the sample.

FIG. 3 shows a stage external view of the present invention.

In FIG. 3, stages Z1, Z2, and Z3 show an example of three independent fulcrums supporting a photomask. The fulcrums of these three stages Z1, Z2 and Z3 are arranged to form a tripod (three-point support) such as an equilateral triangle or an isosceles triangle so as to easily support the photomask. Each stage consists of a piezoelectric element and a built-in capacitive sensor for linear position control, etc. Feedback control is performed by electrical signals from a computer so that it moves in the height direction with accuracy on the order of nanometers. there is Since the piezoelectric element generates a large force and has a very high rigidity, the lateral displacement is small and it is resistant to vibration. Due to these properties, the photomask can be moved straight (straight in the Z-axis direction).

Here, the stage Z1 has a conical supporting portion, a conical hole in the receiving portion, and a degree of freedom of 0 (rotation is possible, which will be described later with reference to FIG. 7).

The stage Z2 has a conical support, a long conical hole as a receiving part, and a degree of freedom of 1 (it can move slightly in the direction of the long conical, which will be described later with reference to FIG. 7).

The stage Z3 has a flat supporting portion, a flat receiving portion, and two degrees of freedom (it can be automatically moved in the X and Y directions on the plane, which will be described later with reference to FIG. 7).

FIG. 4 shows a detailed external view of the stage of the present invention. This is shown in FIGS. 4A, 4B, and 4C, respectively, by extracting the supporting portions of the stages Z1, Z2, and Z3 of FIG. 3 described above for easy understanding.

The support portion Z1 (cone) in FIG. 4A is an extract of the support portion of the stage Z1 in FIG. 3 for easy understanding.

A support portion Z2 (long cone) in FIG. 4B is an extract of the support portion of the stage Z2 in FIG. 3 for easy understanding.

(c) support portion Z3 (flat) in FIG. 4 is an extract of the support portion of stage Z3 in FIG. 3 for easy understanding.

FIG. 4(d) shows an example of the structure inside each support. A piezo-electric element 8 and a capacitive sensor are incorporated in each support, and a voltage is applied to the piezo-electric element to automatically adjust the distance in the Z direction.

FIG. 5 shows an example (surface) of the photomask fixing mechanism of the present invention. This is a pallet that is used to transport photomasks. Pallets are generally made of metal and are electrically conductive. Aluminum, titanium, etc. can be used. In special cases, ceramics or quartz can also be used. In that case, plating or the like is applied to the surface to make it conductive. As shown in the drawing, the pallet has a rotation fixing mechanism for mechanically pressing the photomask in three directions and mounting it, thereby fixing the positional relationship between the photomask and the pallet.

FIG. 6 shows an example (back side) of a photomask fixing mechanism of the present invention. This shows the back side of the pallet, showing the holes and their positions into which the projections of FIG. 5 are inserted. As shown in FIG. 7, which will be described later, one of the conical projections is matched with the male-female conical hole (stage Z1) exactly. Since gravity acts on the pallet, the contact friction between the conical pin (supporting part of stage Z1) and the hole (conical receiving part) becomes very large, and once the pin is fitted in the hole, it can be easily dislodged. Does not come off (gravity drop method). Therefore, it is designed so that it does not come off even if an acceleration of sub-G or less, which is normally generated by acceleration or deceleration of the XY stage 9, occurs. As a result, even if a sudden acceleration/deceleration operation occurs with the movement of the XY stage, accurate positioning is maintained without deviation.

In order to withstand lateral acceleration greater than 1 G, it is desirable to use a material with a greater coefficient of friction for the protrusions and their contact surfaces. Alternatively, there is a method such as making a deep hole and a high protrusion. One way is to increase the weight of the pallet.

On the other hand, the other cone-shaped projection (the cone-shaped projection of stage Z2) has a male-female relationship in one direction and contacts without a gap, but there is a gap in the other direction. conical hole). Since this hole has a clearance in the direction of the conical hole, it can be moved freely with a slight dimensional error in the positional relationship between the pin position and the hole (degree of freedom = 1). In the vertical direction, however, there is no clearance so that the conical pin fits tightly and is accurately positioned. These two holes (conical hole and long conical hole) provide positioning on the order of submicrons. The third protrusion is used simply to keep the pallet horizontal (maintained on the flat part of stage Z3) under the load of the pallet.

As described above, the position of the pallet is determined by mechanical parts. Therefore, it is better to use a material with a small coefficient of thermal expansion for the pins and pallet to ensure accurate positioning in the XXYZ directions regardless of the installation environment of the device. is made.

FIG. 7 shows an explanatory view of the photomask fixing mechanism of the present invention.

FIG. 7(a) shows an example in which a cone-to-cone hole (stage Z1) is provided. This shows an example in which a conical projection (pin) is provided as a stage Z1 on the lower side of the drawing and a conical hole is provided on the upper side as a receiving portion. In this example, the projection (pin) on the lower side fits into the conical hole on the receiving side by gravity and is fixed with 0 degrees of freedom.

FIG. 7(b) shows an example in which a cone-long cone hole (stage Z2) is provided. This shows an example in which a conical protrusion (pin) is provided as a stage Z2 on the lower side of the drawing and an oblong cone hole is provided on the upper side as a receiving portion. In this example, the lower projection (pin) fits into the oblong hole on the receiving side by gravity, but with 0 degrees of freedom in the direction of the rim of the cone and 1 degree of freedom in the direction of the long cone for a short distance. , can move slightly in one direction (see description of FIG. 6).

(c) of FIG. 7 shows an example in which a flat surface (stage Z3) is provided. This shows an example in which a flat protrusion (pin) is provided as a stage Z3 on the lower side of the drawing and a flat receiving portion is provided on the upper side. In this example, the flat portion on the lower side is only pressed by gravity against the flat portion on the receiving side, and is fixed by static friction, so to speak. Although it has two degrees of freedom, it is fixed without moving due to static friction (it is desirable to use a material with as large a static friction as possible and fix it).

FIG. 8 shows a photomask tilt measurement flowchart of the present invention.

In FIG. 8, S1 initializes stage Z (Z1, Z2, Z3) to a neutral height.

S2 moves the stage to predetermined coordinates on the mask. In this step, the stage is sequentially moved horizontally and vertically symmetrically to three or more predetermined coordinates on the mask so as to be as close to the center of the photomask as possible, and S2 is repeated. For example, as shown in a photomask 71 in FIG. 9, which will be described later, the photomask 71 is applied at nine or more points such as (x0, y0), (x1.y0), (x2, y0) . . . (xn, yn). The XY stage 9 (not shown) is moved and positioned (the position is precisely measured by a laser interferometer (not shown)).

S3 measures and records the height with a mask height sensor. In step S2, the XY stage 9 is used to position the sample 7 at a predetermined position on the photomask 71, and the height sensor measures and records the height of the sample 7 at each position. , the height (Z) is recorded in association with the coordinates (X, Y).

S4 decides whether it is the end. In S3, it is determined whether or not the measurement of the height of the sample 7 at each position is finished. If YES, go to S5. In the case of NO, S2 and subsequent steps are repeated.

S5 calculates a photomask tilt approximation function. This calculates the parameters of the approximation function passing through each height position Z of the photomask 71 (sample 7) recorded in S3. Since the photomask 71 is a good plane, it can be accurately approximated by a low-order approximation formula.

Specifically, based on the data of the three or more positions (x, y) recorded in S3 and the height Z of the sample at that time, the parameters of the approximate function (for example, Z = aX + bY + c) are calculated. For example, as shown in FIG. 11, which will be described later, the parameters of the approximation curve of the photomask tilt are fitted.

S6 calculates and records correction amounts to be given to stages Z1, Z2, and Z3.

As described above, the heights at a plurality of positions on the photomask 71 are actually measured by the height sensor (FIG. 10), and the parameters of the photomask inclination approximation curve are fitted based on these. It becomes possible to create a gradient map (FIG. 11). By inputting the position coordinates of the stages Z1, Z2 and Z3 into the photomask tilt approximation curve, the tilt amount of the photomask at the positions of the stages Z1, Z2 and Z3 as shown in FIG. 12 can be calculated.

If the positions of the stages Z1, Z2 and Z3 are sufficiently inside the outer periphery of the photomask 71, the height of each position of the stages Z1, Z2 and Z3 can be directly measured. By doing so, the correction amount may be determined directly without performing approximate calculation.

FIG. 9 shows an example of mask tilt amount measurement points of the present invention. This is for a rectangular photomask 71 of 15 cm x 15 cm, stage Z1, Z2 . An example in which Z3 is in the vicinity of the outer circumference as shown in the drawing and the height is measured for each of nine black circle points is schematically shown. This is shown in FIG. 10, which was recorded by measuring the height of each of nine points.

FIG. 10 shows the height Z (FIG. 9) of each measurement point of the invention relative to the reference point. This is obtained by measuring and recording the height of each of the nine points indicated by the black circles in FIG. 9 described above.

FIG. 11 shows the amount of tilt for each stage position of the present invention. This is obtained by calculating and recording the tilt amounts (height differences ΔZ) of the stages Z1, Z2, and Z3 from the respective heights at the nine positions measured in FIG. 10 described above. Predetermined voltages are applied to the piezoelectric elements Z1, Z2, and Z3 in FIG. 1 so as to correct the tilt amounts (height differences ΔZ) of the stages Z1, Z2, and Z3 (correction to be 0). By doing so, every time the sample 7 (pallet) is replaced, the height is automatically measured and automatically corrected, and the tilt amount can be reduced to zero.

FIG. 12 shows an example of an approximation curve (linear approximation) passing through the measurement points of the present invention. For approximation, since the photomask is generally flat, a linear straight line Z=aX+bY is used, and its parameters a and b are calculated by substituting the amount of inclination (difference in height) in FIG. 11 described above.

FIG. 12A schematically shows the tilt of the photomask before correction, and FIG. 12B schematically shows the tilt of the photomask after correction. After the correction, as shown in the figure, the amount of inclination is corrected, and it can be seen that the image is automatically corrected to be substantially horizontal.

Next, an example in which the heights of the stages Z1, Z2, and Z3 are actually measured and automatically corrected will be described in detail with reference to FIGS. 13 to 17. FIG.

FIG. 13 shows a photomask tilt measurement flowchart of the present invention.

In FIG. 13, S11 initializes the stage Z (Z1, Z2, Z3) to a neutral height.

S12 moves the stages to the coordinates of each stage Z1, Z2, Z3. This sequentially moves the stages so that the photomask is directly above each stage Z1, Z2, Z3. For example, as shown in a photomask 71 in FIG. 14 to be described later, the photomask 71 is positioned at three points of stages Z1, Z2, and Z3 by moving the XY stage 9 (not shown) (the position is determined by a laser interferometer (not shown)). positioning with precise measurement).

In S13, the mask height sensor measures and records the height. In step S12, the XY stage 9 is used to position the sample 7 on stages Z1, Z2, and Z3 on the photomask 71, and the height sensor measures and records the height of the sample 7 at each position. As shown in FIG. 15, heights (Z) are recorded in association with stages Z1, Z2, and Z3.

S14 discriminates whether it is the end. At S13, it is determined whether or not the measurement of the height of the sample 7 at each position is finished. If YES, go to S15. In the case of NO, S12 and subsequent steps are repeated.

S15 calculates and records correction amounts to be applied to the stages Z1, Z2, and Z3.

As described above, the heights Z on the stages Z1, Z2, and Z3 on the photomask 71 are measured (FIG. 15), and based on these, the correction amounts on the stages Z1, Z2, and Z3 can be calculated (FIG. 16).

FIG. 14 shows an example of mask tilt amount measurement points of the present invention. This is for a rectangular photomask 71 of 15 cm x 15 cm, stage Z1, Z2 . Z3 is in the vicinity of the outer circumference as shown, and an example in which the height is measured for each of these three points is schematically shown.

FIG. 15 shows the height Z (FIG. 14) for the stages Z1, Z2, Z3 of the invention. This is obtained by measuring and recording the heights of 3 in FIG. 14 described above.

FIG. 16 shows the amount of correction for each stage position according to the present invention. This is obtained by calculating and recording the correction amounts (height differences ΔZ) at the three stages Z1, Z2, and Z3 measured in FIG. 15 described above.

FIG. 17 shows an example of the amount of correction with the offset (35 microns) of the present invention. Here, since there are an infinite number of combinations of solutions that satisfy the correction amounts in the stages Z1, Z2, and Z3 in FIG. is an offset, and the correction amounts of the stages Z1, Z2 and Z3 are determined as shown in FIG. As a result, the stages Z1, Z2, . Z3 is automatically corrected.

FIG. 18 shows a WD controlled CD measurement flowchart with mask tilt correction according to the present invention. This is the stage Z1, Z2 . FIG. 10 shows an overall flow chart in the case of automatically correcting the height of Z3, automatically adjusting the sample 7 horizontally, and adjusting the side length of the pattern of the mask, which is the sample 7. FIG.

here,
(1) As already described with reference to FIGS. 1 to 7, the apparatus (optical height sensor) of the present invention optically measures the height of the sample 7 almost continuously, regardless of electron beam irradiation. can run. Therefore, height measurement can be performed from the moment the sample 7 to be measured is carried into the vacuum chamber. The focus map expresses the distance from the tip of the objective lens pole piece to the object to be measured (the so-called height of the sample) as a function of the observation position (X, Y) of the object to be measured. Since the objective lens 5 is fixed to the top plate, it is positioned at substantially the same height as long as the top plate does not move due to large atmospheric pressure fluctuations. Alternatively, if the temperature of the device is kept constant, the height remains the same since the length does not change.
(2) On the other hand, when the sample 7 to be measured is placed on the sample holder, it is not necessarily placed horizontally. It has the property of being held at an almost constant value. Moreover, the XY stage 9 for moving the sample 7 has characteristics such as yawing, pitching, and rolling. Since the XY stage 9 with a good coefficient has a very high reproducibility, it can be treated as a kind of eigenvalue. In other words, the change in the height of the sample 7 with respect to the change in stage position can be accurately calculated by synthesizing the above two values. In the present invention, stages Z1, Z2, and Z3 are used to adjust the initial inclination of the pallet, which is the largest component, to 0 as much as possible, and then the height displacement unique to the stages is corrected to make WD constant. technology is disclosed. It is explained below.

In FIG. 18, S21 uses the method described above to correct the tilt of the photomask, level it, and adjust the stages Z1, Z2, and Z3 so that the stages Z1, Z2, and Z3 can obtain appropriate movable ranges. Set the tilt correction value and offset for each.

S22 moves the stage. This moves stages Z1, Z2, Z3 in addition to stages Z1, Z2, Z3 respectively to the desired height.

In S23, the XY stage moves to the CD measurement point of the sample 7. FIG.

S24 applies control voltages to the stages Z1, Z2, and Z3 so that the WD value at the CD measurement point is constant (tilt correction value+offset value+voltage for WD control). In the case of adjacent measurement points, the height hardly changes, so if the previous value is held and the difference from the next applied voltage is made small, the static speed will increase and the height control will become easier. High stability is obtained.

S25 autofocuses. In the state of S24, current is applied to the objective lens 5 (air-core dynamic coil) to change the focus and supply the focal current for the point of just focus.

S26 measures the CD. A just-focused image is obtained, and the width of a predetermined pattern is measured.

S27 decides whether it is the end. If YES, exit. If NO, return to S24 and repeat.

FIG. 18 shows a WD controlled CD measurement flowchart with mask tilt correction according to the present invention. This is the stage Z1, Z2 . FIG. 10 shows an overall flow chart for automatically correcting the height of Z3, automatically adjusting the sample 7 horizontally, and measuring the length of the pattern of the mask, which is the sample 7. FIG.

here,
(1) As already described with reference to FIGS. 1 to 7, the apparatus (optical height sensor) of the present invention optically measures the height of the sample 7 almost continuously, regardless of electron beam irradiation. can run. Therefore, height measurement can be performed from the moment the sample 7 to be measured is carried into the vacuum chamber. The focus map expresses the distance from the tip of the objective lens pole piece to the object to be measured (the so-called height of the sample) as a function of the observation position (X, Y) of the object to be measured. Since the objective lens 5 is fixed to the top plate, it is positioned at substantially the same height as long as the top plate does not move due to large atmospheric pressure fluctuations. Alternatively, if the temperature of the device is kept constant, the height remains the same since the length does not change.
(2) On the other hand, when the sample 7 to be measured is placed on the sample holder, it is not necessarily placed horizontally. It has the property of being held at an almost constant value. Moreover, the XY stage 9 for moving the sample 7 has characteristics such as yawing, pitching, and rolling. Since the XY stage 9 with a good coefficient has a very high reproducibility, it can be treated as a kind of eigenvalue. In other words, the change in the height of the sample 7 with respect to the change in stage position can be accurately calculated by synthesizing the above two values. In the present invention, stages Z1, Z2, and Z3 are used to adjust the initial inclination of the pallet, which is the largest component, to 0 as much as possible, and then the height displacement unique to the stages is corrected to make WD constant. technology is disclosed. It is explained below.

In FIG. 18, S21 uses the method described above to correct the tilt of the photomask, level it, and adjust the stages Z1, Z2, and Z3 so that the stages Z1, Z2, and Z3 can obtain appropriate movable ranges. Set the tilt correction value and offset for each.

S22 moves the stage. This moves stages Z1, Z2, Z3 in addition to stages Z1, Z2, Z3 respectively to the desired height.

In S23, the XY stage moves to the CD measurement point of the sample 7. FIG.

S24 applies control voltages to the stages Z1, Z2, and Z3 so that the WD value at the CD measurement point is constant (tilt correction value+offset value+voltage for WD control). In the case of adjacent measurement points, the height hardly changes, so if the previous value is held and the difference from the next applied voltage is made small, the static speed will increase and the height control will become easier. High stability is obtained.

S25 autofocuses. In the state of S24, current is applied to the objective lens 5 (air-core dynamic coil) to change the focus and supply the focal current for the point of just focus.

S26 measures the CD. A just-focused image is obtained, and the width of a predetermined pattern is measured.

S27 decides whether it is the end. If YES, exit. If NO, return to S24 and repeat.

After automatically correcting the heights of the stages Z1, Z2, and Z3 as described above, it is possible to move the XY stage to the measurement point, perform WD correction, autofocus, and automatically measure the length of the pattern. . At this time, even if the pallet is replaced, the heights of the stages Z1, Z2, and Z3 can be automatically corrected and the horizontality can be automatically corrected. High-speed length measurement is now possible by moving.

FIG. 19 shows a mask tilt measurement flowchart of the present invention. This is the height measurement flow chart of the sample 7 when the height of the sample 7 is actually measured. Here, even one height sensor can measure the height of the photomask, but if two height sensors are used, the height of the portion irradiated with the electron beam can be obtained. A detailed description will be given below.

In FIG. 19, S31 moves the sample to measurement coordinates (Xn, Yn). This moves the sample 7 (photomask) to the indicated measurement point (Xn, Yn) with the XY stage 9 (while performing precise measurement with a laser interferometer).

S32 determines whether the measurement point is within the area measurable by the two sensors. This is because when the two height sensors (Z1, Z2) shown in FIG. 1 are used and both of these two height sensors 151, 152 are within the measurement area, YES in S32 and proceeds to S33. In the case of NO, since it is found that one or both of the two height sensors are not within the measurement area (outside the measurement area), the result of S32 is NO, and the process proceeds to S34.

In S33, the two height sensors 151 and 152 are found to be within the measurement area due to YES in S32, so the average value of the two sensors is determined as the height.

In S34, since it is determined that one or two of the two height sensors are not within the measurement area due to NO in S32, it is further determined whether the height of Z1 is OK. Since it turns out that one or two of the two height sensors are not within the measurement range, it also determines if the height of Z1 is OK (within the measurement range). If YES, the average height of the height Z1 of the OK height sensor and the height Z (X+L, Y) of the position (X+L, Y) of the NG height sensor on the focus map Calculate (Z1+Z(X+L, Y))/2. On the other hand, in the case of NO in S34, the process proceeds to S36.

In S36, since it is determined that the height of Z1 is NG due to NO in S34, it is further determined whether the height of Z2 is OK. This is because the sensor of Z1 out of the two height sensors (Z1, Z2) is found not to be within the measurement range, so it is further determined whether the height of Z2 is OK (within the measurement range). In the case of YES, the height Z2 of the OK height sensor (Z2) and the height Z (XL , Y) and the average height (Z2+Z(XL, Y))/2 is calculated. On the other hand, in the case of NO in S36, both the Z1 and Z2 height sensors are found to be NG, so the height Z of the map values (X, Y) is calculated to be the same.

As described above, the height is measured using the height sensors (Z1, Z2) while the sample 7 (photomask) is moved to the designated measurement point and positioned. Real-time and high accuracy by calculating the average value of the height of the sample measured when only one is possible and the height on the map if the other cannot be measured, and calculating from the height on the map when both cannot be measured. It is possible to measure the height of the position on the photomask indicated by .

FIG. 20 shows an autofocus flowchart using stage Z of the present invention. This is a flow chart for autofocusing the height Z0 of the sample 7 using the stage Z (Z1, Z2, Z3).

In FIG. 20, S41 moves the stage Z to the vicinity of the focus.

S42 moves the stage Z by a predetermined amount.

A step S43 acquires an image for focus evaluation. These S41 to S43 are Z near the height of the surface of the sample 7 fixed to the sample holder 6 on the piezoelectric elements Z1, Z2, and Z3 mounted on the XY stage 9 in FIG. A predetermined voltage is applied to the piezoelectric elements Z1, Z2, and Z3 (stage Z) to move the piezoelectric elements Z1, Z2, and Z3 (stage Z) in the direction (S1). image for focus evaluation) are sequentially obtained (S42, S43). The movement in the Z direction is precisely moved by applying a predetermined voltage to the piezoelectric elements Z1, Z2 and Z3 while measuring in real time using a height sensor. Same below.

S44 discriminates whether it is the end. It is determined whether acquisition of the image for focus evaluation is completed. If YES, go to S45. In the case of NO, S42 and subsequent steps are repeated.

S45 estimates the in-focus distance. This is done by creating a function that expresses the distance in the Z direction from the images before and after the focal point position (position in the Z direction) of the stage acquired in S43, and the degree of focus at that time on the vertical axis. The in-focus distance in the Z direction is calculated from the function and estimated as the in-focus distance.

S46 moves the stage to the in-focus position. For this, the stages Z1, Z2, and Z3 are moved to the in-focus position (distance in the Z direction) estimated in S45.

As described above, by automatically adjusting (autofocusing) the distance in the Z direction at the measurement point (pattern) of the sample 7 using the stage Z, the WD is always kept constant, the magnification is always kept constant, and the electronic The irradiation direction of the beam can always be perpendicular to the surface of the sample 7, or the like.

1 is a configuration diagram of an embodiment of the present invention; FIG. It is an optical height measuring device (bottom view) of the present invention. 1 is a stage overview diagram of the present invention; FIG. FIG. 4 is a detailed overview of the stage of the present invention; It is a photomask fixing mechanism diagram (front side) of the present invention. It is a photomask fixing mechanism diagram (back side) of the present invention. FIG. 4 is an explanatory diagram of a photomask fixing mechanism of the present invention; 4 is a photomask tilt measurement flow chart of the present invention. It is an example of a mask tilt amount measuring point of the present invention. FIG. 9 is an example of the height Z of the measuring point of the present invention (FIG. 9). It is an example of the tilt amount of the stage position of the present invention. It is an example of an approximation aspect (linear approximation) passing through the measurement points of the present invention. 4 is a photomask tilt measurement flow chart of the present invention. It is an example of a mask tilt amount measuring point of the present invention. It is an example of the height of the stage of this invention. There is an example of the correction amount of the stage of the present invention. It is an example of a correction amount to which an offset is added according to the present invention. 4 is a WD control CD measurement flowchart with photomask tilt correction according to the present invention. It is a mask inclination measurement flowchart of this invention. 4 is an autofocus flow chart using the stage of the present invention;

1: electron gun 2: electron beam 3: deflector 4: electron detector 5: objective lens 6: sample holder 7: sample (specimen)
8: Piezo piezoelectric element 9: XY stage 11: Overall control 12: Just focus determination means 13: Image processing means 14: Autofocus control means 15: Sample height information processing means 16: Stage Z control means 17: XY stage control means 151, 152: Sensor (height sensor)
153, 154: reflectors 155, 156: optical fibers Z1, Z2, Z3: stage Z (height stage)

Claims (3)

In an apparatus for correcting the tilt of the sample and adjusting the distance between the sample and the objective lens to be constant ,
an XY stage that moves the sample in the XY directions;
a stage Z consisting of a stage Z1, a stage Z2 and a stage Z3 for moving the sample in the height direction;
a height measuring device that measures the height of any three or more points of the sample fixed on the stage Z;
Based on the result of measuring the tilt in the height direction of the sample fixed on the stage Z by the height measuring device, the tilt in the height direction of the sample generated when the sample is moved by the XY stage height correction means for correcting the tilt to zero, and then independently correcting the heights of the stages Z1, Z2, and Z3 so that the distance becomes constant;
The stage Z has a support portion for holding a sample holder having a receiving portion, and a mechanism having a piezoelectric element that contracts when voltage is applied to correct the height,
The support portion holds the sample holder so as to maintain positioning even when the XY stage moves by fitting into the hole of the receiving portion due to the gravity of the sample holder. Tilt corrector. The height measuring device has a reflecting mirror for reflecting a laser beam on the lower surface of an objective lens that irradiates the sample with a narrow electron beam and scans the sample in a plane. 2. The sample tilt correction according to claim 1, wherein the light is reflected downward in a direction to irradiate the sample at three arbitrary points, and the height is measured based on the light reflected from each of the three points. Device. In the adjustment method for correcting the tilt of the sample and adjusting the distance between the sample and the objective lens to be constant ,
an XY stage that moves the sample in the XY directions;
a stage Z consisting of a stage Z1, a stage Z2 and a stage Z3 for moving the sample in the height direction;
a height measuring device that measures the height of any three or more points of the sample fixed on the stage Z,
Based on the result of measuring the tilt in the height direction of the sample fixed on the stage Z by the height measuring device, the tilt in the height direction of the sample generated when the sample is moved by the XY stage providing a height correction step of correcting the tilt to zero, and then independently correcting the heights of the stages Z1, Z2, and Z3 so that the distance becomes constant;
The stage Z is provided with a support portion for holding a sample holder having a receiving portion, and is provided with a mechanism having a piezoelectric element that contracts when voltage is applied to correct the height,
The support portion holds the sample holder so as to maintain positioning even when the XY stage moves by fitting into the hole of the receiving portion due to the gravity of the sample holder. Tilt correction method.

Images