JP2503508B2 - Transfer device and pattern forming method using the transfer device - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to an apparatus for exposing a mask pattern on a photosensitive substrate, and more particularly to a transfer and exposure apparatus for fine patterns used in the manufacture of semiconductor elements (optical exposure). System, X-ray exposure system, or electron beam exposure system).

[Conventional technology]

In recent years, in the manufacture of semiconductor devices (VLSI, etc.), the minimum line width of patterns handled in the photolithography process has become thinner and narrower, and the line width that can be resolved by a reduction projection type exposure apparatus (so-called stepper) and submicron of 1 μm or less. Has reached the realm of. Furthermore, in an excimer stepper using excimer laser light as illumination light for exposure, it is possible to resolve a line width of about 0.6 μm. Further, an exposure apparatus using a soft X-ray having a wavelength of about a dozen angstroms is also attracting attention as a method for easily obtaining a high resolution. When the line width forming the circuit pattern becomes thin in this way, the accuracy of alignment between the mask (or reticle), which is the original plate, and the photosensitive substrate (semiconductor wafer coated with resist, etc.) naturally becomes a problem. This alignment, that is, the alignment technique has been improved year by year along with the description of photoelectric sensing, and the accuracy has been obtained so far as it is not practically hindered if not sufficient at present. After this alignment operation, the mask (reticle) and the photosensitive substrate are prevented from moving relative to each other, and the pattern of the mask (reticle) is exposed. The time required for this exposure operation is about 0.2 seconds to 1 second for the optical stepper and about 10 to 30 seconds for the X-ray stepper. Therefore, if the relative position between the mask and the photosensitive substrate changes during the exposure time due to some cause, such as external vibration or air fluctuation, the correct resolving power cannot be obtained, and the pattern transferred in the resist cannot be obtained. Deformation (thickness of line width) and displacement will occur.

Therefore, a method of detecting a relative displacement from the alignment completion position during the exposure operation and issuing a warning when the displacement amount exceeds the allowable range is disclosed, for example, disclosed in Japanese Patent Laid-Open No. 61-53724. Has been done.

[Problems to be solved by the invention]

In the above conventional technique, simply a mask (reticle) is used.
Since only the relative positional displacement between the photosensitive substrate and the photosensitive substrate is detected, the relative positional displacement during exposure and the positional displacement of the pattern that can be confirmed after developing the photosensitive substrate do not necessarily exactly match. There was a problem. This means that even if the relative displacement exceeds the permissible range, it occurs only momentarily during the entire exposure time, or if the relative displacement is within the permissible range, the relative displacement is maintained over the entire exposure time. This is because, for example, a case where the positional deviation progresses occurs. For this reason, in the above-mentioned conventional technique, there arises such inconvenience that a warning is issued when it is not necessary, or a warning is not issued when the positional deviation of the pattern becomes large.

[Means for solving problems]

In order to solve the above problems, the present invention detects a positional deviation amount that detects a relative positional deviation amount generated between a mask and a photosensitive substrate during an exposure operation of a mask (or reticle) pattern on the photosensitive substrate. Means (interferometer 7), means for producing positional deviation history information (up / down counter 30, clock generator 32, memory 34, etc.), and photosensitive characteristics of the photosensitive substrate (especially resist) (for example, the resist layer is completely developed after development). Exposure threshold value) and history information,
A deformation detecting means for detecting (or predicting) the deformation of the pattern exposed on the photosensitive substrate by calculation is provided.

[Work]

In the present invention, the change in the light amount distribution of the transfer pattern edge portion during the exposure operation is considered in accordance with the relative positional deviation between the mask and the photosensitive substrate that occurs during the exposure operation,
The deviation amount of the line width of the pattern resolved on the resist, the deviation amount of the central position of the pattern, and the like are predicted based on the change of the light amount distribution and the characteristics of the resist. For this reason, the deformation of the pattern transferred onto the resist is required more accurately than in the conventional method.

〔Example〕

FIG. 1 is a diagram showing the overall configuration of an exposure apparatus according to an embodiment of the present invention, showing the configuration when applied to a wafer stepper that exposes a reticle pattern on a wafer by a step-and-repeat method. Illumination light from an exposure light source (not shown) is incident on a condenser lens (a part of the illumination optical system) 1 through a shutter 2 that can be opened and closed, and is made uniform illumination light here to make the reticle R uniform. Illuminate. The reticle R is held by the reticle holder 3. The image of the circuit pattern or the like formed on the reticle R is image-projected on the wafer W by the projection lens 4. A resist layer having a predetermined thickness is applied to the surface of the wafer W. The wafer W is placed and held on the stage 5 which is two-dimensionally movable, and its movement is controlled by the motor 6. The two-dimensional position of the stage 5 is sequentially detected by a laser light wave interferometer length measuring device (hereinafter referred to as an interferometer) 7 with a resolution of 0.02 μm, for example. Although only a pair of the motor 6 and the interferometer 7 is shown in FIG. 1,
Actually, a pair is provided for each of the x-direction and the y-direction and controls the two-dimensional movement of the stage 5 and the stage 5.
{That is, the coordinate position of the wafer W) is measured.
The alignment optical system 8 detects an alignment mark during relative alignment between the reticle R and one shot area on the wafer W to be exposed, that is, during alignment. The method and configuration of the alignment optical system 8 need not be particularly limited in this embodiment, and any device provided in the wafer stepper will not cause any problem. The main control system 10 integrally controls the apparatus of this embodiment, and inputs coordinate measurement values from the interferometer 7 and outputs drive commands to the motor 6. Further, the main control system 10 exchanges predetermined commands and information with the control unit 12 that controls the opening and closing of the shutter 2, and the pattern deformation detection unit 14 shifts the pattern transfer device on the resist during the exposure operation. Enter information about line width deviation. The positional deviation history detection unit 16 inputs the positional information from the interferometer 7 and the exposure state signal from the shutter control unit 12, and creates positional deviation history information within the exposure time, specifically, a positional deviation histogram. To do.

The pattern deformation detection unit 14 calculates the deviation of the line width, the deviation of the pattern transfer position (line width center position), and the like based on the history information and the known photosensitive characteristics of the resist, and the result is mainly used. Output to control system 10. Based on the result, the main control system 10 determines whether or not the exposure has been performed with sufficient resolution, and if insufficient, issues a warning, display, etc.

FIG. 2 is a circuit block diagram showing a specific configuration of a part of the displacement history detecting unit 16. Up (UP) and down (DOWN) pulses from the interference system 7 are input to the up / down count circuit (UDC) 30. The interferometer 7 generates the up / down pulse each time the stage 5 moves 0.02 μm, for example. The UDC 30 starts counting up / down pulses based on the input of the status signal STS indicating the exposure state from the shutter controller 12. AND gate circuit
Reference numeral 31 is a clock generator that generates clock pulses at intervals sufficiently shorter than the exposure time for one shot area.
Input the pulse signal from and the status signal STS,
The clock pulse is output only during the exposure period. The address counter (CNT) 33 sequentially counts clock pulses based on the input of the clock pulse, and sequentially updates the access address (address value) to the memory (RAM) 34. The RAM 34 sequentially stores the count value from the UDC 30 at the designated address. That is, the RAM 34 stores a history of the relative positional deviation amount during the exposure operation.

FIG. 3 is a graph showing a relative position deviation during exposure operation, that is, history information, where the horizontal axis represents time and the vertical axis represents the amount of positional deviation (for example, the number of up / down pulses from the interferometer 7). ) Is represented. Time ts to time te
Is the exposure time, and the clock pulse from the clock generator 32 is generated at the time indicated by the black circle in the figure.

Further, it is assumed that the count values of UDC30 and CNT33 are both set to zero (or another predetermined value) before time ts. In addition, in FIG. 3, before time ts, the reticle R and the wafer W are aligned within a predetermined error range by the alignment operation and are in a substantially stationary state.

Although not shown in FIG. 2, the history information stored in the RAM 34 is converted into a histogram as shown in FIG. 4 by the arithmetic processing circuit provided in the detector 16. In FIG. 4, the horizontal axis represents the positional deviation amount (pulse number) Δx in the x direction, and the vertical axis represents the frequency of each deviation amount. Normal interferometer 7
Since the output of is a digital value (pulse train), such a histogram expression is realistic, but for the time being after that, in order to simplify the explanation, consider the continuous curve of FIG. 5 instead. To do. The curve of FIG. 5 is also called a positional shift relationship, and in the case of the step-and-repeat method, the shape of the curve is not always constant for each shot. Now, suppose that the pattern formed on the reticle R is one linear light-shielding portion CP formed in the transparent portion PR as shown in FIG. If a positional deviation as shown in FIG. 5 occurs during exposure of this pattern, the calculated light intensity distribution obtained on the resist can be easily obtained by convolving the light and dark information of the original pattern with the positional deviation relationship. To be In fact, since the pattern on the reticle R is a very simple pattern (simple light and dark), it is easy to understand that the pattern shown in FIG. 7 is obtained. FIG. 7A shows a light amount distribution when the reticle R and the wafer W are stationary at the correct positions and no relative position error occurs. When the positional deviation (same as FIG. 5) as shown in FIG. 7 (b) occurs in this, FIG. 7 (c)
A light amount distribution as shown in is obtained. In FIG. 7 (c), the right edge has a waveform obtained by integrating the waveform of FIG. 7 (b) at the position, and the left edge has a waveform obtained by inverting the waveform. A light-shielding portion in which a light amount distribution having a waveform as shown in FIG. 7C actually occurs on the resist
It becomes a projected image of CP.

On the other hand, let us consider the characteristic of the light quantity of the resist versus the residual quantity after development (photosensitive characteristics of the resist). This is generally shown as in FIG. 8 (a), but here it is simply used as in FIG. 8 (b). In FIG. 8, the horizontal axis represents the integrated light amount (dose amount), and the vertical axis represents the remaining resist amount (layer thickness). In this photosensitive characteristic, the light amount P in FIG. 8B represents a boundary value at which the resist is completely removed.
By matching this characteristic with the light amount distribution obtained in FIG. 7 (c), it is possible to know where the actual pattern is obtained on the resist. FIG. 9 is a diagram showing how to obtain it, and FIG. 9 (a) shows a pattern (a light-shielding portion CP having edges E ₁ and E ₂ ) on the assumption that exposure is performed at a correct position. FIG. 9 (b) shows the light quantity distribution (the same as FIG. 7 (c)) when there is a position shift, and FIG. 9 (c) binarizes (b) with the light quantity threshold P of the resist. The expected pattern obtained by FIG. 9 (d) shows the zero point of the displacement curve in correspondence with the left and right edges E ₁ and E ₂ in FIG. 9 (a). In FIG. 9, the position of the right edge E ₂ ′ in (c) is integrated (shaded portion) from the left of the position deviation curve in FIG. 9 (d), and its value matches the light amount threshold value P. The left edge E ₁ ′ is the point at which the position deviation curve of FIG. 9 (d) is integrated from the right and its value matches the threshold value P.

Now, the value to be evaluated now is the deviation of the pattern waveform of FIG. 9 (c) with respect to the pattern waveform shown in FIG. 9 (a), which is generally considered separately for the deviation of the line width and the deviation of the center. Target. If the deviation of the right edge E ₂ ′ shown in FIG. 9 (c) is r and the deviation of the left edge E ₁ ′ is l ( _{−1 in the} figure because it is displaced in the opposite direction to r), the line width deviation ( The line width is 1/2 (r-1), and the center deviation is 1/2 (r + 1).

As described above, r and l are obtained by integrating the displacement curve from the left and right until a predetermined value is reached. This is the fourth
Returning to the histogram of the figure, just right (forward direction)
This is because counting the frequency from the beginning or counting the frequency from the left (in the negative direction).
Easy to achieve with 14. Of course, the result is the same even if the direction of integration is reversed and a point that coincides with (total exposure amount-P) is searched for. Here, the total exposure amount means that the exposure time to be applied to the wafer, the illuminance, etc. are known in advance. Here, it is assumed that the pattern on the reticle R is a "line" formed by the light-shielding portion CP as shown in FIG. 6, but if this is a "slit" that is light-dark inverted, only the sign of the line width deviation result is inverted, The method of finding is exactly the same. Further, since the above value is not affected with respect to the difference in the width of the assumed pattern on the reticle R, this result can be used for evaluating the exposure of any pattern.

Therefore, the operation of this embodiment is summarized as follows. When the wafer W is exposed by the step-and-repeat method, the shutter 2 is released when the relative positioning with the reticle R with respect to one shot area on the wafer W is completed.
Is opened. In response to this, the RA of the misregistration history detection unit 16
The history information of the relative displacement is sequentially stored in M34. When the shutter 2 is closed and the exposure operation is completed, the pattern deformation detection unit 14 detects the line width deviation and the center deviation.
Calculate two quantities. In parallel with this, the stage 5 starts moving for exposure to the next shot area.

When the line width deviation and the center deviation are obtained by the above equation, they are sent to the main control system 10. The main control system 10 stores the deviation amounts in correspondence with the map of the shot area on the wafer W, and gives a warning or display in real time as needed. Alternatively, when the exposure of one wafer W is completed, the shot map on the wafer is displayed on the cathode ray tube together with the map.

As described above, the pattern deformation detecting unit 14 integrates (integrates) the position deviation curves obtained as shown in FIG. 5 [or FIG. 7 (b) and FIG. 9 (d)] from the left and right. It is only necessary to perform an extremely simple calculation such as obtaining the deviation amount (r, l) from the zero point of the point where the integrated value coincides with the light amount threshold P based on the photosensitive characteristics of the resist. Further, in the case of the present embodiment, since the positional deviation information is treated as a discrete histogram as shown in FIG. 4, the predicted value of the amount of line width deviation and center deviation is the measurement resolution of the up / down pulse from the interferometer 7 ( For example, 0.02 μm). The zero point in the positional deviation curve shown in FIG. 5, FIG. 7B, and FIG. 9B of this embodiment is the design of the shot area on the wafer W on which the pattern of the reticle R is exposed. It corresponds to the upper position, that is, the shot address position. It goes without saying that pattern deformation detection is similarly performed for both the x direction and the y direction.

Next, a second embodiment of the present invention will be described. In the above-described first embodiment, when the exposure amount for one shot area on the wafer is exposed, if the line width deviation amount or the center deviation amount exceeds a predetermined value, only a warning or display is given. In the second embodiment, the center deviation is corrected more positively. Therefore, in this embodiment, it is possible to delicately control the overlay accuracy of the pattern newly transferred to the wafer and the pattern already formed on the wafer.

Therefore, in the present embodiment, hardware or software for sequentially calculating the pattern center deviation and the line width deviation at high speed during the exposure operation is required. Here, as a representative, the procedure of software processing will be described based on the flowchart of FIG. In this embodiment, the center deviation is corrected by moving the stage 5 by a small amount, but the correction is performed under certain restrictions. One limitation is that the line width deviation and the center deviation cannot be corrected independently. That is, in order to correct the center deviation,
This means that the amount of line width deviation must be aggravated (increased). Therefore, the correction of the center deviation can be performed within the range when the allowable value of the line width deviation has a margin.

Now, the flowchart of FIG. 10 shows the clock pulse generation interval (for example, 2 mSe) during one shot exposure operation.
It is executed sequentially for each c). In FIG.
When the clock pulse is generated, the coordinate measurement value from the interferometer 7 is read and the difference Δx from the target stationary position (shot position) of the stage 5 is stored in the memory (step 100).
At this time, in order to create a histogram (directly from the difference Δx), a predetermined address in the memory is made to correspond to the target stationary position,
Positive displacement amount (+ Δx) in the direction from which the address increases
Then, the direction in which the address decreases is defined as a negative displacement amount (−Δx), and the increase or decrease of each address (address value) is made to correspond to the number of up / down pulses from the interferometer 7. Then, the content of the address corresponding to the read difference Δx is set to 1 (or a constant value).
Only try to increase. As a result, a histogram is created in the address area having the memory. In the case of the present embodiment, the created histogram corresponds to the history information of the present invention. Next, the histograms being created in the memory are integrated (area calculation) from the left and right as shown in FIG. 9 (d) (step 102). Then, it is judged whether or not the integrated value K becomes larger than the light amount threshold value P (step 104), and if K> P, the program of this flowchart changes from the program of this flowchart to the main program in order to wait until the next clock pulse is generated. To return. Step
When it is determined that K ≦ P in 104, the center deviation amount ΔXP and the line width deviation amount ΔDP are calculated in the same manner as in the previous embodiment.

 That is, ΔXP = 1/2 (r + 1) and ΔDP = 1/2 (r-1).

Next, it is judged whether or not the calculated line width deviation amount ΔDP is smaller than a predetermined allowable value TD (step 10
8) If ΔDP <TD, the stage 5 is finely moved in the x direction or the y direction by a corresponding amount to correct the center deviation (step 110), and the process returns to the main program.
If ΔDP ≧ TD is determined in step 108, the line width deviation amount has reached the allowable value, so it is no longer meaningful to correct the center deviation, and a warning or display is given at this point ( Step 112).

That is, by executing step 112 during the exposure operation, the positional deviation during the exposure can be detected in real time, and the exposure can be stopped at that time. Further, step 110 in FIG. 10 is shown to be always executed if ΔDP <TD during the exposure operation for one shot, but if step 110 is executed for the first time, it may be corrected only once. Good. Further, the same effect can be obtained in the correction in step 110 even if the reticle R is slightly moved. As for the center deviation, the maximum correction amount is determined from the margin of the line width deviation at the time of determination in step 108, and therefore the center deviation may exceed the allowable value TD even if the correction is performed. If it is known, step 112 may be executed to issue a warning. Further, as shown in FIG. 10, when the deviation of the edge position is sequentially calculated during the exposure, the integrated value K of the histogram obtained in step 102 is (total light amount-
The point that coincides with P) may be obtained for each sampling by the clock pulse. In this case, the total light amount means that the exposure time is predetermined under a certain exposure illuminance, for example. Assuming that the total light intensity (set value) is 0.5 seconds, the light intensity threshold value P is 0.3 seconds under the illuminance, and the clock pulse sampling period is 2 mSec, the sampling number corresponding to the total light intensity (after completion of exposure) Is the same as the integrated value of the histogram of) and the number of samplings corresponding to the threshold value P is 150, so the value corresponding to (total light amount-P) is 100. Therefore, when the histograms being created are integrated from the left and right in step 102, the integrated value K is 100.
When it matches with, step 106 and subsequent steps are executed. However, in this case, the difference is that the correspondence between the assumed edge position deviation direction and the integration direction is opposite.

Next, other embodiments of the present invention will be described. In each of the following embodiments, the method of detecting the center deviation and the line width deviation described in the previous embodiment is applied as it is. In the third embodiment, as described in step 106 in FIG. 10,
The center deviation and line width deviation are obtained for each shot during the exposure of one wafer by the step-and-repeat method. By the time the exposure of one wafer is completed, it is counted for each shot whether or not the center deviation amount or the line width deviation amount exceeds an allowable value, and the number of shots exceeding the allowable value is predetermined. If the specified number is exceeded, a warning or display will be issued. Warnings and indications are issued during or after the exposure operation of one wafer. In the third embodiment, when a warning is issued for each of a plurality of wafers to be exposed, it is understood that the static state of the stage is deteriorated due to some instability factor or disturbance (floor vibration, etc.). .
If a warning is issued for a small number of the plurality of wafers, only that wafer can be prevented from being sent to the next process.

In the fourth embodiment, the center deviation or line width deviation is sequentially stored during the step-and-repeat exposure operation in association with the shot position on the wafer. When a plurality of wafers in the same lot are exposed, the same memory is stored for each wafer, and when at least the second and subsequent wafers are exposed, the center shift and the line at a specific shot position on the wafer are performed. A warning display is generated when the width is frequently misaligned. The warning and display may be generated during the exposure operation for the second and subsequent wafers, or after the exposure operation for one lot of wafers is completed.

According to the fourth embodiment, it can be seen that the positional deviation frequently occurs at the specific shot position on each of the wafers in the lot, and therefore the shots to be inspected with priority at the time of the inspection in the subsequent process can be known in advance. become.

In the fifth embodiment, similar to the fourth embodiment, the line width deviation amount is sequentially stored for each shot position on the wafer. Then, the exposure is completed for a plurality of wafers (at least two), and 3
When performing exposure of the first and subsequent wafers, statistically, for shot positions where the line width deviation amount becomes large,
Control is performed so as to lengthen the settling time of the stage when positioning the shot area with respect to the reticle. Normally, after the stage is stepped and the stage is driven within the predetermined positioning error with respect to the target stationary position, the time staying within the error is measured and after the predetermined time (several hundred mSec) has elapsed. When it is within the error, exposure is performed. Therefore, when it is known that the positional deviation (line width deviation) is statistically large at the next shot position, control is performed so that the predetermined time measured after stepping, that is, the settling time is made longer than in the normal case. Therefore, according to the fifth embodiment, in the process of processing a large number of wafers, it is possible to reduce the positional deviation even for a specific shot having a large positional deviation.

As described above, in each of the embodiments of the present invention, the positional deviation is checked during the exposure of the semiconductor element pattern on the wafer, but it is not always necessary during the exposure of the element pattern. The same can be applied at the stage of adjusting and inspecting the positioning accuracy, stepping accuracy, etc.

The embodiments of the present invention have been described above.
In particular, the present invention can be widely applied to devices other than the exposure device using light in the ultraviolet region, such as an X-ray exposure device using a soft X-ray and an electron beam exposure device using an electron beam. In addition to masks for semiconductor device manufacturing, even in a device that projects a simple rectangular pattern on a substrate and processes the substrate surface with an energy beam that is narrowed into a rectangular shape, the projected image of the rectangular pattern and the relative position of the substrate Misalignment can be detected during processing.

〔The invention's effect〕

As described above, according to the present invention, the positional deviation of the pattern projected on the substrate, the dimensional deviation, etc. are taken into consideration during the projection operation or after the projection, and the positional deviation history information and the characteristics of the sensitive layer (resist) are taken into consideration. Since it is predicted and detected, it is possible to accurately identify the pattern abnormality without developing the resist or inspecting using another measuring device. For this reason, in a semiconductor device production site, it is possible to find a shot having a poor resolution among shots on an exposed wafer before sending it to the next step, which is extremely useful in production control.

Further, according to the pattern forming method of the present invention, the shape change (line width change or center position change) of the pattern image transferred in the photosensitive layer on the photosensitive substrate is displaced from the alignment state between the mask and the photosensitive substrate. Since it is calculated based on the history information of the amount, it is possible to make a prediction that substantially matches the amount of deformation of the actual pattern image obtained on the photosensitive substrate after development.
It is possible to issue a warning or display for an exposure shot in which a position shift that is a real problem occurs.

Further, according to the pattern forming method of the present invention, the history information of the positional deviation from the alignment state between the mask and the photosensitive substrate is sequentially generated during the transfer period, and the history information is transferred onto the photosensitive substrate based on the history information. Since the deformation of the pattern image (change in line width and center position) is specified in real time and the relative position of the mask and photosensitive substrate is corrected during the transfer operation, the pattern image transferred to the photosensitive layer is substantially Since the correction works on the true positional deviation that acts so as to deform, more faithful pattern formation is possible.

[Brief description of drawings]

FIG. 1 is a diagram showing a configuration of a projection exposure apparatus according to an embodiment of the present invention, FIG. 2 is a circuit block diagram showing a part of the configuration of a misregistration history detection unit, and FIG. 3 is a graph showing a misregistration history. FIG. 4 is a graph showing a histogram created according to the history of displacement, FIG. 5 is a graph in which the histogram of FIG. 4 is replaced with a continuous curve, and FIG. 6 is an example of a pattern provided on the reticle. FIG. 7 (a),
8 (b) and 8 (c) are charts showing how the light amount distribution changes due to the positional deviation of the transfer pattern formed in the resist, and FIGS. 8 (a) and 8 (b) schematically show the photosensitive characteristics of the resist. Graphs shown in FIGS. 9 (a), (b), (c),
(D) is a chart showing the state of the transfer pattern on the resist obtained by the positional deviation, and FIG. 10 is a flow chart showing the operation according to another embodiment of the present invention. In each figure, R ... Reticle W ... Wafer 5 ... Stage 7 ... Interferometer 10 ... Main control system 14 ... Pattern deformation detection unit 16 ... Position deviation history detection unit

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI technical display location H01L 21/30 541S

Claims (6)

(57) [Claims] 1. Illuminating a mask having a predetermined pattern,
In an apparatus for transferring the pattern onto a photosensitive substrate aligned in a predetermined positional relationship with the mask, a relative positional shift between the mask and the photosensitive substrate that occurs during an operation of transferring the pattern onto the photosensitive substrate. Position deviation detecting means for detecting the amount; means for generating positional deviation history information based on the detected positional deviation amount; and transfer to the photosensitive substrate based on the photosensitive characteristic of the photosensitive substrate and the history information And a deformation detecting means for detecting the deformation of the formed pattern. 2. The means for generating the positional deviation history information includes means for creating a histogram corresponding to the frequency of relative positional deviation amount with reference to an alignment target position at which the mask and the photosensitive substrate should be aligned. 2. The apparatus according to claim 1, wherein the deformation detecting unit calculates a deviation amount of an edge position of the transferred pattern based on a histogram created and a threshold value on the photosensitive characteristic. . 3. The deformation detecting means separately calculates a change in the line width of the transferred pattern and a change in the transfer center position based on the deviation amount of the edge position obtained by the calculation. The device according to claim 2. 4. Illuminating a mask having a predetermined pattern,
A method of forming a pattern on the photosensitive substrate using a transfer device that sequentially exposes the image of the pattern to each of a plurality of shot areas on the photosensitive substrate, comprising: Sequentially detecting the relative positional deviation amount from the alignment state of the mask and the photosensitive substrate that occurs during the transfer operation, and generating positional deviation history information based on the positional deviation amount; A step of calculating change information corresponding to a shape change or a position change of a pattern transferred in the photosensitive layer of the photosensitive substrate, based on the calculated change information; It is determined whether the change in the image of the transferred pattern is within the permissible range, and if there is a shot area that exceeds the permissible range, the photosensitive group of the shot area is determined. Pattern forming method which comprises the step of displaying information representing the position of the above. 5. A plurality of holders on the photosensitive substrate, comprising: a holder for holding a mask on which a predetermined pattern is formed; and a stage for holding a photosensitive substrate exposed by the image of the pattern and moving it two-dimensionally. In the method of forming a pattern on the photosensitive substrate by using a transfer device that sequentially transfers the image of the pattern to each of the shot areas, the pattern image is detected during the transfer period of the image of the pattern to the shot area. The positional deviation history information is sequentially generated based on the relative positional deviation amount from the alignment state between the pattern image and the shot area, and is transferred into the photosensitive layer of the photosensitive substrate based on the historical information. Sequentially calculating the amount of deformation of the image of the pattern; and sequentially determining whether or not the calculated amount of deformation of the image exceeds a predetermined allowable value during the transfer period. As a result of the determination, when the calculated deformation amount of the image does not exceed the allowable value, the position of at least one of the holder and the stage is adjusted so that the calculated deformation amount of the image is comprehensively corrected. And finely correcting before the end of the transfer period. 6. The method according to claim 5, wherein the finely correcting step is performed only once during the transfer period.