JP5080378B2 - Fine processing method using atomic force microscope with excellent height controllability - Google Patents

Description

  The present invention relates to a method of mechanically processing a fine region using an atomic force microscope.

  Nanometer-order microfabrication technology is required for advanced functions and high integration, and research and development of processing technologies such as local anodization and microscratch processing using a scanning probe microscope (SPM) is actively pursued. Has been done. Recently, not only the pursuit of the possibility of fine processing, but also a precise shape and high-precision processing are required as a practical processing machine.

  In recent years, as an example in which an accurate shape and high-accuracy processing are required in an apparatus based on an atomic force microscope (AFM), there is correction of a pattern surplus defect of a photomask (Non-patent Document 1). Photomask surplus defect correction by AFM is performed by imaging in normal AFM contact mode or intermittent contact mode during observation using an AFM probe harder than the material to be processed (material of surplus defects). Recognizing and processing is performed by cutting feedback and fixing a hard probe at the same height as the underlying glass surface and physically removing excess defect portions on the glass surface by scanning. Furthermore, in this case, the problem that it is difficult to observe and process defects due to charge-up in a focused ion beam defect correction apparatus that has been used as a correction apparatus for fine defects in conventional masks is also overcome. I know I can fix it. For this reason, recently, a defect correction method using an AFM-based apparatus has been used at a mask manufacturing site. Since the mask serves as the master for wafer transfer, if the processing accuracy at the correction location is low, or if there is overetching or uncut material, transfer characteristics will be adversely affected, causing device defects on all transferred wafers. Therefore, an accurate shape and high-precision processing are required in the AFM removal processing.

  In defect correction, the deviation of the CD value between the normal part and the defect correction part at the time of defocusing at the time of wafer transfer is an important factor in practical use as well as the transmittance and correction accuracy. This is because the magnitude of the deviation depends on the cutting height of the glass substrate in the processing region, and height controllability is important to ensure the likelihood (Non-patent Document 1).

Furthermore, if there is a drift in the height direction, it takes time to measure the height reference and actually start machining, so a deviation from the target height occurs, reducing the height controllability of the machining location. It was. Such drift in the height direction is caused by the heat caused by the temperature difference between the sample and the internal environment of the device, or when the ball screw stage is used as a coarse motion mechanism, etc. Etc.). In order to ensure the likelihood at the time of the defocus mentioned above, the process excellent in height controllability which considered the influence of drift as much as possible is desired.
T. Amano, M. Nishiguchi, H. Hashimoto, Y. Morikawa, N. Hayashi, R. White, R. Bozak, and L. Terrill, Proc. Of SPIE Vol. 5256 538-545 (2003)

  An object of the present invention is to provide a method for mechanically processing a fine region using an atomic force microscope with good controllability even when there is a drift in the height direction due to heat or the like.

  In particular, it has been found that the thermal drift can be approximated linearly unless the time is short and extremely large. Utilizing this property, the contribution of thermal drift is mainly corrected out of the drift in the height direction. First, the drift velocity in the height direction is calculated from the difference between the measured values of two height measurements with an arbitrary time difference for the same target, and the drift amount predicted according to the calculated drift velocity is calculated in the height direction. As a correction value, the target height position during processing is corrected, the feedback is canceled, and the predetermined workpiece is removed by mechanical processing with a probe that is harder than the workpiece material fixed at the corrected height. Do.

  Alternatively, the thermal drift uses the property of converging with the lapse of the operation time of the apparatus, and waits for the drift speed to become 5 nm / min or less when calculating the drift speed as described above. Next, when the speed is reached, the feedback is canceled, and the workpiece is removed by performing mechanical machining with a probe harder than the workpiece material fixed at the target height position. Note that the processing time at one place is within one minute as the quality currently required for mask correction. Therefore, it is possible to satisfy the quality by processing after the drift velocity has converged to 5 nm / min or less.

  Alternatively, when an arbitrary number of times of machining are performed on the same position, the target height is corrected using the drift amount predicted according to the drift speed described above at the start of machining for each machining time as a correction value in the height direction. After the feedback is released, the workpiece to be recognized is removed by performing mechanical machining with a probe harder than the workpiece material after setting and fixing to the height. The fixed height correction and machining are repeated with the estimated drift amount until the set number of machining operations are completed.

  When the area of the processing region is large, the processing region is divided into blocks of an appropriate size, and the divided blocks are processed one by one. In this case as well, the target height position is corrected using the drift amount predicted according to the drift speed described above at the start of machining for each block as the correction value in the height direction, and the feedback is released after setting the height. Then, the height of the workpiece is fixed, and the workpiece to be recognized is removed by mechanical machining with a probe harder than the workpiece material. Repeat the fixed height correction and machining with the predicted drift amount until machining of all blocks is completed.

  Alternatively, when the area of the processing region is large, the processing region can be divided into lines having appropriate widths, and the divided lines can be processed one by one. In this case as well, the target height position is corrected using the drift amount predicted according to the drift speed at the start of machining for each line as the correction value in the height direction, and the feedback is released after setting the height. Then, the height of the workpiece is fixed, and the portion to be processed set by mechanical processing with a probe harder than the workpiece is removed. Repeats fixed height correction and machining with the predicted drift amount until machining of all lines is completed.

  Regarding the drift in the height direction, in order to predict the drift amount and correct the deviation in the height direction, it is possible to perform processing with extremely excellent height controllability.

  In particular, due to the nature of the thermal drift in the height direction, machining is started after the drift has converged to a predetermined drift speed or less as the operating time of the device elapses, so machining with extremely high controllability is performed. be able to.

  Even when machining the same position any number of times, for drift in the height direction, the drift amount is predicted and the deviation in the height direction is corrected at the start of machining for each machining time. Therefore, it is possible to perform processing with extremely excellent height controllability.

  Furthermore, when the area of the machining area is large, the machining area is divided into blocks of appropriate sizes, and the drift amount in the height direction is estimated for each block by predicting the drift amount in the height direction at the start of machining. Therefore, even if it takes a long time to process all the blocks, it is possible to perform processing with extremely excellent height controllability.

  Alternatively, if the area of the machining area is large, divide the machining area into lines of appropriate width, and for each line, predict the amount of drift in the height direction at the start of machining and eliminate the deviation in the height direction. Because of the correction, even if it takes a long time to process all lines, it is possible to perform processing with extremely excellent height controllability.

  Hereinafter, a photomask defect correcting device will be described as an example of an atomic force microscope fine processing device.

  The photomask defect correction apparatus according to the present embodiment is an apparatus based on an atomic force microscope (AFM), and a probe (for example, diamond) that is harder than the material to be processed so that black defects can be removed by mechanical processing and at the time of processing. It has a thick cantilever with a high spring constant so that the blade tip is not displaced by twisting the cantilever. A probe harder than the work material is sharpened to a tip diameter of 50 nm or less so that a high-resolution AFM image can be obtained. At the same time, it also has a function to capture coordinate information from the defect inspection apparatus, a high-precision scanner and a high-precision XY stage for high-precision foreign object positioning. Detection of defects in the processed part is performed by scanning the processed surface while applying feedback so that the amplitude attenuation rate is constant by resonating a cantilever having a sharpened probe (hereinafter referred to as dynamic mode) This is done by recognizing the defect as a region. The defect area recognized in this way is detected by fixing the probe at the position where the surface position of the glass substrate or the position slightly dug from the glass substrate surface is the target height position, and scanning. Existing portions are selectively removed as black defects. The processing waste generated after the removal may be removed by wet cleaning or dry ice cleaning.

  A photomask in which black defects are found by the defect inspection apparatus is introduced into an AFM-based photomask defect correction apparatus. The XY stage is moved to the position where the black defect is found by the defect inspection device, the probe is brought close to the photomask, a wide visual field of about 10 μm is observed by AFM, black defects are detected, and the AFM observation is performed with a narrow visual field of 1 to 3 μm. Thus, the black defect area to be removed is recognized.

  Here, it is known that the thermal drift can be approximated linearly unless the time is short and extremely large. By utilizing this property, the contribution of the thermal drift is mainly corrected out of the drift in the height direction in the recognized black defect removal processing described above. FIG. 1 shows a flowchart of the microfabrication method of the present invention. First, the drift velocity in the height direction is calculated from the difference in the measured values of the height measurement with respect to the same position of the glass substrate or pattern twice with an arbitrary time difference. Using the drift amount predicted according to the calculated drift speed as a correction value in the height direction, the target height position during processing (the surface position of the glass substrate 2 or a position dug several nm from the glass substrate 2) is corrected. Only the black defect region 3 recognized by scanning the AFM machining probe 4 fixed at the corrected height after canceling the feedback is selectively selected as mechanical machining with a probe harder than the workpiece material. Remove. The operation of the probe in this case is shown in FIG. As described above, the microfabrication method of the present invention enables correction of black defects with extremely excellent height controllability by correcting deviation due to drift in the height direction mainly caused by thermal drift.

  Or the drift of the said height direction can also be correct | amended using the property that a thermal drift converges with progress of the operating time of an apparatus. FIG. 3 shows a flowchart in that case. Thus, when the AFM probe repeatedly contacts in the measurement in the dynamic mode described above, the height at the same position on the glass substrate or pattern is measured, and the difference between the measured heights before and after The drift velocity in the vertical direction is calculated, and it waits for the drift velocity to converge to 5 nm / min or less. Next, after the drift has converged to the speed, the feedback is canceled and the AFM processing probe fixed at a predetermined height position (a surface position of the glass substrate 2 or a position dug several nm from the glass substrate 2). By scanning 4, only the recognized black defect region 3 is selectively removed as mechanical processing with a probe harder than the material to be processed. As described above, the microfabrication method of the present invention is extremely height controllable by processing after reaching a predetermined convergence state while confirming the convergence of the deviation due to the drift in the height direction mainly due to the thermal drift. This enables excellent black defect correction.

  When an arbitrary number of machining operations are performed on the same position, as shown in the flowchart of FIG. 4, the drift amount predicted according to the drift velocity calculated in advance from the drift velocity in the height direction every time machining is started. Is corrected by correcting the target height position with the correction value in the height direction, releasing the feedback, and then setting and fixing it to the height, and performing mechanical machining with a probe that is harder than the workpiece material. The processed part is removed. Repeated height correction and machining with the estimated drift amount until the set number of machining operations are completed. As described above, the microfabrication method of the present invention is extremely excellent in height controllability by sequentially correcting the deviation of the drift in the height direction even when performing the machining at the same position a plurality of times. Enables black defect correction.

  When the area of the processing region is large, the processing region is divided into blocks of an appropriate size as shown in the flowchart of FIG. 5, and the divided blocks are processed one by one. Also in this case, the target height position is corrected using the drift amount predicted according to the drift speed at the start of machining for each block as a correction value in the height direction, and after feedback is canceled, the height is set. -Remove the recognized workpiece by fixing and performing mechanical machining with a probe that is harder than the workpiece material. Repeated height correction and machining by the predicted drift amount until machining of all blocks is completed.

  Alternatively, when the area of the processing region is large, the processing region is divided into lines of an appropriate size as shown in the flowchart of FIG. 6, and the divided lines are processed one by one. Also in this case, the target height position is corrected using the drift amount predicted according to the drift speed at the start of machining for each line as a correction value in the height direction, and the feedback is released and then set to the height -Remove the recognized workpiece by fixing and performing mechanical machining with a probe that is harder than the workpiece material. Repeated height correction and machining with the estimated drift amount until machining of all lines is completed.

  Thus, even when the machining area is large and machining takes time, the drift amount predicted according to the drift speed described above for each divided block or each line is used as a correction value in the height direction. By correcting the target height position, processing with extremely excellent height controllability is possible.

  In addition, the correction based on the prediction of the drift amount described above can be applied not only in the height direction but also in the XY directions if there is an appropriate drift marker or a substitute for it. . In this case, correction by predicting the drift amount is also performed in the height direction and the XY direction at that time, so that the three-dimensional positioning accuracy is improved, so that the black defect can be corrected with higher accuracy. Furthermore, the processing after drift convergence due to the lapse of the operating time of the apparatus is also performed after convergence of drift in the height direction and in the XY direction at that time, thereby enabling more advanced black defect correction.

It is a flowchart which shows the procedure in the case of processing by estimating the drift amount of a height direction. It is a schematic sectional drawing explaining the case where it processes by estimating the drift amount of a height direction. It is a flowchart which shows the procedure in the case of processing, after confirming that the drift amount of a height direction converges. It is a flowchart which shows the procedure in the case of performing a process of multiple times, estimating a drift amount. It is a flowchart which shows the procedure in the case of performing by estimating the drift amount when an area is large. It is a flowchart which shows the procedure in the case of performing by estimating the drift amount when an area is large.

Explanation of symbols

1 Photomask (shading film)
2 Photomask (glass)
3 Black defect 4 AFM machining probe

Claims (4)

A step of performing a first height measurement at an arbitrary reference position by a probe;
A step of performing a second height measurement at the same position as the first time after an arbitrary time after the first measurement;
Calculating a drift velocity in the height direction from the difference between the measured values of the height of the reference position measured at the first time and the second time and the time difference;
A step of correcting the target height position using the drift amount predicted according to the drift speed during processing as a correction value in the height direction;
Releasing the feedback related to position information in the Z direction of the probe after fixing the probe to the corrected height, and fixing the height of the probe ;
A processing method using an atomic force microscope characterized by comprising: In the processing method using the atomic force microscope according to claim 1,
When performing an arbitrary number of machining operations on the same position, the target height position is corrected using the drift amount predicted according to the drift speed at the start of machining for each machining time as a correction value in the height direction, A processing method using an atomic force microscope. In the processing method using the atomic force microscope according to claim 1 or 2,
When the area of the machining area is large, the machining area is divided into an arbitrary number of blocks or lines, and the drift amount predicted according to the drift speed at the start of machining for each block or each line is calculated in the height direction. A processing method using an atomic force microscope characterized by correcting a target height position as a correction value. In the processing method using the atomic force microscope according to claim 1 or 3,
A processing method using an atomic force microscope, characterized in that processing is started after the calculated drift velocity becomes 5 nm / min or less.

Images