Radiant energy – Irradiation of objects or material – Irradiation of semiconductor devices
Reexamination Certificate
2002-09-27
2004-08-10
Wells, Nikita (Department: 2881)
Radiant energy
Irradiation of objects or material
Irradiation of semiconductor devices
C250S492200, C250S492220
Reexamination Certificate
active
06774380
ABSTRACT:
CROSS-REFERENCE TO RELATED APPLICATIONS
This application is based upon and claims the benefit of priority from the prior Japanese Patent Application No. 2001-302639, filed Sep. 28, 2001, the entire contents of which are incorporated herein by reference.
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a variably shaped beam EB (electron beam) writing system.
2. Description of the Related Art
A manufacturing method of an LSI is as follows. A pattern on the mask is transferred to the resist, developed, and etched on the wafer using an optical stepper or a scanner to form the resist pattern. Thereafter, the pattern of one layer is manufactured on the wafer through various semiconductor processes. Next, the LSI is manufactured by repeating the above-mentioned processes by using the mask having another pattern.
The method of manufacturing the mask used for such the manufacture of the LSI is performed by exposing, developing, and etching the resist on the mask with the variably shaped beam EB writing system.
There is the system by which the resist spread on the wafer is exposed directly by charged beams such as electron beams and the ion beams or optical beams by a variably shaped beam pattern writing system which are also used for the manufacturing method of such a mask.
Such the variably shaped beam EB writing system inputs the pattern data to which information in the pattern to be drawn is described and draws the pattern to be drawn by using information on this pattern data.
In general, a predetermined unit is specified explicitly or implicitly regarding to the length or the position, and a digital value based on this unit is described in the pattern data.
For instance, if a unit is 1.25 nm, 500 nm is set as “400”. This unit originates in the design of the variably shaped beam pattern writing system, and, for instance, is an inherent value to system decided by the balance of an electro-optical system and a DAC performance of an analog circuit for the electron beam exposure device.
In general, the position unit and length at the design of the LSI is set independently to the inherent length unit and the position unit of the above-mentioned pattern writing system.
For instance, it is assumed that the pattern transfer to resist is preformed by an optical stepper and the device is designed with 1.25 nm as a unit. When the reduction rate of a used optical stepper is ⅕, a length unit of the design pattern on the mask becomes 6.25 nm.
On the other hand, it is assumed that the unit of the control of the variably shaped beam pattern writing system to write pattern on the mask is 1 nm. At this time, the unit of the pattern on the mask becomes 6.25 nm, and this cannot be represented with the integral multiple of 1 nm, which is the control unit of the variably shaped beam writing system. That is, mismatch between the unit of the pattern and the unit of the device is caused and the following various problems are caused.
The case where a certain pattern becomes a unit and is arranged repeatedly in the LSI pattern will be considered as an example. The pitch of this repetition becomes the integral multiple of 6.25 nm on the mask, too.
For instance, when the pitch is 1237.5 nm (6.25 nm×198), this does not become the integral multiple of 1.00 nm which is the unit of the device. The error is caused at the position of the pattern which is drawn finally when data is made by ignoring the fractional portion of 0.5.
For instance, the shift at the position to which it extends it will be generated in 0.5 nm×100=50 nm in the 100th pattern when there are 100 repetitions.
In avoiding this, there is a method of rearranging the structure of the repetition in data. In this case, there is a problem that the amount of data increases because the data structure of pattern which originally expressed by one array changes more complicated one (such as data structure expressed by several arrays).
The problem of positioning error also appears in the other case, as follows. The original LSI pattern is sometime shrunk in order to enhance the LSI performance. Let say the unit of the design of an original LSI pattern is the integral multiple of that of the variably shaped beam pattern drawing device (for instance, 4 nm on the mask). If the LSI is shrunk by 0.8, then the length unit after reduction becomes 3.2 nm, and does not become the integral multiple of unit 1 nm of the variably shaped beam pattern drawing device, and the above-mentioned problem appears again.
That is, even when the original pattern is 4 nm unit, if the LSI whose pattern is 0.8 times thereof is made, the length unit after reduction becomes 3.2 nm, and does not become the integral multiple of unit 1 nm of the variably shaped beam EB writing system, and the above-mentioned problem appears again.
On the other hand, there is a device which performs the reduction of the above-mentioned entire pattern inside the device according to the variably shaped beam EB writing system. For instance, after the data of the pattern to be drawn is input to the variably shaped beam EB writing system, the reduced pattern data is created from the data by using the computer attached by the device.
However, the calculation time becomes long in a case of reduction rate 0.8, and exceeds the drawing time by far, and the entire system falls into the state of waiting for the end of such a calculation. This remarkably degrades the use efficiency of the entire system.
In addition, this kind of problem is occurred in the development of the new pattern writing system as described below.
For instance, conventionally after putting the device of length unit of 10 nm to practical use, the accuracy of the device has been improved by making the unit a half, such that the unit of first generation's device is 5 nm and the next generation thereof is 2.5 nm.
However, in a new device, it is necessary to draw the pattern data in conventional (Unit 2.5 nm) on the other hand in many cases. Therefore, the necessity for drawing the data with the conventional 2.5 nm unit in addition to the data with 1 nm unit in a new device is caused. Therefore, the above-mentioned problem appears.
Therefore, the mismatch of the unit of the variably shaped beam EB writing system and the unit of the pattern data becomes large, and the above-mentioned problem is caused.
To avoid this, there is a method of assuming the unit to be not 1.25 nm but 1 nm in the next generation's device after the device with 2.5 nm unit. As a result, mismatch can be avoided about the pattern data with 1 nm unit.
On the other hand, a new device should draw in conventional pattern data (2.5 nm unit) in many cases. That is, the necessity for drawing the data with conventional 2.5 nm unit in addition to the data with 1 nm unit is caused in a new device. Therefore, the above-mentioned problem appears.
Various problems as mentioned above have been caused, since the unit of the pattern data and the unit of the pattern writing system shift like this.
Moreover, in a stage continuous movement type Gaussian beam method raster scan device, there is an example of a similar function.
This is assumed that
(1) A standard beam size (fixed value) (of Gaussian beam) is assumed to be an integer multiple (n times) and a big beam is formed, and
(2) It is assume the scanning speed of the beam to be 1
at a standard (fixed) speed.
However, in a case of the EB system which uses the variably shaped beam method, since the shape and the size of the (shot) beam for each shot change, a fixed beam size does not exist unlike the case of the Gaussian beam.
Moreover, the beam position does not change continuously, but changes at random in the vector scanning method. Therefore, a constant beam scanning speed does not exist unlike the raster scanning method.
In addition, since it is necessary to control the beam position on the wide plane in the vector scanning method, in many cases, it becomes necessary to correct the deflection distortion with higher-order (including terms more than second order of x
2
Kabushiki Kaisha Toshiba
Kalivoda Christopher M.
Oblon & Spivak, McClelland, Maier & Neustadt P.C.
Wells Nikita
LandOfFree
Variably shaped beam EB writing system does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Variably shaped beam EB writing system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Variably shaped beam EB writing system will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3288855