Data processing: generic control systems or specific application – Specific application – apparatus or process – Product assembly or manufacturing
Reexamination Certificate
1999-02-24
2002-03-26
Grant, William (Department: 2121)
Data processing: generic control systems or specific application
Specific application, apparatus or process
Product assembly or manufacturing
C700S121000, C250S370070
Reexamination Certificate
active
06363296
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Technical Field
This disclosure relates to semiconductor fabrication tools and more particularly, to an improved system and method for automated defect inspection of photomasks.
2. Description of the Related Art
Semiconductor fabrication processes typically include photolithographic processing to pattern areas of a surface of a semiconductor device. The semiconductor fabrication process typically includes applying a photoresist material to the surface of the semiconductor device. The photoresist is patterned by exposing the photoresist to light, typically ultraviolet light, to crosslink the resist material. This cross linking prevents a reaction with a developer which develops away areas of the photoresist which were not crosslinked by the exposure to the UV light. Other types of photoresists are prevented from crosslinking in when exposed to ultraviolet light.
Photoresists are patterned using a photomask. The photomask functions as a shield to prevent light form passing through it in predetermined areas during photolithography. The photomask typically provides a black or highly absorbent layer of material, usually chromium or a chromium alloy, patterned in accordance with the patterning design to be projected onto the photoresist. The absorbent layer is formed on a substrate, which may include a glass or quartz material.
With decreasing feature sizes of semiconductor components, photomasks are increasingly more difficult to fabricate and inspect to ensure acceptable results. The defect inspection capability of these photomasks is limited to a certain minimum feature size. This minimum feature size is typically the groundrule of the features which photomask is used to produce, i.e, 150 nm minimum feature size for 1×mask magnification or 600 nm at 4×mask magnification).
Since photomasks include a multitude of features below a micron in size, inspections of photomasks are performed using automated inspection devices. Referring to
FIG. 1
, a measuring apparatus
10
includes a stage
14
for positioning a photomask
16
to be measured. An energy source
18
irradiates photomask
16
with a predetermined intensity of light. A photosensitive device or sensor
20
collects reflected and/or transmitted intensities and stores the data in a storage device
22
. A processor
24
is used to perform calculations for determining correct feature sizes according to transmitted and reflected intensities. Processor
24
includes a data set to which to compare intensity profiles of the photomask to be inspected. Primarily two system of inspection are used for inspecting photomasks. One is a die-to-die approach. The die-to-die approach compares features of the photomask against similar features of the same photomask to determine if any defects exist. Typically, an ultraviolet laser is used to transmit light through the photomask to be inspected and the master photomask. Light intensities are measured for transmitted and/or reflected light to compare the patterns of the two photomasks. A second approach includes a die-to-database measurement systems. A reference database computer (RDC) is included in processor
24
to provide a digital image used to compare against the photomask to be inspected. This is the most accurate technique in verifying that the original circuit has been correctly transferred to the photomask. A laser beam is propagated through or reflected from the photomask to be inspected and the intensity at a particular location is compared to the digital image.
Both of the approaches have the inspection capability for complex reticle patterns, including those with narrow geometries, dense optical proximity correction (OPC) and phase shift masks (PSM). OPC helps compensate for lost light to ensure that the precise patterns are formed on a semiconductor wafer. For example, without OPC, a rectangle can end up looking like an oval on the wafer because light tends to round on the edges. OPC corrects this by adding tiny serifs (lines) to the corner to ensure that the corners are not rounded or moving a feature edge so wafer features are sized more accurately. Phase shift masks alter the phase of light passing through the photomask, and permit improved depth of focus and resolution on the wafer. Phase-shift helps reduce distortion on line resolution of wafer surface irregularities.
In conventional automated systems, a certain number of sub-groundrule features may cause a failure in the inspection of the photomask. To image successfully sub-micron features on a silicon wafer, sub-groundrule features on the photomask become increasingly more likely.
In one example, an active area mask (photo mask) is used in the formation of active area (AA) features for trench type dynamic random access memory (DRAM) designs. To equalize shortening of length for features due to corner rounding, asymmetric biases are applied to the designed features on the photomask. Biasing includes shrinking or growing geometries either through data base manipulation (data base biasing) or through process control (process biasing). With decreasing groundrules these biases tend to become larger. As shown in
FIG. 2
, an example of biasing is shown. A design pattern for a line
40
has a groundrule of 175 nm and is desired to be 1050 in length and 350 nm away from another line
42
which is also 1050 nm in length. Data for a photomask is biased by a 200 nm bias for lines
40
′ and
42
′ to achieve the design lengths for lines
40
and
42
. However, the length of 1250 nm forms a space
44
of 150 nm, i.e., a sub-groundrule feature.
Depending on an unused cell design, these biases often result in sub-groundrule feature sizes on the respective photomasks used in the imaging process. For 150 nm groundrules, for example, the space between adjacent AA features along an axis is about 300 nm in data stored for the design dimensions. A length bias of more than 75 nm per shape results in a sub-groundrule feature on the mask (see also FIG.
2
). A similar situation may occur due to the presence of serifs in masks written from data sets corrected for optical proximity as well as due to the presence of assist features in chrome on glass (COG) or phase shift masks.
As shown in
FIG. 3
, serifs
50
and
52
are illustratively shown for a mask. Serifs
50
and
52
and lines
53
are used to produce lines
54
and
56
, respectively. Other assist features may also used, for example assist features
58
,
60
and
62
may be used to provide structures
64
and
66
, respectively as shown in FIG.
4
. Serifs and assist features often produce sub-groundrule features on the photomask.
Due to these sub-groundrule features fabricated on photomasks, the photomasks are not defect inspectable. Therefore, a need exists for a system and method which eliminates defect failures of photomasks due to benign sub-groundrule features on the photomask.
SUMMARY OF THE INVENTION
A method for inspecting photomasks, in accordance with present invention includes the steps of providing a design data set for fabricating a photomask, searching the design data set for sub-groundrule features, eliminating the sub-groundrule features from the data set to form an inspection data set and inspecting a photomask fabricated in accordance with the design data set by employing the inspection data set.
Another method for inspecting photomasks includes the steps of providing a design data set for fabricating a photomask, identifying sub-groundrule features in the design data set which meet predetermined size criteria, forming an inspection data set by eliminating the identified sub-groundrule features from the design data set, providing an inspection tool for comparing a photomask fabricated in accordance with the design data set to the data of the inspection data set, adjusting the inspection tool such that identified sub-groundrule features are overlooked and inspecting the fabricated photomask by employing the inspection data set.
A program storage device readable by machine, tangibly embodying a program of instruction
Bahta Kidest
Infineon - Technologies AG
LandOfFree
System and method for automated defect inspection of photomasks does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for automated defect inspection of photomasks, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for automated defect inspection of photomasks will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2842048