U.S. patent application number 12/848871 was filed with the patent office on 2010-11-25 for method for processing an object with miniaturized structures.
This patent application is currently assigned to CARL ZEISS SMS GMBH. Invention is credited to Tristan Bret, Michael Budach, Klaus Edinger, Thorsten Hofmann.
Application Number | 20100297362 12/848871 |
Document ID | / |
Family ID | 40933181 |
Filed Date | 2010-11-25 |
United States Patent
Application |
20100297362 |
Kind Code |
A1 |
Budach; Michael ; et
al. |
November 25, 2010 |
METHOD FOR PROCESSING AN OBJECT WITH MINIATURIZED STRUCTURES
Abstract
A method for processing an object with miniaturized structures
is provided. The method includes feeding a reaction gas onto a
surface of the object. The method also includes processing the
object by directing an energetic beam onto a processing site in a
region, which is to be processed, on the surface of the object, in
order to deposit material on the object or to remove material from
the object. The method further includes detecting interaction
products of the beam with the object, and deciding whether the
processing of the object is to be continued or can be terminated
with the aid of information which is obtained from the detected
interaction products of the beam with the object. The region to be
processed is subdivided into a number of surface segments, and the
interaction products detected upon the beam striking regions of the
same surface segment are integrated to form a total signal in order
to determine whether processing of the object must be continued or
can be terminated.
Inventors: |
Budach; Michael; (Hanau,
DE) ; Bret; Tristan; (Darmstadt, DE) ;
Edinger; Klaus; (Lorsch, DE) ; Hofmann; Thorsten;
(Rodgau, DE) |
Correspondence
Address: |
FISH & RICHARDSON PC
P.O. BOX 1022
MINNEAPOLIS
MN
55440-1022
US
|
Assignee: |
CARL ZEISS SMS GMBH
Jena
DE
|
Family ID: |
40933181 |
Appl. No.: |
12/848871 |
Filed: |
August 2, 2010 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
PCT/EP2009/001296 |
Feb 24, 2009 |
|
|
|
12848871 |
|
|
|
|
Current U.S.
Class: |
427/585 |
Current CPC
Class: |
C23C 16/047 20130101;
H01J 37/3056 20130101; H01J 2237/31744 20130101; H01J 2237/30466
20130101; G03F 1/74 20130101; H01J 2237/30483 20130101; H01J
2237/31732 20130101; H01J 2237/24495 20130101; H01J 37/304
20130101 |
Class at
Publication: |
427/585 |
International
Class: |
C23C 8/06 20060101
C23C008/06 |
Foreign Application Data
Date |
Code |
Application Number |
Feb 28, 2008 |
DE |
102008011530.4 |
Feb 28, 2008 |
DE |
102008011531.2 |
Claims
1. A method, comprising: a) processing an object by directing an
energetic beam onto a region of a surface of the object to deposit
material on the object or to remove material from the object, a
reaction gas being present at the region of the surface of the
object, the region of the surface of the object comprising surface
segments; b) for a given surface segment, detecting interaction
products of the energetic beam with the object and integrating the
detected interaction products to provide a total signal for the
given surface segment; and c) based on information obtained in b),
deciding whether to continue processing the object or to terminate
processing of the object.
2. The method as claimed in claim 1, wherein the areas of the
surface segments differ from one another by at most 300%.
3. The method as claimed in claim 1, wherein, during b) signals are
detected only when the beam incident on the object is more than a
predetermined minimum distance from an edge of the region of the
surface of the object.
4. The method as claimed in claim 3, wherein the minimum spacing is
selected so that the signal caused by the interaction products
virtually exclusively exhibits material contrast.
5. The method as claimed in claim 1, wherein each of the surface
segments has a minimum size such that, on the basis of the spatial
integration over the surface segments and of a parallel temporal
integration over a number of irradiation cycles, statistical noise
of the detected signal of the interaction products is smaller than
the change in the detected signal of the interaction products which
is to be expected on the basis of the change in material occurring
during processing.
6. The method as claimed in claim 1, wherein a quotient of a square
of a circumference around a surface segment and an area of the
surface segment is less than 30 for at least 90% of the surface
segments.
7. The method as claimed in claim 1, wherein: a) is performed with
a first set of beam parameters of the beam; scanning of the surface
of the object in b) is performed with a second set of beam
parameters for the beam; and the second set of beam parameters is
different from the first set of beam parameters.
8. The method as claimed in claim 7, wherein a processing rate for
the second set of beam parameters is less than a processing rate
for the first set of beam parameters.
9. The method as claimed in claim 7, a pixel dwelltime for the
second set of beam parameters is greater than a pixel dwelltime for
the first set of beam parameters.
10. The method as claimed in claim 7, a sequence in which the beam
strikes different locations of the surface of the object for the
first set of beam parameters is different from a sequence in which
the beam strikes different locations of the surface of the object
for the second set of beam parameters.
11. The method as claimed in claim 1, wherein, during scanning of
the surface of the object in b), feeding the process gas to the
object is reduced when compared with feeding of process gas during
a).
12. The method as claimed in claim 1, wherein, after step c),
processing of the object by directing the beam onto the processing
site on the surface of the object in the presence of the reaction
gas is continued only in regions on the surface of the object in
which adequate processing has not yet been established in c).
13. A method, comprising: a) processing an object by directing an
energetic beam onto a processing site of a surface of the object to
deposit material on the object or to remove material from the
object, a reaction gas being present at the region of the surface
of the object; b) scanning the surface of the object with the
energetic beam and detecting interaction products of the energetic
beam with the object; and c) based on detected interaction product
of the energetic beam with the object, deciding whether to continue
processing the object or to terminate processing of the object,
wherein a first set of beam parameters is used for the energetic
beam during a), a second set of beam parameters is used for the
energetic beam during b), and the second set of beam parameters is
different from the first set of the energetic beam parameters.
14. The method as claimed in claim 13, wherein the first and the
second sets of beam parameters differ from one another so that a
processing rate during a) is greater than a processing rate during
b).
15. The method as claimed in claim 13, wherein a pixel dwelltime
for the second set of beam parameters is greater than a pixel
dwelltime for the first set of beam parameters.
16. The method as claimed in claim 13, wherein a sequence in which
the beam strikes different locations of the surface of the object
for the first set of beam parameters is different from a sequence
in which the beam strikes different locations of the surface of the
object for the second set of beam parameters.
17. The method as claimed in claim 13, wherein, during b), feeding
of the process gas to the object is reduced or stopped.
18. The method as claimed in claim 13, wherein, after step c),
processing of the object by directing the beam onto the processing
site on the surface of the object in the presence of the reaction
gas is continued only in regions on the surface of the object in
which adequate processing has not yet been established in c).
19. The method as claimed in claim 13, wherein a region of the
object region to be processed is subdivided into a number of
surface segments, and the interaction products detected upon the
beam striking regions of the same surface segment are integrated to
form a total signal in order to determine whether processing of the
object is to be continued or can be terminated.
20. The method as claimed in claim 19, wherein the areas of the
surface segments differ from one another by at most 300%.
21. The method as claimed in claim 19, wherein, during b) signals
are detected only when the beam incident on the object is more than
a predetermined minimum distance from an edge of the region of the
surface of the object.
22. The method as claimed in claim 21, wherein the minimum spacing
is selected such that the signal caused by the interaction products
virtually exclusively exhibits material contrast.
23. The method as claimed in claim 19, wherein each of the surface
segments has a minimum size such that, on the basis of the spatial
integration over the surface segments and of a parallel temporal
integration over a number of irradiation cycles, the statistical
noise of the detected signal of the interaction products is smaller
than the change in the detected signal of the interaction products
which is to be expected on the basis of the change in material
occurring during processing.
24. The method as claimed in claim 19, wherein a quotient of a
square of a circumference around a surface segment and an area of
the surface segment is less than 30 for at least 90% of the surface
segments.
25. The method as claimed in claim 1, wherein the energetic beam
comprises a charged particle beam.
26. The method as claimed in claim 1, wherein the energetic beam
comprises an electron beam.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is a continuation of, and claims benefit
under 35 USC 120 to, international application PCT/EP2009/001296,
filed Feb. 24, 2009, which claims benefit of German Application
Nos. 10 2008 011 530.4 and 10 2008 011 531.2, filed Feb. 28, 2008.
International application PCT/EP2009/001296 is hereby incorporated
by reference in its entirety.
FIELD
[0002] The present disclosure relates to a method for processing an
object with miniaturized structures, in particular for repairing
masks that are used in semiconductor fabrication, or for processing
semiconductor circuits.
BACKGROUND
[0003] Electron microscopy is a long established method for
inspecting a surface of an object to be tested. In scanning
electron microscopy, the surface of the object to be tested is
scanned via a fine electron beam. The electrons emitted by the
object as a consequence of the striking of the electron beam and/or
electrons backscattered at the object are detected in order to
compose an electron image of the scanned region.
[0004] Electron microscopes usually have the following components:
an electron beam source for producing an electron beam, an electron
optics for focusing the electron beam onto the object to be tested,
a deflecting optics for scanning the surface of the object with the
electron beam, and at least one detector for detecting electrons
backscattered at the object surface or emerging therefrom.
[0005] In addition to the pure inspection, electron microscopes are
also increasingly being used to process miniaturized structures on
an object and/or to produce an object with miniaturized structures.
In this case, material is deposited or removed selectively and with
high precision by feeding to a site to be processed on the object a
reaction gas that is excited by the electron beam striking the site
to be processed on the object and becomes chemically reactive. In
this way it is possible to deposit material at a selected location
or to remove material from the object at selected locations. In
this procedure, the reaction gas is suitably selected as a function
of the material, which is to be removed, of a miniaturized
structure on a surface of the object, or as a function of the
material to be deposited.
[0006] A particular field of application of this technique is to be
found in the field of mask repair for lithography. Masks continue
to play a prominent roll in the production of miniaturized
structures in the semiconductor industry sector. In the course of
lithography, the (photo) mask is transradiated, and there is
produced on a wafer a reduced image of the mask that exposes a
photoresist applied to the wafer, and thus defines on the wafer
structures to be produced in subsequent processing steps. Masking
defects can consequently have a pronounced disadvantageous effect
on the quality of the miniaturized structures produced with their
aid. Since the mask production continues to be time consuming and
costly, mask repair methods are being applied more and more. In
this case, defects in the mask can be repaired very specifically
and with high precision via the chemical reaction described and
induced by the electron beam.
[0007] Mask repair method, and also in the case of other methods
for producing miniaturized structures, involves detecting an end
point of the deposition or removal of material at which sufficient
material has been deposited or removed. Diverse parameters can be
used for the purpose of detecting this, for example signals of
secondary or backscattered electrons, X-rays, gas components and a
current produced in the object.
[0008] When photomasks are being repaired, the defect to be
repaired is typically identified and its shape is determined. This
shape is scanned with the electron beam, and the desired chemical
reaction is supported by the addition of suitable gases. The result
of this chemical reaction is either that superfluous material is
removed (is etched away), or that missing material is deposited,
depending on which type of defect there is, and on which variant is
therefore involved in repairing the defect. One task in the case of
these operations involves detecting the correct end point of the
chemical reaction, which is determined by virtue of the fact that
sufficient material has been deposited or sufficient material has
been removed, since the substrate is attacked if etching lasts too
long and/or the material layer at the repaired site becomes too
thick if too much material is deposited, and this would then become
noticeable later in the lithography process as a defect of the
mask.
[0009] In order to determine the correct end point of the chemical
reaction, it is customary to detect interaction products, such as
secondary electrons or backscattered electrons, emitted by the
object during the process, and to evaluate the detected signals.
The detection of backscattered electrons is fundamentally
particularly suitable in the case of etching processes and of
deposition processes, since the backscatter efficiency depends
strongly on the atomic mass number of the scattering object, and
the detected signal is therefore strongly dependent on the
material. If the superfluous material, usually chromium or MoSi in
the case of a photomask, has been completely removed, the electron
beam is subsequently scattered at the substrate instead of on the
chromium or MoSi, and this then leads to a change in signal.
[0010] Unfortunately, however, during processing most detector
signals are so weak that they are already very noisy for
statistical reasons, and so noise suppression is involved.
[0011] Spatial frequency filtering for noise suppression is known
from WO 1997001153 A. However, spatial frequency filtering
generally cannot be applied in repairing defects, since defects
usually do not have a specific spatial structure.
[0012] It is proposed in general in WO 2006050613-A to evaluate a
"Region of Interest", which is generally a subset of the region to
be repaired, in order to detect the processing end point.
[0013] US 20070278180 A1 discloses a multistep method for electron
beam induced etching. Furthermore, a multistep method for a CVD
method is disclosed in U.S. Pat. No. 7,220,685 B2.
[0014] U.S. Pat. No. 6,608,305 discloses a deposition method that
uses different scanning speeds.
SUMMARY
[0015] It is the object of the present disclosure to specify a
method for processing an object via a beam induced chemical
reaction in which the end point detection for the processing step
is improved.
[0016] In accordance with a first aspect of the present disclosure,
this object is achieved by virtue of the fact that the object
region to be processed is divided into a number of surface segments
of similar areas. The signals of the reaction products, which are
caused because of the interaction of the incident beam with the
object within regions of the same surface segment, are integrated
to form a total signal.
[0017] This first aspect of the present disclosure can be applied
both when the interaction products are detected during the
processing step, and when the detection of the interaction products
is performed in a measuring step temporarily separate from the
processing step.
[0018] In accordance with a second aspect of the disclosure, the
detection of the interaction products for the end point
determination is performed in a separate step in which one or more
beam parameters of the incident beam have been varied in comparison
with the beam parameters in the processing step.
[0019] It is, of course, also possible to apply both aspects of the
disclosure concurrently.
[0020] A method according to the first aspect of the disclosure can
comprise the following steps: [0021] feeding a reaction gas onto a
surface of the object; [0022] processing the object by directing an
energetic beam onto a processing site in a region, which is to be
processed, on the surface of the object, in order to deposit
material on the object or to remove material from the object,
[0023] detecting interaction products of the energetic beam with
the object, and [0024] deciding whether the processing of the
object is to be continued or can be terminated with the aid of
information which is obtained from the detected interaction
products of the energetic beam with the object.
[0025] The energetic beam can be a light beam, for example a laser
beam with ultrashort light pulses with pulse durations of 10 ps or
less, or a beam of charged particles, in particular an electron
beam.
[0026] When deciding whether the processing of the object is to be
continued or can be terminated, the region to be processed can be
subdivided into a number of surface segments, and the interaction
products detected upon the beam striking regions of the same
surface segment can be integrated to form a total signal.
[0027] The areas of the surface segments can all have a similar
area such that the noise component in all surface segments is to
some extent the same. In a specific embodiment the areas of the
surface segments can differ from one another by at most 300%.
[0028] In a specific embodiment it is possible that only signals
are taken into account during formation of the total signal which
are detected when the beam incident on the object has a distance
from the edge of the region to be processed which is larger than a
pre-defined minimum spacing. In a more specific embodiment the
minimum spacing is selected such that the signal caused by the
interaction products virtually exclusively exhibits material
contrast.
[0029] In another specific embodiment each of the surface segments
can have a minimum size such that, on the basis of the spatial
integration over the surface segments and of a parallel temporal
integration over a number of irradiation cycles, the statistical
noise of the detected signal of the interaction products is smaller
than the change in the detected signal of the interaction products
which is to be expected on the basis of the change in material
occurring during processing.
[0030] In another specific embodiment the surface segments can be
determined such that the quotient of the square of the
circumference around the surface segment and the area of the
surface segment is smaller than 30 for at least 90% of the surface
segments.
[0031] In another specific embodiment the processing of the object
in step b) can be performed with a first set of beam parameters of
the beam, and the scanning of the surface in step c) can be
performed with a second set of beam parameters for the beam, and
the second set of beam parameters deviates from the first set of
beam parameters. In an even more specific embodiment the first set
of beam parameters and the second set of beam parameters can differ
from one another such that the processing rate is smaller for the
second set of beam parameters than for the first set of beam
parameters.
[0032] In another more specific embodiment the first and the second
sets of beam parameters can differ from one another at least with
regard to the (pixel) dwelltime of the beam at a location on the
surface of the object, the dwelltime being greater in the second
set of beam parameters than in the first set of beam
parameters.
[0033] In another more specific embodiment the first and second
sets of beam parameters can differ from one another at least with
regard to the sequence in which the beam strikes different
locations on the surface of the object.
[0034] In another specific embodiment during scanning of the
surface of the object in step c) the feed of the process gas to the
object can be reduced when compared with the feed of process gas
during processing of the object.
[0035] In another specific embodiment after the decision in process
step d) the processing of the object by directing the beam onto the
processing site on the surface of the object while reaction gas is
being fed is continued in and only in those regions on the surface
of the object in which no adequate processing has yet been
established in the decision step d).
[0036] Depending on the type of mask to be processed and the type
of defect to be repaired, the size of the surface segments can be
empirically designed such that regions of the defect that have
different material thicknesses belong to different surface
segments. It is thereby possible, despite the surface integration,
to take due account of typical thickness variation of the material
to be removed, which in turn involves different etching
durations.
[0037] The signals that should be taken into account during
formation of the total signal are those for which the beam incident
on the object is at least at a minimum spacing away from the edge
of the region to be processed. With other words, only signals
should be taken into account during formation of the total signal
which are detected when the beam incident on the object has a
distance from the edge of the region to be processed which is
larger than a pre-defined minimum spacing. It is possible thereby
to ensure that the detected signal is not dominated by signal
artifacts that occur, in particular, at the edge of a defect to be
repaired. Such a signal artifact can be present, for example,
whenever the detected signals are strongly influenced by the object
topography. In particular, the minimum spacing can be selected such
that the signal caused by the interaction products virtually
exclusively exhibits material contrast.
[0038] The individual surface segments can be shaped such that the
quotient of the square of the circumference of the surface segment
and the area of the surface segment is smaller than 20 for at least
50% of all surface segments, and smaller than 30 for at least 90%
of all surface segments. Ideally, the surface segments should have
an approximately circular shape or square shape so that the spatial
resolution of the evaluated signal is to some extent similar in the
two mutually perpendicular spatial directions. However, this is not
possible as a rule, since with circular surface segments (shapes)
surface filling without overlapping is impossible, and the edge of
the defect to be repaired or of the region to be processed does not
generally exhibit an ideally round or straight shape. When the
above quotient is smaller than 20 for at least 50% of all surface
segments and is smaller than 30 for at least 90% of all surface
segments, the deviation from the circular shape or square shape is
still acceptably slight, and so there are still no resulting
remarkable or disturbing direction-dependent differences in the
resolution.
[0039] All surface segments taken together should precisely cover
the area of the entire region to be processed. There should not be
any points remaining in the region to be processed that are not
assigned to a surface segment. On the other hand, the surface
segments should also not be mutually overlapping (such that a point
of the region to be processed is assigned to more than one surface
segment), since otherwise another surface segment would also be
varied upon masking of a surface segment.
[0040] A method in accordance with the second aspect of the
disclosure has the following method steps: [0041] feeding a
reaction gas onto a surface of the object; [0042] processing the
object by directing an energetic beam onto a processing site in a
region to be processed on the surface of the object, in order to
deposit material on the object or to remove material from the
object, [0043] scanning the surface of the object with the
energetic beam and detecting interaction products of the energetic
beam with the object, and [0044] deciding whether the processing of
the object is to be continued or can be terminated with the aid of
information which is obtained from the detected interaction
products of the energetic beam with the object.
[0045] The energetic beam can be a light beam, for example a laser
beam with ultrashort light pulses with pulse durations of 10 ps or
less, or a beam of charged particles, in particular an electron
beam.
[0046] The processing of the object in the second step can be
performed with a first set of beam parameters for the beam, and the
scanning of the surface in the third step can be performed with a
second set of beam parameters for the beam, and the second set of
beam parameters can deviate from the first set of beam
parameters.
[0047] The first set of beam parameters can in this case be
optimized to the processing procedure, while the second set of beam
parameters can be optimized for the detection of the interaction
products. In particular, the first set of beam parameters and the
second set of beam parameters can differ from one another such that
the processing rate is smaller for the second set of beam
parameters than for the first set of beam parameters.
[0048] The first and the second sets of beam parameters can differ
from one another with regard to the (pixel) dwelltime of the beam
at a location on the surface of the object, the dwelltime being
greater in the second set of beam parameters than in the first set
of beam parameters. Owing to the lengthened dwelltime of the beam
at a location, after a short time a depletion of the process gas
occurs at this location, the result being that the chemical process
is stopped, or at least slows down, despite the incident beam of
charged particles.
[0049] The first and second sets of beam parameters can differ from
one another with regard to the sequence in which the beam strikes
different locations on the surface of the sample. Particularly in
the case of a meandering scanning strategy both in the processing
step and in the detection of the end point signal, the spacings of
the meanders can be selected to be smaller in the detection of the
end point signal than in the processing step. A reduction or
slowing down of the chemical process is likewise achieved
thereby.
[0050] During scanning of the surface of the object in the third
step, the feeding of the process gas to the object can be reduced,
or even stopped, when compared with the feeding of process gas
during processing of the object. A reduction or slowing down of the
chemical process is likewise achieved thereby.
[0051] The above-named measures for reducing or slowing down the
chemical process can be applied individually or in combination with
one another.
[0052] After the decision in the fourth processing step, the
processing of the object by directing a beam onto the processing
site on the surface of the object while reaction gas is being fed
can be continued in and only in those regions on the surface of the
object in which no adequate processing has yet been established in
the decision step. It is thereby possible to take account of
spatially different processing thicknesses or processing speeds
such that the processing is continued at each location precisely as
long as desired on the basis of the local properties.
BRIEF DESCRIPTION OF THE DRAWINGS
[0053] Details of the disclosure are explained in more detail below
with the aid of the figures, in which:
[0054] FIG. 1 shows a schematic sketch of a processing apparatus
for processing an object,
[0055] FIGS. 2a to 2d show plan views of a structure with a
defect,
[0056] FIG. 3 shows a flowchart of a method in which the region to
be repaired is broken down into surface segments, and
[0057] FIG. 4 shows a flowchart for segmenting the region to be
repaired.
[0058] Identical reference symbols denote identical components.
DETAILED DESCRIPTION
[0059] The processing system 100 comprises an electron microscope
1, a gas feed arrangement 8 for feeding reaction gas to a site, to
be processed, of an object O held on an object holder 81, and an
electrode arrangement 9.
[0060] In a direction of the electron beam propagation, the
electron microscope 1 comprises an electron beam source 3, first
focusing/deflecting elements 48, a backscattered electron detector
6, an energy selector 7, a secondary electron detector 5 and a
focusing lens 4. Second focusing/deflecting elements 47 are
arranged inside the focusing lens. The focusing lens 4 is a
combination composed of a magnetic lens and an electrostatic
immersion lens. The magnetic lens comprises an inner pole piece 42,
an outer pole piece 41 and a coil 43 arranged there between, a
lower end of the inner pole piece 42 and a lower end of the outer
pole piece 41 forming a substantially axial gap 44 in which, during
induction of a magnetic flux through the pole pieces 41, 42 by a
current flow in the coil 43, a magnetic field is generated that
emerges substantially in the region of the axial gap 44. This
magnetic field leads to a focusing of the electron beam, which is
accelerated toward the object O from the electron beam source 3.
The electrostatic immersion lens comprises a radiation tube 45 that
extends through an inner space of the magnetic lens 4 formed by the
inner pole piece 42 and the outer pole piece 41. The electrostatic
immersion lens furthermore comprises a terminal electrode 46
arranged at a spacing from a lower end of the radiation tube 45. By
applying a suitable electric field between the radiation tube 45
and the terminal electrode 46 with the aid of a voltage source
(indicated schematically, without reference symbol), it is possible
to decelerate the primary electrons to a primary energy of
approximately 1 keV suitable for inspecting photomasks. In the
embodiment illustrated, the radiation tube can, for example, be at
+8 keV while the terminal electrode 46 is grounded.
[0061] The electron microscope 1 is subdivided into four different
vacuum spaces 21, 22, 23, 24 that are partially separated from one
another by pressure stages 25, 26, 27. A first vacuum space 21
contains the electron beam source 3. The first vacuum space 21 is
connected by a first connection 29 to an ion getter pump 37. During
operation of the electron microscope, a pressure in the range from
approximately 10.sup.-9 to 10.sup.-10 mbar, for example, prevails
in the first vacuum space 21. A first pressure stage 25 is formed
by an opening 25 symmetrically surrounding the electron beam path.
A second vacuum space 22 is connected via a second connection 30 to
a second vacuum pump 38, an ion getter pump. A second pressure
stage separates the second vacuum space 22 partially from a third
vacuum space 23. During operation of the electron microscope, the
pressure in the second vacuum space 22 can be in the region of
approximately 10.sup.-7 mbar, for example. The backscattered
electron detector 6 and the energy selector 7 are arranged in the
third vacuum space 23. The third vacuum space 23 is partially
separated from the second and from a fourth vacuum space 22, 24 by
pressure stages 26 and 27, respectively, and has a connection 31
that connects the third vacuum space to a third vacuum pump 29.
During operation, the pressure in the third vacuum space can be in
the region of approximately 10.sup.-5 mbar. The fourth vacuum space
24 is partially separated from the third vacuum space 23 by the
third pressure stage 27. In the exemplary embodiment illustrated,
the third pressure stage 27 comprises the secondary electron
detector 5. In this case, an opening of the third pressure stage 27
is formed by the opening, penetrated by the electron beam, of the
secondary electron detector 5. The secondary electron detector 5 is
held in this case in the interior of the electron microscope 1 such
that a pressure compensation between the partially separated vacuum
spaces 23, 24 can be performed only through the opening in the
secondary electron detector. The fourth vacuum space 24 further has
a gas conducting connection 28 to the interior of the vacuum
chamber 2. The gas conducting connection 28 is provided here by a
simple metal tube. The reactive gas fed by the gas feed is
discharged from the fourth vacuum space 24 to the vacuum chamber 2
through the metal tube which has a relatively large diameter in
order to afford as little resistance as possible to the transport
of gas into the interior of the vacuum chamber 2. In the exemplary
embodiment illustrated, the radiation tube 45 has a lower
cylinder-shaped part in the beam direction a region that expands
conically in the direction of the secondary electron detector 5 and
then extends upward, in the shape of a cylinder of comparatively
large diameter, as far as through the second vacuum space 22. The
radiation tube 45 thus surrounds both the secondary electron
detector 5 and energy selector 7, as well as the backscattered
electron detector 6. The radiation tube 45 is held at a spacing
below the secondary electron detector 5 by a vacuum-tight holder,
for example made from ceramic, and is connected to the lower pole
piece 41 in a vacuum-tight fashion in such a way that the fourth
vacuum space 24 substantially comprises an inner space of the
radiation tube and an intermediate space between the insulation and
the third vacuum space 23 adjoining in the direction of the
electron beam source 3. In the region, that is to say in the
vicinity, of the third pressure stage 27 a pressure in the region
of approximately a few 10.sup.-4 mbar, for example, prevails during
operation in the interior of the fourth vacuum space 24, whereas a
vacuum is achieved in the region of approximately a few 10.sup.-5
mbar, for example, in the interior of the vacuum chamber 2. The
vacuum chamber 2 has a connection 32 that connects the interior of
the vacuum chamber 2 to a fourth vacuum pump 40. The first, the
second, the third and the combination of fourth vacuum space and
vacuum chamber can thus each be evacuated individually, thus
enabling an efficient operation of the electron microscope even
when gas is fed into the vacuum chamber.
[0062] A detection surface 51 of the secondary electron detector is
therefore arranged in the fourth vacuum space 24, while the
backscattered electron detector 6 is arranged in the third vacuum
space 23, in which a better vacuum is achieved. The energy selector
7 is arranged in front of the backscattered electron detector 6 in
such a way that all electrons emitted by the object O or
backscattered thereon passes the energy selector 7 in order to be
able to reach a detection surface of the backscattered electron
detector 6. In the embodiment illustrated, the energy selector 7
comprises a first grid 71, a second grid 72 and a voltage source 73
for producing a suitable electric field between the first and the
second grids, in order to enable the reflection of secondary
electrons emerging from the object surface. The grids are arranged
parallel to one another and enclose in an annular fashion the
electron beam path of the primary electron beam produced by the
electron beam source 3. In the exemplary embodiment illustrated,
the first grid 71 is connected to the voltage source 73, while the
second grid 72 is coupled to the radiation tube 45 and thus is at
the same potential as the latter. It is possible to introduce an
insulating tube into the opening, formed by the grids 71, 72 and
penetrated by the electron beam, in order to protect the primary
electron beam against the influence of the electric field applied
between the two grids 71, 72. The electric field applied via the
voltage source 73 is adapted to the primary electron energy and the
particularities of the inspected and processed sample in such a way
that the backscattered electrons pass through the electric field
and are detected at the backscattered electron detector, while the
secondary electrons are reflected due to their lower kinetic energy
and are therefore not detected. The strength of the electric field,
and thus the level of the detection signal, can be improved by
setting the potential difference applied to the grids.
[0063] The embodiment illustrated further comprises an electrode
arrangement 9 that comprises a shielding electrode 91 that is
arranged annularly around the electron beam path and has a central
opening 92 that enables an undisturbed passage of the primary
electron beam and a largely unimpeded passage of secondary and
backscattered electrons. A suitable voltage source (illustrated
schematically without reference symbol) can be used to apply a
suitable voltage to the electrode 91 in order to shield the primary
electron beam effectively from an electric field produced by
charging the object O.
[0064] A fast deflection element 49 is arranged between the
terminal electrode 46 and the electrode arrangement 9. The
deflection element, which is formed as an electrostatic multipole
element, can be used to deflect the electron beam in the plane of
the object in order to move the electron beam with the desired
scanning strategy over the sample region to be processed or to be
analyzed.
[0065] In a method for automatic end point selection, an object is
inspected in a first step, the object being in the case of a mask
repair a photomask in which, for example, miniaturized molybdenum
structures are applied to a quartz substrate. During the
inspection, defects in the mask are identified and processing steps
are selected in order to eliminate or repair the defect. A site to
be processed on the object O is then brought in a processing step
into the region of the primary electron beam, and the gas feeding
arrangement 8 is used to feed a reaction gas that is excited by the
electrons of the electron beam and thus becomes chemically
reactive. Material can thereby be removed, for example. After a
certain time interval of the material removal, the processed site
is inspected once again. The inspection is performed in this case
by detecting backscattered electrons from which secondary electrons
are separated via the energy selector 7. Secondary electrons
emerging from the surface of the object O enter the interior of the
electron microscope 1 and strike the detection surface 51 of the
secondary electron detector 5 in the fourth vacuum space. Those
secondary electrons that penetrate into the third vacuum chamber
through the opening of the secondary electron detector are
reflected by applying a suitable voltage between the first grid 71
and the second grid 72 of the energy selector 7. Only the more
highly energetic backscattered electrons pass the energy selector 7
and reach the backscattered electron detector 6. The electron image
on the basis of which a decision is taken regarding the reaching of
an end point is produced on the basis of the detected backscattered
electrons. The processing of the object can be stopped if the
electron image produced corresponds to a desired image. Otherwise,
another processing step is performed, accompanied by the feeding of
reaction gas. In a particularly advantageous way, this mode of
procedure enables an automatic end point selection, particularly in
the repair of photomasks.
[0066] FIG. 2a is a schematic of an SEM image of a mask structure
for producing two parallel conductor tracks. To this end, the mask
structure has two strips 101, 102 that absorb light of the wafer
scanners later used and between which a transparent strip 103 is
located, and which are embedded laterally in a transparent
environment. In an SEM image produced by detecting backscattered
electrons, the absorbing strips 101, 102 appear to be brighter
because they consist of chromium or MoSi, and these materials have
a larger backscattering coefficient than the environment 104, 103,
105 in which the electrons are scattered on the quartz
substrate.
[0067] FIG. 2b illustrates the same region of the mask when the
mask structure has a so-called opaque defect. In a region 106, the
two opaque strips 101, 102 are undesirably connected to one
another. Where this mask structure to be used in a wafer scanner,
an undesired electrical short circuit would result in the region
106. Consequently, in a repair process the excessive material in
this region 106 is to be removed.
[0068] FIG. 2c illustrates the region to be repaired as extracted
from the mask image by image analysis. In a subsequent step, the
region 107 to be repaired, which is illustrated in an enlarged
fashion in FIG. 2d, is broken down into a multiplicity of surface
segments 108, 109 that all have an area of approximately equal
size, and which, when assembled, produce the region to be repaired.
Breaking down the region to be repaired into surface segments 108,
109 promotes the end point detection, that is to say it serves to
determine the instant at which the processing of the mask is
stopped in the individual surface segments. To this end, the
intensity of the backscattered electrons is detected either when
repairing the mask, or in a separate measurement step with the aid
of the backscattered electron detector. In order to reduce the
signal noise that cannot be avoided because of the Poisson
statistics of the detected backscattered electrons, the signals
that belong to a single surface segment 108, 109 are respectively
integrated. A single measured value results for each surface
segment as a consequence thereof. If this measured value
undershoots or overshoots a prescribed desired value, the
processing is then stopped at this surface segment. Consequently,
the scanning strategy with which the electron beam is guided over
the object is changed for the further processing such that
subsequently the processing is continued only in the surface
segments in which the signal, integrated over the surface segment,
of the backscattered electrons has not yet reached the prescribed
limiting value.
[0069] The following points of view are taken into account when
defining the individual surface segments. All surface segments have
a minimum size so that the averaging over each of the surface
segments has the desired signal-to-noise ratio. On the other hand,
each of the surface segments also has a maximum size so that
typical variations in the defect, such as different defect heights,
inhomogeneities in the etching or deposition rate, are determined
with sufficient spatial resolution during the end point detection.
In order to optimize the signal-to-noise ratio, the areas of all
the surface segments should be approximately of the same size. In
this process, it is respectively only such pixels or object points
that are at a specific minimum spacing from the edge of the region
107 to be repaired which are counted to the size of the surface
segments. The reason for this is that the backscattered electron
signal is frequently corrupted at the edge of the region to be
repaired by effects that are determined by the surface topography
of the object and/or of the mask. Furthermore, all surface segments
should exhibit an approximately round or square shape such that the
spatial resolution in the mutually perpendicular spatial directions
remains to some extent homogeneous despite the spatial filtering
that results from averaging over all points of the same surface
segment. To this end, the quotient of the square of the
circumference around the surface segment and the area of the
surface segment should be smaller than 30, preferably smaller than
23, for each surface segment.
[0070] Since the concrete parameters depend both on the respective
mask type and on the respective defect type, the parameters for the
individual surface segments are determined empirically in
corresponding test runs, and then input into the control software.
To this end, defective masks can be etched or repaired in test
passes. Pixel for pixel, the signal characteristic of the end point
signal is then determined as a function of the processing cycles
(loops). It is then determined from the signal characteristic to
what distance from the edge of the defect the end point signal is
still influenced by the edge of the defect, and which minimum
spacing from the edge of the defect is therefore to be observed for
a good end point signal. Upon later use of the apparatus, the
appropriate values are then selected by the system after selection
of the mask type and the defect type, and the division into the
surface segments is then undertaken via a software program.
[0071] The process itself is reproduced schematically in FIG. 3. In
a first step 121, a scanning electron microscope image of the
defect--or, more accurately, the mask with the defect to be
repaired--is shown. By taking account of the defect type and the
mask type, which were either previously determined or input by the
user in a step 120, the region to be repaired is then firstly
determined by image analysis in a step 122, and the suitable
breakdown of the region to be repaired into surface segments of
approximately the same size is subsequently undertaken in a step
123. The region to be repaired then simultaneously also defines in
spatial terms the scanning strategy with which the electron beam is
later guided over the object. Thereafter, the repair of the defect
by a radiation induced chemical process then begins in a step 124.
To this end, to the object is fed a suitable gas or gas mixture via
a gas inlet system, and the electron beam is guided over the
individual points of the region to be repaired. The dwelltime for
which the electron beam remains at each location, and at a time at
which the electron beam again irradiates on the next occasion a
region already irradiated earlier (the so-called refresh time) are
determined by the gas chemical process, and are likewise part of
the scanning strategy.
[0072] The end point signal is detected in a step 125, either
during the repair process or in a separate measurement step in
which, by suitable measures such as a change in the dwelltime, in
the refresh time and/or a reduction in the gas flow, the chemical
processing procedure is slowed down when compared with the
situation in the processing step 124, or is stopped. In the
particular case described, the electrons backscattered on the
object are detected for the end point signal. The end point signal
is then integrated over each surface segment in a step 126, that is
to say the step 126 supplies precisely one measured value for each
surface segment. In a subsequent step 127, the measured value
obtained in step 126 for each surface segment is compared with a
previously defined limiting value. If the limiting value is not
reached for any of the surface segments, the system returns to the
step 124 and continues the repair of the defect in all surface
segments. However, if the limiting value is reached in one or more
surface segments, the scanning strategy is changed in a subsequent
step 128. This change in the scanning strategy amounts to the
spatial coordinates of all object points in those surface segments
where the limiting value is reached being removed from the scanning
strategy, that is to say the points or pixels belonging to these
surface segments are later no longer approached by the electron
beam.
[0073] In a subsequent step 129, it is further checked whether at
least one surface segment that involves further processing is still
present after changing the scanning strategy. If this is the case,
the system returns to the processing step 124 with a changed
scanning strategy. Otherwise, the end 130 of the process is reached
when the limiting value has been reached in all surface
segments.
[0074] If the end point signal is obtained in a separate
measurement step in the method previously described with the aid of
FIG. 3, it should be ensured that no, or only a reduced, radiation
induced chemistry takes place in the measurement step. This can be
achieved in general by virtue of the fact that a depletion of the
process gas occurs in the measurement step. One possibility to this
end is, of course, to reduce the gas flow during the measurement
step, or to stop it entirely. However, since the speed at which the
gas flow can be changed is relatively low, it is more sensible to
change the scanning strategy in the measurement step when compared
with the scanning strategy in the processing step. To this end, the
time for which the electron beam dwells at the same location, that
is to say the so-called pixel dwelltime, can be lengthened in the
measurement step, and/or the duration which lies at least between
the approach to the same location on the object, that is to say the
refresh time, can be shortened. The effect of both measures is that
the locally available process gas is depleted at the respective
location after a relatively short time, and consequently the gas
chemical process is slowed down and more or less comes to a
standstill. It is possible in this way to exclude overetching or
overdeposition during the measurement step.
[0075] FIG. 4 illustrates an algorithm for subdividing the region
to be processed into surface segments. It starts after, in step 122
in FIG. 3, the region to be processed is determined in terms of
position and shape. It also presupposes that a minimum number of
pixels that each surface segment has is prescribed. In a following
step 133, all edge points are then firstly masked, that is to say
all points and/or pixels that do not exhibit an empirically
determined minimum spacing from the edge of the region to be
repaired. In a further step 134, an unmasked pixel is then selected
and combined with pixels in the environment to form a surface
segment. In this process, pixels are added until the surface
segment formed has the desired minimum size, that is to say the
minimum number of pixels. In this process, it can either be
attempted, starting from a circle around the pixel, to further
expand the circle, or a start is made from a substantially square
surface that is then expanded.
[0076] If a segment has reached the desired size, the pixels
belonging to this segment are masked in a following step 135, and a
check is made in a step 136 as to whether there are still present
further unmasked pixels that have not yet been assigned to a
surface segment. If this is the case, the system returns to step
134 and selects a new unmasked pixel in order to form a new surface
segment.
[0077] In general, it can come about that individual unmasked
pixels that are not yet assigned to a surface segment are still
left over at the end of the algorithm. These are then added to a
already existing neighboring surface segment.
[0078] Experiments have shown that on the basis of a target number
of 400 pixels per surface segment and of a square basic surface as
desired shape, the above described algorithm leads to a division of
the region to be repaired in the case of which it holds for 50% of
the surface segments that the quotient of the square of the
circumference around the surface segment and of the area of the
surface segment is smaller than 18, and that it holds for 90% of
the surface segments that the quotient of the square of the
circumference around the surface segment and of the area of the
surface segment is smaller than 23. The size distribution of the
surface segments is virtually constant in this case, that is to say
apart from a few exceptions virtually all surface segments have the
prescribed size, and there are only a few surface segments at the
edge of the region to be processed that are larger.
[0079] If the same algorithm is carried out in the form in which
circular surfaces form the basic shape, the circular surfaces then
being expanded by adding pixels, it is seen that it then holds for
50% of the surface segments that the quotient of the square of the
circumference around the surface segment and of the area of the
surface segment is smaller than 15, and that it holds for 90% of
the surface segments that the quotient of the square of the
circumference around the surface segment and of the area of the
surface segment is smaller than 18. However, a relatively large
number of surface segments result which are larger than the pixel
number. This algorithm therefore yields advantages to the effect
that the spatial measurement resolution is relatively homogenous in
mutually perpendicular directions, but at the expense of a larger
number of larger surface segments, and thus at the expense of a
lower spatial resolution overall, or a larger inhomogeneity of the
spatial resolution of the end point detection. The other algorithm,
by contrast, yields relatively uniformly large surface segments,
although with the disadvantage that deviations from the ideal
circular or square shapes are more pronounced so that the spatial
measurement resolution is somewhat inhomogeneous in mutually
perpendicular directions. Further experiments have shown that both
algorithms lead to a complete and an unambiguous division of the
region to be processed into surface segments of which the largest
surface segment, for the requirement that each comprise at least
400 pixel, has less than 800 pixels, the largest surface segment
resulting thus having an area at most twice as large as the
smallest surface segment.
[0080] The effect that is achieved by evaluating the surface
segments during the end point detection can be shown in the
simplest way with the aid of the following numerical examples:
dwelltimes typically used in the electron beam induced chemical
processes lie in the range between 30 and 200 ms, and typical
currents lie in the range from 10 to 100 pA. Typical refresh times,
that is to say times which at least pass until the same pixel is
again irradiated with the electron beam, lie between 50 .mu.s and
10 ns. Given a current of 50 pA and an irradiation duration of 100
ns, it is possible to calculate that approximately 30 primary
electrons are incident on each pixel per pixel and per cycle
(loop). If the signal is integrated over 100 cycles, the result is
3000 primary electrons incident in the 100 cycles per pixel.
[0081] Typical backscattering coefficients, that is to say the
number of the electrons backscattered per incident primary
electron, depend on the material and are approximately 0.29 for
chromium, approximately 0.21 for MoSi, and approximately 0.15 for
quartz. It follows therefrom that for chromium on quartz, the
backscattered electron signal drops by approximately 50% when the
chromium is completely removed at a site. In the case of MoSi on
quartz, the difference is substantially less: with this material
combination a difference of only approximately 20% occurs in the
backscattered electron signal.
[0082] However, not every backscattered electron is detected, since
the electrons are, at all, scattered into the entire half space,
and only a portion thereof are incident on the detector. Typical
detector efficiencies for backscattered electrons are approximately
5*10.sup.-4.
[0083] Multiplying the above named efficiencies by the named number
of electrons per 100 cycles results in 0.3 detected electrons to be
expected per 100 cycles. As a result of the Poisson distribution, a
statistical spread of the results of approximately +-0.55 is to be
expected, that is to say the signal noise is approximately double
the size of the signal to be expected when the electrons are
scattered on chromium and/or MoSi. As a consequence thereof, the
expected change in the signal of 20% and 50%, respectively,
referred to the total signal cannot be detected as a result of the
statistical noise.
[0084] However, there is a considerable change in the situation if
integration is performed over surface segments that comprise 100 to
1000 pixel: for integration over 400 pixels, instead of the 0.3
electrons to be expected the result is approximately 120 electrons
per surface segment and 100 cycles. The statistical spread to be
expected on the basis of Poisson statistics is then 11 events, that
is to say the noise is only approximately 10% of the signal value.
In these circumstances, the difference in signal to be expected of
50% for chromium on quartz or 20% for MoSi on quartz can be
detected, since this signal difference is now at least as large as
the noise, or the noise is smaller than the difference in signal to
be expected. An alternative to the spatial integration by combining
pixels to form surface segments would be an appropriate lengthening
of the temporal integration. However, there would then be a need to
integrate over 100*400=40 000 cycles instead over 100 cycles in
order to reach the same signal-to-noise ratio, and this is
permissible only if no gas chemistry is proceeding during the
signal acquisition for the end point detection, since such a large
number of cycles in conjunction with activated gas chemistry can
certainly correspond to etching depths or deposition thicknesses of
several 100 nm.
[0085] As may be seen, the inventive method provides the freedom of
setting the signal-to-noise ratio of the detected interaction
products via two parameters, specifically via the number of the
cycles over which the signal is acquired, and via the size of the
surface segments over which the signal is spatially integrated. It
is sensible to set these parameters such that the statistical noise
of the detected interaction products is lower than the difference
in signal, to be expected upon processing, for the detected
interaction products on the basis of the material that is changing
during processing and on which the interaction products are
produced. The number of the cycles over which the signal is
temporarily integrated is selected, once again, such that the
attainment of the processing product is checked sufficiently
frequently without too great an enlargement of the entire
processing period. The selection of the appropriate parameters is,
of course, dependent on the respective individual case. It has
proved to be expedient to perform temporal integration over 50 to
1000 cycles. The result, given customary processing rates in the
range from 1000 to 100 000 cycles per 100 nm processing thickness,
is that checking of the end point is carried out statistically
after at most every 2 nm processing thickness, preferably after
each 1 nm processing depth, without unduly lengthening the total
processing period. It is to be understood here, that the shorter
temporal integration period of 50 cycles is applied in the case of
processes with a high processing speed, that is to say for those in
which processing rates of approximately 1000 cycles per removal of
100 nm occur, and longer temporal integration periods of
approximately 1000 cycles are applied for processes of lower
processing speed in which approximately 100 000 cycles per 100 nm
processing thickness are involved.
[0086] The disclosure has been described above using the example in
which the gas chemical process is induced by an electron beam.
However, the disclosure can likewise be applied when the gas
chemical process is initiated by an ion beam or by ultrashort light
pulses. If the radiation induced gas chemical process is to be
initiated by an ion beam, there would be a need either to replace
the electron optical system illustrated in FIG. 1 by an ion optical
system, or to provide an ion optical system--that is to say a
so-called crossbeam system or dual beam system--in addition to the
electron optical system. Ion optical systems differ from electron
optical systems in that an ion source is used instead of the
electron source, and the lenses used are exclusively electrostatic
lenses. Moreover, as a result of the changed polarity of the
particles in ion optical systems, the electrostatic potentials
respectively applied generally have an inverse sign than is the
case in electron optical systems.
[0087] If the radiation induced gas chemical process is to be
induced by a high energy laser beam, the electron optical system
illustrated in FIG. 1 would have to be replaced by a setup
corresponding to a laser scanning microscope, or the optics
corresponding to a laser scanning microscope would have to be
provided in addition. Such a laser scanning microscope should then
have as a light source a high energy laser that emits temporally
ultrashort light pulses, that is to say light pulses with pulse
durations below 10 ps.
[0088] Also, the processing has been illustrated above chiefly by
the example of etching, that is to say the repair of so-called
opaque defects. The same inverse principles also apply, however,
when material is deposited for the repair. In this case, whenever
the surface segment is completely repaired, a rise in signal occurs
by comparison with the signal of the interaction products before
the repair, the rise in signal resulting, just as in the case
described above, from the different backscattering coefficients of
the various materials, and analogous conditions resulting with
regard to the signal-to-noise ratio.
[0089] Finally, the disclosure has been explained referring to the
example of mask repair. However, it can equally well be used when
radiation induced chemical processes are being used for other
purposes, for example in the case of so-called via etching and in
the production of the conductor tracks in semiconductor components
in order to modify the latter for examinations and tests.
[0090] The disclosure has been explained in the context of a
comprehensive system. However the disclosure does not only comprise
the combination of all features of the comprehensive system but
also comprises all individual features which are disclosed as well
as all sub-combinations of all disclosed features.
* * * * *