Scattered incident X-ray photons for measuring surface...

Details Scattered incident X-ray photons for measuring surface... Scattered incident X-ray photons for measuring surface...

1999-03-10

2002-04-23

Picardat, Kevin M. (Department: 2822)

Semiconductor device manufacturing: process

With measuring or testing

Optical characteristic sensed

C438S007000, C438S964000, C378S070000, C378S086000

Reexamination Certificate

active

06376267

ABSTRACT:

BACKGROUND OF THE INVENTION
1 Field of the Invention
This invention relates to integrated circuit manufacturing and more particularly to determining the roughness of an unpatterned surface of a semiconductor topography.
2. Description of the Relevant Art
Fabrication of an integrated circuit is a complex process involving numerous steps. To form a metal-oxide-semiconductor (MOS) transistor, for example, a gate dielectric is formed on a semiconductor substrate which is doped with either n-type or p-type impurities. A gate conductor is formed over the gate dielectric, and dopant impurities are introduced into the substrate to form a source and drain. Such transistors are connected to each other and to terminals of the completed integrated circuit using conductive interconnect lines.
A pervasive trend in modern integrated circuit manufacture is to produce transistors having feature sizes as small as possible. Many modern day processes employ features, such as gate conductors and interconnects, which have less than 1.0 &mgr;m critical dimension. As feature size decreases, the sizes of the resulting transistors as well as that of the interconnects between transistors also decrease. Fabrication of smaller transistors allows more transistors to be placed on a single monolithic substrate, thereby allowing relatively large circuit systems to be incorporated on a single, relatively small die area.
This trend toward reduced feature sizes imposes severe demands on the lithography processes used to define features in integrated circuit fabrication. In a lithography process, a film of a radiation-sensitive material called photoresist is typically formed upon the surface of the material to be patterned. This photoresist film is then exposed through a mask to radiation. Portions of the photoresist which are exposed to the radiation undergo a chemical change, such that subsequent use of a chemical called a developer will exclusively remove either the exposed or unexposed resist portions. In this way, the mask pattern may be transferred to the photoresist. The retained photoresist may then be used as a mask for subsequent etching or doping of the underlying material, thereby transferring the mask pattern to this material. Complications with photolithography processes can occur, however, which cause the feature sizes of the patterned material to be different than those of the mask. Such feature size differences become even more significant as feature size is reduced, and the “error” in the patterned dimension may approach the magnitude of the intended dimension itself.
One source of feature size error in photolithography processes may be scattering of exposing radiation from rough surfaces underlying a photoresist film. A situation in which this may occur is illustrated in FIG.
1
. Photoresist film
2
is formed over material
4
which has a rough upper surface. Exposing radiation
6
is directed into photoresist film
2
through a transparent portion of mask
8
. An enlarged view including the interface between photoresist film
2
and underlying material
4
is shown in FIG.
2
. In
FIG. 2
, incident exposing radiation photons
10
are represented using open arrowheads. Some of incident photons
10
not absorbed by photoresist film
2
or transmitted to underlying material
4
may be scattered at upper surface
14
of material
4
, producing scattered photons
12
, indicated using filled arrowheads. It is postulated that increased roughness of surface
14
increases the variation in scattering angles exhibited by scattered photons
12
. A large variation in scattering angles is in turn believed to increase the likelihood that scattered photons
12
may penetrate portions of photoresist layer
2
external to the portion subtended by mask feature width W
m
. The boundaries of this intended feature width are shown by dashed lines in FIG.
2
.
This scattering of exposing radiation into portions of photoresist film
2
external to the intended feature width results in an increased exposed portion of photoresist film
2
, as shown in FIG.
3
. Exposed portion
16
of photoresist film
2
extends beyond the boundaries of intended feature width W
m
, so that the width W
p
of the feature transferred to the photoresist is larger than W
m
. An error of W
p
−W
m
is therefore incurred in the transfer of the mask feature to the photoresist. This error may be particularly significant for small intended feature width W
m
, since W
p
−W
m
may become comparable to W
m
. For very small intended feature sizes of approximately 0.25 micron or less, root-mean-square (RMS) roughness values on the order of tens of angstroms or less are believed to be capable of producing significant error. Monitoring of roughness values for surfaces of layers to be patterned may therefore be extremely important.
Atomic force microscopy (AFM) is a currently-used technique for determining surface roughness. AFM involves high-resolution scanning (angstrom resolution) of a probe held extremely close to a surface. The chemical force between the probe tip and surface is measured during the scan, producing a high-resolution scan of the surface topography from which RMS roughness may be obtained. AFM can produce accurate measurements of RMS roughness values as low as about 0.5 angstroms, but this accuracy comes at a high cost in measurement time. An AFM scan of an area a few microns on a side may require tens of minutes, or even longer, to perform. To obtain a roughness value characteristic of the overall surface of a semiconductor topography, which is typically formed upon a substrate having a diameter of 8 inches or more, would require sampling multiple areas and could take many hours to obtain by AFM.
In addition to the long measurement times needed for AFM measurement of roughness, sample damage using AFM is also a possibility. Although AFM is theoretically non-destructive, the very close proximity of the measurement tip to the sample allows the possibility of “tip crashes”, in which the tip comes into forceful contact with the sample surface during a measurement. Such contact can cause scraping and/or scratching of the surface. In one mode of AFM operation, known as contact mode, the probe tip is in constant contact with the sample surface, which may cause sample damage.
It would therefore be desirable to develop a method for relatively rapid and non-destructive determination of roughness on surfaces of materials used in semiconductor manufacturing. The method should allow determination of roughness over larger areas of a surface than are conveniently scanned using AFM.
SUMMARY OF THE INVENTION
The problems outlined above are addressed by a method using a glancing-angle X-ray fluorescence (XRF) technique in combination with calibration standards to determine the roughness of a target surface. In conventional XRF techniques, a beam of primary X-rays is directed at the surface of a semiconductor wafer, and the energies (or corresponding wavelengths) of resultant secondary X-rays emitted by atoms of elements on and just under the surface of the wafer are measured. Atoms of elements in target materials emit secondary X-rays with uniquely characteristic energies. Thus the elemental compositions of materials on and just under the surface of the wafer may be determined from the measured energies of emitted secondary X-rays. In the glancing-angle XRF technique recited herein, a signal containing primary X-rays scattered from the sample surface is analyzed, in addition to a secondary X-ray beam. A plot of the emitted secondary beam strength versus scattered primary beam strength exhibits a characteristic slope. This slope corresponds to a particular value of RMS roughness, which may be determined using a direct imaging or profiling technique such as AFM. After a calibration curve or table relating slope to RMS roughness is generated, subsequent roughness measurements may be performed using only XRF and comparison to the calibration data, without a requirement for further time-consuming AFM measurements.
The glancing-angle XRF mea

Affiliated with

Also associated with

Rating

Scattered incident X-ray photons for measuring surface... does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Scattered incident X-ray photons for measuring surface..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Scattered incident X-ray photons for measuring surface... will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2887508