26

Phase Measurement Tools for PSM

Hal Kusunose

CONTENTS

26.1 Introduction
26.2 Direct Measurement of Phase-Shift and Transmittance at Exposure

Wavelength

26.3 Phase-Shift Measurement
26.4 Transmittance Measurement
26.5 MPM Series
26.6 Interferometer Optics
26.7 Fringe Scan
26.8 Transmittance Measurement
26.9 Dependence of Phase-Shift Measurement Value on Pattern Size
26.10 Automatic Operation Function
26.11 Issue of Long-Term Stability
26.12 Summary
References

26.1

Introduction

Shortening exposure wavelength is an effective method to delineate a finer pattern in
lithography. However, from an economical standpoint, it is desirable to postpone the
change of the exposure wavelength and to extend the life of the already existing lithog-
raphy tools. Phase-shift mask (PSM) technology is an already accomplished method for this
purpose. Many types of PSMs have been already developed. Attenuated-type embedded
phase-shift masks (EPSM) are mainly used for hole patterns, and alternating phase-shift
masks (APSM) or chrome-less type masks are used for interconnection.

26.2

Direct Measurement of Phase-Shift and Transmittance
at Exposure Wavelength

A PSM is a type of mask that controls the phase of transmitted light. Phase-shift must be
controlled correctly because a PSM that has a phase-shift deviating from the designed

Rizvi / Handbook of Photomask Manufacturing Technology DK2192_c026 Final Proof page 577 7.3.2005 6:38pm

© 2005 by Taylor & Francis Group.

value gives not only lower-than-expected resolution value, but also causes such harmful
effects as focus offset, variation in pattern sizes, etc. Fabrication of PSMs and verification
of phase-shift require a phase-shift measurement tool, and it is desirable that this tool be
capable of measuring both phase-shift and transmittance with high precision because it is
necessary to control both phase-shift and transmittance simultaneously, especially for
EPSM. In EPSM production, phase-shift and transmittance of EPSM are virtually depen-
dent on a shifter layer deposition process, and mask blanks makers definitely need a
phase-shift and transmittance measurement tool. There is always a possibility that optical
properties and film thicknesses change during pattern formation and the cleaning pro-
cess. Thus, mask shops also need phase-shift and transmittance measurement.

26.3

Phase-Shift Measurement

There are two possible methods for phase-shift measurement. One is to measure directly
phase-shift and transmittance by the light of the same wavelength as that of wafer
exposure. The other is to obtain phase-shift and transmittance by calculation after the
phase-shift measurement by the light of a wavelength different from that of the wafer
exposure. In the case of EPSM, film structure is not uniform, and there are cases where the
inner structure is very complicated or where multi-layers exist when the layer is strictly
viewed. The phase-shift can be induced, not only by shifter layer but also by slight etching
of the quartz pattern itself. In these cases, there is a possibility that an error may be
produced in the phase-shift calculated by conversion, using refractive indices of the
shifter material at measurement and exposure wavelengths. A measurement result
obtained by the light with another wavelength may be different from one obtained by
exposure light. It is widely believed that the usage of mask transmission light with the
same wavelength as that of exposure light is very important to avoid this error.

26.4

Transmittance Measurement

Transmittance has conventionally been measured using a spectrophotometer in the mask
blank production process. In this measurement method, the transmittance, which is the
result of the combination of both substrate and film, has been calculated by referencing the
transmittance of air (air reference transmittance). On the other hand, the definition of
transmittance value should be the relative transmittance of the shifter area by referencing
that of the opening area, which is quartz reference transmittance. Exposure light intensity is
controlled, so that the delineated linewidth of the resist pattern on a wafer may become the
designed value in the lithography process. For this reason, the number of cases in which
the definition of transmittance for EPSM is changed to the quartz reference is increasing.

Meanwhile, conversion of the air reference to the quartz reference can be performed

assuming that the transmittance of the substrate remains constant. However, error can
take place during the wafer exposure process, the cause of which is that the transmittance
of the patterned area may virtually drop when the etched surface on the substrate
becomes coarse during the pattern etching process. In this case, the scattered light portion
in the light transmitted through the patterned opening increases, resulting in the decrease
of the intensity of the light that transmits through the lens pupil of the exposure tool. This

© 2005 by Taylor & Francis Group.

kind of phenomenon suggests that even APSM shifter patterns generated by etching
substrates can undergo change in effective shifter transmittance. It might be necessary
to measure the transmittance of the shifter area by quartz reference using a measuring
tool that has a lens pupil with the same degree of NA as that of the wafer exposure tool in
a strict sense.

26.5

MPM Series

MPM series tools are the only kinds of tools commercially available in the world. MPM
series tools are capable of directly measuring the phase-shift and transmittance of the
PSM by the same light wavelength as that of the wafer exposure tools. MPM100 [1,3], the
tool with measuring wavelengths of 436 and 365 nm, MPM248 [2] with 248 nm, and
MPM193 with 193 nm, have been available and are now used by many mask blanks
makers and mask makers for process development and quality control. The MPM series
is also being used for incoming inspection of masks before applying masks to wafer
fabrications.

Figure 26.1

shows the external view of MPM248.

PSM is also the fundamental technique for F2 lithography, and MPM157, the most

advanced model in the MPM series, with a measurement wavelength of 157 nm, was
completed and is being released commercially. There are some tools available from other
companies that provide phase-shift by complicated calculation. However, the accuracy of
phase-shifts by these tools is not established yet. So these tools must be calibrated with
MPM, and they are not used for quality assurance applications. In the measurement
of phase-shift and transmittance in this field, MPM series tools are now de facto standard
equipment.

26.6

Interferometer Optics

All the models in the MPM series employ a common structure, which is a compact Mach–
Zehnder-type shearing interferometer that is placed behind the objective lens.

Figure 26.2

shows the optical system of MPM248 as an example. The light source for illumination is a
Hg–Xe lamp for MPM100 and MPM248. A deuterium lamp is used for MPM193. The
mask to be measured is illuminated by a light of the same wavelength as that of the wafer
exposure and the illumination light is obtained by selecting the light from the emission
spectrum by using an interference filter and a spectroscopic prism. Transmitted light
mask pattern images are enlarged by an objective lens and laterally shifted by the Mach–
Zehnder type shearing interferometer, and then images are projected and overlapped by a
pinhole and a camera. In the case of phase-shift measurement, the images from the phase-
shift pattern area and from the non-shift pattern area are made to overlap at the image
plane that is also the photo-detecting position. In such image overlapping, light intensity
at the photo-detector is determined by phase-shift that is the sum of phase-shift between
light transmitted through a shifter patterned area and light transmitted through a non-
shifter patterned area, and intrinsic phase-shift in an interferometer. For the instruments
that utilize two beams interference, fluctuation of optical length at split path in
an interferometer can cause significant deterioration of the measurement precision.

© 2005 by Taylor & Francis Group.

Disturbance of airflow and temperature fluctuation mainly affects the measurement ac-
curacy. MPM has the Mach–Zehnder-type shearing interferometer behind the objective
lens, and this configuration makes the light paths near the mask common paths. The result
is that the split path, which is especially sensitive to precision, is kept only on the inside of
the interferometer. Furthermore, the interferometer structure is made of SiC ceramic that
has a very small thermal expansion coefficient and high mechanical stiffness, which brings
about very stable measurement even under temperature fluctuation and mechanical
vibration. All the models in the MPM series guarantee phase-shift measurement repeat-
ability of 3s ¼ 0.58 for short-term measurement.

Figure 26.3

shows interferometer stability

versus ambient temperature drift after the lamp is turned on. The horizontal axis in
this figure is elapsed time, the vertical axis on the right is ambient temperature, and the

FIGURE 26.1
A view of MPM248 system.

© 2005 by Taylor & Francis Group.

Image detector

Variable
phase-shifter

Tube lens

PSM

Objective

Grating

Condenser

M

M

B.S.

B.S.

Pinhole mirror

Ellipsoidal mirror

Xe

−

Hg arc lamp

Filter

Beam homogenizer

Photo-multiplier
tube

FIGURE 26.2
Interferometer microscope optics.

Interferometer system drift (deg.)

Ambient temperature (deg.)

−

0.1

−

0.05

0

0.05

0.1

−

0.15

0.15

21

22

23

24

25

20

26

Elapsed time (h:min.)

0:00

0:20

0:40

1:00

1:20

1:40

2:00

Interferometer system drift (deg.)

Ambient temperature (deg.)

FIGURE 26.3
Interferometer system drift.

© 2005 by Taylor & Francis Group.

vertical axis on the left is the averaged phase-shift after measuring five times the phase-
shift on the area where phase-shift does not exist. This measurement result shows that
interferometer stability stays within +0.18 against the temperature fluctuation of 18C,
meaning that this tool is very stable against temperature drift.

26.7

Fringe Scan

The basic operation for measurement of phase-shift and transmittance employs a tech-
nique called fringe scan, which modulates one of the optical lengths within the interfer-
ometer. The specific fringe scan method employed for the MPM series is to measure the
interference signal intensity when an optical wedge (variable phase-shifter) that is placed
on one side of the split beam in the interferometer is moved in the perpendicular
direction against the optical axis. The light path length of one of the beams is modulated
by one fringe. Figure 26.4 shows the various signal waveforms that result when fringe
scan is performed at different places on a EPSM. In order to measure the phase-shift value
on a PSM, phase difference between the light beams, one transmitted through the phase-
shift pattern image portion and the other through the non-shift pattern image portion, is
obtained first by performing fringe scan. Next, the phase-shift (intrinsic phase) generated
by the difference of the light path length of two beams inside the interferometer is
eliminated from the phase-shift value mentioned above. Phase-shift generated inside of
the interferometer is usually obtained by performing fringe scan using the area on the
mask where the phase-shift is zero. Fringe scanning at the mask area where the phase-
shift is zero is not necessary for every measurement. The phase-shift measurement
procedure in the MPM series employs the cancellation of the phase-shift generated inside
the interferometer. It requires two phase-shift measurements. One is obtained by meas-
uring the phase-shift on the overlapped image by placing a phase-shift pattern image on
the left and a non-phase-shift pattern on the right. The other measurement is obtained by
an opposite positioning, placing a non-phase-shift pattern image on the left and a phase-
shift pattern image on the right. The resulting phase-shift difference is two times larger
than that of a phase-shift generated in the actual phase-shifter, so MPM obtains the
accurate phase-shift by dividing this number by 2.

Figure 26.5

demonstrates this method.

Phase-shift obtained by the method described here corresponds to the average value of
two measurements, and short-term repeatability becomes high, accordingly. The time

40

30

20

10

50

0

50

100

0

150

Sample: attenuating-type PSM
transmittance: 10.5%

VPS position (

µ

m)

Signal level (A.U)

Interference of

quartz/quartz
quartz/shifter
shifter/shifter

FIGURE 26.4
Fringe scan signal.

© 2005 by Taylor & Francis Group.

required for measurement of either phase-shift or transmittance is about 30 s, which meets
the measurement speed requirement for general applications.

26.8

Transmittance Measurement

Transmittance by referencing quartz is defined as the ratio of the transmitted light
intensities of two lights, one transmitted through the shifter area and the other transmit-
ted through the non-shifter area. Generally speaking, there are many cases when the
transmittance of a small area is influenced by the peripheral patterns. This phenomenon is
called the Schwarzschild–Villiger effect [4], and the origin of this phenomenon is stray
light in the optical system.

Figure 26.6

shows an example of this phenomenon. MPM has

made it possible to effectively reduce this stray light by combining low coherent light
irradiation and white light interferometer [5]. Unwanted stray lights that are scattered or
reflected on the lens surface go through a light path longer than the normal light path. As
a result, the optical path length difference between that of normal light and that of stray
light becomes larger than the coherent length, and the stray light loses its coherency
toward normal light. The stray light becomes a constant component in the interference
signal obtained by performing a fringe scan under this situation. In other words, MPM
measures the transmittance by the ratio of two signal amplitudes of interference signals
both from shifter/shifter (EPSM layer area) and quartz/quartz (pattern opening area), as
can be seen in

Figure 26.4.

This method eliminates a stray light effect by utilizing low

temporal coherency. In contrast with the method of simply comparing the light inten-
sities, MPM has made it possible to measure transmittance close to the pattern edge,
namely, transmittance of a small pattern.

Figure 26.7

shows three kinds of methods to

measure transmittance: 1) the method to simply compare the transmitted light brightness;
2) the method to utilize the amplitude of the interference signal without lateral shift; and

f

=

f

1

−

f

0

=

f

0

−

f

2

=

Wave front 1

Wave front 2

(

f

1

−

f

0

) + (

f

0

−

f

2

)

2

=

2

f

1

−

f

: conventional

: improved

1st

2nd

f

0

: resident phase-shift

in an interferometer

f

1

φ

f

2

f

FIGURE 26.5
Improved phase measurement sequence.

Rizvi / Handbook of Photomask Manufacturing Technology DK2192_c026 Final Proof page 583 7.3.2005 6:38pm

© 2005 by Taylor & Francis Group.

3) the method to utilize the amplitude of the interference signal with lateral shift. The
measurement positions are in close vicinity of the pattern edge. This figure clearly shows
that the measurement value near the pattern edge is changed when there is influence by
stray light. However, the effect of stray light is most effectively eliminated when an
interferometer is used and additionally when the image is laterally shifted. Such an
illumination condition improves the lateral resolution of transmittance measurement. In
fact, the MPM series is designed in such way that a light beam consisting of a light
component of spatially coherent light illuminates a mask in the lateral direction, and
interference occurs when an image is displaced by a specific distance in the horizontal
direction by a shearing interferometer. As an alternative method, it has been reported that
reduction of the S-V effect can be achieved without using an interferometer by making the
illumination area small enough. However, transmittance measurement error occurs be-
cause detected light decreases due to focus error, and this could be a problem in such
cases. The MPM series is free from the effect of focusing error because incident light
illuminates a large area on the mask. Based on this feature, all the models in the MPM
series guarantee transmittance measurement repeatability of 3s ¼ 0.2%.

26.9

Dependence of Phase-Shift Measurement Value on Pattern Size

The accuracy of the phase-shift measurement value for small patterns depends on the
focusing error.

Figure 26.7

shows the relationship between focus height and phase-shift

measurement value. This phenomenon corresponds, in reverse manner, to the instance in
which focus offset occurs in the wafer exposure process when a PSM with its phase-shift
deviates from 1808. The phase-shift measurement value changes relative to the focus

MoSiON
HT shifter

Quartz

L

Measurement
spot (0.35

µ

m)

Distance from the boundary

L

(

µ

m)

−

100

Transmitance (%)

0

100

200

300

400

500

20

95

15

100

Conventional method
(light intensity ratio)

Interferometer method
(interfringe signal amplitude ratio)

Without shearing 10

µ

m Shearing

FIGURE 26.6
Transmittance near boundary between an opening and attenuating phase-shifter pattern.

© 2005 by Taylor & Francis Group.

point deviation, when patterns with a size near the resolution limit are measured. Thus,
when measuring the phase-shift of the actual patterns that are used in wafer exposure, the
distance between the mask and objective lens must be strictly controlled. At the same time,
for the phase-shift measurement of such a small pattern, the measurement result is directly
affected not only by the effect of focus deviation but also by the optical aberration of the
lens, since the proportion of higher-order diffraction light is large. An aberration, which is a
phase-shift between the 0-order diffraction and the higher order of diffracted light in a
small pattern, superposes the phase-shift of the phase-shift pattern itself. It is impossible to
separate this superposition. Focus deviation also can be regarded as a kind of wave front
aberration, and it is virtually impossible to measure the phase-shift of the patterns with an
actual exposure pattern size correctly. Meanwhile, there is a region where the phase-shift
measurement value does not depend on focus height near the focus point in a relatively
large pattern region, as can be seen in Figure 26.7. This means that the problem of
measurement deviation dependent on the focus does not occur when this region is utilized.
For this reason, phase-shift measurement result for EPSM is guaranteed by using monitor
patterns of about 7.5 mm, which is large enough to avoid this measurement deviation.
A phase-shift difference between exposure pattern and monitor pattern could occur due to
a micro-loading effect during the etching process, leading to an error factor. It is necessary
to optimize the etching process by confirming the dependence of pattern size on etching
depths using AFM to know the difference of the etching rate quantitatively, so

170

180

190

160

200

5-

µ

m hole

4-

µ

m hole

3-

µ

m hole

2.5-

µ

m hole

2.0-

µ

m hole

1.5-

µ

m hole

1.0-

µ

m hole

Phase-shift (deg.)

Focus position (

µ

m)

EPSM, Transmittance=6.3%

−

5

0

5

−

10

10

FIGURE 26.7
Focus latitude of phase-shift measurement with MPM248.

© 2005 by Taylor & Francis Group.

that correction of the micro-loading effect becomes possible by the management of target
value offsets against the monitor pattern.

26.10

Automatic Operation Function

MPM is equipped with an automatic operation function utilizing the automatic recogni-
tion capability of a simple pattern edge. It provides continuous and unattended meas-
urement of phase-shift and transmittance over the entire mask area. Teaching of
measurement points is performed either by memorizing the operator’s measuring pro-
cedure and subsequently measuring based on that memorized data, or by inputting the
measurement point coordinates from a data file.

26.11

Issue of Long-Term Stability

Differing from those optical microscopes that simply provide observation of optical
image, for those optical measurement tools like MPM that output measured numerical
numbers, such as phase-shift and transmittance, the effect of contamination on optical
systems is recognized as a measurement value change that is directly readable. Thus, the
effects of chemical contamination in the atmosphere where tools are installed must be
very clearly recognized as a problem or issue of long-term measurement repeatability.
Especially in those conventional clean rooms where countermeasures to chemical con-
tamination are not applied, chemical substances such as siloxane, Di-Octyl Phthalate
(DOP), and hydrocarbons in the atmosphere are deposited by the irradiation of DUV
light, and therefore frequent cleaning of optical systems and parts becomes necessary. To
avoid this, it is important to keep the atmosphere around the optical system chemically
clean and to install tools such as MPM248 and MPM193 that use DUV light in a chem-
ically clean booth employing activated carbon filter. Unfortunately, the effect of these
anticontamination arrangements is not good enough and Lasertec is carrying out the
development of a purge mechanism using mechanical parts with minimum out-gas and
chemical filters for limited space. The area around the interferometer is especially sensi-
tive to contamination, and it is necessary to make both temperature fluctuation and
turbulence as small as possible to achieve good short-term measurement repeatability.
Our current goal is to make cleaning of the interferometer area unnecessary even when
tools are installed in conventional clean rooms.

26.12

Summary

This chapter describes principles of phase-shift and transmittance measurement with
MPM. There are many unique technologies employed in the system, such as the combin-
ation of low temporal coherent light illumination with an interferometer and periodically
and spatially coherent illumination. They are effective in improving lateral resolution and

© 2005 by Taylor & Francis Group.

accuracy of the measurements. MPM series tools have been released with all wavelengths
commonly used in photolithography, and play a very important role in meeting the
requirements of PSM applications.

References

1. H. Fujita, H. Sano, H. Kusunose, H. Takizawa, K. Miyazaki, N. Awamura, T. Ode, and

D. Awamura, Performance of i-/g-line phase-shift measurement system MPM100, Proc. SPIE,
2793, 497–512 (1996).

2. H. Kusunose, N. Awamura, H. Takizawa, K. Miyazaki, T. Ode, and D. Awamura, Direct phase-

shift measurement with transmitted deep-UV illumination, Proc. SPIE, 2793, 251–260 (1996).

3. H. Kusunose, A. Nakae, J. Miyazaki, N. Yoshioka, H. Morimoto, K. Murayama, and K. Tsuka-

moto, Phase measurement system with transmitted UV light for phase-shifting mask inspection,
Proc. SPIE, 2254, 294–301 (1994).

4. K. Schwarzschild and W.Villiger, Astroyhys. J., 23 (1906).
5. H. Takizawa, H. Kusunose, N. Awamura, T. Ode, and D. Awamura, Transmittance measurement

with interferometer system, Proc. SPIE, 2793, 489–496 (1996).

Rizvi / Handbook of Photomask Manufacturing Technology DK2192_c026 Final Proof page 587 7.3.2005 6:38pm

© 2005 by Taylor & Francis Group.