analysis of a scheme for de-magnified talbot lithography · 2013-02-21 · (received 22 june 2011;...
TRANSCRIPT
Analysis of a scheme for de-magnified Talbot lithographyL. Urbanski, M. C. Marconi, A. Isoyan, A. Stein, C. S. Menoni et al. Citation: J. Vac. Sci. Technol. B 29, 06F504 (2011); doi: 10.1116/1.3653507 View online: http://dx.doi.org/10.1116/1.3653507 View Table of Contents: http://avspublications.org/resource/1/JVTBD9/v29/i6 Published by the AVS: Science & Technology of Materials, Interfaces, and Processing Related ArticlesThermally reflowed ZEP 520A for gate length reduction and profile rounding in T-gate fabrication J. Vac. Sci. Technol. B 30, 051603 (2012) Formation of large-area GaN nanostructures with controlled geometry and morphology using top-downfabrication scheme J. Vac. Sci. Technol. B 30, 052202 (2012) Resist–substrate interface tailoring for generating high-density arrays of Ge and Bi2Se3 nanowires by electronbeam lithography J. Vac. Sci. Technol. B 30, 041602 (2012) Real-time measurements of plasma photoresist modifications: The role of plasma vacuum ultraviolet radiationand ions J. Vac. Sci. Technol. B 30, 031807 (2012) Electron beam lithography on vertical side faces of micrometer-order Si block J. Vac. Sci. Technol. B 30, 041601 (2012) Additional information on J. Vac. Sci. Technol. BJournal Homepage: http://avspublications.org/jvstb Journal Information: http://avspublications.org/jvstb/about/about_the_journal Top downloads: http://avspublications.org/jvstb/top_20_most_downloaded Information for Authors: http://avspublications.org/jvstb/authors/information_for_contributors
Downloaded 23 Aug 2012 to 129.82.233.89. Redistribution subject to AVS license or copyright; see http://avspublications.org/jvstb/about/rights_and_permissions
Analysis of a scheme for de-magnified Talbot lithography
L. Urbanski and M. C. MarconiNSF ERC for Extreme Ultraviolet Science & Technology and Electrical and Computer Engineering,Colorado State University Fort Collins, Colorado 80523
A. IsoyanSynopsys, Inc., Hillsboro, Oregon 97124
A. SteinCenter for Functional Nanomaterials, Brookhaven National Laboratory, New York 11973
C. S. Menoni and J. J. RoccaNSF ERC for Extreme Ultraviolet Science & Technology and Electrical and Computer Engineering,Colorado State University Fort Collins, Colorado 80523
(Received 22 June 2011; accepted 23 September 2011; published 10 November 2011)
The authors describe a photolithographic scheme based on the replication of a periodic transparent
mask in a photoresist utilizing the coherent self-imaging Talbot effect. A periodic two-dimensional
diffractive structure (or Talbot mask) composed of unit tiles distributed in a square matrix was
illuminated by a coherent extreme ultraviolet (EUV) beam from a table top EUV laser. The
illumination beam was reflected in a spherical mirror and the Talbot mask was placed in the path
of the convergent beam. At designed locations determined by the Talbot distance, reduced replicas
of the mask were obtained and used to print the slightly de-magnified copies of the mask on the
surface of a photoresist. Experimental results showing the de-magnification effect are in good
agreement with the diffraction theory. The limits of the technique are discussed. VC 2011 AmericanVacuum Society. [DOI: 10.1116/1.3653507]
I. INTRODUCTION
In recent years, novel applications of nano-structures that
are now impacting many fields have been demonstrated. Of
particular interest are the periodic structures with nanometer
dimensions replicated over large areas, that allowed the
implementation of unique devices such as stimulated surface
Raman scattering detectors, plasmonic structures capable to
focus light beyond the diffraction limit or “superlenses,”
novel approaches to nano-lithography based on plasmon in-
terference, etc.1–4 The rapid advancement of nano-
technology that has allowed the practical demonstration of
these new devices and techniques, along with the develop-
ment of new applications made possible by nanostructures
will be favored by easy access to a versatile method used to
produce such nano-structures in a cost effective manner.
We have previously reported the feasibility of generating
periodic nano-structures of arbitrary design arranged in a
two-dimensional structure on the surface of a photoresist by
a photolithographic method based on self-imaging using a
table top EUV laser.5 A similar approach was recently used
by Zanke et al. to print large areas of periodic circular holes
in a hexagonal lattice in what was reported as coherent dif-
fraction lithography.6 The approach allows the fabrication
of periodic nanostructures patterned in the photoresist by
the exact replication of a diffractive transparent mask pro-
duced by the coherent self-imaging effect also known as
the Talbot effect.7 In our former experiment we extended
the original concept of the Talbot effect to arbitrary diffrac-
tive units (or tiles) that we referred to as generalized Talbot
imaging.5
In this work we demonstrate that it is possible to replicate
and simultaneously scale the size of the printed nanostructure
by adequately illuminating the diffractive mask with a conver-
gent coherent beam. The de-magnification is a consequence
of the extra phase introduced in the illuminating beam by the
reflection on a spherical mirror. It depends on the optics used
and the geometry of the setup. A theoretical description of the
self-imaging effect and the de-magnification factor is pro-
vided based on the Kirchhoff–Fresnel formalism. The limita-
tions for this method are further discussed.
The experimental setup is schematically shown in Fig. 1. It
is composed of a concave multilayered mirror with the peak
reflectivity centered at 46.9 nm, for an incidence angle of 18�
and a focal length of f¼ 25 cm. The Talbot mask is placed at a
distance, s, from the mirror in the path of the converging beam.
The distance between the focal plane, F, and the Talbot planes
where the samples to be printed are placed is denoted as z.The Talbot effect that produces the self-imaging of periodic
structures can be explained using the Kirchhoff–Fresnel diffrac-
tion theory. To simplify the formalism, let us consider the sim-
ple case of a one-dimensional structure. Let T(x) be the one-
dimensional transmission function of a periodic object with a
period, p, such that T(x)¼T(xþ p). The transmission function
can be expressed in terms of its Fourier series as
TðxÞ ¼X
m
Cm exp i2pmx
p
� �; (1)
where Cm is the amplitude of the mth Fourier component.
After the reflection in the spherical mirror, the wave-front
acquires an extra phase given by
06F504-1 J. Vac. Sci. Technol. B 29(6), Nov/Dec 2011 1071-1023/2011/29(6)/06F504/4/$30.00 VC 2011 American Vacuum Society 06F504-1
Downloaded 23 Aug 2012 to 129.82.233.89. Redistribution subject to AVS license or copyright; see http://avspublications.org/jvstb/about/rights_and_permissions
M �exp ik
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiðx� x0Þ2 þ f 2
q� �� �
f; (2)
where x and x0 are the coordinates in the mirror’s plane and
in the mask’s plane, respectively, and f is the mirror’s focal
distance.
Assuming a plane wave illumination with unitary
amplitude, the transmitted optical field, UT, after the mask
will be,
UT ¼ �1
ik
ðþ1�1
TðxÞM exp �ikRf gR
dx; (3)
where R ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiðx� x0Þ2 þ z2
q:
The transmitted optical field can be calculated by solving
Eq. (3). Replacing the expressions from Eqs. (1) and (2), the
transmitted field will be proportional to,
UT /X
m
Cm exp i2pðf � sÞ
pzðx� x0Þ
� �
exp i2pðf � sÞkðf � s� zÞ
2p2z
� �: (4)
The original transmission function, T(x), is reproduced when
the last term in (4) is equal to 1, with a scale factor given by
d¼ z/(f – s). The locations of the Talbot planes will be given
by the condition,
ðf � sÞðf � s� zÞz
¼ 2mp2
k; (5)
where m is an integer number. A Talbot image with a de-
magnification, d, will be obtained at distances, d, given by
d ¼ f � s� z ¼ 2mp2ðf � sÞ2mp2 þ kðf � sÞ : (6)
At these distances, the Talbot image will have a period
p0given by
d ¼ p0
p¼ zj j
f � s: (7)
For this experiment we used the Talbot mask which was
described in more detail in Ref. 5. The Talbot mask was
written by e-beam lithography on an �60 nm thick layer of
hydrogen silsesquioxane (HSQ) resist deposited on an ultra-
thin (30 nm) SiN membrane. The thin membrane allowed for
�13.5% transmission at 46.9 nm, the wavelength of the table
top EUV laser used in this test, while the 60 nm thick layer
of HSQ was sufficiently opaque to provide a good contrast
for the diffractive mask. The units of the mask were standard
test patterns with a period of 4.845 lm. An electron micro-
scope micrograph of the mask is shown in Fig. 2.
The illumination source was a highly coherent EUV table
top laser developed at Colorado State University. The energy
per pulse provided by the laser was approximately 0.1 mJ
and 1.2 ns full width at half maximum (FWHM) duration. It
was operated at the repetition rate of 1 Hz. The laser wave-
length, k¼ 46.9 nm, corresponds to the 3s1P1-3p1S0 transi-
tion in Arþ8. The laser emission is produced by exciting Ar
with a fast discharge through a ceramic capillary. The cur-
rent pulse was approximately 21 kA, a 10–90% rise time of
approximately 55 ns, and a first half-cycle duration of
approximately 135 ns.8,9 The relative band width of the
EUV laser is approximately Dk/k¼ 3.5� 10�5, yielding a
coherence length of approximately 840 lm. The spatial co-
herence length of the beam improves with longer capillaries
reaching a fully spatially coherent illumination for the 36 cm
capillary. In this experiment it was calculated to be 740 lm
at the distance where the exposure took place.
The EUV laser impinges a spherical multilayer mirror
that focuses the beam and simultaneously spectrally filters
the EUV plasma background. The mirror has a reflectivity of
approximately 40%, optimized at an incidence angle of 18�.After the mirror, the Talbot mask and the sample coated
with photoresist were placed in a translation stage which
allows for moving the set mask-sample along the path of the
converging beam. The distance between the Talbot mask
and the sample was adjusted for the first Talbot plane, in this
case 960 lm. The sample was a 0.5 mm thick silicon wafer
covered with a 100 nm thick layer of poly-methyl-methacry-
late photoresist deposited by spin coating. The recording of
the photoresist required an exposure of approximately 300
shots. The laser was set to operate at 1 Hz, yielding an
FIG. 1. (Color online) Scheme of the experimental set up. The set mask-
sample is placed at different distances, s, from the focusing mirror. Solidly
moving these two elements along the direction of the convergent beam
(positions 1 and 2 in the figure) allows for different de-magnification
factors.
FIG. 2. Scanning electron micrograph of the Talbot mask. The mask was
defined in 50 nm thick HSQ deposited on a 30 nm thick SiN membrane.
06F504-2 Urbanski et al.: Analysis of a scheme for de-magnified Talbot lithography 06F504-2
J. Vac. Sci. Technol. B, Vol. 29, No. 6, Nov/Dec 2011
Downloaded 23 Aug 2012 to 129.82.233.89. Redistribution subject to AVS license or copyright; see http://avspublications.org/jvstb/about/rights_and_permissions
exposure time of 5 min in order to record one image. This
time can be reduced if the laser is operated at a higher repeti-
tion rate, at a maximum of 4 Hz. After exposure, the samples
were developed in a mixture of methyl-isobuthyl-ketone and
isopropanol for 1 min. Subsequently, they were rinsed with
pure isopropanol and dried using ultra high purity nitrogen.
The developed samples were scanned using an atomic
force microscope. Figure 3 shows the images of the prints
obtained at different positions along the converging beam.
The set mask-sample was solidly moved along the path of
the converging beam to distances of d1¼ 22 cm, d2¼ 24.4
cm, and d3¼ 24.5 cm from the mirror. These distances yield
calculated de-magnification factors of d1¼ 0.96, d2¼ 0.85,
and d3¼ 0.82, respectively.
The scan size of all images is 20� 20 lm2. From the
images, the period of the structures was measured to be
p1¼ 4.69 lm, p2¼ 4.35 lm, and p3¼ 4.24 lm, yielding to
measured de-magnification factors, dM1¼ 0.97, dM2¼ 0.89,
and dM3¼ 0.87, respectively.
The discrepancy between the measured and calculated
de-magnification values may stem from two facts. One of
them being the measurement error of the s distance, that in
our set up was estimated to be (1 mm), while the other is due
to the variation in the effective focal length of the mirror due
to the natural divergence of the illumination beam (�5
mrad). Considering these two factors, the calculated and
measured values are within the experimental errors.
The dependence of the de-magnification with the distance
from the mirror is governed by expression (7) and changes
very rapidly when approaching the focal plane. This depend-
ence is shown in Fig. 4, where the de-magnification factor
p0/p is plotted as a function of the normalized distance from
the mirror, s/f. The solid curve is the analytical expression of
the de-magnification as a function of the mirror–mask dis-
tance, s, normalized to the focal distance, f. The circles are
the actual data points. The graph shows that the de-
magnification changes very little throughout almost all of
the range, s, and only close to the focus (s/f¼ 1), does the de-
pendence becomes more pronounced.
One limiting factor for this method is the spherical aber-
ration introduced by the mirror that produces slightly differ-
ent focal distance values for the tangential and sagittal
directions. This is clear in the prints obtained, which showed
a slightly different period in the two directions indicating an
astigmatic image. This problem can be solved by replacing
the spherical mirror by an off-axis parabolic mirror or, alter-
natively, designing the Talbot mask with the appropriate cor-
rection for the spherical aberration.
The main limitation on setting a large value of the demag-
nification factor arises from its strong dependence with the
distance when the Talbot plane is located close to the focal
plane. We may assume that to have a clear print the mask
should be positioned within the depth of focus (DOF) which
is typically a few microns. Even for this small distance, close
to the focal plane the de-magnification factor variation is so
strong that the print is degraded. We can see this effect if we
consider that the error in determining the demagnification
factor is given by
Dd ¼ zT
Z2ðDOFÞ;
where the depth of focus is DOF ¼ k½ð1þ 2p2=kWÞ2�, W is
the size of the mask, and Z is the distance from the mask to
FIG. 3. (Color online) Atomic force micrographs of the prints obtained at
different distances from the focusing mirror. The period decreases when the
set mask-sample moves towards the focus. The de-magnification factors and
distances measured from the focusing mirror are: (a) dM1¼ 0.97, d1¼ 22
cm, (b) d M2¼ 0.89, d2¼ 24.4 cm, and (c) d M3¼ 0.87, d3¼ 24.5 cm.
06F504-3 Urbanski et al.: Analysis of a scheme for de-magnified Talbot lithography 06F504-3
JVST B - Microelectronics and Nanometer Structures
Downloaded 23 Aug 2012 to 129.82.233.89. Redistribution subject to AVS license or copyright; see http://avspublications.org/jvstb/about/rights_and_permissions
the focal plane. The error in positioning the mask diverges
when reaching the focal plane. It is thus not possible to set
large demagnification factors and this method is only re-
stricted to modest de-magnification values.
II. SUMMARY
In summary, we have analyzed the effect that a conver-
gent beam illumination has on the size of the printed fea-
tures utilizing Talbot lithography. The de-magnification
effect becomes apparent when the sample is located close
to the focal plane of the focusing mirror. In the vicinity of
the focal plane the de-magnification factor changes very
fast with the distance, and the error in determining its value
with respect to the distance from the mask to the focal
plane diverges, limiting the range that can practically be
achieved. This method utilizes a periodic mask and focus-
ing optics, which makes the setup relatively simple to
implement at EUV wavelengths. It is noncontact, high fi-
delity, and capable of printing over relatively large areas
with high resolution.
ACKNOWLEDGMENTS
The authors are indebted to the early contributions of F.
Cerrina to this project. This work was supported by the
National Science Foundation (Award No. ECCS 0901806),
the NSF ERC for Extreme Ultraviolet Science and Technol-
ogy (Award No. EEC 0310717). This research was carried
out in part at the Center for Functional Nanomaterials, Broo-
khaven National Laboratory, which is supported by the U.S.
Department of Energy, Office of Basic Energy Sciences,
under Contract No. DE-AC02-98CH10886.
1K. Kneipp, Y. Wang, H. Kneipp, L. T. Perelman, I. Itzkan, R. Dasari, and
M. S. Feld, Phys. Rev. Lett. 78, 1667 (1997).2S. Vedantam, H. Lee, J. Tang, J. Conway, M. Staffaroni, E. Yablonovitch,
Nanoletters 9, 3447 (2009).3Z. W. Liu, Q. H. Wei, and X. Zhang, Nano Lett. 5, 957 (2005).4X. G. Luo and T. Ishihara, Appl. Phys. Lett. 84, 4780 (2004).5A. Isoyan, F. Jiang, Y. C. Cheng, P. Wachulak, L. Urbanski, J. J. Rocca, C.
S. Menoni, M. C. Marconi, and F. Cerrina, J. Vac. Sci. Technol. B 27, 2931
(2009).6C. Zanke, M. H. Qi, and H. I. Smith, J. Vac. Sci. Technol. B 22, 3352
(2004).7H. F. Talbot, Philos. Mag, 9, 401 (1836).8C. D. Macchietto, B. R. Benware, and J. J. Rocca, Opt. Lett. 24, 1115
(1999).9B. R. Benware, C. D. Macchietto, C. H. Moreno, and J. J. Rocca, Phys.
Rev. Lett. 81, 5804 (1998).
FIG. 4. De-magnification factor, d, as a function of the normalized distance,
s/f. The solid curve corresponds to the predicted values according to the
Kirchhoff–Fresnel theory and the red circles correspond to the measured
values.
06F504-4 Urbanski et al.: Analysis of a scheme for de-magnified Talbot lithography 06F504-4
J. Vac. Sci. Technol. B, Vol. 29, No. 6, Nov/Dec 2011
Downloaded 23 Aug 2012 to 129.82.233.89. Redistribution subject to AVS license or copyright; see http://avspublications.org/jvstb/about/rights_and_permissions