"Minimizing spatial-dispersion-induced birefringence in
crystals used for precision optics by using mixed crystals of materials with
the opposite sign of the birefringence"
John Burnett,
Zachary Levine,
and Eric Shirley
National Institute of Standards and Technology
Gaithersburg, MD 20899
Abstract
We recently measured and calculated an intrinsic birefringence in
CaF2, SrF2, and BaF2 cubic crystals in the
ultraviolet (UV). These results present serious problems for use of these
crystalline materials for precision optics in the UV, e.g., for UV optical
lithography. Since we found that the intrinsic birefringence of CaF2
has an opposite sign to that of SrF2 and BaF2, we propose a
correction solution in which a mixed crystal
Ca1-xSrxF2 or
Ca1-xBaxF2 is made which
effectively zeroes out the birefringence of the crystal at a given wavelength,
for appropriately chosen values of x. The appropriate x values
are determined approximately by the mixtures where the linearly interpolated
birefringences go to zero. Optical components such as lenses and beam splitters
could be made of these mixed crystals that would be free of intrinsic
birefringence aberrations at a chosen wavelength. These could be used as
components of precision optical systems at this wavelength, such as
photolithography projection systems.
We have recently measured and calculated for the first time an intrinsic,
spatial-dispersion-induced birefringence n=n<-110> - n<001> near
193 nm and 157 nm in CaF2 and similar intrinsic
birefringences in SrF2 and BaF2, but
with an opposite sign [1-2]. Our measured values are shown in
Table 1 below, along with our calculated values. The stated uncertainties
of the measurements are standard uncertainties due to statistical errors. The
calculated values have estimated relative uncertainties of 15 %, but
absolute standard uncertainties no smaller than 2 × 10-7.
The stated theoretical uncertainty takes into account issues of convergence
with respect to several numerical cutoffs. However, it does not take into
account the systematic uncertainties of the overall theoretical framework,
which remain to be assessed by comparison to measurement as this study
continues.
Table 1. Our measurements and calculations of
n=n<-110> - n<001>
for CaF2, SrF2, and BaF2.
| Material |
n × 107(193 nm) |
n × 107(156 nm) |
| (measured) |
(calculated) |
(measured) |
(calculated) |
| CaF2 |
-3.4 ± 0.3 |
-1.3 |
-11.8 ± 0.4 |
-18 |
| SrF2 |
+6.6 ± 0.2 |
+9.8 |
+5.7 ± 0.3 |
+7.3 |
| BaF2 |
+19 ± 2 |
+27 |
+33 ± 3 |
+52 |
In the references [1-2] we suggested one method to partially
correct the effect in optical systems by combining <111> oriented lenses
with transverse crystal axes rotated relatively by 60°. Here we propose a
method to null out the effect at a given wavelength in each optical element.
CaF2, BaF2, and SrF2 all have the same
fluorite crystal structure (space group Fm3m). Mixed crystals retaining the
cubic symmetry can be made from these, e.g.,
Ca1-xSrxF2,
Ca1-xBaxF2, and
Ca1-x-yBaxSryF2,
using the vacuum Stockbarger technique or gradient freeze technique, by mixing
CaF2, SrF2, and/or BaF2 powders in the premelt
material. The result, for example for
Ca1-xSrxF2, is a crystal with
optical properties intermediate between those of CaF2 and
SrF2, and the variation should be roughly linear with the x
value. The intrinsic birefringence for
Ca1-xSrxF2 would also be
expected to be intermediate between that for CaF2 and
SrF2. We showed that the spatial-dispersion-induced intrinsic
birefringence depends approximately on a single parameter [1].
Since the value of this intrinsic birefringence for SrF2 has an
opposite sign from that for CaF2 in a wavelength range including
193 nm and 157 nm, and n
should be a continuous function of x at a given wavelength, then a
value x can be found for which n=0 at a given wavelength in the range. If the intrinsic
birefringence varies approximately linearly with x between the two
endpoints, then the value of x required to null out the effect should be
determined approximately by the mixture where the interpolated by birefringence
goes to zero, x | n(CaF2)/[ n(SrF2) - n(CaF2)]|. From the
values in the Table, this gives for
Ca1-xSrxF2
x 0.34 near
157 nm. It has been demonstrated that single-crystal, solid solutions
Ca1-xSrxF2
can be formed for any value of x [3]. Similarly,
Ca1-xBaxF2 solid solutions can
be formed for at least some values of x. A value of x for the
mixture Ca1-xBaxF2 may be found
to null out n at a given UV wavelength
in the range. Using the quaternary mixture
Ca1-x-yBaxSryF2
allows a nulling at two wavelengths, or a broadband correction. This is
possible because the n for
SrF2 and BaF2 have different wavelength dependencies.
Numerous other mixed crystal fluoride combinations that zero out
n are similarly possible.
The mixed crystal materials with intrinsic birefringence nulls targeted for a
given wavelength can be ground and polished into optical elements such as
lenses and beam splitters, just as with CaF2. These elements however,
will have minimal aberrations associated with the intrinsic birefringence at
the target wavelength. Thus they are appropriate for high-precision optics for
that wavelength, where such aberrations are unacceptable. For example, for
157 nm lithography optics, values of n for the optical materials higher than ~1 × 10-7
are considered unacceptable. Our measured value at 156 nm for
CaF2, the principal material being considered for the optics in
these systems, is n = -11.8 × 10-7,
making CaF2 unacceptably birefringent for 157 nm lithography
applications. The same is also true for 193 nm lithography. Optics
fashioned from the mixed crystals as discussed above, will have intrinsic
birefringences and the consequent aberrations within the acceptable range and
meet the design requirements of 157 nm lithography systems.
To
view documents which are "pdf files," a reader is required and
noted here by .
This software
may be downloaded without charge. |
|
|
| [1] | |
John H. Burnett, Zachary H. Levine, and Eric L. Shirley,
"Intrinsic birefringence in
calcium fluoride and barium fluoride,"
(PDF 70 kB ),
Phys. Rev. B 64, 241102(R) (2002). |
|
[2] | |
John H. Burnett, Zachary H. Levine, and Eric L. Shirley,
"Intrinsic Birefringence in 157 nm Materials," in
Proceedings of the Second International Symposium on
157 nm Lithography, ed. by R. Harbison (International SEMATECH,
Austin, 2001). |
|
[3] | |
E.G. Chernevskaya and G.V. Anan'eva, "Structure of Mixed Crystals
Based on CaF2, SrF2, and BaF2," Sov. Phys.
Solid State 8, 169 (1966). |

Inquiries or comments:
john.burnett@physics.nist.gov
Online: July 12, 2001
- Last update: April 2003
|