Journal of Crystal Growth 649 (2025) 127949

Contents lists available at ScienceDirect

Journal of Crystal Growth

journal homepage: www.elsevier.com/locate/jcrysgro

Ultraviolet optical properties of CaSrF2 crystal grown by the cone die

Czochralski method
Yugo Akabe a,*, Kotaro Enomoto a , Keito Shinohara a , Marilou Raduban b, Toshihiko Shimizu a ,
Nobuhiko Sarukura a , Rui Okumura a , Ryosuke Kodama a, Kazuya Takahashi c , Toru Kawamata d,
Kazumasa Sugiyama d , Tsuguo Fukuda c
a
Institute of Laser Engineering, Osaka University, 2–6 Yamadaoka, Suita, Osaka 565–0871, Japan
b
Unitec Institute of Technology, 139 Carrington Road, Mount Albert, Auckland 1025, New Zealand
c
Fukuda Crystal Laboratory Co., Ltd., 6-6-3 Minami-Yoshinari, Aoba-ku, Sendai 989-3204, Japan
d
Institute for Materials Research, Tohoku University, 2-1-1 Katahira, Aoba-ku, Sendai 980-8577, Japan

A R T I C L E I N F O A B S T R A C T

Communicated by Jeong Ho Ryu Fluoride crystals with extremely wide band gaps are ideal optical materials in the UV wavelength range. Large 4-
inch diameter calcium strontium fluoride (Ca0.582Sr0.418F2) single crystal was grown using the Czochralski
method with a Cone-shape Die (CD-CZ). The refractive index and relative transmittance of the crystal was
evaluated by cutting it into a triangular prism and polished. The direct measurement of the refractive index and
relative transmittance was done by using a spectrograph to image the refraction of light as it passes through a
dual prism set-up consisting of a SQ prism as reference and either CaF2 or Ca0.582Sr0.418F2 as the material under
evaluation. Characterization results showed that Ca0.582Sr0.418F2 has excellent refractive index dispersion and
transmittance in the UV region, confirming the applicability of the dual prism with spectrograph setup to
measurements of the refractive index and relative transmittance in the UV region.

1. Introduction exposure to high energy laser light. In recent years, the demand for more
durable optical materials has increased as lasers have become more
Fluoride single crystals are highly transparent to wavelengths powerful. Furthermore, large crystals of at least 2 in. in diameter are
ranging from the infrared to the vacuum ultraviolet (VUV) regions. required for practical use. However, pure CaF2 has several limitations
Therefore, they are valuable as optical elements for many applications. that must be overcome, including high cleavage, F defects, structural
In terms of crystal growth, fluoride crystals have a lower melting point phase transitions at high temperatures, and birefringence in the short
and are easier to manufacture compared to oxide crystals because they wavelength region. The problem with birefringence can be solved by
do not require crucibles of precious metals such as iridium and platinum. using mixed crystals [12]. In a recent paper, we reported on the growth
However, fluorides are relatively soft and easily crack, making post- and properties of large Ca0.582Sr0.418F2 single crystals with a diameter of
growth processing such as machining and polishing difficult. In addi­ 4 in. [3]. The success in the growth of the Ca0.582Sr0.418F2 single crystals
tion, fluoride crystals must have a low concentration of impurities to is due to the use of a resistance-heated Czochralski (CZ) furnace and a
prevent light absorption and emission. It should also be free of bubbles conical graphite die (Cone Die) that is placed inside the crucible used for
and inclusions that scatter light. Because of these constraints, produc­ melting the raw material. This growth method is called the Czochralski
tion of fluoride crystals is particularly difficult for UV applications, technique using Cone-shape Die (CD-CZ method) [34]. In this paper, we
which require high quality crystals. This has resulted to low productivity evaluate the optical quality of large Ca0.582Sr0.418F2 single crystals
of fluoride crystals for UV applications. grown by the CD-CZ method. The Ca0.582Sr0.418F2 crystal was processed
Calcium fluoride (CaF2) is mainly used as an optical material in the into prisms. The wavelength-dependent refractive index of
VUV region. CaF2 is particularly useful in photolithography using a 190 Ca0.582Sr0.418F2 was then evaluated using a technique that we developed
nm ArF laser because of its optical durability to withstand prolonged with a dual prism configuration combined with a spectrometer.

* Corresponding author at: Institute of Laser Engineering, Osaka University, 2–6 Yamadaoka, Suita, Osaka 565–0871, Japan.
E-mail address: [email protected] (Y. Akabe).

https://doi.org/10.1016/j.jcrysgro.2024.127949
Received 31 July 2024; Received in revised form 10 October 2024; Accepted 11 October 2024
Available online 19 October 2024
0022-0248/© 2024 Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
Y. Akabe et al. Journal of Crystal Growth 649 (2025) 127949

The combination of the spectrometer and the dual prism configura­ width, and height of 1 cm. The lattice constants estimated from powder
tion allows the refractive index in a broad wavelength range to be ob­ X-ray diffraction (powder XRD) measurements for each Sr concentration
tained in a single measurement. Conventional methods for measuring in the starting material shows an excellent linear fit between the lattice
refractive index use degree deviation, ellipsometry, Brewster’s method, constants of CaF2 and SrF2, as reported previously [3]. This indicates a
and interferometry. All these existing methods use several mono­ uniform crystal growth without any change in the crystal structure and
chromatic light sources to measure the refractive index at discrete without significant segregation. The composition of both the top and
wavelengths, and then perform fitting using the Sellmeier equation or bottom sections of the ingot, as well as the remaining material in the
other methods to obtain the refractive index dispersion parameters. On crucible, was analyzed using energy dispersive X-ray fluorescence
the other hand, such discrete refractive index measurements may miss spectroscopy (Rigaku NEXCG) to explore the variation in composition
steep refractive index changes due to the presence of specific absorption within the ingot and deviations between the ingot and the melt. The Sr
peaks. In contrast, the combination of the spectrometer and the dual distribution across the crystals, the feedstock, and the residual melt
prism configuration, which calculates the refractive index from changes showed comparable trends. Deviations from expected composition lines
in the position of direct light irradiation, uses a broad white light source could be attributed to experimental factors, such as Ca dissolution and
to simultaneously obtain the refractive index for all wavelengths. In seed crystal evaporation [3]. The experimental setup is shown in Fig. 1
addition, while existing refractive index measurement methods require (a.) A detailed description of the dual prism configuration and spec­
precise and complex setups, our technique has several advantages, trometer setup can be found elsewhere [5]. A laser-driven fiber pigtail
including flexibility in the choice of illumination source, alignment lamp (Energetiq EQ99CAL) emitting at 210–800 nm was used as a high-
stability, scalability, and the ability to measure a variety of samples. brightness light source. The light was collimated using lens 1 (f = 100
Thus, measurements of new material samples where absorption peaks mm) and lens 2 (f = 100 mm) and then focused onto a diffuser screen by
may be present can be performed very quickly and easily. lens 3 (f = 500 mm). The prism surfaces were uncoated and polished.
Using this dual-prism system, we have previously performed The two prisms were aligned using acetone as the contact solution. To
refractive index and absorptivity measurements on borosilicate glass, measure the deflection of light from the prisms, the axial distance L from
but mainly in the visible region [5]. Moreover, the method used to the end face of the sample prism to the diffusion screen was fixed.
analyze the results was problematic, and the refractive index dispersion Perpendicularity of the prisms to the incident beam was maintained by
information obtained contained many skipped values and absorption monitoring the back reflection from the front surface of the reference
peaks. In order to show that the dual-prism system can determine the prism and passing it through a narrow pinhole. In addition to the
refractive index dispersion and absorption peaks over a wider wave­ Ca0.582Sr0.418F2 prism, a CaF2 and SQ prism of the same size were pur­
length range, this experiment used a UV imaging camera to measure the chased for reference. Light incident on the prism pair is deflected by an
optical parameters of the CD-CZ-grown Ca0.582Sr0.418F2 single crystal for angle δ in Fig. 1 (b). To measure this deflection, a UV Nikkor lens (f =
the wavelength range from 200 to 400 nm. The process of analyzing the 105 mm) was attached to the incident slit of a UV spectrometer (Spec­
results has been improved by taking the ratio of the refractive index trometer, KV100-1) with a focal length of 100 mm to image the vertical
dispersion parameter to the transmittance compared to CaF2 to deter­ position of the beam. The grating density of the spectrometer is 400
mine the presence or absence of absorption peaks. mm− 1. The wavelength resolution of the spectrograph is 0.26 nm. The
spatial position of the spectral components was imaged with a 2040 ×
2. Experiment 2048 CCD sensor.
Given the refractive index n1 of the reference glass, the wavelength-
A part of the 4-inch Ca0.582Sr0.418F2 single crystal grown by the CD- dependent refractive index n2 of the target sample (CaF2 or
CZ method was processed to create a triangular prism with a length, Ca0.582Sr0.418F2) is derived from Equation (1). More precisely, X should

Fig. 1. (a) Schematic diagram for measurement of the refractive index in optical glass based on double-prism configuration with imaging spectrograph. (b) Optical
geometry of light exiting the two-element prism pair.

2
Y. Akabe et al. Journal of Crystal Growth 649 (2025) 127949

be calculated including the change in position d due to refraction within difference in refractive index of CaF2 and Ca0.582Sr0.418F2. As shown in
the prism, but in this experiment this value is very small and X and x can the figure, the contrast at one wavelength is spatially broad. Therefore,
be approximated to be equal. we previously calculated the center position of the light at each wave­
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ length using direct Gaussian fitting to such an image. However, due to
n2 = 2sin2 δ + n21 + 2n1 sinδ (1) sharp peaks in the noise, the Gaussian fitting did not work well, and the
final refractive index dispersion graphs were generally mixed with
The dual prism configuration combined with a spectrometer is limited to skipped values. Furthermore, due to the limitations in the grating den­
cases where the refractive index of the material being measured is not sity of the spectrometer, it was necessary to change the center wave­
large. The refractive index and dispersion graph of the Ca0.582Sr0.418F2 length of the imaging monochromator multiple times within the 400-
prism is expected to lie between that of CaF2 and SrF2, especially in the 650 nm range and then taking multiple images. This created a discon­
UV region, so the position of the dispersed light projected onto the tinuous structure on the refractive index dispersion curve. To address
diffuser screen should not be too far from the non-refracted position. If that issue, this work employed an imaging spectrograph that can cover
the refraction is too great such that the refracted light exceeds the angle all wavelengths from 200 to 400 nm such that it was not necessary to
of view of the imaging spectrometer, the setup must be realigned, change the center wavelength of the imaging monochromator and only
compromising the reliability of the data. In addition, increasing the one image per prism pair is obtained. The center position at each
number of lenses and other components to fit within the angle of view wavelength was calculated by smoothing to remove noise before per­
complicates the system and decreases the amount of light reaching the forming Gaussian fitting. To ensure that the parameters of the Gaussian
imaging camera from the light source. Therefore, two patterns of mea­ fit were set more precisely than before, the initial center position was set
surements were performed; one using a prism pair of SQ + to the point of maximum light intensity after smoothing. We estimated
Ca0.582Sr0.418F2 and another one using a prism pair of SQ + CaF2. CaF2 the accuracy of the wavelength-dependent refractive index of this sys­
has a larger refractive index than SQ and a refractive index that is closer tem to be 10-2 [5]. Since the variation of the refractive index values was
to that of Ca0.582Sr0.418F2. Therefore, originally, the refractive index was less than 10-4, the effect of smoothing is negligible.
derived from the difference in the position of the light when the prism The obtained refractive index dispersion diagram is shown in Fig. 3.
was not present such that the light traveled straight ahead and the po­ Information about the refractive indices of CaF2 and SrF2 was taken from
sition of the light when it travelled through the prism pair [5]. This work the database. The calculated refractive indices of Ca0.582Sr0.418F2 are
introduced an innovation to that original method by measuring two intermediate between those of CaF2 and SrF2. Although the
different prism pair patterns and the difference was used to measure the Ca0.582Sr0.418F2 sample has a higher percentage of Ca than Sr, the
refractive index. This difference is derived from the difference in the refractive indices of Ca0.582Sr0.418F2 are closer to SrF2 than to CaF2. The
refractive indices of CaF2 and Ca0.582Sr0.418F2. Therefore, the refractive three-term Sellmeier dispersion equation was fitted to the refractive
index data obtained from the SQ + CaF2 prism pair were used to indices of the measured samples. The six coefficients for the absorption
calculate the absolute refractive index of Ca0.582Sr0.418F2 and the band strength and wavelength of maximum absorption, respectively,
parameter when approximated by the Sellmeier equation. The relative were derived by the least-squares fitting method [678].
transmittance difference between CaF2 and Ca0.582Sr0.418F2 was
0.56759 0.47134 3.84888
measured by comparing the amount of light at each wavelength detected n2 − 1 = + + (2)
λ2 − 0.02526 λ2 − 0.01005 λ2 − 1200.67
by the imaging spectrometer.
The dispersion equation is shown in Equation (2). The relative trans­
3. Results mittance of Ca0.582Sr0.418F2 relative to CaF2 was then calculated from
the light intensity at each wavelength. Considering that light passed
Fig. 2 shows the images obtained by the imaging spectrometer. Fig. 2 through the center of the prism pair and acetone was used as the contact
(a) is the image obtained with the SQ + CaF2 prism pair while Fig. 2 (b) liquid, the ratio of the amount of light transmitted through SQ and
shows the image obtained with the SQ + Ca0.582Sr0.418F2 prism pair. The acetone and incident on CaF2 and Ca0.582Sr0.418F2 was calculated. Fig. 4
horizontal axis shows wavelengths from 200 nm to 400 nm, and the shows the results. The ratio of transmittances ranged between 0.95 and
vertical axis shows spatial information. Although the angle of view of 1.1. A wavy structure can be seen in the ratio of transmittances between
the imaging camera was not changed in Figures (a) and (b), it can be 200 nm and 400 nm. This may be due to interference between the prism
seen that the position of the curves in the two images is different. This is pairs. The refractive index of acetone is about 1.37 in the UV region,
due to the difference in projection position that originates from the which is lower than that of SQ, CaF2, and Ca0.582Sr0.418F2. Therefore, the

Fig. 2. The images of imaging spectrograph. (a)SQ and CaF2. (b) SQ and Ca0.582Sr0.418F2.

3
Y. Akabe et al. Journal of Crystal Growth 649 (2025) 127949

Fig. 3. Dispersion curves of Ca0.582Sr0.418F2, CaF2 and SrF2.

Fig. 4. Transmittance ratio of Ca0.582Sr0.418F2 relative to the CaF2 reference glass.

effect of light reflection at the interface is expected to be larger. The necessary to obtain the absolute transmittance. Nevertheless, the results
transmittance of Ca0.582Sr0.418F2 exceeded that of CaF2 on the long of this experiment indicate that our dual prism and spectrometer
wavelength side (transmittance ratio is greater than one) and was lower refractive index measurement system can be applied not only in the
than that of CaF2 on the short wavelength side. On the other hand, the visible region (as previously reported [5]) but also in the UV region. Our
imaging spectrometer curve in Fig. 2 shows a slight noise-like contrast in dual prism and spectrometer refractive index measurement system is
the Ca0.582Sr0.418F2 image, especially at the long wavelength side. expected to be one of the methods for quickly and easily determing the
Although this result confirms the absence of a large absorption peak, the optical parameters for the development of new materials such as
selection of the contact liquid and improvement of the system are Ca0.582Sr0.418F2 which is attracting attention as an optical material for

4
Y. Akabe et al. Journal of Crystal Growth 649 (2025) 127949

the UV region. Furthermore, it was confirmed that both the value and Investigation, Conceptualization. Kotaro Enomoto: Validation, Inves­
shape of the refractive index dispersion of Ca0.582Sr0.418F2 are close to tigation. Keito Shinohara: Validation, Investigation. Marilou Radu­
that of CaF2, and that there is no large absorption peaks in the UV region. ban: Writing – review & editing. Toshihiko Shimizu: Project
These results may lead to the future production of stronger administration. Nobuhiko Sarukura: Project administration, Method­
Ca0.582Sr0.418F2 materials that can replace CaF2 not only as window ology. Rui Okumura: Investigation. Ryosuke Kodama: Supervision.
materials but also as lenses, prisms, and various other optical elements. Kazuya Takahashi: Resources. Toru Kawamata: Investigation. Kazu­
In addition, although this experiment was conducted on a material that masa Sugiyama: Supervision. Tsuguo Fukuda: Supervision.
is expected to maintain a high transmittance over a wide wavelength
range, we plan to evaluate the dispersion near the sharp absorption edge Declaration of competing interest
of ion-doped optical materials and to verify that it is possible to rapidly
measure steep refractive index dispersion, which has been difficult to The authors declare that they have no known competing financial
identify and resolve using conventional techniques. In addition, the interests or personal relationships that could have appeared to influence
resistance of Ca0.582Sr0.418F2 to optical damage may also be measured the work reported in this paper.
using this technique and will be investigated in further detail in future
experiments. Data availability

4. Conclusion Data will be made available on request.

In conclusion, large 4-inch diameter Ca0.582Sr0.418F2 single crystal References

was grown using the CZ method using cone die (CD-CZ method). No
foreign matter or bubbles were observed. The refractive index of the [1] J.H. Burnett, Z.H. Levine, E.L. Shirley, Hidden in plain sight: Calcium fluoride’s
intrinsic birefringence, Photonics Spectra (2001) 88–92.
Ca0.582Sr0.418F2 single crystal was evaluated by processing part of the as- [2] J.H. Burnett, Z.H. Levine, E.L. Shirley, Intrinsic birefringence in calcium fluoride
grown crystal into a triangular prism that was used in conjunction with a and barium fluoride, Phys. Rev. B 64 (2002) 241102.
reference SQ prism, thereby also allowing the transmittance ratio to be [3] K. Takahashi, M. Cadatal-Raduban, N. Sarukura, T. Kawamata, K. Sugiyama,
T. Fukuda, Crystal growth and characterization of large Ca0. 582Sr0. 418F2 single
determined. The refraction of light as it passed through the prism pair crystal by Czochralski method using cone die, J. Cryst. Growth 628 (2024) 127541.
was visualized with a spectrograph unit. The refractive index of [4] M. Cadatal-Raduban, T. Shimizu, N. Sarukura, K. Takahashi, T. Fukuda, Crystal
Ca0.582Sr0.418F2 was observed to be between that of purely CaF2 and SrF2 growth of ultra-large MgF2 and LiCaAlF6 single crystals by a double-crucible
Czochralski technique, J. Cryst. Growth 571 (2021) 126260.
crystals. The transmittance of Ca0.582Sr0.418F2 exceeded that of CaF2 on [5] Youwei Lai, Jacque Lynn Gabayno, Marilou Cadatal-Raduban, Kohei Yamanoi,
the long wavelength. These results may lead to the future production of Toshihiko Shimizu, and Nobuhiko Sarukura, Direct measurement of refractive index
stronger Ca0.582Sr0.418F2 materials that can replace CaF2 not only as and dispersion of optical glass by dual-prism configuration with imaging
spectrograph, Japanese J. Appl. Phys. (2019) 58.
window materials but also as lenses, prisms, and various other optical
[6] H. Malitson, Interspecimen comparison of the refractive index of fused silica, J. Opt.
elements. Soc. Am. 55 (1965) 1205–1208.
Funding [7] H.H. Li, Refractive index of alkaline earth halides and its wavelength and
This research did not receive any specific grant from funding temperature derivatives, J. Phys. Chem. Ref. Data 9 (1980) 161–289, and references
therein.
agencies in the public, commercial, or not-for-profit sectors. [8] L. Rodríguez-de Marcos, J.I. Larruquert, J.A. Aznárez, M. Fernández-Perea, R. Soufli,
J.A. Méndez, S.L. Baker, E.M. Gullikson, Optical constants of SrF2 thin films in the
CRediT authorship contribution statement 25–780-eV spectral range, J. Appl. Phys. 113 (2013) 143501.

Yugo Akabe: Writing – original draft, Visualization, Validation,