Czochralski growth of oxides

doi:10.1016/j.jcrysgro.2003.12.044

Journal of Crystal Growth

Volume 264, Issue 4, 31 March 2004, Pages 593-604

https://doi.org/10.1016/j.jcrysgro.2003.12.044 Get rights and content

Abstract

The first oxide grown using the Czochralski technique (CaWO4) was in 1960. Since that time, the Czochralski technique has become the method of choice for the growth and production of many bulk oxide materials used in the electronics and optical industries, e.g. lasers, substrates, scintillators, nonlinear and passive optical devices. This paper will trace the development of the initial process from its beginnings in the early 1960's to its present state and highlight some of the significant advances over the years that have allowed the oxide crystal growth industry to develop. It will conclude with an assessment of the current status and outline directions for further research and development.

Introduction

The use of the Czochralski [1] process to grow semiconductor crystals (Si and Ge) was well established by the mid-1950s [2], [3], [4]. The first reported oxide material grown using the Czochralski technique (CaWO₄) was in 1960 [5]. This initial paper was followed quickly by numerous other papers dealing with the growth of a variety of oxide materials such as LiNbO₃ [6], LiTaO₃ [7], BGO [8], YAG (Y₃Al₅O₁₂) [9], Nd:YAG [10] and Al₂O₃ (Sapphire) [11], [12]. By the mid-1960s, the Czochralski process for the growth of oxide materials was becoming well established. Today, the Czochralski technique, also known as crystal pulling, has become the method of choice for the growth and production of many bulk oxide materials used as components for the electronics and optical industries, e.g. lasers, substrates, scintillators, nonlinear and passive optical devices. To achieve this progress, numerous advances to the understanding of the role that liquid and crystal composition, interface shape/fluid dynamics, thermal geometry and puller design have on the quality of the resulting crystal had to be determined. This paper will trace the development of the initial process from its beginnings in the early 1960s to its present state and highlight some of the significant advances over the years that have allowed the oxide crystal growth industry to develop so that today the size and variety of materials has expanded to include 125 to 150 mm diameter oxide single crystal that can weigh as much as 50 kg. It will then conclude with an assessment of the current status and outline directions for further research and development.

Section snippets

Crystal composition

One of the first problems encountered during the growth of various crystals was that the properties were found to vary through the crystal length as well as from crystal to crystal and from one researcher to another. It was always assumed that the composition of these oxide materials was stoichiometric; hence, the initial starting melt composition was also stoichiometric. The first oxide material identified to show variations in properties due to stoichiometric variations of the liquid was LiTaO

Interface shape/fluid dynamics

With the growth of oxide materials in the early 1960s, very little attention was paid to the shape of the growth interface and its influence on the quality of the resulting crystal. The growth systems were small and the crystals usually were 1–2 cm in diameter. Given such small growth systems, most early crystals were grown with a high thermal gradient and the shape of the growth interface was conical. Furthermore, it was assumed that a conical interface was necessary to maintain stable growth

Diameter control

Of the many mechanical advances that were made in the growth system and furnace construction for oxide materials, the one that has made the most contribution to the growth and quality of oxides has been the development of a useful method for active control of the diameter of the growing crystal. Initial control systems were based solely on control of the power input based on the assumption that if the power input into the furnace was constant then the resulting temperature during growth would

Future directions

As the demand for oxide crystals and their uses increases, the need for the growth of larger crystals (>75 mm diameter) with improved quality will also increase, i.e. the oxide crystal growth industry will follow the same path as the semiconductor industry with cost becoming the main driver. This has already happened for several materials such as Nd: YAG, Sapphire (Al₂O₃), LiNbO₃ and several garnets (GGG and SGGG). However, unlike the semiconductor industry, adequate models for the design of

Summary

Over the past 40+ years of Czochralski oxide growth, significant progress has been made in understanding the various factors involved in the growth of these materials. Each decade has addressed a series of pressing problems that have advanced the Czochralski growth of oxide materials (Table 1). We have progressed from the growth of small (1 cm diameter×3 cm length) 100 g crystals suitable for research purposes to the growth of large oxide crystals that now are 75 mm or larger in diameter that can

References (63)

S.L. Blank et al.
J. Crystal Growth
(1972)
C.D. Brandle et al.
J. Crystal Growth
(1972)
C.D. Brandle et al.
J. Crystal Growth
(1972)
C.D. Brandle et al.
J. Crystal growth
(1974)
V.J. Fratello et al.
J. Crystal Growth
(1987)
C.D. Brandle et al.
J. Crystal growth
(1973)
D. Mateika et al.
J. Crystal Growth
(1979)
N. Iyi et al.
J. Solid State Chem.
(1984)
H.L. Glass et al.
Mater. Res. Bull.
(1973)
K. Takagi et al.
J. Crystal Growth
(1976)

J.R. Carruthers

J. Crystal Growth

(1976)

J.C. Brice et al.

J. Crystal Growth

(1977)

C.D. Brandle

J. Crystal Growth

(1977)

S. Miyazawa

J. Crystal Growth

(1980)

D.C. Miller et al.

J. Crystal Growth

(1978)

D.C. Miller et al.

J. Crystal Growth

(1981)

C.D. Brandle

J. Crystal Growth

(1982)

J.J. Derby et al.

J. Crystal Growth

(1987)

D.F. O’Kane et al.

J. Crystal Growth

(1972)

K.J. Gartner et al.

J. Crystal Growth

(1972)

W. Bardsley et al.

J. Crystal Growth

(1972)

A.E. Zinnes et al.

J. Crystal Growth

(1973)

A.J. Valentino et al.

J. Crystal Growth

(1974)

W. Bardsley et al.

J. Crysral Growth

(1974)

J.A.S. Ikeda et al.

J. Crystal Growth

(1988)

V.J. Fratello et al.

J. Crystal Growth

(1993)

X. Chen et al.

J. Crystal Growth

(1999)

P.F. Paradis et al.

J. Crystal Growth

(2003)

C.D. Brandle et al.

J. Crystal Growth

(1974)

C.D. Brandle

J. Crystal Growth

(1982)

J. Czochralski

J. Phys. Chem.

(1918)

Cited by (91)

Optical and scintillation properties of Yb-doped La<inf>2</inf>Hf<inf>2</inf>O<inf>7</inf> crystal grown by core heating method for fiber reading remote-dosimetry system
2023, Optical Materials
(La_0.99Yb_0.01)₂Hf₂O₇ (Yb:LaHO) crystal has been developed by the core heating method as a near-infrared-emitting oxide scintillator for a fiber reading remote-dosimetry system under the high-dose-rate region. This system consists of a scintillator, a long silica optical fiber and a detector unit. Near-infrared-emitting photons can be separated from significant fiber-derived noise mainly observed in the UV–blue range. The noise is generated in the optical fiber itself by gamma-rays and charged particles. The Yb:LaHO crystal had a maximum emission peak at 975 nm when excited by X-rays and a transmittance of approximately 29% at 977 nm. The afterglow intensity of the Yb:LaHO crystal under X-rays was 11% that of Cr:α-Al₂O₃ (ruby) crystal. We demonstrated the applicability of the remote-dosimetry system with the Yb:LaHO crystal irradiated with gamma rays from a ⁶⁰Co source with an activity of ∼77 TBq. The signal-to-noise ratio (SNR) of the Yb:LaHO crystal was 1.9 times higher than that of the ruby crystal at ∼ 1 kGy/h, and the lowest limit of the dynamic range was comparable between the Yb:LaHO and ruby crystals. The results of the SNR and afterglow intensity showed that the Yb:LaHO crystal is a more suitable scintillator than the ruby crystal for fiber reading remote-dosimetry systems.
Defect generation behavior in Czochralski-grown ScAlMgO<inf>4</inf> crystal using synchrotron X-ray topography
2023, Journal of Crystal Growth
Synchrotron X-ray topography was employed to characterize the defect structure and investigate the defect generation process in Czochralski (CZ)-grown ScAlMgO₄ (SAM) crystals. In the X-ray topographs of the < $\bar{2}$ 0 2 0>, < $\bar{2}$ 0 2 22>, and 0 0 0 24 diffractions of a SAM(0 0 0 1) in-plane wafer, straight- and stream-type dislocations were analyzed using the extinction rule. These dislocations were ascertained to have both < $\bar{1} \bar{1}$ 2 0> and <0 0 0 1> components of the Burgers vector. A SAM(0 $\bar{1}$ 1 0) cross-section wafer was then observed by X-ray topography to investigate the generation process of the defect structure. Straight-type dislocations along [ $\bar{1}$ 0 1 0], [0 1 $\bar{1}$ 0], and [1 $\bar{1}$ 0 0] were generated at the early-growth stage, to release accumulated large thermal strain derived from the solid–liquid phase change. Dislocation-free regions and low-dislocation density regions, including stream- and edge-type dislocations, appeared in the late-growth stage. The weak contrast of the growth-striation in the late-growth stage indicates the reduction in strain. The reduced strain is considered to have generated dislocation-free and low-dislocation density regions. The dislocation structure in SAM reflected the behavior of the thermal strain during the CZ-growth.
Distribution coefficients for rare-earth doping in Y<inf>3</inf>Ga<inf>5</inf>O<inf>12</inf> garnet: Comparison with Y<inf>3</inf>Al<inf>5</inf>O<inf>12</inf>
2022, Journal of Crystal Growth
Citation Excerpt :
The cubic lattice parameter a is obtained as 12.268 Å from the positions of eight peaks in 45° < 2θ < 80°, and this value agrees well with those reported previously on polycrystalline samples [17]. It should be noted that crystals pulled from the melt have larger a, due to partial substitution of Y3+ for Ga3+ at the octahedral B site [4,25]. Such defect is much less severe in flux-grown crystals [4], which is also evidenced in thermal conductivity data [26].
We present the distribution coefficient k in Y₃Ga₅O₁₂ garnet (YGG) for a series of rare-earth dopants (RE = La, Ce, Nd, Eu, Tb, Er, and Lu), and compare the results with our previous data on RE-doped Y₃Al₅O₁₂ garnet (YAG). Under our condition of flux growth, single crystals of undoped YGG show the shape of pure trapezohedron with {2 1 1} faces, whereas RE doping stabilizes {1 1 0} faces. This is to be compared with the case of YAG, where {1 1 0} was the dominant face in both undoped and RE-doped crystals. In RE-doped YGG, k remains near unity when the size mismatch between the dopant and host Y³⁺ ions is within about 3–4%, and drops off rapidly for larger dopants. This is in contrast to RE-doped YAG, where k decreases continuously from the smallest (Lu³⁺) to the largest (La³⁺) dopant. We find such distinct behaviors to be correlated with the reported solution energies for the series of RE ions in YAG and YGG.
The kinetic parameters of the main thermoluminescence glow peak of Al<inf>2</inf>O<inf>3</inf>:C,Mg: A critical evaluation of different analytical methods
2022, Journal of Luminescence
Carbon and magnesium co-doped aluminum oxide (Al₂O₃:C,Mg) is a highly sensitive luminescence dosimeter with promising use in a wide spectrum of radiation-related applications, including neutron dosimetry and as a fluorescent nuclear track detector (FNTD). The goal of this work is to critically evaluate diverse methods and approaches for the determination of the kinetic parameters using Al₂O₃:C, Mg as a case study. Al₂O₃:C, Mg was beta irradiated with doses from 0.1 to 0.6 Gy. Besides thermoluminescence (TL) peaks at 325, 350, and 375 K, the analysis of the activation energy, frequency factor and order of kinetics focused exclusively on the main TL peak at 450 K. Analysis by curve fittings used a number of freeware, namely, Glowfit, TLAnal, the Thermoluminescence Glow Curve Deconvolution (TGCD) package, and the spreadsheet TLDecoxcel. The results from the computational approaches were compared with results obtained by analyzing experimental data using conventional methods as initial-rise, whole glow peak, variable heating rate and peak shape. The performance of the computational methods is satisfactory as the values found are consistent with the ones determined by the methods listed above.
Optical and spin inhomogeneous linewidths in <sup>171</sup>Yb<sup>3+</sup>:Y<inf>2</inf>SiO<inf>5</inf>
2022, Optical Materials: X
Single crystals of ¹⁷¹Yb $^{3 +}$ :Y $_{2}$ SiO₅ are promising systems for quantum technologies as an efficient optical to spin interface. This is due to the long optical and spin coherence lifetimes, or narrow homogeneous linewidths, observed in this material at zero magnetic field. Inhomogeneous linewidths are also of importance since they affect peak absorption, available bandwidth or spin–spin interactions. In this work, we investigate inhomogeneous linewidths in a series of ¹⁷¹Yb $^{3 +}$ :Y $_{2}$ SiO₅ single crystals grown by the Czochralski method with the purpose of identifying the broadening mechanisms in the regime of ultra-low doping, i.e. in the ppm range. In a series of samples obtained with varying doping level, growth atmosphere, and cut at different locations in a crystal boule, optical inhomogeneous linewidths between 0.6 and 1.3 GHz are observed. This study suggests that these variations are due to changes in dislocation densities. Spin inhomogeneous linewidths, recorded using electron paramagnetic resonance, did not show the differences between samples observed in the optical domain. They were otherwise found to vary as a function of magnetic field orientation, enabling determination of directions that minimize e.g. detrimental spin–spin interactions. These results suggest ways to reduce optical inhomogeneous linewidths in ¹⁷¹Yb $^{3 +}$ :Y $_{2}$ SiO₅, which is of interest for high efficiency quantum memories and optical to microwave quantum transducers.
Observation of defect structure in ScAlMgO<inf>4</inf> crystal using X-ray topography
2022, Journal of Crystal Growth
We examined the defect structure in a ScAlMgO₄ (SAM) crystal intended for use as a substrate of GaN using synchrotron X-ray topography. No dislocations were observed in 70% of the 40 mm SAM wafer, indicating a wide dislocation-free region. The remaining 30% of the wafer was composed of high and ultra-high dislocation density regions. Two types of dislocations, straight-type and stream-type dislocations, along the [0 1 $\bar{1}$ 0], [1 $\bar{1}$ 0 0], and [1 0 $\bar{1}$ 0] directions, were predominantly present in the high dislocation density region. The dislocation colonies with only straight-type or stream-type dislocations were observed and the high dislocation density region was occupied by these colonies. The lattice strains of the wafer center region that had no dislocations and the wafer edge region that had an ultra-high dislocation density weren’t uniform. The SAM crystal had unique structures such as a wide dislocation-free region, straight-type and stream-type dislocations, formation of dislocation colonies, and lattice strains that weren’t uniform.

View all citing articles on Scopus

View full text

Czochralski growth of oxides

Abstract

Introduction

Section snippets

Crystal composition

Interface shape/fluid dynamics

Diameter control

Future directions

Summary

J. Crystal Growth

J. Crystal Growth

J. Crystal Growth

J. Crystal growth

J. Crystal Growth

J. Crystal growth

J. Crystal Growth

J. Solid State Chem.

Mater. Res. Bull.

J. Crystal Growth

J. Crystal Growth

J. Crystal Growth

J. Crystal Growth

J. Crystal Growth

J. Crystal Growth

J. Crystal Growth

J. Crystal Growth

J. Crystal Growth

J. Crystal Growth

J. Crystal Growth

J. Crystal Growth

J. Crystal Growth

J. Crystal Growth

J. Crysral Growth

J. Crystal Growth

J. Crystal Growth

J. Crystal Growth

J. Crystal Growth

J. Crystal Growth

J. Crystal Growth

J. Phys. Chem.