JPS6278197A - Manufacture of homogeneous lithium niobate - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for producing lithium niobate crystals and to the production of devices constructed from lithium niobate crystals.

Background of the Technology Lithium niobate is widely used in a variety of applications because of its electrical and dielectric properties and the relative ease with which large single crystals of high quality can be produced. Typical uses include:
There are uses in surface acoustic wave devices and as substrates for integrated optical devices. Because of its low dielectric constant and high acoustic velocity, lithium niobate is particularly important as surface acoustic wave devices, and because of its relatively large electro-optical effect and high Curie temperature, it is also useful as a variety of optical devices. be.

Single-crystal lithium niobate is usually produced by pulling it from a melt of lithium niobate by a method commonly called the so-called Tsouklaski method. Previous studies on the growth of lithium niobate indicate that the components are not stoichiometrically harmonically fused. Harmonic melting is a state in which the composition of the melt and the composition of the solid in equilibrium with the melt are the same.

Harmonic melt composition is very important for crystal growth procedures such as lithium niobate.

With this melt composition, the composition of the pulled crystal does not vary along its length, making it possible to obtain a crystal with a constant composition. The crystal has identical optical and acoustic properties along the length of the crystal.

For this reason, many studies have been conducted to determine the harmonic melt composition of lithium niobate. Lerner (L
The paper “Stoichiometry” by Erner et al.
Death monocristo demetaniobate de lithium (
Stoechi-ometrie des Mon
ocristaux de Metaniobat
ede Lithium) J, Journal of Crystal
Al Growth), 3.4, (1968), pages 231-235, the range of solubility for single phase lithium niobate at about 12O0C is determined. At this temperature, the lithium content in single-phase LiNbO3 ranges from 44.1 molar percentage Li 2O to 50
．． 1 mole percent Li2O, where 50.
0 mole percentage is the stoichiometric composition.

At this point, it is necessary to mention a little about nomenclature. Conventionally, LiNbO3 is considered to consist of 50 mole percent Li2O and 50 mole percent Nb2O5. Deviations from stoichiometry are expressed as deviations from 50 mole percent Li 2 O.

In the paper cited above, Lerner et al. also performed measurements of lattice parameters and differential thermal analysis to determine interfaces in the phase diagram, as well as measurements of transition temperatures for various compositions. ,
These measurements indicate that the harmonic melt composition is between 48 and 49 mole percent Li2O.

Since the harmonic melt composition is very important, more detailed studies are being conducted on the phase diagrams of lithium oxide and niobium oxide systems.

A particularly detailed study was carried out by Carruthers.
[Nonstoichio-met and crystal growth of lithium niobate]
ry and Crystal Growth of
Lithium N1o-bate) J, Journal of Applied Physics (Journal
of ApHied Physics), 42, A5
．． Published in May 1971, pages 1846-1851. They discuss the growth and percentage of lithium niobate crystals, as well as the effects of crystal composition on the properties (particularly optical properties) of lithium niobate crystals.

Carruthers et al. also performed a series of measurements to determine the harmonic melt composition. These experiments include measurements of Curie temperatures for a range of ceramic compositions and comparisons of these results with Curie temperatures obtained from single crystals grown from known melts. cucumber one temperature and ceramic
By comparing the compositional behavior of both the sample and the single crystal grown from the melt, the harmonic melt composition is determined to be 48.6 mole percent Li2O. Furthermore, 102
In the process for determining the range of solid solution at O
It is said to be within the range of The data obtained by them further showed the Curie temperature as a function of the near-stoichiometric composition of lithium niobate, the partition coefficient between the liquid and the solid lithium niobate at equilibrium, and the stoichiometry of lithium niobate. It provides other valuable information such as the phase diagrams of lithium oxide and niobium oxide around the composition.

Strong evidence for harmonic melt composition can be obtained from numerous scientific studies reported in scientific journals. for example,
F, R, Natush (F, R.

[Effect of Optical Inhomogeneity on Phase Matching of Nonlinear Crystals] et al.
ical Inhomogeneities on
Phase Matching in Non
1inear Crystals), Journal of Applied Physics (Journal of A
ppliecLPPhysics), V o l,
41, &6 (May 1970), page 2564
-2576, the phase matching conditions for the second harmonic generation are measured for various intercepts along the pull axis of lithium niobate. This result is associated with an unusually large variation in the refractive index of the crystal, and it is concluded that this variation is due to compositional variations along the length of the crystal.

By measuring the birefringence of the crystal along the growth direction, they give the conclusion that the harmonic melt composition is 48.6 mole percent LiO. Similar measurements were carried out in the paper “Growth of high-quality LiNb crystals from harmonic melts” by R,L Byer et al.
High-Qua eyes ty LiNb03Crysta
ls from the Congruent Mel
t) J, Journal of Applied Phys.
sic's), Vol. 41, A6 (May 1970), pages 232O-2325. Carruthers et al. and Nash and Byer
In their study, the harmonic melt composition was 48.6 mole percent Li.
It is concluded that 2O. Otsucho (K, C
The Congru-entry Melt of LjNbO3
ing Composition) J, Material Research Preten (Material Re5ea)
rchBulletin) further provides evidence that 48.6 mole percent Li2O is the harmonic melting composition of lithium niobate.

These experiments now lead crystal growers and others involved in the production of crystals to believe that the harmonic melt composition of lithium niobate is 48.6 mole percent Li2O. This composition is used in the production of

SUMMARY OF THE INVENTION The present invention is based on the very surprising discovery that harmonic melt compositions are significantly different from those hitherto accepted and used. The present invention relates to a process for growing lithium niobate crystals by pulling them from a melt having a melt composition of 48.45 mole percent Li2O.

Although it is very important to precisely control the melt composition to obtain a homogeneous composition throughout the crystal, this method allows for variations up to ±0.08 mole percent. However, preferably ±0.04 or less, more preferably ±00
A variation of less than 2 mole percentages is required. According to the invention, very good lithium niobate crystals with high homogeneity, constant composition and constant properties such as refractive index, dielectric constant, Curie temperature, nonlinear optical properties, birefringence, etc. Obtainable. The invention includes an optical device constructed from lithium niobate grown in accordance with the invention. This device has excellent properties because the composition of the lithium niobate crystal is known and uniform, and the optical properties are constant. This is particularly suitable for optical integrated circuits with lithium niobate substrates.

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a process for producing lithium niobate crystals from melts of lithium and niobium oxides. Here, the concentration of lithium expressed as a molar percentage of Li2O is 4845±0.08. This melt usually consists of high purity lithium and niobium oxides (e.g. 0.11 or 0.05, or even 0.0
containing impurity concentrations of up to 3 mole percent), with small amounts (usually up to 1.0 mole) of other elements (other than lithium and niobium) for the purpose of adjusting various physical properties such as Curie temperature, refractive index, etc. (up to a percentage) can also be added.

The present invention is applicable to all processes for producing crystalline lithium niobate using the equilibrium state between harmonically melted lithium niobate and solid lithium niobate. Typically, lithium niobate is produced from molten lithium niobate by a crystal pulling method. This crystal pulling method is often referred to as the Tsouklaski method and consists of introducing lithium niobate seeds into a lithium niobate solution and slowly pulling it up while rotating. C, D, Brandle and A, J, Valentino.

Growth), December 1972, page 3
-8 [Czochraski Growth of Rare Earth Gallium Garnet]
of Rare Earth Gallium Gar
A typical crystal pulling apparatus and process is shown in J.

FIG. 1 shows a typical crystal pulling apparatus 10 used in carrying out the present invention. This device is a heat source (rr coil 1
1), platinum crucible 1 containing lithium niobate melt 13
2, and a lithium niobate crystal 14 to be pulled. The platinum crucible 12 is equipped with platinum after a heater 15 and a platinum lid 16 to control the temperature gradient inside the pulling device, and is further equipped with a ZrO□ base 17, a ZrO2 support 18
, ZrO2 disk 19, ZrO2 tube 2O, Zr
It has a heat insulating structure consisting of an O2 log 21, a ZrO2 lid 22, a ZrO2 felt 23, a platinum reflector 24, and a crystal sleeve 25. An initial melt level 26 is shown, which drops as the lithium niobate crystals 14 are pulled out of the melt 13.

An important aspect of the invention is the determination of harmonic melt composition. The method for determining the harmonic melt composition consists in accurately measuring the Curie temperature (Tc) of a lithium niobate crystal of known composition.

The Curie temperature was determined by mixing known amounts of analyzed raw materials and measuring the resulting crystal composition using high-precision differential thermal analysis (DTA). In this way, a melt whose initial composition was known was sampled before and after crystal growth. Since the upper and lower cross sections of the grown crystal are in equilibrium with the melt, the distribution coefficient between the melt and the crystal was also measured by this method.

Perkin Elmer DTA Model 1700 (Pe
rkin-Elmer DAT Model 1700
) 'r Fluke Model 2452 (Fluke
DTA data were obtained by use with a data acquisition system based on Model 2452).

This measurement was carried out over a temperature range of 1000° C. to 1200° C. Heating and cooling rates were expressed as 20° C./min, and measurements were performed in a flowing nitrogen atmosphere. In a process where aluminum oxide was used as a reference material, a platinum crucible was used as a container for this material. Gold melting curves were used to calibrate the device.

The crystals used in this experiment were grown by the Tsouklaski method. Starting materials (Li2CO3 and Nb2O5
) are both 99.999% pure and manufactured by Johnson Matthey.
Inc.). Prior to the preparation of the charge, each of the constituent oxides was heated from 12 to 1 at a temperature of 500°C.
It was dried for 4 hours. After drying, the starting powder was weighed and thoroughly mixed. Weighing was performed with an accuracy of ±0.01 grams. The mixed powder is then heated to 1000°C from 12 to 1
It was reacted for 4 hours. Weight loss was checked after heating to ensure complete decomposition of the lithium carbonate.

This reaction powder was then placed in a crucible in a process for compaction into pellets. In all experiments, the final reacted and compressed powder pellet weight was 240.0±0.1 grams.

The compressed LiNbO3 pellets were then placed in the (RF heated) furnace shown in FIG. 1, described above. The molten charge (240 grams) was placed in a 2 inch outside diameter x 2 inch height platinum crucible. The crystal growth parameters used in this experiment are as shown in Table 3. Once the charge was completely melted, the melt was stirred for several minutes to ensure thorough mixing before proceeding with growth.

Previous experiments have shown that complete melting of the charge does not necessarily mean that the melt is completely homogeneous. After stirring,
Quickly insert polycrystalline Al2O30 into the melt,
Initial melt samples were collected by drawing. This process was repeated 2.3 times until approximately 0.5 gram sample was obtained. This sample is then Tc
It was used as a composition standard for measurements.

Table ■ Crystal growth parameters Growth direction “C” axis (0001) Pulling speed
0.15 inch/hour (0.38 cm/hour) Rotation speed 15 RPM Atmosphere Air crucible 2 inch outside diameter x 2 inch depth Made of platinum Initial charge 240.0 grams Typical crystal weight 150 to 60 grams crystal Length of approximately 10 cm Crystal diameter 2O to 21- After completion of crystal growth, the final bath melt sample was collected in a manner similar to that in which the initial melt sample was collected. However, in this case, the crucible and residual melt were subsequently heated in the process for the furnace to cool to room temperature for crystal removal. As with the initial sample, the melt was thoroughly stirred before collecting the sample.

Single crystal samples were collected from the top and bottom of each crystal ball. The initial crystal sample used for Tc measurements is typically 0.5 mm thick, 20 mm in diameter, and weighs 0.
7 to 1 gram. For all initial crystal samples? (Ratio of solidified melt) was ≦001.

The final crystal samples were of similar size, but all had one side that was a crystal interface. The Tc of the upper and lower sections of the crystal ball was measured by DTA.

The result of Tc measurement shows that the molar percentage of L1□0 in the melt is 4.
7.0 to 490, and most samples showed a mole percentage range of 480 to 490. When this data is expressed by the linear least squares method, it is found that the following relationship exists between the Curie temperature and the composition.

T c = -637,30 + 36.70 C or C = 17.37 + 0.02725 T c where T c is Curie f in degrees Celsius? A fK 'c, and C represents the Li2O content expressed in hundred moles of the sample. Careful observation of the Curie temperature data for this composition reveals the presence of nonlinearity as the L1□0 melt component increases. Therefore, this data is expressed as T c = a + b (+ c C
” Expressed by the least squares method using the form, the following relationship: T c = 9095.2-369.050 +4.2
28C2, which turns out to give a better approximation than the linear model.

The Li2O content of individual samples was determined using this relationship and by measuring the Tc of the final liquid sample, the top of the crystal, and the bottom of the crystal. Assuming perfect mixing of the liquid and negligible diffusion into the solid, the following relationship holds for each sample pair (solid-liquid pair): s ko = keff = − I Here, ko(keff) represents the partition coefficient of Li2Q, and C3 and Ct represent the concentration in solid and liquid, respectively.

The results for several single crystal growth experiments are shown graphically in FIG. 2 and in summarized form in Table H. The graph in FIG. 2 shows the distribution coefficient between crystal and melt as a function of melt composition. Table 1 shows the results for several samples with initial compositions ranging from 47.00 to 49.00 mole percent. These results from this experiment indicate that the harmonic composition of LiNbO3 (Ko = 1.000
) has a molar percentage of Li2Q of 48.4 to 48.5.
It shows that it is within the range of .

00 Ch (Oto+-+Otoaudience 2 -4v-s ws 6 0 0 孝拉 子 連 連 Ｇ Ｇ Ｇ Ｇ Ｇ Ｇ Ｒ Ｇ Ｇ Ｇ Ｇ Ｒ Ｇ Ｇ Ｇ Ｇ Ｇ ! Scroll 48, 45 for the purpose of experimentally confirming the dimensionally harmonic melt composition molar percentage L
An initial melt containing i 2 O was prepared and crystals were grown. In this growth, the fraction of liquid crystallized was 0.72. DTA heating curves were taken for the initial melt, final melt, crystal top, and crystal bottom. This data is as summarized in table 111. It can be seen that the Tc for these four different samples are equal within ±2° C. when corrected with the Au calibration.

Thus, the harmonic composition is 48.45 mole percent LI2O
It was confirmed that this Tc was 1137°C.

Table ■ 48.45 mole percentage Li2O Initial melt 1138 Final melt 1137 Upper crystal 1140 Lower crystal 1137 Tc of Li-rich LiNb0 exists where the solidus line is below Tc, and the enthalpy of fusion is smaller than this Tc thermal change. It is difficult to measure by DTA because it hides the

To determine the composition at the interface of LiNbO3 and Li5Nb04, the increase in weight measured upon complete mineralization is added to the initial crystal composition measured by DTA in the process of converting to molar percentage LiO. Ta.

These measurement results are for LiNbO3 and Li3NBO
The lithium niobate composition at the interface with 4 is 1060℃
49.96±0.03 molar percentage Li 2O.
0.03 mole percentage of 49°89 at 1100°C,
It was shown that the mole percentage of Li 2O was 49.81±003 at 1150°C.

Since the crystal plates produced from single crystal balls grown from the same melt composition have the same single composition, they are
It is particularly useful for manufacturing ridge device elements. It is particularly useful as a substrate for optical integrated circuits where the same process is used to manufacture a large number of devices.

A typical procedure for manufacturing optical waveguide circuits and the like is as follows. Optical waveguides are fabricated by diffusing titanium into a lithium niobate substrate. The titanium pattern is produced by a conventional photosensitive one-layer process. As shown in FIG. 3(a), a layer of photoresist is formed over the lithium niobate.

A pattern of photoresist is created by drilling holes where the waveguides are desired using photoresist technology. (See Figure 3(b)).

Titanium is then deposited to cover the entire surface of the substrate (FIG. 3(C)). When this photosensitive corrosion resistant film is removed, a pattern of metallic titanium is left on the surface (FIG. 3(d)). A diffusion process is performed to diffuse metallic titanium into a lithium niobate substrate to fabricate optical integrated circuits (FIG. 3(e)). Preferably, this diffusion process is carried out in a closed container so that the composition and homogeneity of the lithium niobate is not compromised during the diffusion of titanium, and the lithium niobate has the same lithium activity as the lithium niobate substrate. Placed near the powder bed held. Typically, the degree to which this titanium diffusion process is performed is determined by the requirements of the titanium diffusion process, as the powder bed maintains the composition and homogeneity of the lithium niobate substrate. Diffusion is typically carried out at temperatures ranging from 1050 to 1100°C for 1 to 10 hours.

A typical optical integrated circuit 40 is shown in the 41st column I. The optical integrated circuit 40 shown here is a directional dual-balancer. The device is fabricated from a substrate 41 of melt-grown lithium niobate according to the present invention. Optical waveguide section 42
and 43 are formed by the diffusion process described above. This optical waveguide section j extends through this substrate to the other end.

Electron cables 44 and 45 are used to make connections between optical waveguides.
is provided. In one mode of operation, i[
When no voltage is applied to the pole, it exits the waveguide 42. When a suitable potential is applied to electrode 45 through terminal 46 and resistor 47, the light remains and exits waveguide 43. Various other integrated optical circuits can be manufactured in accordance with the present invention. The main advantages of the substrate produced according to the invention are that it always has the same composition and properties (e.g. optical properties) and that the process can be adjusted to optimize the performance of any given device. be.

[Brief explanation of drawings]

Figure 1 shows a reservoir device for pulling crystals from cerebrospinal fluid; Figure 2 shows in graphical form the partition coefficient between crystals and melt as a function of melt composition; Figure 3 4 shows a diagram of a structure used in the manufacture of an optical integrated circuit; and FIG. 4 shows an optical integrated circuit having a lithium niobate substrate manufactured according to the present invention. [Explanation of symbols of main parts] 13. ...Lithium niobate melt 14...
...Lithium niobate crystal 41...Lithium niobate substrate Applicant: American Telephone
And Telegraph Company FIG, / FIG, 2 (Sire) ← St. Li2O FIG, 3

Claims (1)

[Claims] 1. In a process for manufacturing a device comprising at least one lithium niobate crystal, the lithium niobate is manufactured by solidifying the crystal from a melt, and the melt is A process consisting essentially of molten lithium niobate, characterized in that the molten lithium niobate consists of 48.45±0.08 mole percent Li_2O and the balance Nb_2O_5. 2. In the process described in claim 1,
A process characterized in that the molten lithium niobate consists of 48.45±0.04 mole percent Li_2O and the balance Nb_2O_5. 3. In the process described in claim 2,
A process characterized in that the molten lithium niobate consists of 48.45±0.02 mole percent Li_2O and the balance Nb_2O_5. 4. In the process described in claim 3,
A process characterized in that the crystals are pulled from the melt. 5. In the process described in claim 1,
A process characterized in that the crystals are pulled from the melt. 6. In a device composed of lithium niobate crystals, the lithium niobate crystals are produced by solidifying the crystals from a melt, the melt consists essentially of molten lithium niobate, and the lithium niobate crystals are made by solidifying the crystals from a molten liquid; Li_2O with 48.45±0.08 mole percentage of lithium niobate
and the remaining Nb_2O_5. 7. In the device according to claim 6,
A device characterized in that the crystals are pulled from the melt.