JP6578995B2 - Waveguide type optical device - Google Patents

Description

  The present invention relates to a method for driving a waveguide type optical element including an optical waveguide and an electrode for controlling a light wave propagating through the optical waveguide, and more particularly, a waveguide including a bias electrode for compensating for so-called drift. The present invention relates to a type optical element.

In the fields of optical communication and optical measurement, Mach-Zehnder type light using a ferroelectric crystal lithium niobate (LiNbO ₃ ) (also referred to as “LN”) as a substrate as a waveguide type optical element such as an optical modulator. Modulators are widely used.

  The Mach-Zehnder type optical modulator includes an incident waveguide for introducing light from the outside, a branching unit for propagating the light introduced by the incident waveguide in two paths, and a branching unit after the branching unit. Mach-Zehnder type optical waveguide composed of two parallel waveguides for propagating each of the generated light and an output waveguide for combining the light propagated through the two parallel waveguides and outputting them to the outside Is provided.

  The Mach-Zehnder type optical modulator includes an electrode for changing and controlling the phase of the light wave propagating in the parallel waveguide using the electro-optic effect. In general, the electrodes include an RF (high frequency) signal electrode (hereinafter referred to as an “RF electrode”) formed on or near the parallel waveguide and a ground electrode that is spaced apart from the RF electrode. It is configured.

  Further, in a Mach-Zehnder type optical modulator using LN as a substrate, in order to prevent changes in modulation characteristics due to drift phenomena such as so-called DC drift and temperature drift, in general, in addition to the RF electrode, a parallel waveguide is used. A bias electrode is formed along the line, and a voltage difference is generated between the parallel waveguides by applying a voltage to the bias electrode, thereby compensating for the voltage shift amount due to the drift phenomenon (Patent Document 1).

  Since the bias electrode for drift compensation is formed alongside the RF electrode along the length direction of the parallel waveguide, the RF electrode is lengthened to reduce the voltage required for the modulation operation, and the half-wave voltage is reduced. When trying to reduce (Vπ), the length of the bias electrode is shortened, and a higher voltage may have to be applied to the bias electrode.

  As a result, a phenomenon (acceleration of DC drift) in which an increase in the amount of DC drift easily proceeds due to application of a high electric field to the LN substrate via the bias electrode may occur.

JP-A-5-224163

  From the above background, in a waveguide type optical element having a bias electrode, it is desired to effectively prevent acceleration of a DC drift phenomenon that occurs due to application of a high electric field to the substrate via the bias electrode.

One aspect of the present invention is a waveguide-type optical element, which propagates through the optical waveguide, a substrate having an electro-optic effect, two optical waveguides formed on the substrate, and the like. An RF electrode for controlling a light wave; and a DC bias electrode for applying a DC voltage. The RF electrode and the DC bias electrode require a voltage required for the DC bias electrode during operation. A non-conductive layer made of a non-conductive material is formed on the substrate, and a charge dispersion layer made of a conductive material is formed on the non-conductive layer. Each of the RF electrode and the bias electrode includes a base layer formed on the charge distribution layer and an upper layer formed on the base layer, and the base layer of the bias electrode includes the RF electrode. Before Underlayer thickness is formed thicker than, and, before Symbol nonconductive layer comprises silicon dioxide (SiO2), the charge distribution layer is silicon, the base layer comprises titanium, the upper layer is Au including.
According to another aspect of the present invention, the DC bias electrode is configured such that an electric field to be generated in the substrate by the DC bias electrode during operation is in a range of 1 V / μm to 2 V / μm.
According to another aspect of the invention, the thickness of the underlayer of the bias electrode is 50 nm or more.

It is a figure which shows the structure of the waveguide type optical element which concerns on one Embodiment of this invention. It is an AA cross-sectional arrow view of the waveguide type optical element shown in FIG. FIG. 2 is a cross-sectional view taken along the line BB of the waveguide type optical element shown in FIG. 1. It is CC sectional view taken on the line of the waveguide type optical element shown in FIG. It is a figure which shows an example of the experimental result regarding the dependence to the electrode base layer material of DC drift amount in the case of using a charge dispersion film. It is a figure which shows an example of the experimental result regarding the dependence to the electrode base layer material of DC drift amount when not using a charge dispersion film. It is sectional drawing of the waveguide type optical element used for the experiment of FIG.

  Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a diagram showing a configuration of a waveguide optical device according to an embodiment of the present invention. 2 is a cross-sectional arrow view of the waveguide type optical element AA shown in FIG. 1, FIG. 3 is a cross-sectional arrow view of the BB, and FIG. 4 is a cross-sectional arrow view of the CC.
The waveguide optical device 10 is a Mach-Zehnder optical modulator in which a Mach-Zehnder (MZ) optical waveguide 102 is formed on the surface of a substrate 100.

  The substrate 100 is a substrate made of lithium niobate (LN), which is an electro-optic material, and is, for example, a Z-cut LN substrate. The MZ type optical waveguide 102 formed on the surface of the substrate 100 has parallel waveguides 104 and 106.

  On the surface of the substrate 100, a non-conductive layer 120 made of a non-conductive material and a charge distribution layer 122 for eliminating the localization of space charges in the non-conductive layer 120 are formed ( (See FIG. 2 etc.)

The non-conductive layer 120 is, for example, a so-called buffer layer provided for the purpose of avoiding light loss caused by absorption of a light wave propagating through the MZ type optical waveguide 102 by an electrode 108 or the like described later. For example, it is made of a material having a dielectric constant lower than that of the substrate 100 (for example, silicon dioxide (SiO ₂ )). The charge distribution layer 122 may be a silicon (Si) layer, for example.

  On the charge distribution layer 122 immediately above the parallel waveguides 104 and 106, radio frequency (RF) electrodes 108 and 110 are formed along the parallel waveguides 104 and 106, respectively (FIG. 2). The ground electrodes 112, 114, and 116 are formed so as to sandwich the RF electrodes 108 and 110 with a predetermined distance from each of the RF electrodes 108 and 110. High-frequency signals for controlling light waves propagating through the parallel waveguides 104 and 106 are applied between the RF electrode 108 and the ground electrodes 112 and 116 and between the RF electrode 110 and the ground electrodes 112 and 114, respectively. The By these high frequency signals, for example, light input from the left end of the MZ type optical waveguide 102 is modulated (for example, intensity modulated) and output from the right end of the figure.

  A bias electrode 150 for compensating for the drift phenomenon is also formed on the charge dispersion layer 122 (FIG. 3). The bias electrode 150 is disposed immediately above the parallel waveguides 104 and 106 at a predetermined distance from the operation electrodes 152 and 154 formed along the parallel waveguides 104 and 106, and the operation electrodes 152 and 154, respectively. The reference electrodes 160, 162, and 164 are provided so as to sandwich the working electrodes 152 and 154 at a distance.

  A reference potential is applied to the reference electrodes 160, 162, and 164, and a positive voltage or a negative voltage with respect to the reference potential is applied to the working electrodes 152 and 154. As a result, a refractive index difference is generated between the parallel waveguides 104 and 106, and a voltage shift amount (that is, a voltage shift amount necessary for an optical modulation operation by the RF electrodes 108 and 110) caused by the drift phenomenon described above is generated. Compensated.

  Each of the electrodes 108, 110, 112, 114, 116, 152, 154, 160, 162, and 164 described above is formed on the charge dispersion layer 122. As shown below, the charge dispersion layer The base layer is formed on the base layer 122 and the upper layer formed on the base layer. In particular, in the waveguide type optical element 10 according to the present embodiment, the RF electrodes 108 and 110 and the ground electrodes 112 to 116 to which a high-frequency voltage for light modulation is applied, and a high voltage for drift compensation are applied. The thickness of the underlying layer is different from that of the bias electrode 150.

  In FIG. 2, each of the RF electrodes 108 and 110 and the ground electrodes 112, 114, and 116 is formed on the underlying layers 208 a, 210 a, 212 a, 214 a, and 216 a formed on the charge distribution layer 122, and on these underlying layers, respectively. The upper layers 208b, 210b, 212b, 214b, and 216b are formed respectively. The underlying layers 208a, 210a, 212a, 214a, 216a are, for example, titanium (Ti), and the upper layers 208b, 210b, 212b, 214b, 216b are, for example, gold (Au).

  Similarly, in FIG. 3, each of the working electrodes 152 and 154 and the reference electrodes 160, 162 and 164 constituting the bias electrode 150 is an underlying layer 252 a, 254 a, 260 a, 262 a, 264 a formed on the charge distribution layer 122. And upper layers 252b, 254b, 260b, 262b and 264b respectively formed on these base layers. The underlying layers 208a, 210a, 212a, 214a, 216a are, for example, titanium (Ti), and the upper layers 208b, 210b, 212b, 214b, 216b are, for example, gold (Au).

  As described above, in the waveguide type optical element 10 of the present embodiment, the RF electrodes 108 and 110 and the ground electrodes 112 to 116 to which a high-frequency voltage for light modulation is applied, and a high voltage for drift compensation are used. The thickness of the underlayer differs depending on the bias electrode 150 to be applied. For example, the base layers 208a, 210a, 212a, 214a, and 216a of the RF electrodes 108 and 110 and the ground electrodes 112 to 116 depend on the high frequency band applied to the RF electrodes 108 and 110, for example, the high frequency is 25 Gbaud. In this case, the underlying layer 252a, 254a, 260a, 262a, 264a of the bias electrode 150 is, for example, dependent on the electric field applied to the parallel waveguides 104, 106 by the bias electrode 150, for example, When the electric field is 1 to 2 V / μm, for example, it is 50 μm.

  According to the knowledge of the inventors of the present invention, when the electric field induced in the substrate 100 by the bias electrode 150 is high, the increasing tendency (increase rate) of the DC drift induced in the MZ type optical waveguide 102 is as follows. Depending on the thickness of the base layers 252a, 254a, 260a, 262a, 264a (hereinafter collectively referred to as “underlayer 252a etc.”), when the thickness of the underlayer 252a etc. falls below a certain threshold, Increased acceleration is significantly increased.

FIG. 5 is a diagram showing the relationship between the thickness of the underlayer obtained by experiment and the DC drift amount (voltage shift amount) per 24 hours. The experiment was performed using a Z-cut LN substrate in the same manner as the optical element 10 described above. When a Z-cut LN substrate is used, the problem of temperature drift due to charges generated during temperature changes due to the pyroelectric effect is significant. Therefore, a countermeasure for forming the charge dispersion layer 122 on the non-conductive layer 120 I took. A semiconductive material mainly composed of silicon Si was used as the charge dispersion film, and gold (Au) having excellent conductivity and chemical stability was used as the upper electrode material. The optical waveguide was formed by a Ti diffusion method. The non-conductive layer 120, using the SiO ₂ doped with _In 2 _{O 3} and TiO _2. The concentrations of In ₂ O ₃ and TiO ₂ are approximately 5 mol% and 6 mol%, respectively, and the film thickness of the nonconductive layer 120 is 0.55 μm. Further, a Si film was formed as a charge dispersion film on the buffer layer by sputtering. Ti and Ni having excellent adhesion are used for the underlayers 252a, 254a, 260a, and 264a. The upper layer of the electrode was gold (Au) and formed by electroplating. The positional relationship between the bias electrode and the optical waveguide is the same as in FIG. The distance between the electrode 152 and the ground electrodes 160, 162, 164 is 15 μm.

DC drift is measured using a bias tracking method that maintains a constant bias state (H.Nagata, K.Kiuchi, S.Shimotsu, J.Ogiwara, and J.Minowa, "Estimation of direct current bias and drift of Ti: LiNbO ₃ optical modulators ", J. Appl. Phys., vol. 76, no. 3, pp. 1405-1408 (1994)). The initial applied voltage is ± 10 V, and the measurement environment temperature is 85 ° C. The initial applied voltage to the bias electrode is 10 V, and the difference between the bias applied voltage after 24 hours and the initial applied voltage is shown as the amount of bias fluctuation.

  From the results shown in FIG. 5, the DC drift amount depends on the Ti film thickness of the underlayers 252a, 254a, 260a, and 264a both when Ti is used and when Ni is used. The DC drift amount tends to be small. It can be seen that the effect of reducing the DC drift amount is almost saturated at a thickness of 50 nm or more. Although the results are not shown, the same tendency can be seen when Cr, Al, TaN, TiN or the like is used as the underlayer.

This phenomenon can be explained as follows.
The DC drift is caused by the generation and movement of space charges in the nonconductive layer 120, the substrate 100, or the optical waveguide 104. The acceleration factor of the DC drift amount is attributed to carrier injection into the non-conductive layer 120, the substrate 100, and the optical waveguides 104, 106, an external electric field, stress / strain, etc., in addition to an increase in ambient temperature and an increase in incident light intensity. The generation of charge carriers can be considered. Carrier injection is a phenomenon that depends on the material of the electrode underlayer, but in our test results, the dependence of the DC drift fluctuation amount on the work function of the electrode underlayer material is unclear. On the other hand, a strong dependence on the film thickness of the underlayer was observed. In any underlayer material such as Ti, Cr, Ni, etc., those in which an underlayer having a large underlayer thickness was formed tended to have a small amount of bias fluctuation. Furthermore, after the electrodes were short-circuited and held at a high temperature in order to diffuse the space charge unevenly distributed and fixed on the non-conductive layer 120, the substrate 100, and / or the optical waveguide 104, the evaluation was performed again by the DC drift bias tracking method. A device having an underlayer film thickness of 50 nm or more showed almost the same characteristics as the first test result, but a device having a film thickness of less than 50 nm tended to be worse than the first test result.

  When we analyzed a device in which the DC drift fluctuation amount exceeded 15 V (that is, a device in which the applied voltage reached 25 V or more), a heterogeneous phase that appeared to be gold silicide was detected in part of the lower portion of the bias electrode. The shape is a granular or columnar shape of submicron to 2 microns, part of which is needle-shaped, and the generation location is limited to the lower part of the bias electrode. In the case where an underlayer having a film thickness of 50 nm or more is formed, there is no detection of a different phase of gold silicide, but the amount of the different phase is much smaller than that of a thickness less than 50 nm, and the particle size of the different phase is also small. I understood it. From these results, it is highly probable that the occurrence of heterogeneous phases of gold silicide induces acceleration of DC drift, and if the film thickness of the underlayer is 50 nm or more, it becomes a barrier layer for generating gold silicide. It can be estimated that the practical effect of the above is obtained.

  In some devices where the amount of bias drift exceeds 15 V, a large number of micro parts (micro domains) in which the polarization direction is inverted are observed in the optical waveguide part under the electrode to which the voltage is applied. Silicides are harder than metals and have toughness, so that they are used as conductive probes. From the observation examples, it is estimated that there are cases where gold silicide is generated, and bias drift is accelerated due to local dielectric breakdown or precursor injection of carriers in the buffer layer due to the spike effect of gold silicide.

  Even if the device has an underlayer thickness of less than 50 nm, gold silicide is not detected at all before the test or when the applied voltage is approximately 20 V or less. Much less than From this, it can be determined that Au silicide is generated and grown by applying a bias voltage. In addition, a case (see Japanese Patent Application Laid-Open No. 1995-084228) in which gold silicide greatly stretches and grows to cause a short circuit between the electrodes was not found. This is presumably because water or water vapor that promotes the silicidation solid-phase reaction is excluded from the housing of the device.

In this test using our LiNbO ₃ crystal for the substrate 100, the dependence of the DC drift amount on the material of the underlayer of the electrode is unclear, and the possibility is pointed out in Japanese Patent Application Laid-Open No. 2011-118438. The occurrence of the electron injection phenomenon was not confirmed. This is presumably because the work function of Si is around 4. Ti, Ni, Cr, Al. Since the work function of metals such as TiN and TaN is 4.1 or higher, even when the base layer is used, the influence of the work function of the base layer metal upon carrier injection into the non-conductive layer, the substrate 100 or the optical waveguide 104 is small. it is conceivable that.

  In other words, when a material containing Si as a main component is used as the charge dispersion layer, there is a work material such as Ti or Cr that is a metal material that is concerned about electron injection into the dielectric crystal in the underlayer of the electrode. A metal material of less than 5 eV can be used. However, it is necessary to prevent the generation of gold silicide particles, and the film thickness must be 50 nm or more. Any metal such as Ti, Ni, Cr, etc., can suppress the generation of gold silicide at a thickness of about 50 nm because it is in a substantially continuous film state with a film thickness of about 50 nm, and a barrier layer for the solid phase reaction between Si and Au. It is estimated that In any film of Ti, Ni, and Cr having a thickness of about 50 nm or 75 nm, pinholes are detected when the transmitted light test is performed, so it cannot be said to be a complete continuous film. However, since the film thickness cannot be measured by the ellipsometry method with a film thickness of about 30 nm, it is considered that the film is in a continuous film state that hardly transmits light.

  Therefore, by increasing the thickness of the base layer 252a and the like in the bias electrode 150 to be larger than a predetermined thickness depending on the intensity of the electric field induced in the substrate 100 by the bias electrode 150, the acceleration of the increase in the DC bias is increased. The rise can be suppressed.

  On the other hand, with respect to the RF electrodes 108 and 110, the voltage applied to these RF electrodes 108 and 110 is generally much smaller than the voltage applied to the bias electrode 150, and the RF electrodes 108 and 110 allow the substrate 100 to be used. The amount of increase in DC drift due to the electric field induced therein is negligible as compared with the DC drift due to the electric field induced in the substrate 100 by the bias electrode 150. Further, from the viewpoint of the high frequency applied to the RF electrodes 108 and 110, metals such as Ti, Ni, and Cr that are suitable as the underlayer are metals that have a lower conductivity than AU, Ag, Cu, and the like. The thinner the thickness of 252a, the better.

  In order to realize efficient operation and wideband operation, it is essential to reduce the attenuation of the millimeter waveband signal at the RF electrode, and so-called skin effect must be taken into consideration. For example, in an electrode using a good conductor Au, the skin depths at 30 GHz, 50 GHz, and 100 GHz are 437 nm, 338 nm, and 239 nm, respectively, and most of the current flows through shallower portions (portions close to the skin). Therefore, the electrode loss of only about 50 nm from the skin is simply replaced by the good conductor Au from Ti, Ni, Cr, etc., the conductor loss increases, and the characteristic deterioration becomes remarkable. It is necessary to keep the thickness of the base film below one-tenth or less of the skin depth in order to secure the operation characteristics in the millimeter wave band. For example, in a modulator that operates in a millimeter wave band of 30 GHz or more, it is desirable that the thickness of the base film of the control electrode portion be 40 nm or less and the base electrode of the bias electrode portion be 50 nm or more.

  For this reason, the configuration of the optical waveguide element 10 for suppressing the acceleration phenomenon of DC drift accompanying the drift compensation operation while suppressing the amplitude voltage of the high-frequency signal required for the optical modulation operation is small, such as the base layer 252a and the upper layer 252b. The base layer 252a and the like in the bias electrode 150 are at least as thick as the base layer 208a and the like in the RF electrodes 108 and 110, depending on the type of metal used for the RF electrode 108 and the like, and the band of the high-frequency signal applied to the RF electrode 108 and the like. It is effective to increase the thickness.

  As a more specific example, as a specific condition in a general optical waveguide device that is put to practical use, for example, when the electric field induced in the substrate 100 by the bias electrode 150 is 1 to 2 V / μm In this case, when the base layer 252a or the like is made of Ti and the upper layer 252b or the like is made of Au, by increasing the thickness of the base layer 252a or the like to 50 nm or more, an increase in acceleration due to an increase in DC bias can be effectively suppressed.

  As described above, in the optical waveguide device 10 according to this embodiment, the thickness of the underlayer 108a and the like that respectively constitute the RF electrodes 108 and 110 and the corresponding ground electrodes 112 to 116 is below the bias electrode 150. The stratum 252a and the like are formed thick. Thereby, in the optical waveguide device 10, it is possible to effectively suppress an increase in an increase in DC drift due to application of a high electric field accompanying drift compensation while reducing the high-frequency signal amplitude voltage necessary for the optical modulation operation. .

  In the above-described embodiment, the Z-cut LN substrate 100 is used as an example. However, the present invention is not limited to this, and an X-cut LN substrate may be used.

  Note that when the charge dispersion film 122 is made of a material other than Si or Si-based material, or when the charge dispersion film 122 is not formed, the dependency of the DC drift amount on the base material of the electrode becomes clear. Care must be taken in selecting the electrode base material. FIG. 6 shows the material dependence of the electrode underlayer of the DC drift amount in the device configured not to form the charge dispersion film shown in FIG.

In the device shown in FIG. 7, X-cut LiNbO ₃ was used as the substrate 700 as the substrate 700 in order to eliminate as much as possible the influence of temperature drift due to pyroelectricity. The optical waveguides 702 and 704 were formed by a Ti diffusion method. As the non-conductive layer 710, a SiO ₂ film was formed by a vacuum deposition method. The film thickness of the nonconductive layer 710 is 0.55 μm, and no charge dispersion film is formed on the nonconductive layer 710. The set film thickness of each of the underlying layers 722a, 732a, and 742a of the bias electrodes 720, 730, and 740 is 50 nm. Gold (Au) of the upper layers 722b, 732b, and 742b of the bias electrodes 720, 730, and 740 was formed by an electroplating method. The distance d70 between the signal electrode 730 and the ground electrode 720 shown in FIG. 7 is 15 μm, the distance d72 between the signal electrode 730 and the ground electrode 740 is 30 μm, and the electric field strength to each waveguide is not equal. . The method for measuring the DC drift amount is the same as described above.

From FIG.
-There is a difference in the DC drift phenomenon between materials with a work function of 5 eV or more and materials with less than 5 eV.
-The amount of DC drift is smaller when a material having a work function of 5 eV or more is used,
-When a material having a work function of less than 5 eV is used, the amount of DC drift when a negative voltage is applied is large.
It can be understood that carrier injection phenomenon, probably electron injection, is occurring. In addition, since the thickness of the underlayers 722a, 732a, and 742a is 50 nm and is a substantially continuous film, the influence of Au that is a material of the upper electrodes 722b, 732b, and 742b having a work function of about 4.9. Is considered small.

  In addition, here, the underlayers 722a, 732a, and 742a were prototyped by setting the film thickness to 50 nm, which is a substantially continuous film. However, a material having a work function of 5 eV or more such as Ni, Ge, TiN, and TaN is used as the underlayer. When used, it is considered that acceleration of the drift phenomenon can be effectively suppressed to some extent even if it is formed in an island shape instead of a continuous film. However, since Ni, Ge, TiN, TaN, and the like are relatively hard conductive materials, if they are only partially formed, it is considered that the same phenomenon as the above-described spike effect due to gold silicide is likely to occur. Drift characteristics may be deteriorated. At the initial stage of sputter deposition of Ni, Ge, TiN, and TaN, the crystal grows in a columnar crystal form with a wide gap (that is, as a discontinuous film), and the covering ratio of the substrate is about 50% with a film thickness of about 20 nm. In order to avoid the spike effect, it is desirable that the thickness of the underlayer be 20 nm or more.

  In the above-described embodiment, Au is used for the upper layers 208b, 210b, 212b, 214b, and 216b of the bias electrode 10. However, the bias electrode does not need characteristics as a high-frequency line. For this reason, the upper layer of the bias electrode does not need to be made of a material having a particularly small electric resistance, and the thickness and the cross-sectional area of the upper layer are highly flexible. Needless to say, materials such as Cr, Ni, Al, Cu, Ag, Pt, Ir, Ge, Ti, Ta, TiN, and TaN may be used as the material of the upper layer. Alternatively, the upper layer may be formed using a material having a work function of 5 eV or more without forming the base layer. However, considering the effects of electrical resistance, chemical stability of electrode materials, ease of formation, ease of wiring such as wire bonding, and stress strain on the substrate, gold Au and its alloys are comprehensive. Is a suitable material for the upper layer.

In addition, the DC bias electric field strength and the environmental temperature that clarify the dependence of the DC drift characteristics on the work function of the underlayer and the polarity of the applied voltage are determined by the materials and compositions of the charge dispersion layer and the non-conductive layer, Depends on film thickness, film formation method, etc. When a Si ₃ N _4-x- based material is used for the charge dispersion film 122, the dependency on the base layer material is relatively clear even if the DC bias electric field is small, but the charge dispersion film 122 is also non-conductive layer 120. If they are not formed together, their dependence is not clear. However, in our experience, a tendency similar to that in the case where the above-described charge dispersion film 122 is not formed is recognized. Therefore, the selection of the underlayer material is considered to have a certain effectiveness in improving the DC drift characteristics.

  It is also known that the difference between the bias voltage polarities of the DC drift amount depends on the material composition and film forming method of the non-conductive layer 120 and the charge dispersion film 120 and the crystal orientation of the substrate 100. Therefore, when the DC bias is adjusted by differential driving or when a bipolar DC bias is applied, in addition to the selection of the electrode base layer, the material composition and the film formation method of the non-conductive layer 120 and the charge dispersion film 122 Also, it is necessary to consider the crystal orientation of the substrate 100.

  DESCRIPTION OF SYMBOLS 10 ... Waveguide type optical element, 100 ... Substrate, 102 ... MZ type optical waveguide, 104, 106 ... Parallel waveguide, 108, 110 ... RF electrode, 112, 114, 116 ..Ground electrode, 120 ... non-conductive layer, 122 ... charge dispersion layer, 150 ... bias electrode, 152, 154 ... operating electrode, 160, 162, 164 ... reference electrode, 252a, 254a, 260a, 262a, 264a... Underlayer, 252b, 254b, 260b, 262b, 264b... Upper layer.

Claims (3)

A substrate having an electro-optic effect;
Two optical waveguides formed on the substrate;
An RF electrode for controlling a light wave propagating through the optical waveguide;
A DC bias electrode for applying a DC voltage;
With
The RF electrode and the DC bias electrode are configured such that a voltage required for the DC bias electrode during operation is higher than a voltage required for the RF electrode,
A non-conductive layer made of a non-conductive material is formed on the substrate, and a charge dispersion layer made of a conductive material is formed on the non-conductive layer,
Each of the RF electrode and the bias electrode is composed of a base layer formed on the charge distribution layer and an upper layer formed on the base layer,
The underlayer of the bias electrode is formed thicker than the underlayer of the RF electrode , and
The non-conductive layer includes silicon dioxide (SiO 2), the charge distribution layer is silicon, the base layer includes titanium, and the upper layer includes Au.
Waveguide type optical element. The DC bias electrode is configured such that an electric field to be generated in the substrate by the DC bias electrode during operation is in a range of 1 V / μm or more and 2 V / μm or less.
The waveguide type optical device according to claim 1. A thickness of the underlayer of the bias electrode is 50 nm or more;
The waveguide type optical device according to claim 1 .

Images