JP7822147B2 - Electroabsorption Modulator - Google Patents

Description

The present invention relates to electroabsorption modulators, particularly to the shape of electrodes.

Optical modulation technology is used in many applications, including optical fiber communications, free-space optical communications, and information processing using optical pulses, creating a demand for compact optical modulators with high-speed response. One such optical modulator is the electro-absorption modulator (EA (Electro-Absorption) modulator), which utilizes the electro-absorption effect of semiconductors. EA modulators have a p-i-n structure with a multi-quantum well layer as the i-layer, and light intensity modulation can be achieved by applying a voltage to the p-i-n structure. EA modulators are known to have a structure based on an InP substrate with a multi-quantum well layer made of InGaAsP or similar. Other known structures use a Si base with an i-layer made of, for example, SiGe. EA modulators are also often used as semiconductor optical devices integrated with semiconductor lasers.

Patent Document 1 discloses a structure in which the p-side electrode of an EA modulator is composed of a mesa-top electrode along the mesa, a pad electrode to which an externally connected wire is bonded, and an extraction electrode that connects the pad electrode and the mesa-top electrode. Patent Document 2 discloses that the carriers (photocurrent) generated when the EA modulator absorbs light are distributed in a direction along the mesa of the EA modulator. Furthermore, a structure is shown in which the connection between the extraction electrode and the mesa-top electrode is located on the light input end side to improve heat dissipation during this process.

JP 2014-145973 A JP 2009-222965 A

T. Ido, S. Tanaka, M. Suzuki, M. Koizumi, H. Sano, H. Inoue, "Ultra-high-speed multiple-quantum-well electro-absorption optical modulators with integrated waveguides," Journal of Lightwave Technology, vol. 14, no. 9, pp. 2026-2034, Sept. 1996. R. Sahara, K. Morito, K. Sato, Y. Kotaki, H. Soda, N. Okazaki, "Strongly improved frequency response at high-optical input powers from InGaAsP compensated strain MQW electroabsorption modulators," IEEE Photonics Technology Letters, vol. 7, no. 9, pp. 1004-1006, Sept. 1995.

Non-Patent Document 1 reports that the primary cause of the reduced response speed of EA modulators is the modulator's capacitance. Furthermore, Non-Patent Document 1 reports that shortening the modulator length improves response speed. On the other hand, shortening the modulator length reduces the total amount of light absorption, lowering the extinction ratio, which can contribute to an increase in the communication error rate in optical transmission. Therefore, achieving a high extinction ratio in an EA modulator with a short modulator length requires increasing the amount of light absorption per unit length. However, increasing the amount of light absorption leads to an increase in the number of carriers (photoexcited carriers) generated in the EA modulator. While this would not be a problem if the generated photoexcited carriers were smoothly extracted from the EA modulator, there are cases where they are not sufficiently extracted. In such cases, the carriers remain within the EA modulator, particularly in the multiple quantum well layer. This results in carrier pileup in the multiple quantum well layer. Carrier pileup leads to a decrease in the EA modulator's characteristics, such as its extinction characteristics and response speed. Non-patent document 2 shows that the extinction characteristics and response speed decrease depending on the input light intensity and light absorption amount to the EA modulator.

As disclosed in Patent Document 2, the photocurrent is larger on the optical input side of an EA modulator than on the optical output side. Assuming that a substantially constant voltage is applied in the optical axis direction of the EA modulator, the photocurrent distribution can be said to indicate the tendency of the distribution of photoexcited carriers. In other words, this indicates that there are more carriers that must be extracted from the multiple quantum wells on the optical input side than on the optical output side. Even if carriers are extracted smoothly when the voltage applied to the EA modulator is DC or low-speed (less than kHz), carriers may not be extracted when an AC voltage of GHz or higher is applied. If carriers are not extracted quickly, the extinction characteristics deteriorate, and frequency characteristics, such as bandwidth, decrease. This effect is particularly noticeable when the amount of light absorption per unit length is large, such as in short modulators or at high temperatures where light absorption increases.

The present invention was made in consideration of these issues, and aims to provide an electroabsorption modulator with excellent high-speed response.

(1) The electroabsorption modulator according to the present disclosure comprises a substrate, a mesa structure provided on a first surface of the substrate, the mesa structure including a first conductivity type cladding layer, a multiple quantum well layer, and a second conductivity type cladding layer, a first conductivity type electrode electrically connected to the first conductivity type cladding layer, and a second conductivity type electrode provided on a second surface of the substrate, the first conductivity type electrode having a mesa electrode arranged along the extension direction of the mesa structure, a pad electrode to which an external electrical signal is input, and a lead electrode connecting the mesa electrode and the pad electrode, the mesa structure comprising an optical input end to which light is input from the outside and an optical output end opposite the optical input end, the connection position between the center position of the lead electrode in the short direction and the mesa electrode being close to the optical output end in the longitudinal direction of the mesa electrode, and the connection position being less than 50% from the optical output end side with respect to the longitudinal length of the mesa electrode.

The present invention provides an electroabsorption modulator with excellent high-speed response.

1 is a top view of an electroabsorption modulator integrated laser according to a first embodiment of the present invention; 2 is a cross-sectional view taken along line AA of the electroabsorption modulator shown in FIG. 1. 2 shows the distribution of photocurrent in the optical axis direction of the electroabsorption modulator shown in FIG. 1. 2 shows the results of a simulation of the electric field strength in the multiple quantum well layer of the electroabsorption modulator shown in FIG. 1. 2 shows the results of a simulation of the electric field strength in the multiple quantum well layer of the electroabsorption modulator shown in FIG. 1. The electric field intensity ratio and photocurrent distribution at 10 GHz in the optical axis direction are shown. The electric field intensity ratio and photocurrent distribution at 20 GHz in the optical axis direction are shown. The electric field intensity ratio and photocurrent distribution at 40 GHz in the optical axis direction are shown. The electric field intensity ratio and photocurrent distribution at 60 GHz in the optical axis direction are shown. 10 is a simulation result showing the average electric field strength ratio for each length of the electroabsorption modulator. FIG. 10 is a top view of an electroabsorption modulator integrated laser according to a second embodiment of the present invention. FIG. 10 is a top view of an electroabsorption modulator integrated laser according to a third embodiment of the present invention. 10 is a cross-sectional view of the electroabsorption modulator shown in FIG. 9 taken along the line CC. FIG. 10 is a top view of an electroabsorption modulator integrated laser according to a fourth embodiment of the present invention.

Embodiments of the present invention will be described in detail below with reference to the drawings. Note that in all figures used to explain the embodiments, components having the same functions will be given the same reference numerals, and repeated explanations will be omitted. Note that the figures shown below are merely for explaining examples of the embodiments, and the size of the figures does not necessarily correspond to the scale described in these examples.

[First embodiment]
FIG. 1 is a top view of an electroabsorption modulator-integrated laser 1 according to a first embodiment of the present invention. The electroabsorption modulator-integrated laser 1 is a semiconductor integrated device in which an EA modulator 2 (electroabsorption modulator), a waveguide connection section 3, and a semiconductor laser 4 are integrated on a common substrate. The electroabsorption modulator-integrated laser 1 is an integrated device in which the semiconductor laser 4, the waveguide connection section 3, and the EA modulator 2 are optically connected to each other in this order. The semiconductor laser 4 emits continuous light, and the waveguide connection section 3 transmits the light emitted from the semiconductor laser 4 to the EA modulator 2. The EA modulator 2 includes a multiple quantum well layer 10 that absorbs light corresponding to the oscillation wavelength of the semiconductor laser 4. The continuous light that passes through the waveguide connection section 3 and enters the EA modulator 2 is intensity-modulated by the EA modulator 2 and converted into a modulated optical signal such as a binary or quaternary signal. The modulated optical signal output from the EA modulator 2 is output from a front end face 101. A dielectric non-reflective coating (not shown) is formed on the front end facet 101. A dielectric highly reflective coating (not shown) is formed on the rear end facet 102, which is the end face on the opposite side of the semiconductor laser 4. Note that although an example of an integrated type is shown in this embodiment, the effects of the present invention can be obtained even when the present invention is applied to a single EA modulator.

The semiconductor laser 4 is, for example, a DFB (Distributed Feedback) laser that oscillates in the 1.3 μm band. The oscillating wavelength band may be the 1.55 μm band or another wavelength band. Furthermore, the semiconductor laser 4 is not limited to a DFB laser; it may also be a FP (Fabry-Perot) laser, DBR (Distributed Bragg Reflector) laser, or DR (Distributed Reflector) laser. The semiconductor laser 4 is equipped with a semiconductor laser p-side electrode 5 (first conductivity type electrode) for conducting current. The semiconductor laser 4 oscillates continuous light by injecting current between the n-side electrode 14 (second conductivity type electrode) (described below) and the semiconductor laser p-side electrode 5.

The EA modulator 2 includes a p-side electrode 105 for the EA modulator. Figure 2 is a schematic diagram of the EA modulator 2 taken along the A-A cross section perpendicular to the optical axis. A mesa structure is provided on an n-type n-InP substrate 9, with semi-insulating InP burying layers 12 disposed on both sides of the mesa structure. The mesa structure is formed by stacking, from the bottom, a portion of the n-InP substrate 9, an i-type multiple quantum well layer 10, and a p-type InP cladding layer 11. These layers form a p-i-n structure. Although not shown, an n-type optical confinement layer may be disposed between the multiple quantum well layer 10 and the n-InP substrate 9. Similarly, a p-type optical confinement layer may be disposed between the upper side of the multiple quantum well layer 10 and the p-type InP cladding layer 11. Although not shown, a p-type contact layer is disposed on the p-type InP cladding layer 11. A passivation film 13 is disposed on the top surface of the burying layer 12. The p-side electrode 105 for the EA modulator is in contact with a p-type contact layer (not shown), allowing current to flow through the p-i-n structure. An n-side electrode 14 is disposed on the surface of the n-InP substrate 9 opposite the surface on which the mesa structure is formed. The n-side electrode 14 is connected to GND potential. As shown in FIG. 1 , the p-side electrode 105 for the EA modulator is composed of a mesa-top electrode 6 disposed along the mesa structure, a pad electrode 8 to which an external current-carrying means (e.g., a wire) is connected, and an extraction electrode 7 connecting the mesa-top electrode 6 and the pad electrode 8. An external electrical signal is input to the pad electrode 8, and the electrical signal is input to the mesa-top electrode 6 via the extraction electrode 7. The individual electrodes constituting the p-side electrode 105 for the EA modulator are integrally formed and all have the same configuration. The p-side electrode 105 for the EA modulator is composed of three layers: Ti/Pt/Au, from the p-type contact layer side. Here, the extraction electrode 7 is connected to the tip of the mesa electrode 6 on the light output end side.

The multiple quantum well layer 10 is provided in the region extending to the end of the mesa-top electrode 6 on the light output end side, and a waveguide with a mesa structure (not shown) is provided between this end and the front end face 101. Note that instead of a waveguide with a mesa structure, a window structure without a mesa structure may also be provided.

Figure 3 shows the distribution of photocurrent along the optical axis when the EA modulator 2 absorbs light. Photocurrent refers to the current that flows when a reverse bias is applied to the EA modulator 2 while the output light from the semiconductor laser 4 is being transmitted to the EA modulator 2. The magnitude of the photocurrent is roughly proportional to the amount of light absorbed by the EA modulator 2. In other words, the left side of Figure 3 represents the semiconductor laser 4 side, and the right side represents the front facet 101 side. The photocurrent is larger at position B, i.e., the semiconductor laser 4 side. This indirectly indicates that a large number of photoexcited carriers are generated at the light input end. The photocurrent decreases from the light input end to the light output end. In other words, in the EA modulator 2, the photocurrent distribution is not constant along the optical axis. The photocurrent is proportional not only to the intensity of the input light but also to the magnitude of the voltage applied to the EA modulator 2. The higher the voltage applied to the EA modulator 2, the greater the photocurrent. The voltage applied to the EA modulator 2 is applied to the pad electrode 8 via an external current-carrying means. The voltage applied to the pad electrode 8 is transmitted to the mesa electrode 6 via the lead electrode 7. The voltage is then transmitted from the mesa electrode 6 to the p-type contact layer and p-type InP cladding layer 11, and is applied to the multiple quantum well layer 10.

When a voltage is applied, the electric field strength along the optical axis of the mesa structure is constant when the applied voltage is a DC voltage. However, we found that the electric field strength along the optical axis of the mesa structure is not constant when a high-speed AC signal is applied. Figure 4 shows the results of a simulation of the electric field strength distribution in the multiple quantum well layer 10 at the B-B' cross section of this structure, obtained through electromagnetic field analysis. The simulation conditions were a length of 150 μm and a width (narrower width) of the lead electrode 7 along the optical axis of 10 μm. In this case, the center of the lead electrode 7 is located 3% of the length of the entire mesa electrode 6 from the optical output end (B' side). As a comparative example, we also performed a simulation of a structure in which the center of the lead electrode 7 coincides with the center of the mesa electrode 6. In this structure, the center of the lead electrode 7 is located 50% of the length of the entire mesa electrode 6 from the optical output end (B' side). Hereinafter, the connection position of the lead electrode 7 will be referred to as the 3% position, the 50% position, etc., based on the ratio of the distance from the optical output end to the center of the lead electrode 7. Figure 4 shows simulation results when a 40 GHz high-frequency electrical signal is applied to the pad electrode 8. The lengths of the mesa electrode 6 and the multiple quantum well layer 10 are assumed to be the same. The solid line represents the electric field intensity distribution of this embodiment (3% position), and the dotted line represents the electric field intensity distribution of the comparative example (50% position). As shown in Figure 4 , the electric field intensity of the comparative example is generally constant at any position, whereas the electric field intensity of the EA modulator 2 of this embodiment varies depending on the position. Specifically, the electric field intensity at the optical input end (side B) is higher than that of the comparative example, and the electric field intensity at the optical output end (side B') is lower than that of the comparative example. The distribution in electric field intensity is due to the fact that the connection position between the lead electrode 7 and the mesa electrode 6 is on the optical output end (side B'). A high-frequency electrical signal applied to the pad electrode 8 is transmitted through the lead electrode 7 to the mesa electrode 6. The electrical signal applied to the mesa electrode 6 is transmitted to the mesa electrode 6 and also to the multiple quantum well layer 10 via the p-type contact layer. At this time, the mesa electrode 6 is considered to function electrically as an open stub from the point of contact between the mesa electrode 6 and the lead electrode 7. As a result, a distribution of electric field strength is presumed to occur on the mesa electrode 6. At this time, a strong electric field can be applied to the optical input end, where the photocurrent (i.e., photoexcited carriers) is large. Furthermore, the electric field strength gradually decreases from the optical input end to the optical output end. However, the amount of photoexcited carriers (amount of photocurrent) also decreases. Because there are few carriers to be extracted, carriers can be extracted sufficiently even when the electric field strength decreases, making carrier pileup unlikely to occur. As a result, carrier pileup can be suppressed throughout the EA modulator 2. The greater the applied voltage, the more photoexcited carriers can be extracted from the multiple quantum well layer 10, and the EA modulator 2 according to this embodiment has a structure with excellent carrier extraction capabilities. Because of its excellent carrier extraction capabilities, it is possible to suppress carrier pile-up and provide an EA modulator 2 with excellent extinction characteristics and high-frequency characteristics.

Figure 5 shows the simulation results of the electric field strength of the EA modulator 2 when high-frequency electrical signals ranging from 10 GHz to 60 GHz are applied to the pad electrode 8. Note that because the absolute value of the electric field strength varies with frequency, it is normalized based on the position B'. At 10 GHz, the electric field strength does not vary significantly between the optical input terminal side B and the optical output terminal side B'. However, as the frequency increases, the electric field strength at the optical input terminal side B increases relative to the optical output terminal side B'. In other words, the effect of applying a stronger electric field to the optical input terminal side B, which has a larger photocurrent, increases with increasing frequency. Higher drive frequencies tend to make pileup more likely, and the present invention can suppress pileup, particularly at high drive frequencies, making it effective for EA modulators compatible with 56 Gbps.

The electric field strength distribution due to differences in the connection position of the lead electrode 7 was examined. The connection positions for this embodiment were set at 3% (the forefront), 10%, 20%, and 33% (see the fourth embodiment and Figure 11), and for the comparative example, at the 50% position (center), and a similar simulation was performed. Other conditions were the same as above. Figure 6A shows the results for 10 GHz, Figure 6B for 20 GHz, Figure 6C for 40 GHz, and Figure 6D for 60 GHz. The horizontal axis in each figure indicates the position in the optical axis direction. The line plotted as Iph indicates the photocurrent when a DC voltage is applied, and corresponds to the right-hand axis. The other plots correspond to the normalized electric field strength ratio for each of the above-mentioned connection positions, with the electric field strength when the connection position is at 50% as the reference, and correspond to the left-hand axis.

For the 10 GHz electrical signal shown in Figure 6A, the electric field strength tends to be roughly equivalent regardless of the connection position, being roughly equivalent at any position along the optical axis compared to the 50% position. Furthermore, there is little difference in electric field strength due to the connection position. For the 20 GHz electrical signal shown in Figure 6B, the electric field strength ratio tends to be larger at the optical input end side B and smaller at the optical output end side B'. This tendency is more pronounced at higher frequencies, as is clear from Figure 6C for 40 GHz and Figure 6D for 60 GHz. Furthermore, as for differences depending on the connection position, the electric field strength ratios at the 3%, 10%, and 20% positions, which are closer to the optical output end side B', are roughly equivalent, but the rate of increase with frequency at the 33% position is smaller. These results demonstrate that by positioning the connection between the lead electrode 7 and the mesa electrode 6 closer to the optical output end side B' than to the center of the mesa electrode 6, it is possible to increase the electric field strength ratio at the optical input end side B. The effect is particularly pronounced when the ratio is set to 33% or less. Furthermore, in order to obtain a stable effect taking into account manufacturing variations, it is more preferable to set the ratio to 20% or less, at which point the electric field strength ratio becomes roughly the same.

Regardless of the connection position, the electric field intensity ratio is less than 1 at a position slightly closer to the optical output end B' than the center position in the optical axis direction. Areas where the electric field intensity ratio is less than 1 indicate a lower electric field intensity compared to the conventional 50% connection, resulting in a decrease in the extraction efficiency of photoexcited carriers. For example, as shown in Figure 6D, the electric field intensity ratio becomes less than 1 at a certain point. A field intensity ratio less than 1 indicates a lower electric field intensity compared to the conventional structure with a 50% connection position. A lower electric field intensity indicates a decrease in the extraction efficiency of photoexcited carriers. Therefore, it can be said that areas where the electric field intensity ratio is less than 1 may exhibit a decrease in high-frequency characteristics compared to the conventional structure. However, as shown in Figure 6D, the photocurrent amount is not constant between the optical input end B and the optical output end B', but rather exhibits a decreasing distribution toward the optical output end B'. The total photocurrent amount in areas where the electric field intensity ratio is greater than 1 accounts for 95% of the total photocurrent amount. Conversely, the total photocurrent amount in areas where the electric field intensity ratio is less than 1 accounts for only 5% of the total photocurrent amount. Even if the electric field strength in the region where 5% of the total photocurrent is distributed is smaller than in the conventional structure, the effect on photocurrent discharge is small. In fact, increasing the electric field strength in the region where 95% of the total photocurrent is distributed significantly improves the extraction of photocurrent (proportional to photoexcited carriers), resulting in overall improvements in extinction characteristics and high-frequency characteristics. In other words, a feature of the present invention is that by shifting the connection position of the extraction electrode 7 from the center (50%) of the mesa electrode 6 to the optical output end side B', the distribution of the electric field strength along the optical axis of the mesa structure is changed. While the electric field strength along the optical axis of the mesa structure in the conventional structure is approximately constant, in the present invention, the electric field strength at the optical output end side B', where the photocurrent is small, is reduced, and instead the electric field strength at the optical input end side B, where the photocurrent is large, is increased. As a result, an EA modulator with excellent high-speed response can be provided.

Next, we investigated the relationship between the length of the multiple quantum well layer 10, i.e., the modulator length Lmod, and the effects of the present invention. To support high-speed driving of 25 Gbaud or more, it is effective to shorten the modulator length and reduce parasitic capacitance. Taking into consideration the need to support driving of 56 Gbaud or more, simulations were performed for modulator lengths of 100 μm and 125 μm in addition to the above-mentioned 150 μm. Furthermore, as with Figures 6A to 6D, simulations were performed for each connection position between the lead electrode 7 and the mesa electrode 6. The width of the lead electrode 7 was not changed.

Figure 7 shows the simulation results. This figure shows the results when a 60 GHz electrical signal was applied. The horizontal axis indicates the connection position of the lead electrode 7. As with Figures 6A to 6D, the vertical axis indicates the electric field strength ratio normalized with respect to the electric field strength when the connection position is at 50%. Furthermore, as is clear from Figures 6A to 6D, the electric field strength ratio varies with position in the optical axis direction. Therefore, in Figure 7, the vertical axis represents the average electric field strength ratio in the region on the optical input end side B of the mesa electrode 6, where 95% of the total photocurrent is distributed. As shown in Figure 7, an improvement in electric field strength can be achieved regardless of the modulator length. Furthermore, the improvement in the average electric field strength ratio is generally proportional to the modulator length. At least for modulator lengths in the range of 100 μm to 150 μm, the above effect can be achieved by shifting the connection position of the lead electrode 7 closer to the optical output end side B'. However, when the connection position of the lead electrode 7 is 33% and the modulator length is 100 μm, the improvement effect on the average electric field strength ratio is small. In order to obtain an improvement effect on the average electric field strength ratio regardless of the modulator length, it is preferable that the connection position of the lead electrode 7 be 20% or less. Furthermore, when it is 10% or less, the improvement effect on the average electric field strength ratio reaches its peak, so it is also preferable to connect it to this position.

In this embodiment, calculations were performed assuming that the lengths of the multiple quantum well layer 10 and the mesa electrode 6 were the same, but in reality, the multiple quantum well layer 10 may be slightly longer due to manufacturing variations. Even in this case, the effects of the present invention can be fully obtained.

Second Embodiment
FIG. 8 is a top view of an electroabsorption modulator integrated laser 201 according to a second embodiment of the present invention. The only difference from the electroabsorption modulator integrated laser 1 shown in the first embodiment is the structure of the EA modulator 202. In the optical axis direction, the center of the pad electrode 8 coincides with the center of the mesa electrode 6. The lead electrode 7 is tilted at 45 degrees with respect to the optical axis. The mesa electrode 6 is rectangular, and the mesa electrode 6 and the lead electrode 7 are connected at a position where the top right vertex of the mesa electrode 6 (the front end face 101 side and the pad electrode 8 side of the mesa electrode 6) coincides with the top vertex of the lead electrode 7. The length of the mesa electrode 6 is 150 μm, and the narrower width of the lead electrode 7 is 10 μm. Therefore, the center of the lead electrode 7 is connected to a position approximately 5% of the distance from the optical output end of the mesa electrode 6. A window structure is formed between the mesa electrode 6 and the front end face 101. It goes without saying that this structure also achieves the same effects as the first embodiment. Furthermore, compared to the electroabsorption modulator integrated laser 1 shown in the first embodiment, the pad electrode 8 can be positioned closer to the inside of the element (lower side in FIG. 8), thereby shortening the overall length of the electroabsorption modulator integrated laser. Here, the lead-out electrode 7 is inclined at 45 degrees with respect to the optical axis, but this is not a limitation and the angle of inclination is arbitrary. The lead-out electrode 7 is also not limited to a straight line and may be curved (see the fourth embodiment). However, because the lead-out electrode 7 can cause parasitic capacitance, it is preferable that its area be as small as possible.

[Third embodiment]
FIG. 9 is a top view of an electroabsorption modulator integrated laser 301 according to a third embodiment of the present invention. FIG. 10 is a schematic diagram of a CC cross section perpendicular to the optical axis of the EA modulator 302. The only difference from the electroabsorption modulator integrated laser 1 shown in the first embodiment is the structure of the EA modulator 302; the semiconductor laser 4 has the same structure. Of the n-type electrodes 305 for the EA modulator, which are electrodes on the front side of the EA modulator 302, the mesa-top electrode 6 is located on top of the mesa structure. The pad electrode 8 is rectangular. The lead electrode 7 connects the mesa-top electrode 6 and the pad electrode 8 and is approximately L-shaped. As shown in FIG. 10, the lead electrode 7 is connected to the mesa-top electrode 6 along the side of the mesa structure. An insulating film (not shown) is located between the side of the mesa structure and the lead electrode 7.

As shown in Figure 10, the multilayer structure is composed of a p-InP substrate 109 (p-type substrate), a multiple quantum well layer 10, and an n-InP cladding layer 110 (n-type cladding layer). A p-side electrode 314 is formed on the back side where the mesa structure is not formed. Here, the p-InP substrate 109 also functions as a p-type cladding layer. The n-InP cladding layer 110 also functions as a contact layer. The mesa structure is composed of a portion of the p-InP substrate 109, the multiple quantum well layer 10, and the n-InP cladding layer 110. Unlike the EA modulator 2 shown in the first embodiment, no buried layer 12 is disposed on the sides of the mesa structure. While the minimum configuration is shown here, a contact layer may be disposed on the top layer of the mesa structure to reduce optical confinement or resistance, and a passivation film to protect the semiconductor layer may also be disposed as needed.

In this embodiment, too, the lead electrode 7 is connected to the mesa-top electrode 6 closer to the optical output end side B' than the center. Even with this configuration, the effect of increasing the electric field strength at the optical input end side B shown in the first embodiment can be achieved. In other words, in an EA modulator with a p-i-n structure, the effect of the invention can be achieved if the EA modulator electrode connected to either the p-side or n-side semiconductor has the following configuration. That is, the EA modulator electrode includes a mesa-top electrode and a lead electrode arranged on the mesa structure, and the lead electrode is connected to the optical output end side of the mesa-top electrode. Furthermore, because the lead electrode 7 is approximately L-shaped, it can be easily manufactured using photolithography.

[Fourth embodiment]
FIG. 11 is a top view of an electroabsorption modulator-integrated laser 401 according to a fourth embodiment of the present invention. The only difference from the electroabsorption modulator-integrated laser 1 shown in the first embodiment is the structure of the EA modulator 402. The lead electrode 7 of the p-side electrode 405 for the EA modulator is curved toward the optical input end. The lead electrode 7 is connected 33% of the way from the optical output end. The mesa electrode 6 of the p-side electrode 405 for the EA modulator is 150 μm long, and the narrower width of the lead electrode 7 is 10 μm. A window structure is formed between the mesa electrode 6 and the front end facet 101. As shown in FIGS. 6C and 6D , this structure with a 33% connection point can, of course, achieve the same effects as the first embodiment. Furthermore, compared to the electroabsorption modulator-integrated laser 1 shown in the first embodiment, the pad electrode 8 can be positioned closer to the inside of the device (the lower side in FIG. 11 ), thereby shortening the overall length of the electroabsorption modulator-integrated laser. However, since the lead electrode 7 is a factor that generates parasitic capacitance, it is preferable that the area thereof be as small as possible.

The present invention is not limited to the above examples and is applicable to EA modulators in general. If the lead electrode is connected to the mesa electrode not in the center but offset toward the optical output end, the electric field intensity distribution changes, resulting in a bandwidth improvement effect. In other words, this invention improves characteristics by the connection position to the mesa electrode, regardless of the shape of the pad electrode or the lead electrode. The p-type and n-type may also be reversed. Furthermore, the above describes a case where the connection position is 3% from the optical output end, and is the end of the mesa electrode on the optical output end side. However, since this ratio is determined by the width of the lead electrode 7 and the length of the mesa electrode 6, the ratio is not limited to 3% as long as the end of the lead electrode 7 and the end of the mesa electrode 6 on the optical output end side are aligned.

1,201,301,401 Electroabsorption modulator integrated laser, 2,202,302,402 EA modulator, 3 Waveguide connection portion, 4 Semiconductor laser, 5 P-side electrode for semiconductor laser, 6 Mesa electrode, 7 Lead electrode, 8 Pad electrode, 9 n-InP substrate, 10 Multiple quantum well layer, 11 p-type InP cladding layer, 12 Buried layer, 13 Passivation film, 14 N-side electrode, 101 Front end facet, 102 Rear end facet, 105,205,405 P-side electrode for EA modulator, 109 p-InP substrate, 110 n-InP cladding layer, 305 N-side electrode for EA modulator, 314 P-side electrode.

Claims (5)

A substrate;
a mesa structure provided on a first surface of the substrate, the mesa structure including a first conductivity type cladding layer, a multiple quantum well layer, and a second conductivity type cladding layer;
a first conductivity type electrode electrically connected to the first conductivity type clad layer;
a second conductivity type electrode provided on a second surface of the substrate,
the first conductivity type electrode includes a mesa electrode disposed along an extension direction of the mesa structure, a pad electrode to which an external electric signal is input, and a lead electrode connecting the mesa electrode and the pad electrode;
the mesa structure includes an optical input end to which light is input from the outside, and an optical output end on the opposite side of the optical input end;
the mesa-top electrode has a first end closer to the optical input end than to the optical output end, and a second end closer to the optical output end than to the optical input end;
a connection position between the lead electrode and the mesa-up electrode is an end of the mesa-up electrode on the light output end side;
the electric field strength at the first end of the mesa electrode is greater than the electric field strength at the second end;
a portion of the pad electrode is disposed between the second end of the mesa electrode and the light output end;
the lead-out electrode extends in a direction perpendicular to a longitudinal direction of the mesa electrode and is connected to the pad electrode;
a width of the lead-out electrode in the longitudinal direction of the mesa electrode is narrower than a width of the pad electrode;
Electroabsorption modulator. 2. The electro-absorption modulator according to claim 1,
the length of the multiple quantum well layer in the longitudinal direction is 100 μm or more and 150 μm or less;
Electroabsorption modulator. 3. The electro-absorption modulator according to claim 1,
Further, buried layers are provided so as to sandwich the mesa structure.
Electroabsorption modulator. 2. The electro-absorption modulator according to claim 1,
the mesa structure further comprises a contact layer on the first conductivity type cladding layer.
Electroabsorption modulator. 2. The electro-absorption modulator according to claim 1,
The external electrical signal includes an electrical signal of 20 GHz or higher.
Electroabsorption modulator.