JP2008301374A - Oscillator manufacturing method and oscillator - Google Patents

Abstract

An object of the present invention is to provide an oscillator manufacturing method capable of improving output by reducing electromagnetic wave loss in a resonance state, and the oscillator.
According to an oscillator manufacturing method, a gain medium 202 having a gain with respect to an oscillating electromagnetic wave and a resonator structure for resonating the electromagnetic wave are provided, and the resonator structure confines a direction perpendicular to the propagation direction of the electromagnetic wave. An oscillator having a waveguide structure is manufactured. A semiconductor multilayer film including the gain medium 202 is epitaxially grown on the first substrate 200. A groove 205 having a shape such that the cut-off frequency defined by the cross-sectional shape of the groove 205 is smaller than the frequency of the electromagnetic wave is formed on the first substrate 200 or the second substrate 209 different from the first substrate. To do. The first substrate and the second substrate are bonded together so that the groove 205 is sandwiched between the first substrate 200 and the second substrate 209 to form a waveguide structure.
[Selection] Figure 1

Description

The present invention relates to an oscillator manufacturing method and an oscillator. In particular, the present invention relates to a method of manufacturing an oscillator (millimeter wave / terahertz wave oscillator) that oscillates electromagnetic waves in a frequency band in the frequency range from the millimeter wave band to the terahertz band (30 GHz to 30 THz), and the oscillator.

In the frequency region from the millimeter wave band to the terahertz wave, one of oscillators known so far is an oscillator in which a negative resistance element is monolithically formed on a substrate. Patent Document 1 discloses an oscillator in which a resonator structure and a gain medium are made monolithic by covering a semiconductor substrate having a negative resistance element with a metal film or the like. Since the semiconductor substrate having a relatively high dielectric constant is provided inside the resonator, the resonator can be miniaturized.

FIG. 7 is a cross-sectional view illustrating the oscillator of Patent Document 1. In FIG. In the oscillator of Patent Document 1, the negative resistance element formed by the multilayer semiconductor 12 and the contact layers 11 and 13 generates an electromagnetic wave in the frequency domain. 15 and 16 are electrodes for injecting a current into the negative resistance elements 11, 12 and 13. The metal film 17 forms a resonator structure including the negative resistance elements 11, 12, 13, the high resistance region 14, and the semi-insulating substrate 10 therein. The electromagnetic wave that resonates in such a resonator structure is extracted to the outside through the window 18. In FIG. 7, R represents the dimension of the cylindrical resonator, and Patent Document 1 describes that it is, for example, 800 mm for millimeter wave oscillation.
Japanese Patent Laid-Open No. 8-116074

However, in the conventional oscillator described above, electromagnetic wave loss in a semiconductor substrate or the like provided inside the resonator weakens the resonance and decreases the output of the oscillator. Here, dielectric loss and free carrier absorption loss can be considered as electromagnetic wave loss. Regarding the former, for example, it is known that a typical semiconductor substrate such as GaAs or InP has a phonon pole around 10 THz (8 THz for GaAs and 9 THz for InP). Near the pole, electromagnetic wave loss is very large. In addition, the bottom of the dielectric loss extending from such a pole is considered to extend to the vicinity of 100 GHz in the millimeter wave band on the low frequency side and cannot be ignored. The latter is due to the finite concentration of free carriers remaining in the semiconductor substrate, and becomes larger at the lower frequency side.

In view of the above problems, the method of manufacturing an oscillator according to the present invention has the following characteristics. A gain medium having a gain with respect to an oscillating electromagnetic wave and a resonator structure for resonating the electromagnetic wave, the resonator structure having a waveguide structure for confining at least a direction perpendicular to the propagation direction of the electromagnetic wave It is a manufacturing method of an oscillator. The manufacturing method includes at least a first step, a second step, and a third step. In the first step, a semiconductor multilayer film including the gain medium is epitaxially grown on a first substrate. In the second step, the cut-off frequency defined by the cross-sectional shape of the groove is smaller than the frequency of the electromagnetic wave on the first substrate or the second substrate different from the first substrate. A groove having a shape is formed. In the third step, the first substrate and the second substrate are bonded so that the groove is sandwiched between the first substrate and the second substrate to form the waveguide structure.

Moreover, in view of the said subject, the oscillator of this invention has the following characteristics. A gain medium having a gain with respect to an oscillating electromagnetic wave and a resonator structure for resonating the electromagnetic wave, the resonator structure having a waveguide structure for confining at least a direction perpendicular to the propagation direction of the electromagnetic wave . A semiconductor multilayer film including the gain medium is epitaxially grown on the first substrate, and the cutoff defined by the cross-sectional shape of the groove is formed on the first substrate or the second substrate different from the first substrate. A groove having a shape whose frequency is smaller than the frequency of the electromagnetic wave is formed. Further, the groove structure is sandwiched between the first substrate and the second substrate to constitute the waveguide structure. Typically, the frequency of the electromagnetic wave includes a part of a frequency range of 30 GHz to 30 THz.

According to the method for manufacturing an oscillator of the present invention, it is possible to select a configuration that does not include a semiconductor substrate having a relatively large electromagnetic wave loss in the resonator. Can be improved. It is also possible to give various variations to the resonator structure. Furthermore, the output of the oscillator of the present invention having a configuration that can be easily manufactured by the manufacturing method of the present invention can be improved.

EMBODIMENT OF THE INVENTION Below, the form for implementing this invention is demonstrated with reference to drawings.

(First embodiment)
FIG. 1 shows a cross-sectional structure of a first embodiment of a millimeter wave / terahertz wave oscillator to which the present invention can be applied, and a manufacturing method thereof. FIG. 1B shows the structure in the longitudinal direction of the waveguide structure perpendicular to the paper surface including the gain medium 202 of FIG. 1A, and FIG. 1A shows the structure of the cross section perpendicular to the longitudinal direction. Indicates.

In this embodiment, the gain medium is a resonant tunneling diode (RTD) based on a photon-assisted tunnel composed of a III-V group compound semiconductor. As the resonator structure, a cavity secured by bonding an Si substrate to an RTD is used. RTD is a preferable gain medium example as a gain medium constituting a millimeter wave / terahertz wave oscillator to which the present invention can be applied.

In FIG. 1, the gain medium 202 is an RTD as described above, and has a configuration such as spacer layer / barrier layer / well layer / barrier layer / well layer / barrier layer / spacer layer. InGaAs that lattice-matches to the InP substrate 200 can be used as a well layer, and InAlAs that lattice-matches or non-matched AlAs can be used as the barrier layer. More specifically, in order from the emitter side to the collector side, InGaAs 5.0 nm / AlAs 1.3 nm / InGaAs 7.6 nm / InAlAs 2.6 nm / InGaAs 5.6 nm / AlAs 1.3 nm / InGaAs 5.0 nm (the number indicates the thickness) A semiconductor multilayer structure can be used.

If any layer is undoped intentionally without carrier doping, the free electron absorption loss in the gain medium 202 can be ignored. RTD has a gain in a frequency region from a millimeter wave band to a terahertz band based on a phenomenon called a photon assist tunnel. For example, the RTD configured as described above has a peak current density of about 90 kA / cm ² when an electric field of 0.3 V is applied, and exhibits a negative resistance from 0.3 V to 0.6 V.

Such a gain medium 202 is sandwiched between the electrical contacts 211 and 213. The electrical contacts 211 and 213 of the gain medium 202 are made of, for example, an n-InGaAs 50 nm semiconductor film lattice-matched to the InP substrate 200. Here, Si is used as a dopant, and the electron concentration is set to 2 × 10 ¹⁸ cm ⁻³ .

Further, the gain medium 202 is sandwiched between the conductive layers 201 and 203, and each of them has an n-InGaAs 400 nm (thickness of the conductive layer 201) and 100 nm (thickness of the conductive layer 203) lattice matched to the InP substrate 200. It is composed of a semiconductor film. Here, the electron concentration of these semiconductor films is set to 1 × 10 ¹⁹ cm ⁻³ to make ohmic contact with electrodes 204 and 206 such as Ti / Au.

The above is an example of the configuration on the InP substrate 200, and the present invention is not limited to this. For example, a semiconductor multilayer structure such as InAs / AlAsSb or InAs / AlSb on an InAs substrate, GaAs / AlAs on a GaAs substrate, or Si / SiGe on a Si substrate can be considered.

Furthermore, the structure of the other part of a present Example is demonstrated. In FIG. 1A, reference numeral 209 denotes a Si substrate provided with a groove 205 by MEMS processing or the like. The surface of the Si substrate 209 is a conductive layer 208, for example, an Au plating layer of 1 mm. However, a part of them is formed by being bonded so as to be in electrical contact with the electrode 204. Accordingly, the groove 205 becomes a cavity formed by being surrounded by the conductive layers 201 and 203, the electrodes 204 and 206, and the conductive layer 208 on the surface of the Si substrate 209. That is, a conductive layer composed of a metal, a carrier-doped semiconductor, or a metal and a carrier-doped semiconductor is formed on the surface of the groove 205 to form a cavity. The groove 205 may be filled with, for example, air with extremely small electromagnetic wave loss. Reference numeral 207 denotes a dielectric, for example, SiO ₂ having a thickness of about 300 nm that is sufficiently thin with respect to the cavity scale.

As described above, a bias circuit for applying a bias electric field for injecting a current into the gain medium 202 is configured. As long as a bias electric field can be applied, a conductive material may be used for the portion 207. When a voltage of 0.3 V to 0.6 V is applied between the electrode 204 (or the conductive layer 208) and the electrode 206 by an external electric field applying means (not shown), the gain medium 202 generates an electromagnetic wave in the cavity 205. The cavity 205 forms a rectangular waveguide, and the cut-off frequency is determined based on the cross-sectional shape having dimensions of width W and height H.

Assuming that the gain medium 202 on the InP substrate 200 is a sufficiently long mesa shape, feeding by a rectangular waveguide probe (FED
with Probe). therefore,
f _c = √ ((n ₁ p / W) ² + (n ₂ p / H) ² ) × C ₀ / 2p
Becomes the cut-off frequency. C ₀ is the speed of light in vacuum. Here, TE01 (that is, n ₁ = 1, n ₂ = 0) is the lowest mode that can be excited (David for FED with Probe)
M. Pozer Author ^{"Microwave Engineering 3 rd -Ed",} John Wiley & Sons, Inc. (2003), see Chapter 4). Therefore, when the cut-off frequency f _c = 300 GHz is selected, W = 500 mm. At this time, H needs to have the same size as W. If H = W / 2 or more, the conductor loss in the TE01 mode can be kept relatively small. Here, W = 500 mm and H = W / 2 = 250 mm. In this way, the oscillation frequency can be set to 300 GHz or more.

The above configuration will be described with reference to FIG. 1B showing the structure of the waveguide structure in the longitudinal direction. The oscillation frequency f depends on the length L of the rectangular waveguide. In the present embodiment, open ends 221 and 222 that also serve as electromagnetic wave extraction windows are provided. Therefore, when the gain medium 202 is positioned at the antinode of the standing wave of the electric field reflected and resonated by the open ends 221 and 222, it is easy to obtain an oscillator output.

For example, in FIG. 1B, when the gain medium 202 is located at the center in the longitudinal direction of the cavity 205, f = 2mC ₀ / (2L × √e _eff ). Here, m is an integer of 1 or more. 2 before m is an additional factor caused by the gain medium 202 being positioned at the antinode of the standing wave. √e _eff is a correction factor for correcting the guide wavelength in the rectangular waveguide, and √e _eff = b / k ₀ (b is a propagation constant, k ₀ is a wave number in a vacuum). For simplicity, the correction factor is approximated to 1 and the case of L = 2.4 mm is considered. In this case, it is the beginning of the integer m = 3 is f> f _c. Therefore, when the oscillation frequency is drawn to m = 3, it is estimated that the oscillation frequency f = 375 GHz. However, more generally, since the reactance component in the gain medium 202 or the like is capacitive or inductive, the actual oscillation frequency is slightly deviated from the estimated oscillation frequency f.

It is preferable to use a semiconductor process technique and a MEMS (Micro Electro Mechanical Systems) processing technique for manufacturing the structure as described above. That is, the semiconductor process technology is used on the InP substrate 200 side, and the MEMS processing technology is used on the Si substrate 209 side.

First, the InP substrate 200 side. First, an InP substrate 200, which is the first substrate, is formed on an n-InGaAs layer 201, 211, an InGaAs / AlAs or InGaAs / InAlAs multi-quantum well 202, an n-InGaAs layer 213 by a molecular beam epitaxy (MBE) method or the like. , 203 is epitaxially grown. Ti / Au 204 is deposited on the surface as an electrode, and the n-InGaAs layer 201 is etched into a mesa shape as shown in FIG. Etching uses photolithography and ICP (inductively coupled plasma) dry etching. Ti / Au206 is deposited on the etched surface as an electrode. Then, after depositing SiO ₂ 207 by plasma CVD or the like, the mesa-like epitaxial layer is exposed to a thickness of 300 nm as described above. Thus, a dielectric is included in the portion of the waveguide structure where electromagnetic waves propagate. The above process corresponds to a process including a first process of epitaxially growing a semiconductor multilayer film including a gain medium on a first substrate.

Next is the Si substrate 209 side. First, a Si substrate 209 having a thickness of 500 mm, which is a second substrate, is prepared, and a resist pattern is formed at a position where a cavity is desired to be secured using photolithography. Here, commercially available AZ4990 (manufactured by Clearant) was used as the photoresist. Next, a groove 205 having a depth of 250 μm is formed by ICP-RIE using a Bosch process. Here, SF ₆ and C ₄ F ₈ were used as the etching gas. Both sides of the groove 205 are defined by a wall portion extending in the longitudinal direction of the waveguide structure, and a protruding portion for sandwiching the gain medium 202 is left in the groove. Further, an Au / Ti layer (for example, 100 nm / 20 nm) serving as a seed layer (not shown) is formed on the entire front and back surfaces of the Si substrate 209. After that, an Au plating layer 208 is formed by plating a 1 μm Au film using electroplating. In the above process, a groove having such a shape that the cut-off frequency defined by the cross-sectional shape of the groove is smaller than the frequency of the electromagnetic wave is formed on the first substrate or a second substrate different from the first substrate. This corresponds to a step including the second step of forming.

Finally, bonding by pressure bonding is performed between the electrode 204 and the Au plating layer 208 on the Si substrate 209. Heat fusion using solder such as AuSn may be used. In addition to the bonding by bonding, the periphery of the InP substrate 200 and the Si substrate 209 may be fixed with an epoxy resin (not shown) or the like. The above process corresponds to a process including a third process in which the first substrate and the second substrate are bonded so that a groove is sandwiched between the first substrate and the second substrate to form a waveguide structure. To do.

The above manufacturing method includes a gain medium having a gain with respect to an electromagnetic wave to be oscillated and a resonator structure for resonating the electromagnetic wave, and the resonator structure guides at least a direction perpendicular to the propagation direction of the electromagnetic wave. An oscillator having a tube structure is produced.

As a modified configuration of the waveguide structure, examples such as those shown in FIGS. 2, 3, and 4 can be employed by using a well-known microwave technique. In the example of FIG. 2, the cavity 205 in the structure of FIG. 1 is deformed to form a cavity 305 that forms a shape called a ridge-type waveguide. The cavity 305 is formed of a substrate 309 processed into an E-shaped cross section as shown in FIG.

In the example of FIG. 3, the cavity 305 has the same shape as the example of FIG. 2, and the gain medium 402 is formed on the substrate 400 in a discrete manner in the longitudinal direction of the waveguide structure. Each gain medium 402 is preferably provided at an antinode of a standing wave of an electric field that resonates. Accordingly, the electrical contacts 411 and 413, the conductive layer 403, and the electrode 404 are also distributed and formed on the substrate 400. The conductive layer 401, the electrode 406, and the dielectric 407 are formed as shown in FIG. According to this structure, the oscillator outputs from the distributed gain medium 402 can be combined.

In the example of FIG. 4, the ridge waveguide in the example of FIG. 2 is provided with a taper as shown in FIG. A cavity 505 having a shape called a fin line is formed by the Si substrate 509. In a typical case, impedance matching with the gain medium 202 is easily performed, and the oscillator output is more easily extracted. The taper is formed only on the central protrusion of the substrate 509 for sandwiching the gain medium 202 portion.

Also in these modifications, any of the above-described MEMS processing can be produced by applying the above-described MEMS processing to the Si substrates 309 and 509 having the gold plating layers 308 and 508.

According to the manufacturing method of the present embodiment, it is possible to select a configuration in which the resonator does not include a semiconductor substrate having a relatively large electromagnetic wave loss. Therefore, the electromagnetic wave loss in the resonance state is reduced as compared with the conventional example, and the output of the oscillator can be improved. The oscillator manufactured in this way has a preferable configuration in which only a gain medium such as a negative resistance element is included in the resonator without including a semiconductor substrate in which electromagnetic wave loss cannot be ignored. In this case, the size of the resonator in the frequency region from the millimeter wave band to the terahertz band is several millimeters to several microns, but such a configuration can be easily formed by using the current MEMS processing technology. Further, if necessary for miniaturization, as described above, a low-loss dielectric material with small electromagnetic wave loss can be included in the resonator.

(Second embodiment)
FIG. 5 shows a cross-sectional structure of a second embodiment of a millimeter-wave / terahertz wave oscillator to which the present invention can be applied, and a manufacturing method thereof. The present embodiment is an example in which a part of the waveguide structure of the first embodiment includes a low-loss dielectric and is easy to manufacture.

In FIG. 5, the gain medium 602 is an RTD on the same InP substrate 600 as in the first embodiment for simplicity. The electrical contacts 611 and 613 are composed of n-InGaAs 50 nm with an electron concentration of 2 × 10 ¹⁸ cm ⁻³ , and the conductive layers 601 and 603 are n-InGaAs 100 nm semiconductor films with an electron concentration of 1 × 10 ¹⁹ cm ^−3. Consists of. Further, Ti / Au or the like is used for the electrodes 604 and 606.

In FIG. 5, for simplification, 609 is a Si substrate provided with a groove 605 by the same MEMS processing as in the first embodiment. In this embodiment, the cavity 605 formed by being surrounded by the conductive layers 601 and 603, the electrodes 604 and 606, and the conductive layer 608 on the surface of the Si substrate 609 includes a dielectric 607 to form a rectangular waveguide. Here, the dielectric 607 and Si0 ₂ having a thickness of 300 nm. Si0 ₂ 607 has a phonon pole in the mid-infrared region (> 30 THz) (see Phys. Rev. B., vol. 38, 1255 (1988)) and relatively low dielectric loss. In addition, the SiO ₂ 607 thus formed also serves as an insulating layer on the side wall of the RTD 602. When W = 500 mm and H = W / 2 = 250 mm as in the first embodiment, only 300 nm of 250 mm in the height direction is SiO ₂ 607.

The oscillation frequency f at this time is only slightly shifted from the oscillation frequency f ₀ when the dielectric 607 is not included. To estimate the shift, this is perturbed by the cavity resonator dielectric (Cavity
Perturbation) is as follows.
(ff ₀ ) / f ₀ =-(e _r -1) × t / 2H
(David for Cavity Perturbation
M. Pozer Author ^{"Microwave Engineering 3 rd -Ed",} John Wiley & Sons, Inc. (2003), see Chapter 6). Here, e _r = 3.8 is the dielectric constant of SiO ₂ 607, and t = 300 nm is the thickness of SiO ₂ 607. When calculated, f = (1−0.00168) × f ₀ , and the influence of SiO ₂ 607 is slight. This embodiment can be combined with the configuration of the modification of the first embodiment. The other points are the same as in the first embodiment.

(Third embodiment)
FIG. 6 shows a third embodiment of a millimeter wave / terahertz wave oscillator to which the present invention can be applied. This embodiment has a leaky wave coupling type array configuration in which a plurality of millimeter wave / terahertz wave oscillators are provided in parallel as in the first embodiment. In this embodiment, the gain medium 702 is a resonant interband tunnel diode (RITD) based on a photon-assisted tunnel made of a group IV semiconductor. The resonator structure includes a cavity 705 secured by bonding a Si substrate 709 to a gain medium 702.

In FIG. 6, the gain medium 702 is RITD as described above, and has, for example, a configuration using a Type II heterojunction such as an n-type δ-doped portion / spacer layer / Type II spacer layer / p-type δ-doped portion / Type II p-type layer. Have In this configuration, Si that is lattice-matched to the Si substrate 700 can be used as a spacer layer, and SiGe that is not matched to the Si substrate 700 and forms a Type II heterojunction can be used for the Type II spacer layer and Type II p-type layer. More specifically, a semiconductor multilayer structure of n-Si (δ-dope) /Si1.0nm/SiGe2.0nm/p-SiGe (δ-dope) /p-SiGe1.0nm in order from the emitter side to the collector side (Refer to Appl. Phys. Lett, Vol. 83, 3308 (2003) for the structure of the gain medium 702). Here, the spacer layer and the Type II spacer layer are undoped.

Such a gain medium 702 is sandwiched between electrical contacts 701 and 703 that also serve as a conductive layer. The electrical contact 701 is made of, for example, an n-Si 100 nm semiconductor film lattice-matched to the Si substrate 700. Here, P is used as a dopant, and the electron concentration is set to 5 × 10 ¹⁹ cm ⁻³ . The electrical contact 703 is made of, for example, a p-Si 100 nm semiconductor film lattice-matched to the Si substrate 700. Here, the electron concentration is set to 5 × 10 ¹⁹ cm ⁻³ using B as a dopant. In this way, the electrical contacts 701 and 703 are in ohmic contact with electrodes 704 and 706 such as Al.

The RITD 702 has a gain in the frequency region of the millimeter wave band based on a phenomenon called a photon assist tunnel. For example, the RITD configured as described above exhibits a peak current density of about 40 kA / cm ² when an electric field of 0.3 V is applied, and exhibits a negative resistance when an electric field of 0.3 V to 0.4 V is applied.

In FIG. 6, the Si substrate 709 is provided with a groove 705 by MEMS processing or the like. Similar to the first embodiment, the cavity 705 formed by being surrounded by the conductive layers 701 and 703, the electrodes 704 and 706, and the conductive layer 708 on the surface of the Si substrate 709 constitutes a rectangular waveguide. Reference numeral 707 denotes a dielectric such as SiO ₂ , and the thickness thereof is, for example, SiO ₂ and about 200 nm that is sufficiently thin with respect to the scale of the cavity 705.

Such a sufficiently thin dielectric 707 also has a slight electromagnetic evanescent component (referred to as a leaky wave). Adjacent cavities 705 in FIG. 6 are coupled by this leakage wave and can share a resonance state with each other. This can be expected when the length of the coupling portion 707 in the width direction (left-right direction in FIG. 6) is shorter than the attenuation length of the leaky wave. For example, the attenuation length is in the order of millimeters near the frequency of 300 GHz. Therefore, in this embodiment, the length of the coupling portion 707 in the width direction is 100 mm.

Further, like the first embodiment, the dielectric 707 also functions as a bias circuit element for applying a bias electric field for injecting a current into the gain medium 702. In this way, the bias circuits are arranged in parallel. In this case, it is not affected by instability due to negative resistance. By using such a method, it is possible to combine the leaky waves in the frequency region from the millimeter wave band to the terahertz band to synthesize the oscillator output. Of course, it is also possible to form an array structure by coupling electromagnetic waves by a coupling method using TEE junctions well known in the microwave technology, quasi-optical injection locking using a reflecting mirror, or the like. This embodiment can also be combined with the configuration of the modification of the first embodiment. Other points are the same as in the above embodiment.

In the above embodiments, RTD and RITD have been described as examples of the gain medium. However, for example, an Esaki diode, a Gunn diode, or other negative resistance diode can be replaced. In addition, although the example in which the cavity is filled with air has been described, in order to further reduce electromagnetic wave loss due to water vapor contained in the air, the cavity may be filled with dry air or nitrogen using a sealing material or the like. It may be a low vacuum. Furthermore, with regard to securing the cavity, the Si substrate, which is the first substrate provided with grooves by MEMS processing, has been described as an example, but the substrate, which is the second substrate having a gain medium, is etched to ensure the cavity. May be.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a cross-sectional view for explaining a configuration of a millimeter wave / terahertz wave oscillator of a first embodiment to which the present invention is applicable and a method for manufacturing the same. Sectional drawing explaining the modification of a 1st Example. Sectional drawing explaining the other modification of a 1st Example. Sectional drawing explaining the further another modification of a 1st Example. Sectional drawing for demonstrating the structure of the millimeter wave and terahertz wave oscillator of 2nd Example which can apply this invention, and its manufacturing method. Sectional drawing for demonstrating the structure of the millimeter wave and terahertz wave oscillator of 3rd Example which can apply this invention, and its manufacturing method. Sectional drawing which shows the structure of the oscillator of a prior art example.

Explanation of symbols

200, 400, 600, 700 ... first substrate
201, 203, 401, 403, 601, 603 ... conductive layer
211, 213, 411, 413, 611, 613, 701, 703 ... Electric contact (conductive layer)
202, 402, 602, 702 ... gain medium (negative resistance diode)
204, 206, 404, 406, 604, 606, 704, 706 ... electrodes
205, 305, 505, 605, 705 ... cavity (groove)
207, 407, 607, 707 ... dielectric
208, 308, 508, 608, 708 ... conductive layer on the surface of the second substrate
209, 309, 509, 609, 709 ... second substrate (Si substrate)

Claims (9)

A gain medium having a gain with respect to an oscillating electromagnetic wave and a resonator structure for resonating the electromagnetic wave, the resonator structure having a waveguide structure for confining at least a direction perpendicular to the propagation direction of the electromagnetic wave An oscillator manufacturing method comprising:
A first step of epitaxially growing a semiconductor multilayer film including the gain medium on a first substrate;
A groove having a shape such that a cut-off frequency defined by a cross-sectional shape of the groove is smaller than the frequency of the electromagnetic wave is formed on the first substrate or a second substrate different from the first substrate. A second step;
A third step of bonding the first substrate and the second substrate so that the groove is sandwiched between the first substrate and the second substrate to form the waveguide structure;
Including at least
An oscillator manufacturing method characterized by the above. 2. The method of manufacturing an oscillator according to claim 1, wherein the groove is formed by etching the second substrate in the second step. 3. The method of manufacturing an oscillator according to claim 2, wherein the second substrate is a Si substrate, and the second step includes a step of subjecting the Si substrate to MEMS processing. 4. The method of manufacturing an oscillator according to claim 1, wherein a dielectric is included in a portion where the electromagnetic wave propagates in the waveguide structure. 5. The method for manufacturing an oscillator according to claim 1, wherein a conductive layer composed of a metal, a carrier-doped semiconductor, or a metal and a carrier-doped semiconductor is formed on the surface of the groove. 6. The method of manufacturing an oscillator according to claim 1, wherein, in the first step, the gain medium is distributed along a propagation direction of the electromagnetic wave of the resonator structure. 7. The method for manufacturing an oscillator according to claim 1, wherein a plurality of oscillators are formed in parallel adjacent to each other in a direction perpendicular to the propagation direction of the electromagnetic wave of the resonator structure. A gain medium having a gain with respect to an oscillating electromagnetic wave and a resonator structure for resonating the electromagnetic wave, the resonator structure having a waveguide structure for confining at least a direction perpendicular to the propagation direction of the electromagnetic wave An oscillator,
A semiconductor multilayer film including the gain medium is epitaxially grown on the first substrate,
A groove having a shape such that a cut-off frequency defined by a cross-sectional shape of the groove is smaller than the frequency of the electromagnetic wave is formed on the first substrate or a second substrate different from the first substrate. ,
The groove structure is sandwiched between the first substrate and the second substrate to form the waveguide structure.
An oscillator characterized by that. 9. The oscillator according to claim 8, wherein the frequency of the electromagnetic wave includes a part of a frequency region from 30 GHz to 30 THz.

Images