JP6794508B2 - Oscillator element and oscillator using it - Google Patents

Description

The present invention relates to an oscillating element having a negative resistance element and an oscillator using the oscillating element.

Negative resistance elements are useful in the application field of electromagnetic wave oscillating elements with a resonator. So far, a negative resistance element can be used to generate an electromagnetic wave (hereinafter, simply referred to as "terahertz wave") including at least a part of the frequency band from the millimeter wave band to the terahertz band (30 GHz or more and 30 THz or less). It has been known. As one of them, Non-Patent Document 1 discloses an oscillator in which a negative resistance element and a resonator are integrated on a substrate to make it monolithic. A slot antenna is integrated on a semiconductor substrate having a negative resistance element, and the resonator structure and the gain medium are monolithically configured.

FIG. 9 shows the oscillator of Non-Patent Document 1. In this oscillator, the negative resistance element uses a resonance tunnel diode (Resonant tunneling diode) S-RTD91 provided with a Schottky barrier on the collector side. The resonator is a slot antenna. The slot antenna is composed of a metal pattern 92 on a semiconductor substrate and has capacities 93 and 94 at its ends. The oscillator of Non-Patent Document 1 also includes a rectifier diode 95.

Here, the rectifier diode 95 forms a stabilizing circuit for suppressing parasitic oscillation, which is a problem in an oscillator using a negative resistance element. Parasitic oscillation refers to parasitic oscillation in a frequency region lower than the desired oscillation frequency _fosc . Since such parasitic oscillation lowers the oscillation output at the oscillation frequency _fosc, it can be said that the existence of a stabilizing circuit is important in an oscillator using a negative resistance element.

Therefore, in Non-Patent Document 1, in order to reduce the parasitic oscillation, the position of 1/4 within a wavelength lambda _osc corresponding to the oscillation frequency f _osc toward the power supply side as seen from the S-RTD91, a negative resistance element A low-impedance stabilization circuit is placed. Specifically, as a stabilizing circuit, a shunt rectifier diode 95 is integrated at a position within λ _osc / 4 toward the power supply 96 side when viewed from the S-RTD 91. The resistance 97 is the sum of the internal resistance of the power supply 96 and the resistance of the electric wire. In this way, parasitic oscillation can be reduced by lowering the impedance on the power supply side that supplies the bias voltage in the frequency region of DC or more and less than _fosc .

Rudy, IEEE Electronics Letters, Vol 18, 218 (1997)

The conventional stabilization circuit for suppressing parasitic oscillation as described above is a circuit that uses a diode or a resistor to attenuate the amplitude of parasitic oscillation in a low frequency region where oscillation is not desired. In an oscillating element or oscillator that oscillates electromagnetic waves in the millimeter wave band or terahertz band, the frequency range in such a low frequency region is relatively wide, so it is necessary to attenuate the amplitude of parasitic oscillation over a very wide band. Therefore, there is a possibility that the design becomes complicated and the oscillation of a desired frequency is also attenuated.

In view of the above problems, an object of the present invention is to provide an oscillating element and an oscillator capable of narrowing the frequency range for attenuating parasitic oscillation as compared with the conventional case.

The oscillating element as one aspect of the present invention is an oscillating element that oscillates an electromagnetic wave, and has a negative resistance element and a resonator having a first conductor and a second conductor. The resistance element and the resonator are arranged on a substrate, and the negative resistance element is electrically connected to the first conductor and the second conductor, and the first conductor and the first conductor are described. the second conductor has been capacitively coupled, series resonance frequency f ₁ of the resonator, the capacitance between the first conductor and the second conductor C, the first conductor and the Assuming that the inductance of the second conductor is L ₁ , the velocity of the oscillated electromagnetic wave in vacuum is C ₀ , the relative dielectric constant of the substrate is ε _r , and the diagonal length of the substrate is d, the following equation is obtained. Fulfill.

According to the oscillating element and the oscillator as one aspect of the present invention, the frequency range for attenuating the parasitic oscillation can be narrowed as compared with the conventional case.

FIG. 5 is a cross-sectional view of the oscillating element of the first embodiment. Impedance plots and Smith charts to illustrate the configuration of the resonator. The schematic diagram which shows the oscillator using the oscillation element of 1st Embodiment. FIG. 5 is a cross-sectional view of the oscillating element of the second embodiment. The schematic diagram of the oscillator using the oscillator element of 2nd Embodiment. The schematic diagram and the sectional view of the oscillating element and the oscillator which concern on Example 1. FIG. The figure which showed the simulation result of the oscillating element which concerns on Example 1. FIG. The schematic diagram and the sectional view of the oscillating element and the oscillator which concern on Example 2. FIG. The schematic diagram of the structure of the oscillator of the prior art.

(First Embodiment)
An oscillator that oscillates a terahertz wave has an oscillating element having a negative resistance element and a distributed constant type resonator, and a power supply that supplies a bias voltage to the negative resistance element. The bias voltage from the power supply is supplied to the negative resistance element via the bias supply unit including the electric wire and the conductor. Parasitic low-frequency oscillation (parasitic oscillation) in a terahertz wave oscillator is often generated by the structure associated with this bias supply unit. Therefore, in the oscillating element 100 of the present embodiment, the bias supply unit is characterized. That is, a structure is adopted in which a part of the bias supply unit is included in a part of the distributed constant type resonator 110 and is capacitively coupled.

Specifically, the resonator 110 is a two-conductor type resonator in which two conductors 104 and 106 are capacitively coupled. Seen from the terahertz wave that resonates the resonator 110, they form an integral resonator. Such a structure is the result of diligent research by the inventors, and the process leading to the configuration of the present embodiment will be described below with reference to FIG.

FIG. 2A is a schematic diagram of the configuration of the resonator 2 composed of one conductor and an impedance plot of the resonator 2, and FIG. 2D is a Smith chart of the resonator 2. The impedance of the distributed constant type resonator 2 composed of one conductor becomes zero in the DC limit as shown in FIG. 2A. In the Smith chart of FIG. 2D, since the impedance in the resonator 2 turns clockwise as the frequency increases, the next resonance point is the lowest-order parallel resonance point 11 of the resonator 2.

The parallel resonance point 11 of the resonator 2 is the oscillation frequency _fosc itself of the electromagnetic wave oscillated by the oscillating element or the oscillator. For these reasons, the phases do not match in a wide band of DC or more and less than _fosc , and oscillation is not possible in this frequency region. Therefore, it is considered that parasitic low frequency oscillation is impossible. However, since both ends of the negative resistance element 1 are single conductors having the same potential, a bias voltage cannot be applied.

FIG. 2B is a schematic diagram of a distributed constant type resonator 3 having two conductors and an impedance plot of the resonator 3, and FIG. 2E is a Smith chart of the resonator 3. As can be seen from the impedance plot of the resonator 3 shown in FIG. 2 (e), the impedance of the resonator 3 becomes ∞ in the DC limit. Therefore, in a wide band of DC or more and less than _fosc , an unintended resonant circuit loop 20 occurs, as shown in the Smith chart of FIG. 2 (e). This is typically caused by parasitic capacitance between the two conductors, the inherent junction capacitance of the negative resistance element 1, the wiring of the bias supply section (not shown), the inductance of the wire, and the like. Since the loop 20 usually includes a parallel resonance point 21, parasitic low frequency oscillation is induced. However, since both ends of the negative resistance element 1 are two conductors having different potentials, a bias voltage can be applied.

FIG. 2C shows a distributed constant type resonator 4 having two conductors coupled with capacitance, which is a basic structure in the configuration of the present embodiment. FIG. 2C is a schematic view of the resonator 4 of the present embodiment and an impedance plot of the resonator 4, and FIG. 2F is a Smith chart of the resonator 4.

As shown in FIG. 2 (f), the impedance of the resonator 4 becomes ∞ in the DC limit, but the situation is close to that of the resonator 2 by changing the capacitance C → ∞ between the two conductors capacitively coupled. Become. To be precise, by sufficiently lowering the series resonance frequency f ₁ (hereinafter referred to as "frequency f ₁ ") formed by the inductance L ₁ of the two conductors and the capacitance C, the situation becomes close to that of the resonator 2. .. The frequency f ₁ is represented by the equation (1).
f ₁ = 1 / {2π√ (L ₁ C)} (1)
In reality, the capacitance C → ∞ is difficult, but it is possible to narrowerly limit the frequency domain that suppresses parasitic oscillation, which has been a problem, from DC or more and less than _fosc . That is, since the phases do not match in the band of f ₁ or more and less than _fosc , the frequency domain in which the parasitic oscillation should be suppressed can be narrowed down to DC or more and less than f ₁ . If a narrow frequency range, it becomes easy to control the loop 22 on the Smith chart in FIG. 2 (f), the it is easy to erase the resonance point of f less than ₁ or more DC. Of course, since two conductors having different potentials are arranged at both ends of the negative resistance element 1, a bias voltage can be applied.

Accordingly, in the resonator 4, as in FIG. 2 (c), it is set to be less than a predetermined frequency which has been predetermined frequency f _1. This can be achieved by adjusting the size of the capacitive coupling. The following describes how to set the configuration and the frequency f ₁ of the resonator 4 below a predetermined frequency.

FIG. 1 is a cross-sectional view of an oscillating element (semiconductor die) 100.

The oscillating element 100 has a negative resistance element 101 (hereinafter, “element 101”) and a resonator 110. The element 101 and the resonator 110 are arranged on the substrate (on the substrate 105). The resonator 110 is a first conductor 106 (hereinafter, “conductor 106”) including a first conductor layer 102 (hereinafter, “conductor layer 102”) and a second conductor layer 103 (hereinafter, “conductor layer 103”). ), And a second conductor 104 (hereinafter, “conductor 104”) including a third conductor layer. The element 101 is electrically connected to each of the conductor 106 and the conductor 104. The conductors 104 and 106 may have a plurality of conductor layers, or may be formed of one conductor each. Further, the conductor 104 and the substrate 105 may be integrated. Since the conductor 104 of the present embodiment is composed of only the third conductor layer, the conductor 104 may be referred to as the conductor layer 104 in the following description.

In the "resonator" of the present specification, of the conductor 104 and the conductor 106 arranged on the substrate 105, the portion forming the resonance region by the conductor 104 and the conductor 106 and the conductor 104 and the conductor 106 face each other. The portion that is capacitively coupled is referred to as a resonator 110. In the present embodiment, the resonator 110 has a portion forming a resonance region 108 in which electromagnetic waves resonate, and a portion in which the conductor layer 103 and the conductor layer 104 face each other. Therefore, the conductor 104 also exists outside the resonator 110. The conductor 106 and the conductor 104 and the substrate 105 may be hollow or may have a dielectric.

The substrate 105 is a conductive substrate, which is a semiconductor die cut to a size necessary and sufficient for producing the oscillating element 100. The substrate 105 is in contact with the conductor layer 104 on the substrate 105, and a bias voltage can be applied to the element 101. The conductor layer 102 and the conductor layer 103 are electrically short-circuited and connected, and a bias voltage is applied to the other pole of the element 101. The conductor layer 102 and the conductor layer 103 can be treated as one conductor having the same potential, and a bias voltage can be supplied to the other pole of the element 101.

The resonator 110 is formed with a resonance region 108 surrounded by the conductor 106 and the conductor 104. The resonance region 108 is hollow, and the terahertz wave oscillated from the element 101 resonates. The resonance region 108 is not limited to the cavity, and may be a distributed constant type resonator such as being filled with a dielectric. The resonator 110 has an inductance L ₁ between the conductor 106 and the conductor 104. The inductance L ₁ is an inductance mainly caused by a relatively elongated portion of the conductor layer 102 forming the side surface of the region 108.

In the present embodiment, the main inductance is generated in the conductor layer 102, but since the inductance is caused by the shape of the elongated conductor, if the conductor layers 103 and 104 have elongated portions, the inductance is the sum of the series. Therefore, since, collectively inductance L ₁ may be referred to as a "resonator inductance."

Further, the conductor 106 and the conductor 104 are capacitively coupled. Specifically, the conductor layers 103 and the conductor layers 104, which are arranged so as to face each other at intervals of 1 μm or less, are capacitively coupled. The resonator 110 of the present embodiment has a relatively large capacitance C between the conductor 106 and the conductor 104. Therefore, during the two poles of the element 101, and a inductance L ₁ and the capacitor C in series.

A typical element 101 oscillates at a parallel resonance frequency (oscillation frequency) _fosc (hereinafter, referred to as "frequency _fosc "). The structure that brings about this parallel resonance is the distributed constant type resonator 110. Further, the element 101 does not oscillate at the frequency f ₁ defined by the inductance L ₁ of the resonator 110 caused by the conductor layer 102 and the capacitance C between the conductor layer 103 and the conductor layer 104.

Impedance of the structure of the resonator 110, since repeating the parallel resonance and series resonance alternately on the frequency axis, the frequency f ₁ or more of the same structure, in less than a frequency f _osc, be theoretically phase matched Absent. In other words, oscillator 100 of the present embodiment, without attenuating the oscillation at the desired frequency _{f osc,} can suppress the parasitic oscillation frequency region of less than _{f 1} or _{f osc.} This is as described in the above description of FIG.

The frequency f ₁ is set to be less than the frequency of parasitic oscillation caused by wiring. This is because the structure associated with the bias supply unit that causes parasitic low-frequency oscillation can be roughly divided into two, the electric wire 132 from the power supply 131 to the oscillation element 100 and the wiring of the oscillation element 100. The "wiring" in the present specification is a conductor that passes between the bias voltage from the power supply 131 being supplied to the oscillating element 100 via the electric wire 132 and being supplied to the element 101. In this embodiment, it refers to conductors 104 and 106. In the present embodiment, the bias voltage from the power supply 131 is electrically connected to the oscillating element 100 via the electric wire 132, but the present invention is not limited to this, and the power supply 131 and the oscillating element 100 are electrically connected. It suffices to have a conductor to connect to the target. That is, it is not limited to the electric wire, and may be plate-shaped or the like.

The electric wire 132 mainly causes parasitic oscillation of 10 MHz or more due to an inductance proportional to its length or a distributed constant circuit. This is because the length of the electric wire 132 is several mm to several m, and the electric wire of several m becomes a distributed constant type resonator of about 10 MHz. The wiring causes parasitic oscillation due to an inductance proportional to the length of the wiring of the oscillating element 100 or a distributed constant circuit. For the frequency of parasitic oscillation, when the length of the wiring is approximated to the length of the diagonal line of the semiconductor die (board 105) having a size sufficient for manufacturing the oscillating element 100, the velocity C of the electromagnetic wave in the wiring is set to that of the semiconductor die 105. It is equal to or greater than the value (C / d) divided by the diagonal length d. The speed C of electromagnetic waves in wiring is slower than the speed C ₀ of electromagnetic waves in vacuum. The velocity C of the electromagnetic wave when the wiring is located on the surface of the semiconductor die 105 is expressed by Eq. (2), where the relative permittivity of the semiconductor die 105 is ε _r .

Thus, parasitic oscillation caused by the wiring for depends on the size of the oscillator 100, so that the parasitic distributed constant less than the resonant frequency that may occur in the semiconductor die 105, is appropriately determined frequencies f ₁ Is desirable. The frequency f ₁ is set so as to satisfy the equation (3).

For example, when the oscillating element 100 has a semiconductor die of 20 mm square or less, assuming that the length of the wiring is the same as the length of the diagonal line, the oscillator becomes a distributed constant resonator of about 3 GHz or more. Since an oscillating element that oscillates a terahertz wave typically has a semiconductor die of 20 mm square or less at the maximum, it is desirable to set the frequency f ₁ to less than 3 GHz.

If a structure in which the conductor 104 and the conductor 106 are capacitively coupled is used as the resonator 110 having a part of the bias supply unit and the frequency f ₁ is set to less than the frequency f ₂ , the latter parasitic oscillation cannot be oscillated. can do. Therefore, the frequency f ₁ may be set in any way as long as it is less than the frequency f ₂ . Further, when the conductor layer 103 and the conductor layer 104 are arranged so as to face each other on the substrate 105 as in the present embodiment, the capacity that can be integrated on the same substrate 105 is about 100 nF at the maximum. Therefore, it is desirable to select a frequency f ₁ of 100 MHz or higher. In this embodiment, a case where the frequency f1 is set to about ₁ GHz will be described as an example.

As a method to lower the frequency f _1, there is a method for increasing the capacitance C resonator 110 has. Although due to the shape of the structure and region 108 of the cavity 110, typically, for the lowest-order frequency _{f osc} of the following resonator 0.1 THz, if the capacitance C is more than 0.1 nF, the frequency About f ₁ = 1 GHz can be achieved. The frequency _{f osc} is 1nF or more for the following resonator 1 THz, if the frequency _{f osc} at least 10nF for following the resonator 10 THz, can achieve a frequency of about _f 1 = 1 GHz.

This is because the inductances of the two conductors 104 and 106 forming the resonator 110 are on the order of the reciprocal of the frequency _{fosc in} the SI unit system, although it depends on the configuration and shape thereof. That, 10 pH orders if the resonator 110 of the frequency _{f osc} is 0.1 THz, the frequency _{f osc} is 1pH order if the resonator 110 of 1 THz. Further, a 0.1pH order if the resonator 110 of the frequency _{f osc} is 10 THz, the capacitance C necessary depending on the parallel resonance frequency _{f osc} is different.

The oscillator 150 using the oscillator 100 of the present embodiment will be described with reference to FIG. FIG. 3 is a diagram illustrating the configuration of the oscillator 150. By using the oscillator 100, it is possible to suppress the parasitic oscillation of less than _{f 1} or _{f osc.} Furthermore, the oscillator 150 includes thereby suppressing the parasitic oscillation in _{f 1} than in the frequency domain or DC.

The oscillator 150 includes an oscillator 100 and a power supply circuit 130 (hereinafter, referred to as “circuit 130”). The circuit 130 has a power supply 131 for supplying a bias voltage to the oscillating element 100 and an electric wire 132, and may cause parasitic oscillation in a frequency region of DC or more and less than f ₁ , particularly 10 MHz or more and less than f ₁ . In this embodiment, the shunt rectifying diode (shunt element) 133 (hereinafter, shunt element) 133 (hereinafter, shunt element) is located at the position of the electric wire, which is 1/4 or less of the wavelength corresponding to f ₁ = 1 GHz, that is, 7.5 cm or less toward the power supply side when viewed from the oscillating element 100. , Called "diode 133"). In other words, the diode 133 and the electric wire 132 are connected between the power supply 131 and the oscillating element 100. The length of the electric wire 132 between the position where the electric wire 132 and the oscillating element 100 are connected and the diode 133 is 7.5 cm or less.

Using this method, it is possible to attenuate the amplitude of the parasitic oscillation in f ₁ than in the frequency domain or DC. Insertion of the diode 133 is equivalent to lowering the Q value of resonance at the parallel resonance point of the loop 120 on the Smith chart in FIG. 2 (f). As a result, the loop 120 moves to the left side on the chart. Therefore, in the oscillator 150 of the present embodiment, only the parasitic low frequency oscillation can be suppressed without attenuating the amplitude of the terahertz wave having a desired frequency _fosc . It is not limited to the rectifying diode, and any shunt element having a resistor or the like can have the same effect.

Alternatively, an oscillator (not shown) in which the oscillator 100 and the power supply circuit 130 are modularized using a short electric wire (not shown) of 7.5 cm or less, which is sufficiently smaller than 1/4 of the wavelength corresponding to f ₁ = 1 GHz, is configured. You may. Also in this method, parasitic low frequency oscillation can be suppressed without attenuating the amplitude of the terahertz wave having a desired frequency of _fosc . The method of suppressing parasitic oscillations in f ₁ than in the frequency domain or DC is not limited to this, it various methods are applied.

The oscillation element 100 of the first embodiment and the oscillator 150 using the oscillation element 100 have been described above. According to the oscillating element 100, it is possible to provide an oscillating element and an oscillator capable of narrowing the frequency range for attenuating parasitic oscillation as compared with the conventional case. In particular, oscillator 100, it is possible to suppress the parasitic oscillation of less than frequency _{f 1} or frequency _{f osc.} That is, if oscillators are structured by using the oscillator of the present embodiment, since the parasitic oscillation of less than f ₁ or f _osc is suppressed by the oscillator, band to attenuate the amplitude of the parasitic oscillation, such as by construction of the power supply circuit It gets narrower. Therefore, the suppression of parasitic oscillation becomes easy, and the attenuation of oscillation at a desired frequency _fosc can be further suppressed.

(Second Embodiment)
The oscillating element 200 according to this embodiment will be described with reference to FIG. FIG. 4 shows a cross-sectional view of the oscillating element (semiconductor die) 200.

The oscillating element 200 includes a negative resistance element 201 (hereinafter, “element 201”) and a resonator 210. The element 201 and the resonator 210 are arranged on the substrate 205. The resonator 210 is a first conductor 206 (hereinafter, “conductor 206”) including a first conductor layer 202 (hereinafter, “conductor layer 202”) and a second conductor layer 203 (hereinafter, “conductor layer 203”). ), And a second conductor 204 (hereinafter, “conductor 204”) including a third conductor layer. The element 201 is electrically connected to each of the conductors 204 and 206. In the following description, the conductor 204 may be referred to as the conductor layer 204.

The substrate 205 is in contact with the conductor layer 204. A bias voltage is supplied to one pole of the element 201 via the conductor 204. The conductor layer 202 and the conductor layer 203 are electrically short-circuited and connected to form another pole of the element 201. That is, the conductor layers 202 and 203 can be treated as one conductor having the same potential, and a bias voltage is supplied to the other pole of the element 201 via the conductor 206.

The resonator 210 is a distributed constant type resonator having a region 208 surrounded by the conductor layer 202, the conductor layer 203, and the conductor layer 204. The region 208 may be hollow or may be filled with a dielectric. Among the distributed constant type resonators, in the present embodiment, the element 201 sandwiched between the conductor layer 202 and the conductor layer 204 is a waveguide type resonator extending in a direction perpendicular to the paper surface of FIG. It is what forms.

Thus, without changing the product of the negative differential resistance _-R d junction capacitance Cd and element 201 of the element 201 _(R d> 0), the resonator due to conductive layer 202 inductance _{L 1} and the series resistance _{R s} Can be made smaller than that of the first embodiment. However, in the low frequency region where the entire structure of the oscillating element can be represented by a lumped constant element, it is difficult to satisfy the equation (4), which is a sufficient condition for amplitude attenuation, for the following reasons. Equation (4) can be obtained by solving the differential equation.
R _d > R _s > L / C _d R _d ... (4)
Although it depends on the type of the element 201, in a typical element of an oscillating element that oscillates a terahertz wave, the C _d R _d product that does not depend on the shape is about ^10-11 sec for the Esaki diode and the resonance tunnel diode ^10-12. It will be about sec. Further, the typical series resistance R _{s in} the case of the waveguide type resonator 210 is typically about 0.1 to 1 Ω. Thus, element 201 is the case of resonant tunneling diodes, the inductance _{L 1} of the resonator 210 ^{10 -12} H or ^{10 -13} H or less, ^{10 -11} H or ^{10 -12} H or less Esaki diode is required. However, it is not easy to design such a small inductance structure.

However, even if the equation (4) is not satisfied, the capacitance C between the opposing conductor layer 203 and the conductor layer 204 is decoupled, and the portion of the resonator 210 inside the element 201 as viewed from the element 201 is configured as described later. It should be.

The impedances of the conductor layers 202, 203, and 204 alternately repeat series resonance and parallel resonance on the frequency axis. Therefore, the frequency _{f 1} or more resonators 210, not be matching principle phase in less than the frequency _{f osc.} Therefore, in the frequency region of less than _{f 1} or _{f osc,} parasitic oscillation is suppressed. This is as described in the description with reference to FIG. 2 of the first embodiment. Therefore, in the frequency domain of DC or more and f ₁ or less, if the lumped constant circuit on the power supply side that does not include the resonator 210 is configured to satisfy the sufficient condition of the amplitude damping in the low frequency region, as a result, it is DC or more and less than _fosc. It is possible to suppress parasitic oscillation in the frequency range of.

In this embodiment sets the frequency _{f 1} below 200 MHz. This is because the upper limit of the frequency at which the power supply circuit can be regarded as a lumped constant element rather than as a distributed constant circuit is 200 MHz.

Since the value of the frequency f ₁ is low, capacitor C is required is relatively large. Since the resonator 210 of the present embodiment has a waveguide type structure extending in the direction perpendicular to the paper surface of FIG. 4, the conductor layer 203 and the conductor layer 204 are capacitively coupled by being arranged to face each other. It is easy to form a relatively large capacity C between the two, which is suitable for the present embodiment. Similar to the first embodiment, the capacity that can be integrated on the same substrate 205 is about 100 nF at the maximum, so it is more desirable that the frequency f ₁ is 100 MHz or more.

The oscillator 250 using the oscillator 200 of the present embodiment will be described with reference to FIG. FIG. 5 shows the configuration of the oscillator 250. The oscillator 250 has an oscillator element 200 and a power supply circuit 230 (hereinafter, referred to as “circuit 230”). Circuit 230 includes a power supply 231 and the wire 232, particularly above 10MHz frequency range of _{f 1} lower than the frequency range above DC, can cause parasitic oscillation.

Therefore, in the present embodiment, the shunt element is connected to the electric wire 232 that connects the oscillation element 200 and the power supply 231 to suppress the parasitic oscillation. In the wavelength region where the resonator 210 is considered to be sufficiently small, that is, in the low frequency region where the entire structure can be represented by a lumped constant element, if the sufficient condition of the amplitude attenuation of the equation (4) is extended, as shown in the equation (5). It is represented by.
R _d > R _s2 > L ₂ / (C _d + C) R _d ... (5)
Each series resistor _{R s} and the inductance _{L 1} of the resonator 210 due to the conductor layer 202 is replaced with a series resistor _{R s2} of the electric wire 232 and the inductance _{L 2.} The junction capacitance C _d of the element 201 is replaced by the sum of the junction capacitance C _d and the capacitance C, that is, C _d + C. Since the decoupling capacity C is overwhelmingly larger than C _d, it is approximated as C _d + C = C.

Here, the capacitance C can be adjusted by the area of the portion where the conductor layer 203 and the conductor layer 204 face each other. Further, the negative resistance −R _d (R _d > 0) can be adjusted depending on the width of the waveguide of the element 201. Therefore, depending on the design of the oscillating element 200, the CR _d product can be significantly delayed, such as ^10-7 sec. Further, when the electric wire 232 having the series resistance R _s2 of 0.1 Ω is used, the requirement for the inductance L _{2 of the} electric wire 232 is ^10-8 H or less. It is easy to design the electric wire 232 having such a series resistance R _s2 and an inductance L ₂ .

If the sufficient condition of the amplitude attenuation shown in the equation (5) is satisfied, the loop 22 on the Smith chart in FIG. 2 (f) becomes smaller and moves toward the left side or the lower side on the chart. That is, by forming the oscillator 250 with an oscillation device 200, a parasitic oscillation frequency lower than _{f 1} or _{f osc} is suppressed by the oscillator 200. Therefore, it is possible to provide an oscillating element and an oscillator whose frequency range for attenuating parasitic oscillation can be narrowed as compared with the conventional case.

In particular, oscillator 100, it is possible to suppress the parasitic oscillation of less than frequency _{f 1} or frequency _{f osc.} Therefore, if oscillators are structured by using the oscillator of the present embodiment, since the parasitic oscillation of less than f ₁ or f _osc is suppressed by the oscillator, band to attenuate the amplitude of the parasitic oscillation, such as by construction of the power supply circuit It gets narrower. Therefore, the suppression of parasitic oscillation becomes easy, and the attenuation of oscillation at a desired frequency _fosc can be further suppressed.

(Example 1)
The oscillator 350 according to the first embodiment will be described with reference to FIG. FIG. 6A shows a schematic view of the oscillator 350, and FIG. 6B shows a cross-sectional view of the oscillating element.

The oscillator 350 uses resonance tunnel diodes (RTDs: Ready-to-tunneling diodes) as the negative resistance elements 301a and 301b. In this embodiment, since two distributed constant type resonators 310 are used, two RTDs 301a and 301b (301a and 301b) are provided. The resonator 310 has a first conductor 306 (hereinafter referred to as "conductor 306") and a second conductor 304 (hereinafter referred to as "conductor 304"), respectively.

The first conductor layer 302, the second conductor layer 303, and the third conductor layer (second conductor) 304 are metal layers using Ti / Pd / Au metal having a thickness of 200 nm, respectively. Hereinafter, they are referred to as a metal layer (first metal layer) 302, a metal layer (second metal layer) 303, and a metal layer (third metal layer) 304, respectively. The substrate 305 is a conductive n-InP substrate. The conductor 304 is arranged with the metal layer 304 in contact with the substrate 305, and supplies a bias voltage to one pole of each of the RTDs 301a and 301b.

In the conductor 306 arranged to face the conductor 304, a dielectric benzocyclobutene (BCB: Benzocyclobutene) 32 is arranged between the metal layer 302 and the metal layer 303. Further, the metal layer 302 and the metal layer 303 are electrically short-circuited and connected at the BCB channel portion 315 to form another pole of the negative resistance element 301.

The two resonators 310 have a region formed by being surrounded by metal layers 302, 303, 304 and are separated by a metal filled in the BCB channel portion 315. Each region 308 of the resonator 310 becomes one distributed constant type resonance region, and the distance between the BCB channel portion 315 and the end of the metal layer 302 opposite to the BCB channel portion 315 (length of the resonator 310) l ₁ Correspondingly, an electromagnetic wave having a wavelength λ satisfying l ₁ = λ / 4 resonates.

The wavelength λ here is not the wavelength λ _{0 in} vacuum, but an effective wavelength that is affected by the dielectric constant of BCB 32 and the shapes of the metal layer 302 and the metal layer 303 to shorten the wavelength. Therefore, the metal layers 302 and 303 and the metal layer 304 surrounding the resonance region 308 have an inverted F antenna structure. The two resonators 310 have symmetrical shapes centered on the BCB channel portion 315, in order to cause the two resonators to vibrate coupled with each other at a certain oscillation frequency. The resonator 310 may use a resonator having a different shape, or may not have a resonator on one side.

In this embodiment, a silicon nitride (SiN) film 33 having a thickness of 100 nm is provided between the metal layer 303 and the metal layer 304 to form a capacitance C. Between the two poles of the RTD 301a and 301b, the frequency f ₁ (f ₁ = 1 /) due to the inductance L ₁ in the resonator 310 caused by the metal layer 302 and the capacitance C between the metal layer 303 and the metal layer 304. {2π√ (L ₁ C)}) exists. The frequency f ₁ can be adjusted to a frequency lower than a predetermined frequency by adjusting the area where the metal layer 303 and the metal layer 304 overlap.

Each of the RTDs 301a and 301b has a multiple quantum well structure composed of InGaAs / InAlAs and InGaAs / AlAs on the InP substrate 305. As the multiple quantum well structure, a triple barrier structure is used here. More specifically, it is composed of a semiconductor multilayer film structure of AlAs (1.3 nm) / InGaAs (7.6 nm) / InAlAs (2.6 nm) / InGaAs (5.6 nm) / AlAs (1.3 nm). Of these, InGaAs is a well layer, and lattice-matched InAlAs and unmatched AlAs are barrier layers. These layers are intentionally undoped layers that are not carrier-doped.

Such a multiple quantum well structure is sandwiched between an electrical contact layer made of n-InGaAs having an electron concentration of 2 × 10 ¹⁸ cm ^-3 . In the current-voltage I (V) characteristics of the structure between the electrical contacts, the peak current density is about 280 kA / cm ² , and the negative differential resistance region is from about 0.7 V to about 0.9 V. In this embodiment, a mesa structure having a diameter of 2 μmΦ is used as the RTD301a. A peak current of 10 mA and a negative differential resistance of -20 Ω can be obtained for each of the RTDs 301a and 301b.

The metal layer 302 is a terahertz wave resonance circuit using an inverted F antenna. In this embodiment, the length l ₁ of the resonator 310 designed to have a resonance frequency of about 0.6 THz is 75 μm, which is perpendicular to this. The length in the above direction (hereinafter referred to as the vertical direction) is 150 μm. The RTD301a is arranged outward by x = 40 μm in a direction orthogonal to the vertical direction (hereinafter referred to as a horizontal direction) from the BCB channel portion 315.

The thickness of BCB32 is about 3 μm in this example. Oscillation frequency _{f osc} is, RTD301a influence the terahertz wave band is increased, by the amount of reactance of 301b shifts from the resonance frequency in the resonance region 308, respectively, is about 0.5 THz. The dimensions described above are all design elements and can be appropriately changed according to the desired oscillation frequency _fosc , power, power efficiency, and the like.

The area where the metal layer 303 and the metal layer 304 overlap is about 0.75 mm × about 0.75 mm in this embodiment. When the capacitance C = about 0.33 nF between the metal layer 303 and the metal layer 304 is secured by using such means, the frequency f ₁ = about 3 GHz described above is obtained. In this way, the oscillator 350 using the oscillating element 300 can narrow down the frequency domain in which parasitic oscillation should be suppressed to DC or more and less than 3 GHz.

The oscillating element 300 can be manufactured by the following manufacturing method. First, a semiconductor multilayer film is epitaxially grown on an n-InP substrate 305 by a molecular beam epitaxy (MBE) method, a metalorganic vapor phase epitaxy (MOVPE) method, or the like. That is, n-InP / n-InGaAs, an RTD structure composed of InGaAs / InAlAs, and n-InGaAs are epitaxially grown in this order. Next, the RTD 301 is etched into a circular mesa shape. For etching, EB (electron beam) lithography and dry etching by ICP (inductively coupled plasma) are used. Photolithography may be used.

Subsequently, a Ti / Pd / Au metal film (third metal layer) 304 is formed on the etched surface by a lift-off method, and a silicon nitride film of 100 nm is formed by a sputtering method. If the film is formed by using the plasma CVD method, it is also suitable as a passivation for protecting the side wall in the RTD 301. Further, the Ti / Pd / Au metal film (second metal layer) 303 is formed by the lift-off method, and the capacity C is completed. The unnecessary silicon nitride film that does not contribute to the capacitance is removed by etching in the same pattern as the second metal layer 303.

After that, RTD301 is embedded in BCB32 as a dielectric and flattened until the outermost layer of RTD301 is exposed. Subsequently, only the BCB channel portion 315 is dry-etched to expose the Ti / Pd / Au metal film (second metal layer) 303. At this time, the BCB channel portion 315 becomes a forward taper shape when etching is performed while retreating the photoresist while mixing oxygen with the etching gas using dry etching. Finally, the Ti / Pd / Au metal film (first metal layer) 302 is formed by the lift-off method, and the dielectric BCB 32 is removed to complete the oscillating element 300 of this embodiment.

The oscillator 350 in which the power supply circuit 330 (hereinafter referred to as “circuit 330”) is connected to the oscillating element 300 may use the first embodiment. The power supply 331 included in the power supply circuit 330 prepares a voltage source of about 0.8 V so that the negative differential resistance region of the RTDs 301a and 301b can be biased from about 0.7 to about 0.9 V. The electric wires 332 may be connected to the metal layer 303 and the metal layer 304 by wire bonding or the like, respectively, and the connection location may be anywhere. For example, it may be connected near the end of the oscillating element (semiconductor die).

The shunt resistor (shunt element) 333 is arranged at a position within about 2.5 cm from the position where the electric wire 332 and the metal layer 303 or the metal layer 304 are connected toward the power supply 331 side. In order to make this easier, an integrated resistor may be provided between the metal layer 303 and the metal layer 304 near the connection position. If the resistance value of the shunt resistor 333 is about 10Ω or less, which is equal to or less than the absolute value of the combined resistance of about -10Ω of the two negative resistance elements 301a and 301b, parasitic oscillation in the frequency range of DC or more and less than 3GHz is performed. It can be surely suppressed.

The simulation calculation result of the oscillating element 300 is shown in FIG. FIG. 7A is a Smith chart in which the impedance of the resonator that also serves as the bias supply structure outside both ends of the RTDs 301a and 301b is plotted from about 1 GHz to about 1200 GHz. FIG. 7B is an enlarged view of the Smith chart of FIG. 7A in the vicinity of 1 GHz.

From FIGS. 7 (a) and 7 (b), it can be seen that the intersection of the straight line with zero reactance (imaginary part of impedance) of the Smith chart and the curve showing the frequency dependence of the impedance of the resonator 310 is 3 GHz. .. That is, it can be seen that the series resonance point of the resonator 310 has a frequency of about 3 GHz, and the next resonance point has a frequency of about 0.6 THz, which is the lowest next parallel resonance point of the resonator 310. The simulation is based on the three-dimensional finite element method Ansys high-frequency electromagnetic field simulator HFSS ver. 13 was used.

According to the oscillating element of this embodiment, the frequency range for attenuating the parasitic oscillation can be narrowed as compared with the conventional case.

(Example 2)
The oscillator 400 and the oscillator 450 according to the second embodiment will be described with reference to FIG. FIG. 8A shows a schematic view showing the oscillator 450, and FIG. 8B shows a cross-sectional view of the oscillating element 400.

The negative resistance element 401 of this embodiment is an RTD (hereinafter referred to as “RTD401”). The difference from the first embodiment is that the shape of the RTD 401 is not a mesa shape but a striped shape. In this embodiment, the distributed constant type resonator 410 is a gain waveguide type resonator along the longitudinal direction (hereinafter referred to as the longitudinal direction) of the striped element 401, and electromagnetic waves are applied to the resonator 410. , Nodes, antinodes, etc. are distributed as standing waves. Since the end face is an open end, the antinode of the electric field of the electromagnetic wave is standing at the open end.

Since both ends of the open end in this embodiment, the first metal layer (first conductive layer) 402 (hereinafter, referred to as "metal layer 402") the longitudinal length of the a l _3, the lowest-order resonance Is an electromagnetic wave having a wavelength satisfying l ₃ = λ / 2. The second-order resonance is an electromagnetic wave having a wavelength satisfying l ₃ = 2 × λ / 2, and the third-order resonance is an electromagnetic wave having a wavelength satisfying l ₃ = 3 × λ / 2, and the resonance frequency increases as the order increases.

Other components are the metal layer 402, the second metal layer (second conductor layer) 403 (hereinafter referred to as "metal layer 403"), and the third metal layer (third conductor layer) 404 ( Hereinafter, the n-InP substrate 405 (referred to as “metal layer 404”) is the same as in Example 1. The RTD 401 sandwiched between two metal layers may be used as in the second embodiment, and the substrate 405 at that time does not have to be a lattice matching system with the RTD 401.

The metal layer 402 and the metal layer 403 are connected by sandwiching the BCB 42 as a dielectric and electrically short-circuiting the BCB channel portions 415 and 416. The wavelength λ is not the wavelength λ _{0 in} vacuum, but an effective wavelength that is affected by the dielectric constant of BCB42 and the shapes of the metal layer 402 and the metal layer 403 to shorten the wavelength.

Since the resonator 410 is a gain waveguide type, it depends not only on the shape of the metal layer 402 and the metal layer 403 but also on the semiconductor multilayer film of the RTD 401, and has a relatively strong wavelength shortening effect. This is because the TM0 mode in the RTD 401, that is, the quasi-TEM mode is selected as the waveguide mode. In the case of this embodiment, the resonator 410 is filled with BCB 42 and closed with BCB channel portions 415 and 416. Therefore, in the region where the BCB 42 is filled, the magnetic field lines are in the TE mode having the component H _{z in the} vertical direction.

Therefore, the distance l ₂ between the RTD 401 and the BCB channel portions 415 and 416 is designed as follows. First, component E _x perpendicular to the above vertical and horizontal longitudinal component H _z and the electric flux lines of the magnetic force lines in the sidewall region BCB42 is filled (6), can be expressed as (7) .. Here, the longitudinal direction (longitudinal direction) of the negative resistance element 401 is the z direction, and the orthogonal direction (horizontal direction) is the y direction. (6), (7) is a longitudinal component _{H z} and component _{E x,} the imaginary unit j, the frequency of the electromagnetic wave omega, time vibration component due to the time t exp (j? T), y-direction of the wave Is β _y , and the wave number in the z direction is β _z , and is expressed in a general solution form with coefficients A and B.
H _z = Aexp (jωt + jβ _z z + jβ _y y) + Besp (jωt + jβ _z z-jβ _y y) ... (6)
Ex = ωμβ _y / (k ₀ ^2- β _z ² ) × {Aexp (jωt + jβ _z z + jβ _y y) + Besp (jωt + jβ _z z-jβ _y y)} ... (7)

In this embodiment, since the structure is symmetrical in the y direction, A = B. From equations (6) and (7), the wave number β _{y in the} y direction has a relationship with the wave number β _{z in the} z direction as shown in equation (8). β _y ² = k ₀ ^2- β _z ² ... (8)

In the case of this embodiment in which the wavelength shortening β _z / k ₀ is relatively large, the first term on the right side is sufficiently smaller than the second term, and thus can be expressed as in Eq. (9).
β _y ² = −β _z ² ... (9)

Therefore, the spread of the electromagnetic wave in the y direction is about 1 / β _z using the wave number in the z direction, so that it is possible to oscillate in the waveguide mode of β _z , which is desired by designing l ₂ > 1 / β _z. Become. In the case of the lowest-order resonance, l ₃ = λ / 2 = π / β _z , so a design in which l ₂ > l ₃ / π is desired. A typical aspect ratio is l ₂ : l = 1: 3 or less. In this way, since l ₂ cannot be designed to be small, the inductance L ₁ has a lower limit, but the frequency f ₁ can be adjusted by adjusting the area where the metal layer 403 and the metal layer 404 overlap. it can.

In this embodiment, a gain-guided oscillator with an oscillation frequency of 0.3 THz is designed, the length of the first metal layer 402 in the vertical direction is l ₃ = about 100 μm, and the distance between the RTD 401 and the BCB channel portions 415 and 416. l ₂ = about 35 μm. Since power is supplied from two locations from the BCB channel portions 415 and 416, the counts of both the inductance L ₁ and the series resistance R _{s of} the resonator 410 due to the metal layer 402 are halved. ^However, there remains the inductance _{L 1} of the order of ^{10 -11} H.

Therefore, the area where the second metal layer 403 and the third metal layer 404 constituting the resonator 410 overlap is adjusted, and the capacitance C is designed so as to secure, for example, 10 nF. Specifically, a Hafnia film 43 having an overlapping area of about 0.1 × about 5 mm and a thickness of about 20 nm is used. When securing the capacitance C by using such means, the frequency _{f 1} becomes 500 MHz. In this way, by using the oscillating element 400, the band in which the parasitic oscillation should be suppressed can be narrowed down to the frequency region of DC or more and f ₁ = less than 500 MHz.

The second embodiment may be used for the oscillator 450 in which the power supply circuit is connected to the oscillating element 400. The power supply 431 prepares a voltage source of about 0.8 V so that the negative resistance region of the RTD 401 can be biased from about 0.7 to 0.9 V. Wire 432, using the relatively thick wire 0.1Ω series resistance _{R s2} bake, a request to the inductance _{L 2} and ^{10 -8} H below. Therefore, the electric wire 432 and the oscillating element 400 may be connected by using wire bonding or the like of about ^10-9 H, and of course, a plurality of electric wires 432 may be used in parallel.

By doing so, it is possible to suppress parasitic oscillation in a frequency region of DC or more and f ₁ = less than 500 MHz. Further, a power supply circuit using a shunt element as in the first embodiment may be used.

According to the oscillating element of this embodiment, the frequency range for attenuating the parasitic oscillation can be narrowed as compared with the conventional case.

Although the preferred embodiments of the present invention have been described above, the present invention is not limited to these embodiments, and various modifications and modifications can be made within the scope of the gist thereof.

For example, an imaging system may be configured that includes the oscillating elements of the above-described embodiments and examples, an oscillator using the same, and an image forming apparatus that detects millimeter waves and terahertz waves as images. In the frequency domain from the millimeter wave band to the terahertz band, unlike the infrared region, the energy of background blackbody radiation is small, so active illumination by such an oscillator or oscillator is usually used.

The electromagnetic wave image forming apparatus may be an electronic device having a rectifying element such as a Schottky barrier diode or FET, or a heat conversion device such as a microbolometer, a pyrodetector, or a Golay cell. The terahertz wave that illuminates the subject and has the transmitted or reflected information of the subject is acquired by the image forming apparatus. At that time, if an objective lens is provided between the image forming apparatus and the subject, the present imaging system becomes a focal plane array type and can perform imaging with one shot.

The oscillating elements of the above-described embodiments and examples and the oscillator using the oscillators can be used for manufacturing control, medical image diagnosis, safety management, etc. as an irradiation means for performing active illumination by millimeter waves and terahertz waves, or at ultra-high speed. It can be expected to be applied as a transmission means for communication devices.

101 Negative resistance element 104 Second conductor 106 First conductor 110 Resonator

Claims (12)

An oscillating element that oscillates electromagnetic waves
It has a negative resistance element and a resonator,
The resonator has a first conductor layer, a second conductor layer, and a third conductor layer.
The first conductor layer and the second conductor layer are short-circuited via a connecting portion .
The negative resistance element is connected to the first conductor layer and the third conductor layer.
The second conductor layer and the third conductor layer are arranged so as to face each other.
A dielectric is arranged between the second conductor layer and the third conductor layer.
In a plan view, the connecting portion, the first conductor layer, the second conductor layer, and the first region in which the third conductor layer is overlapped are arranged .
In a plan view, the first conductor layer that does not overlap with the connection portion, the second conductor layer that does not overlap with the connection portion, and the third conductor layer that does not overlap with the connection portion are arranged overlapping. An oscillating element characterized by having a second region, which is defined as. A claim characterized in that, in a plan view, the second conductor layer and the third conductor layer overlap, and the first conductor layer and the second conductor layer do not overlap each other. Item 1. The oscillating element according to item 1. The negative resistance element and the resonator are arranged on a substrate.
The series resonance frequency f ₁ of the resonator has a capacitance between the second conductor layer and the third conductor layer C, the first conductor layer, the second conductor layer, and the third conductor layer. Assuming that the inductance of the conductor layer is L ₁ , the velocity of the oscillated electromagnetic wave in vacuum is C ₀ , the relative permittivity of the substrate is ε _r , and the diagonal length of the substrate is d.


The oscillating element according to claim 1 or 2 , wherein the oscillating element satisfies. The oscillation frequency _{f osc} of the electromagnetic wave is not more than 0.1 THz,
The oscillating element according to any one of claims 1 to 3 , wherein the capacitance C between the second conductor layer and the third conductor layer is 0.1 nF or more. The oscillation frequency _{fosc of the} electromagnetic wave is 1 THz or less.
The oscillating element according to any one of claims 1 to 3 , wherein the capacitance C between the second conductor layer and the third conductor layer is 1 nF or more. The oscillation frequency _{f osc} of the electromagnetic wave is not more than 10 THz,
The oscillating element according to any one of claims 1 to 3 , wherein the capacitance C between the second conductor layer and the third conductor layer is 10 nF or more. The oscillating element according to any one of claims 1 to 6 .
It has a power supply circuit that supplies a bias voltage to the oscillating element.
The power supply circuit is an oscillator including a power supply and a conductor that electrically connects the power supply and the oscillating element. The oscillator according to claim 7 , wherein the conductor that electrically connects the power source and the oscillating element is an electric wire. Has a shunt element
The shunt element is connected to the electric wire between the power supply and the oscillating element.
Wherein said length of wire between the oscillator and the shunt element, oscillator according to claim 8, wherein the at 1/4 or less of the wavelength corresponding to the series resonance frequency f ₁ of the oscillating element .. In the power supply circuit, assuming that the negative resistance of the negative resistance element of the oscillating element is −R _d , the series resistance of the electric wire is R _s2 , and the inductance of the electric wire is L ₂ , then R _d > R _s2 > L ₂ The oscillator according to claim 8 or 9 , wherein the oscillator satisfies / CR _d . The oscillating element according to any one of claims 1 to 6 .
An imaging system comprising: a detecting means for detecting an electromagnetic wave from a subject irradiated with the electromagnetic wave oscillated from the oscillating element. The imaging system according to claim 11 , wherein the detection means is an image forming apparatus that detects electromagnetic waves from the subject as an image.