Semiconductor device, communication apparatus, imaging system, and radar apparatus

Abstract

The present invention comprises: a reference signal oscillator that generates a first signal having a first frequency; a modulation unit that generates a third signal having a third frequency from the first signal and from a second signal having a second frequency corresponding to a signal supplied from a baseband circuit; and an injection synchronous oscillator that generates a fourth signal having a fourth frequency in the terahertz band in synchronization with the third signal.

Description

Semiconductor devices, communication equipment, imaging systems, and radar equipment 【0001】 This disclosure relates to semiconductor devices, communication equipment, imaging systems, and radar equipment. 【0002】 Development of oscillators and detectors for wireless communication systems using terahertz waves is underway. Patent document 1 shows a phased array radio that transmits and receives signals at frequencies of 300 GHz or higher. 【0003】 Japanese Patent Publication No. 2023-010120 【0004】 M. Asada et. al. , “Resonant Tunneling Diodes for Sub-Terahertz and Terahertz Oscillators”, Jpn. J. Appl. Phys. , Vol. 47, No. 6 (2008), pp. 4375-4384M. Asada et. al. ,”Theoretical analysis of coupled oscillator array using resonant tunneling diodes in subterahertz and terahertz range”, J. Appl. Phys. , Vol. 103, 124514 (2008) 【0005】 Generally, as the frequency increases, the wavelength decreases, resulting in a smaller antenna size, but the circuit size increases due to increased harmonic loss. In phased array antennas, the spacing between antennas must be less than the wavelength to suppress grating lobes. To achieve miniaturization and arraying of antennas, miniaturization of the oscillation and modulation circuits is required. 【0006】 This disclosure aims to provide technology advantageous for communication using terahertz waves. 【0007】In view of the above issues, a semiconductor device according to an embodiment of the present disclosure is characterized by comprising: a reference signal oscillator that generates a first signal of a first frequency; a modulation unit that generates a third signal of a third frequency from the first signal and a second signal of a second frequency corresponding to a signal supplied from a baseband circuit; and an injection-synchronous oscillator that generates a fourth signal of a fourth frequency in the terahertz band in synchronization with the third signal. 【0008】 This disclosure provides a technology that is advantageous for communication using terahertz waves. 【0009】 Other features and advantages of the technical ideas derived from this disclosure will become apparent from the following description with reference to the attached drawings. In the attached drawings, the same or similar components are given the same reference numeral. 【0010】The attached drawings are included in the specification and constitute a part thereof, illustrating embodiments in this disclosure and used to explain the technical ideas derived from this disclosure together with their descriptions. Diagram showing an example configuration of the semiconductor device of this embodiment. Diagram showing an example configuration of the semiconductor device of this embodiment. Diagram showing an example configuration of the semiconductor device of this embodiment. Diagram showing an example configuration of the semiconductor device of this embodiment. Diagram showing an example of the injection synchronization characteristics of the RTD oscillator of this embodiment. Diagram showing an example of the injection synchronization characteristics of the RTD oscillator of this embodiment. Diagram showing an example of the configuration of the semiconductor device of this embodiment. Top view showing an example configuration of the semiconductor device of this embodiment. Cross-sectional view showing an example configuration of the semiconductor device of this embodiment. Cross-sectional view showing an example configuration of the semiconductor device of this embodiment. Diagram showing an example configuration of the filter circuit of this embodiment. Diagram showing an example configuration of the filter circuit of this embodiment. Diagram showing an example configuration of the filter circuit of this embodiment. Diagram showing an example configuration of the filter circuit of this embodiment. Diagram showing an example configuration of the filter circuit of this embodiment. Diagram showing an example of the equivalent circuit of the semiconductor device of this embodiment. Diagram showing an example configuration of the semiconductor device of this embodiment. Diagram showing an example configuration of the semiconductor device of this embodiment. Diagram showing an example of the equivalent circuit of the semiconductor device of this embodiment. A diagram showing an example of the equivalent circuit of the semiconductor device of this embodiment. A diagram showing an example of the configuration of the semiconductor device of this embodiment. A diagram showing an example of the configuration of the semiconductor device of this embodiment. A diagram showing an example of the configuration of the semiconductor device of this embodiment. A diagram showing an example of the configuration of the semiconductor device of this embodiment. A diagram showing an example of the configuration of the semiconductor device of this embodiment. A diagram showing an example of the configuration of the semiconductor device of this embodiment. A diagram showing an example of the configuration of the semiconductor device of this embodiment. A diagram showing an example of the configuration of the semiconductor device of this embodiment. A diagram showing an example of the configuration of an imaging system using the semiconductor device of this embodiment. A diagram showing an example of the configuration of a communication system using the semiconductor device of this embodiment. 【0011】 The embodiments will be described in detail below with reference to the attached drawings. Note that the following embodiments do not limit the scope of the claims. While the embodiments describe multiple features, not all of these features are necessary, and the features may be combined in any way. Furthermore, in the attached drawings, identical or similar configurations are given the same reference numerals, and redundant descriptions are omitted. 【0012】 A semiconductor device according to an embodiment of the present disclosure will be described with reference to Figures 1A to 11B. The following description will focus on the case where the semiconductor device is used as a transmitter in a communication system (communication device). However, it is not limited to this, and the semiconductor device of this embodiment can also be used as a receiver in a communication system (communication device). Furthermore, in this disclosure, terahertz waves refer to electromagnetic waves in the frequency range of 100 GHz or higher and 10 THz or lower. In addition, in this disclosure, terahertz waves may refer to electromagnetic waves in the frequency range of 300 GHz or higher and 3 THz or lower. In the description of each embodiment, the description of components that are the same as those in other embodiments may be omitted. Furthermore, embodiments can be modified or combined with other embodiments as appropriate. 【0013】 The semiconductor device 100 of the first embodiment of this disclosure will be described using Figures 1A to 5. Figure 1A is a conceptual diagram showing the wiring of the functional block of the semiconductor device 100 of this embodiment. Figure 1B is a diagram showing an overview of the resonant tunnel diode (RTD) resonant section 102 in the semiconductor device 100 of this embodiment. Figure 1C is a diagram showing a further overview of the functional block in the semiconductor device 100 of this embodiment. 【0014】 As shown in Figure 1A, the semiconductor device 100 comprises a reference signal oscillator 101, an RTD resonant section 102, a mixer 103, and an antenna 104. The reference signal oscillator 101 generates a first signal (reference signal) at a first frequency (frequency f1). The semiconductor device 100 is configured to radiate or receive terahertz wave TW from the antenna 104. The terahertz wave TW is generated on the semiconductor device 100 and is a wireless signal for wireless communication with a transceiver (not shown). 【0015】In the conceptual diagram of FIG. 1A, the functional blocks described above are schematically represented, and the form of connection between each component is shown. The semiconductor device 100 has a semiconductor structure for generating a terahertz wave TW, and as this semiconductor structure, an RTD oscillator 110 including a resonant tunneling diode (RTD) is used. The semiconductor device 100 includes an RTD resonance section 102 that combines the RTD oscillator 110 and a resonance structure. The RTD resonance section 102 is an injection synchronization oscillator connected to the reference signal oscillator 101 via a mixer 103. The RTD resonance section 102 receives the supply of a third signal (RF signal 106) having a third frequency (frequency f3) in the terahertz band. The RF signal 106 having the frequency f3 in the terahertz band is a signal obtained by mixing (mixing) the reference signal 108 having the frequency f1 supplied from the reference signal oscillator 101 and the second signal (data signal 105) having the second frequency (frequency f2) in the mixer 103. The RTD resonance section 102 is an injection synchronization oscillator, oscillates at the fourth frequency (frequency f4) in the terahertz band in synchronization with the supplied RF signal 106, and outputs a fourth signal (THz signal 107) having the frequency f4. An antenna 104 is connected to the RTD resonance section 102, and the THz signal 107 output from the RTD resonance section 102 is radiated into space as the terahertz wave TW from the antenna 104. When the semiconductor device 100 is used as a receiver, the terahertz wave TW is incident from the antenna 104. The fourth frequency can be a frequency higher than the third frequency (frequency f4 > frequency f3), and the third frequency can be a frequency equal to or higher than the first frequency (frequency f3 ≥ frequency f1). 【0016】 The mixer 103 is connected to an IF connection port 121 for inputting and outputting a data signal 105 that is an Intermediate Frequency (IF) signal having the frequency f2. The data signal 105 can be supplied from, for example, a baseband circuit not shown or to the baseband circuit. Further, the mixer 103 is connected to the reference signal oscillator 101 and the RTD resonance section 102. In the present embodiment, the mixer 103 is arranged as a modulation / demodulation section (when used as a transmitter, it can be a modulation section, and when used as a receiver, it can be a demodulation section). 【0017】Mixer 103 has the function of mixing signals of different frequencies. Specifically, mixer 103 mixes the frequencies of a reference signal 108, which is a local oscillator (LO) signal, and a data signal 105 to synthesize an RF signal 106 with a frequency f3 in the terahertz band, and outputs it. When semiconductor device 100 functions as a transmitter, mixer 103 receives input from the reference signal 108 and the data signal 105, and outputs the RF signal 106. When semiconductor device 100 functions as a receiver, mixer 103 receives input from the reference signal 108 and the RF signal 106, and outputs the data signal 105. 【0018】 When the semiconductor device 100 functions as a transmitter, the antenna 104 radiates the THz signal 107 output from the mixer 103 into space as a terahertz wave TW. When the semiconductor device 100 functions as a receiver, the antenna 104 receives the terahertz wave TW propagating through space, and the signal received by the antenna 104 is supplied to the mixer 103 as the THz signal 107 (RF signal 106). 【0019】Using FIG. 1B, the RTD resonance section 102 included in the semiconductor device 100 will be described. The RTD oscillator 110 arranged in the RTD resonance section 102 is a negative resistance element having a negative resistance, and is used as a high-frequency source of the RTD resonance section 102. The RTD resonance section 102 includes a resonance conductor 111 and a ground (GND) conductor 116 (details will be described later using FIG. 3B, etc.) that form a resonance structure with the RTD oscillator 110. An RF signal 106 with a frequency f3 is input to the RTD resonance section 102 from the mixer 103. The RTD resonance section 102 oscillates self-excitedly at a frequency f4 even when alone, but uses the RF signal 106 as a master and operates the RTD resonance section 102 as a slave. As a result, since the RTD resonance section 102 oscillates in a state of being injection-locked to the RF signal 106, the oscillation frequency and phase noise of the RTD resonance section 102, which is the slave, follow the accuracy of the RF signal 106, which is the master. Thus, the RTD resonance section 102 acts as an injection-locked oscillator. By this method, it is possible to stabilize the oscillation frequency and reduce the phase noise of the RTD resonance section 102. The reference signal oscillator 101 is a wave source for synchronizing the oscillation timing at the frequency f4 in the terahertz band of the RTD resonance section 102. Therefore, the RF signal 106 with a frequency f3 in the terahertz band obtained by mixing (mixing) the frequency f1 of the reference signal 108 and the frequency f2 of the data signal 105 may be a subharmonic frequency that is 1 / N (N is a natural number) times the oscillation frequency f4 of the terahertz wave of the RTD resonance section 102. In the present embodiment, the frequency f3 will be described as being 1 / 2 (N = 2) of the frequency f4. 【0020】 The power injected from the reference signal oscillator 101 into the RTD resonance section 102 via the mixer 103 is the output P of one RTD oscillator ＲＴＤIt may be equal to or greater than ((3 / 16) cos ωτ × ΔIΔV). Here, ω is the angular frequency of the frequency f2 in the terahertz band, and τ is the carrier transit time in the semiconductor layer constituting the RTD oscillator 110. Also, ΔI and ΔV are the current difference and voltage difference between the current peak and current valley in the negative resistance region of the RTD oscillator 110, respectively. There is no limit to the number of RTD oscillators 110 arranged in the RTD resonance section 102, and the power obtained by adding up the number of arranged RTD oscillators 110 may be injected from the reference signal oscillator 101 into the RTD resonance section 102. Therefore, the output of the reference signal oscillator 101 is (the number of RTD oscillators 110 arranged in the RTD resonance section 102) × P ＲＴＤ In addition, the output may be greater than the transmission loss from the reference signal oscillator 101 to each RTD oscillator 110 arranged in the RTD resonance section 102. In this embodiment, the number of RTD oscillators 110 arranged in the RTD resonance section 102 will be described as one. 【0021】 Also, the output from the reference signal oscillator 101 is P ＲＴＤ It is also possible to lock even when it is smaller than. Specifically, even if the injected signal is a small signal about one ten-thousandth of the output of one RTD oscillator 110, the RTD resonance section 102 may synchronize with the signal supplied from the signal of the reference signal oscillator 101. Specifically, injection synchronization is possible with a small signal up to about r = 5 × 10 －４ . Here, r is called the injection ratio and is represented by V ｉｎｊ / V ｏｓｃ . The injection ratio r is the voltage amplitude ratio between the voltage amplitude (V ｉｎｊ ) of the injection signal and the voltage amplitude (V ｏｓｃ ) of the oscillation signal. When the injection ratio r becomes smaller, the effect of reducing phase noise by injection synchronization weakens, but it can be adjusted according to the degree of phase noise required for the semiconductor device 100. Therefore, even when a reference signal oscillator 101 having an output smaller than P ＲＴＤ ((3 / 16) cos ωτ × ΔIΔV) is used in the circuit to which the present disclosure is applied, it is possible to synchronize the RTD resonance section 102 while maintaining the frequency accuracy and the like output from the reference signal oscillator 101. Therefore, the signal intensity output from the reference signal oscillator 101 is P ＲＴＤIt can be set without being heavily dependent on the value of [the variable]. 【0022】 A bias voltage is supplied to the RTD resonant section 102 from a bias section (not shown). The bias section supplies power to the bias pattern 112, and the bias voltage is supplied to the resonant conductor 111 via a bias supply via 113. The bias section supplies the power necessary to drive the RTD oscillator 110 and adjusts the bias voltage applied to the RTD oscillator 110. The bias voltage is selected from a voltage that falls within the differential negative resistance region of the RTD oscillator 110 and is applied from the bias section via the bias pattern 112. An AC shunt 115 may also be connected to the bias pattern 112. The AC shunt 115 acts as a filter that shorts signals in a specific frequency band by connecting to GND (e.g., GND conductor 116) using resistance and capacitance. The AC shunt 115 can be appropriately placed in various locations on the bias pattern 112 or on the bias section to suppress parasitic oscillations, which are oscillations of the RTD resonant section 102 at frequencies other than a predetermined frequency. 【0023】 The AC shunt 115 can be any filter that suppresses parasitic oscillations, and other configurations such as a DC shunt may be used instead of the AC shunt 115. Also, the bias power supply that drives the RTD, which is applied from the bias section to the RTD oscillator 110, is DC. Therefore, the bias section and the bias pattern 112 may be directly connected, or they may be connected via a low-pass filter (LPF) or the like. 【0024】The RTD resonant section 102 may be configured to include a microstrip line having an inherent impedance for oscillation in the terahertz band, which is the RTD oscillator 110. The resonant structure of the RTD oscillator 110 using the microstrip line may consist of a resonant conductor 111, which is a linear upper conductor, a GND conductor 116, which is a broad lower conductor, and a dielectric 141 placed between the resonant conductor 111 and the GND conductor 116. The detailed layer configuration will be described later in conjunction with the explanation of Figure 3B. The characteristic impedance of the resonant conductor 111 depends on the thickness and material of the dielectric 141. Typically, the thickness of the dielectric 141 is designed to be sufficiently thin compared to the self-oscillation wavelength of the RTD resonant section 102, and it is known that the thicker the dielectric, the higher the characteristic impedance, and the thinner the dielectric, the lower the characteristic impedance. It is also known that the smaller the dielectric constant of the dielectric material 141, the higher the characteristic impedance, and the larger the dielectric constant, the lower the characteristic impedance. 【0025】 The frequency f4 of the terahertz wave emitted from the RTD resonant section 102 can be determined as the resonant frequency of the all-parallel resonant circuit, which combines the resonant conductor 111 and the reactance of the RTD oscillator 110. Specifically, based on the description of the equivalent circuit of the oscillator in Non-Patent Literature 1, the frequency that satisfies the amplitude condition of equation (1) and the phase condition of equation (2) for the resonant circuit, which combines the admittance of the RTD oscillator 110 and the resonant conductor 111, can be determined as the oscillation frequency f4. Re[YRTD] + Re[YOSC] ≤ 0 ... (1) Im[YRTD] + Im[YOSC] = 0 ... (2) Here, Re[YRTD] is the admittance real part of the RTD oscillator 110 and has a negative value. Re[YOSC] represents the admittance real part of the resonant conductor 111. Im[YRTD] represents the imaginary admittance part of the RTD oscillator 110. Im[YOSC] represents the imaginary admittance part of the resonant conductor 111. 【0026】 The resonant conductor 111 has a wire length (resonator length) indicated by arrow 118 in Figure 1B that corresponds to the effective wavelength λ of the LO signal oscillating within the RTD resonant section 102. ｆ４ It is set to be a multiple of 1 / 2 for λ. ｆ４λ is the effective wavelength of the terahertz wave oscillating in the RTD resonant section 102 in the dielectric 141, and the wavelength of the terahertz wave in vacuum is λ ０ The relative permittivity of dielectric 141 is ε ｒ１ Then λ ｆ４ = λ ０ ×ε ｒ１ －１／２ It is represented as follows: For example, in this embodiment, λ ｆ４ Its frequency is 0.35 THz, and its wavelength in a vacuum is λ ０ It is 0.86 mm. Since the relative permittivity on the resonator is about 2, λ ｆ４ The electrical length of / 2 is 0.215 mm. The resonant conductor 111 can be designed based on this wire length. 【0027】 As a result, the resonant conductor 111 resonates at a frequency in the terahertz band, and a resonant electric field is generated as a standing wave within the resonant conductor 111. The resonant electric field is generated along the direction of arrow 118, with antinodes (points of maximum amplitude) at both ends of the resonant conductor 111 and nodes (points of zero amplitude) in the center of the resonant conductor 111. 【0028】 Furthermore, multiple RTD oscillators 110 are λ ｆ４ When arranged in a resonant conductor 111 of λ / 2, the RTD oscillators are positioned opposite each other with respect to the nodes of the resonant electric field. For example, when two RTD oscillators 110 are mutually injection-synchronized with their phases inverted (opposite phase), the two RTD oscillators oscillate in push-pull mode. Also, when two RTD oscillators 110 are mutually injection-synchronized with their phases in the same phase, the two RTD oscillators oscillate in push-push mode. Specifically, the self-oscillating frequency f4 of the RTD resonant section 102 can be determined by considering the mutual injection synchronization in a configuration in which two separate RTD resonant sections are coupled, as disclosed in Non-Patent Literature 2. The thickness of the dielectric 141 is, for example, λ ｆ４ / 10 or more and λ ｆ４ It may be less than or equal to 3. 【0029】The resonant conductor 111 may include an open stub 114 extending in the direction of arrow 117 shown in Figure 1B. One end of the open stub 114 is connected to the central part of the resonant conductor 111, and the other end is open. The length of the open stub 114 indicated by arrow 117 is equal to the effective wavelength λ of the THz signal 107. ｆ４ It is composed of 4 elements and plays the role of fixing the nodes of the standing waves oscillating within the resonant conductor 111 to the central part of the resonant conductor 111. 【0030】 Furthermore, the width of the connection portion of the bias pattern 112 connected to the via 113 for bias supply is smaller (narrower) than the width in the direction of arrow 118 where the standing wave of the resonant conductor 111 is generated. Also, this width is equal to the effective wavelength λ of the terahertz band THz signal 107 that is present in the resonant conductor 111. ｆ４ Less than 1 / 10 of (λ ｆ４ It may be less than or equal to 10. This is because arranging the bias supply via 113 and bias pattern 112 to a size and position that does not interfere with the resonant electric field in the resonant conductor 111 is suitable for improving the resonance efficiency. In addition, the bias supply via 113 is located in the center, which is a node of the standing wave in the resonant conductor 111. By making the connection point of the bias supply via 113 a node of the standing wave in the resonant conductor 111, the impedance of the oscillation signal at the connection point is maximized. Therefore, losses due to leakage of the oscillation signal from the resonant conductor 111 to the bias line from the bias supply via 113 can be minimized. The bias section is adjusted as appropriate to efficiently synchronize the reference signal oscillator 101 and the RTD resonant section 102. 【0031】 Furthermore, the RTD resonant section 102 self-oscillates at terahertz frequencies due to the wiring impedance determined by the structure of the RTD oscillator 110, the resonant conductor 111, and the open stub 114. This stub structure may also be a spiral inductor or an interdigital capacitor structure made of microstrip lines, and self-oscillation can be performed at any desired frequency by appropriately designing the inductance and capacitance. 【0032】The connection relationships of the semiconductor device 100 will be explained using Figure 1C. In the configuration shown in Figure 1C, the reference signal oscillator 101, together with the phase adjuster 145, constitutes the synchronization signal source 144. However, this is not the only configuration, and the phase adjuster 145 may be omitted. The phase adjuster 145 has the function of adjusting the phase of the reference signal 108 output from the reference signal oscillator 101, and the phase-adjusted reference signal 108 is output from the synchronization signal source 144. The synchronization signal source 144 (reference signal oscillator 101) is electrically connected to the mixer 103 via a filter 146 (first filter) that allows signals of a predetermined bandwidth to pass through. 【0033】 The reference signal 108 output from the synchronization signal source 144 (reference signal oscillator 101) is supplied to the mixer 103 through the filter 146. The frequency f1 of the reference signal 108, which is the LO signal, is mixed with the data signal 105 with frequency f2 in the mixer 103 and modulated to a frequency f3 in the terahertz band. The modulated signal is output from the mixer 103 as an RF signal 106. The mixer 103, which functions as a modulation unit when the semiconductor device 100 is used as a transmitter, is electrically connected to the RTD resonant unit 102 via a filter 147 (second filter) that allows signals of a predetermined bandwidth to pass through. The RF signal 106 output from the mixer 103 has its input to the synchronization signal source 144 (reference signal oscillator) reduced by the filter 146 and is supplied to the RTD resonant unit 102 through the filter 147. 【0034】 The frequency f3 of the RF signal 106 is upconverted to a terahertz frequency f4 by injection-synchronous oscillation of the RTD resonator 102. The upconverted signal is output from the RTD resonator 102 as a THz signal 107. The RTD resonator 102 is connected to the antenna 104 via a filter 148 (third filter) that allows signals of a predetermined frequency band to pass through. The THz signal 107 output from the RTD resonator 102 has its input to the mixer 103 suppressed by the filter 147 and is supplied to the antenna 104 through the filter 148. 【0035】Filters 146-148 may also be called low-frequency band filters (LPF), intermediate-frequency band filters (MPF), or high-frequency band filters (HPF), depending on the frequency range of the signal they can pass through. MPFs and similar filters may also be called Band-Pass Filters (BPF). 【0036】 Next, injection synchronization between the reference signal oscillator 101 and the RTD resonant section 102 will be explained using Figures 2A to 2C. Figure 2A is a graph with the bias voltage applied to the RTD oscillator 110 on the horizontal axis and the frequency oscillated by the RTD resonant section 102 on the vertical axis. The process of supplying a reference signal from the reference signal oscillator 101 to the RTD resonant section 102 via the mixer 103 and stabilizing the frequency of the RTD resonant section 102 is called injection synchronization. In this embodiment, the RTD resonant section 102 utilizes the harmonic component of the RF signal 106 signal with a frequency f4 in the terahertz band, which is obtained by mixing the frequency f1 of the reference signal 108 output by the reference signal oscillator 101 and the frequency f2 of the data signal 105. The solid line 201 in the graph plots the correlation between the oscillation frequency of the RTD resonant section 102 and the bias voltage when a reference signal 108 is supplied from the reference signal oscillator 101, i.e., when injection synchronization is performed. In this embodiment, a reference signal of 0.175 THz, which is the frequency of the subharmonic half of 0.35 THz, is supplied to the RTD resonant section 102 from the reference signal oscillator 101 via the mixer 103. The dashed line 202 shown in Figure 2A plots the correlation between the oscillation frequency and the bias voltage in the RTD resonant section 102 when injection synchronization is not performed. Here, the frequencies described in this embodiment are just examples, and the frequency of the signal output by the reference signal oscillator 101 is not limited to the above example, as long as it is the same as or contains harmonic components of the RTD resonant section 102. 【0037】First, regarding the self-oscillation frequency of the RTD oscillator 110, looking at the dashed line 202 in Figure 2A, when the bias voltage is 0.7V, the RTD oscillator 110 oscillates at 0.35THz. It can be seen that lowering the bias voltage lowers the self-oscillation frequency, and raising the bias voltage raises the self-oscillation frequency. Next, let's look at the solid line 201 when injection synchronization is performed. When the bias voltage is 0.7V, the oscillation frequency of the RTD oscillator 110 is 0.35THz, the same as the frequency in the dashed line 202 without injection synchronization. However, even when the bias voltage is lowered to 0.69V, the oscillation frequency remains at 0.35THz. Similarly, even when the bias voltage is raised to 0.71V, the oscillation frequency of the RTD oscillator 110 remains at 0.35THz. This is a phenomenon in which the oscillation frequency of the RTD resonant section 102 is locked by injection synchronization from the reference signal oscillator 101 via the mixer 103. Thus, when the oscillation frequency of the RTD resonant section 102 is fixed, it is called "locking," and when the frequency changes from a fixed state to an unfixed state, it is called "unlocking." In the graph of Figure 2A, it can be seen that the region in which the oscillation frequency of the RTD resonant section 102 is locked is the region in which the bias voltage is from 0.69V to 0.71V, and this region is called the locking range. Within the locking range, the oscillation frequency of the RTD resonant section 102, which is an injection-synchronous oscillator, is fixed (locked) depending on the frequency of the reference signal oscillator 101. It is also known that the range of this locking range is such that the stronger the signal supplied from the reference signal oscillator 101 to the RTD resonant section 102 via the mixer 103, the wider the bias voltage range, and conversely, the weaker the signal, the narrower the voltage range. In Figure 2A, an example of when the lock is released with a change of 0.01V was given to illustrate the case, but the value and amount of change of the bias voltage are not limited to this, and there may be cases of finer changes or coarser changes (for example, 0.1V). These can be arbitrarily set depending on the configuration of the RTD oscillator 110, the RTD resonant section 102, and the bias section used. 【0038】Figure 2B is a graph showing the bias voltage applied to the RTD oscillator 110 when injection-synchronized, with the phase difference between the injected RF signal 106 and the signal from the RTD resonant section 102 on the vertical axis. At 0.7V, the phase difference is 0°. This indicates that the injection synchronization frequency and the self-oscillation frequency are the same, and therefore the phases of the signal from the reference signal oscillator 101 and the signal from the RTD resonant section 102 are in sync (RF signal 106 frequency f3 × N = RTD resonant section 102 oscillation frequency f4). When the bias voltage applied to the RTD resonant section 102 decreases, the phase changes in the direction of -90°, and when the bias voltage increases, the phase changes in the direction of +90°. This phenomenon is understood to be due to the difference between the self-oscillation frequencies of the reference signal oscillator 101 and the RTD resonant section 102, which manifests as a phase difference when locked. This phase difference can be calculated using a theoretical formula, f ＲＴＤ If φ is the self-excited oscillation frequency of the RTD resonant section 102, and f is the frequency at which the RTD resonant section 102 oscillates due to injection synchronization, then φ = sin －１ (Q√(P ｏ / P ｉ ) × (f ＲＴＤ -f) / (f ＲＴＤ )) ... (3) Here, φ is the oscillation phase of the RTD resonant section 102, and P ｉ P is the oscillation power. ｏ F is the oscillation power in the harmonic component of the reference signal input from the reference signal oscillator 101 and synchronized with, and Q is the Q value of the RTD resonant section 102. The Q value increases as the frequency spectrum becomes sharper and is used as an output indicator of the oscillation circuit. In this embodiment, f ＲＴＤ This is twice the frequency f3 of the reference RF signal 106, where f is the frequency at which oscillation occurs by injection synchronous operation, i.e., the frequency f4 of the THz signal 107 output from the RTD resonant section 102. As can be seen from equation (3), this phase difference changes within the range of -90° to +90°. In other words, the phase difference within the locking range changes within the range of -90° to +90°, and the width of the locking range changes according to the ratio of the oscillation power to the injection power of the RTD resonant section 102, which is an injection-synchronous oscillator. 【0039】 Figure 2C is a graph plotting the frequency f3 of the RF signal 106 supplied from the mixer 103 and the frequency f4 of the THz signal 107 output by the RTD resonant section 102 on the horizontal axis, with the respective oscillation intensities on the vertical axis. The dotted line 203 shown in Figure 2C is a graph plotting the frequency f3 of the RF signal 106 supplied from the mixer 103. Looking at the dotted line 203, it can be seen that a subharmonic frequency of 0.175 THz, which is half of 0.35 THz, is supplied to the RTD resonant section 102. The solid line 201 is a graph plotting the frequency f4 of the THz signal 107 output by the RTD resonant section 102 when injection synchronization is performed. The dashed line 202 is a graph plotting the self-oscillation frequency output by the RTD resonant section 102 when injection synchronization is not performed. These frequencies are examples only, and the frequency of the signal supplied from the mixer 103 is not limited to this example, as long as it is the same as or a harmonic component of the self-oscillating frequency of the RTD resonator 102. At 0.175 THz, there is a signal peak of the reference signal 108 output by the reference signal oscillator 101, which coincides with the injection synchronization frequency and the subharmonic frequency of 0.175 THz, which is half the self-oscillating frequency of 0.35 THz when a bias voltage of 0.7 V is applied to the RTD resonator 102. 【0040】When the RTD resonator 102 is not injection-synchronized, i.e., when it is unlocked, the RTD resonator 102 performs self-oscillation with a relatively low Q value, as shown by the dashed line 202. Therefore, the spectrum shows a peak of a reference signal with an intensity of about 10 dB, shown by the dotted line 203, and an un-injection-synchronized peak of about 30 dB, shown by the dashed line 202. When the RTD resonator 102 is injection-synchronized (locked), a peak of the reference signal shown by the dotted line 203 and an injection-synchronized peak shown by the solid line 201, corresponding to its Q value, are produced. The peak of the spectrum when injection synchronization is performed shows 50 dB, which is higher than the 30 dB in the asynchronous case. The higher the Q value of the reference signal, the higher the frequency accuracy of the injection-synchronized RTD resonator 102, and the higher the Q value at which it oscillates. This frequency accuracy is represented by the width of the spectrum that is 3 dB lower than the peak of the solid line 201, which shows the spectrum when injection-synchronized. This is one of the effects of injection synchronization, which improves the frequency stability of the reference signal and makes it less susceptible to frequency fluctuations caused by external factors. From the relationship with the locking range described above, it can be seen that in order to obtain a stable injection-synchronized spectrum, it is necessary to supply a reference signal with a high Q value spectrum to the RTD resonant section 102 and lock the RTD resonant section 102. The spectral intensity described above is just an example and can be arbitrarily set by the locking range of the RTD resonant section 102. 【0041】 Figure 3A is a top view showing an example configuration of the semiconductor device 100 of this embodiment. Figure 3A shows the configuration from the LO port 120 to the antenna 104 for supplying the reference signal 108 to the RTD resonant section 102. 【0042】Although not shown in Figure 3A, the reference signal oscillator 101 is an oscillator that outputs a reference signal 108 with frequency f1. For example, a general Phase Locked Loop (PLL) circuit may be used for the reference signal oscillator 101. The PLL circuit is configured to oscillate a low-frequency oscillation signal supplied from an oscillation source with an even lower frequency than the reference signal 108, such as a high-precision low-frequency oscillation source using a crystal oscillator, at a desired high frequency. The reference signal 108 output from the reference signal oscillator 101 is input from the LO port 120. 【0043】 The LO port 120 can be connected using, for example, a three-terminal connection via a coplanar line. In a coplanar line, two parallel GND lines are arranged on either side of a single signal line, flanking the signal line. When the reference signal oscillator 101 is located on a semiconductor substrate different from the one on which the components shown in Figure 3A are located, these three wiring patterns (GND line, signal line, GND line) are arranged in parallel and connected to the LO port 120 of the semiconductor substrate on which the mixer 103 and RTD resonator 102 are located. The connection may be made using bumps such as flip chips, or it may be made by wire bonding from a common interposer substrate. The LO port 120 is provided with three terminal pads for connection, and the signal input to the substrate is converted and transmitted to a microstrip line. The reference signal 108 is a high-frequency signal of several GHz or higher, which is higher than the signals used in ordinary electrical circuits. Therefore, circuit configurations used for microwaves and millimeter waves may be used as connections in the LO port 120. 【0044】 The microstrip line connected from the LO port 120 is connected to the mixer 103, which is the modulation unit. In this embodiment, a mixer 103 using diodes is used. In this embodiment, the mixer 103 includes an LO filter 122, an IF filter 123, a diode unit 125, and an RF filter 124. It can also be said that the mixer 103, which is the modulation unit, is equipped with a microstrip line having a unique impedance. 【0045】The LO filter 122 is designed to allow the reference signal 108, which is an LO signal at frequency f1, to pass through, while blocking the data signal 105 at frequency f2 and the RF signal 106 at frequency f3. The LO filter 122 is the same as the filter 147 described above, which allows only frequency f1 to pass through. The LO filter 122 is configured to supply the reference signal 108 from the reference signal oscillator 101 to the mixer 103, and to suppress the backflow of the RF signal 106 and data signal 105 from the mixer 103 to the reference signal oscillator 101. The LO filter 122 is used to prevent a degradation in the quality of the reference signal 108. 【0046】 The data signal 105 is supplied to the semiconductor device 100 from the IF connection port 121. IF signals are often used in heterodyne methods to remove noise components from the data signal to be modulated. In cases where the data signal is modulated directly, such as in a direct conversion method, the data signal can be used directly. The data signal 105 input from the IF connection port 121 is a data signal modulated at frequency f2 and is combined with the reference signal 108 through the IF filter 123. The IF filter 123 acts as an LPF (low-pass filter) that separates high-frequency signals, such as the reference signal 108 and the RF signal 106, from the IF signal line. 【0047】The mixer 103 in this embodiment has a single-diode mixer configuration, in which the cathode of the diode section 125 is connected to GND. The mixer 103 mixes the reference signal 108, which is a LO signal with frequency f1, and the data signal 105, which is an IF signal with frequency f2, input to the diode section 125, utilizing the nonlinearity of the diode section 125, and performs frequency conversion to an RF signal 106. As a result of mixing, the RF signal 106, which is a composite wave with frequency f3, is output from the diode section 125. Frequency f3 is a composite wave including the difference frequency and sum frequency of f1±f2, and in particular, in the heterodyne method, the high-frequency sum frequency component f1+f2 can be selectively used. In this case, the difference frequency component f2-f3 is filtered using an MPF ​​or HPF placed at the output terminal of the mixer. Similarly in this configuration, the difference frequency component f1-f2 is blocked by the RF filter 124 at the output terminal of the mixer 103, and the sum frequency component f1+f2 is supplied to the RTD resonant section 102. 【0048】 The RF signal 106 generated by the mixer 103 passes through the RF filter 124 and is supplied to the RTD resonant section 102. The RF filter 124 is the same as the filter 147 described above and plays the role of impedance matching in the circuit connecting the mixer 103 and the RTD resonant section 102. The RF filter 124 is a necessary component for reducing reflected waves to the mixer 103 and for the efficient transmission of the RF signal 106 to the RTD resonant section 102. In this embodiment, a single-ended mixer is used as the mixer 103, but any element with a configuration commonly referred to as a mixer is acceptable. For example, the mixer 103 may be composed of a single-balanced mixer using multiple diodes, a double-balanced mixer, or a bidirectional diode mixer also called a harmonic mixer. It is also possible to appropriately arrange a mixer using transistors or the like. 【0049】An RF signal 106, which is a master signal with a terahertz frequency f3, is injected into the RTD resonant section 102. This master signal is a mix of the reference signal 108 output by the reference signal oscillator 101 and the data signal 105. The RTD resonant section 102 is an injection-synchronous oscillator that oscillates synchronously with respect to the injected RF signal 106 at a terahertz frequency of 0.35 THz. An RTD oscillator 110 is arranged on the resonant conductor 111. The resonant conductor 111 is connected to the mixer 103. In this embodiment, the resonant conductor 111 is T-shaped, but it may have other shapes. For example, the resonant conductor 111 may be configured to have a resonant structure including a resistor R, an inductance L, and a capacitance C that resonates with the oscillation signal. Furthermore, the number of RTD oscillators 110 arranged on the resonant conductor 111 is not limited to one, but may be two or more. By arranging multiple RTD oscillators 110 and generating a negative resistance sufficient to compensate for the resonance of the resonant conductor 111 and the losses in the circuit, oscillation in the terahertz band becomes possible. 【0050】 Since the RTD resonant section 102 is an injection-synchronous oscillator, when the semiconductor device 100 is used as a transmitter, if a signal flows back from the antenna 104, the reversed signal causes injection synchronization, and the accuracy of the frequency f4 of the THz signal 107 decreases. For this reason, the bandwidth of the filter 148 between the RTD resonant section 102 and the antenna 104 (corresponding to the overlapping portion of the resonant conductor 111 and the antenna connection line 143 in Figure 3A) can be designed to realize a narrow-band filter with a frequency of approximately f4 ± 5 ​​GHz. 【0051】Antenna 104 comprises an antenna connection line 143 and an antenna conductor 140. The RTD resonant section 102 is electrically connected to the antenna connection line 143. The THz signal 107, upconverted from the RF signal 106, is supplied from the RTD resonant section 102 to the antenna conductor 140 via the antenna connection line 143. The antenna connection line 143 may have a bent meander strip line configuration to reduce the circuit area of ​​the semiconductor device 100. Antenna 104 comprises a GND conductor 116, an antenna conductor 140, and a dielectric 142 placed between them (described later using Figure 3C), and is a planar patch antenna that radiates terahertz waves vertically upward from the GND conductor 116. The size of the radiating element is such that the frequency f4 of the radiated terahertz wave propagates through the radiating element at the same frequency as the electromagnetic wave f4 in the direction in which the resonant electric field is generated as a standing wave. ｆ４ λ taken into account ｆ４ The length can be set to approximately 2 / 2. In this embodiment, the direction in which standing waves are generated is parallel to the dashed line A-A' shown in Figure 3A. By adjusting the connection position between the antenna connection line 143 and the antenna conductor 140 to an appropriate position inside the antenna conductor 140, the impedance between the antenna conductor 140 and the RTD resonant section 102 can be matched, and the THz signal 107 can be transmitted efficiently. 【0052】Next, the cross-sectional structure of the semiconductor device 100 will be described using Figures 3B and 3C. Figure 3B is a cross-sectional view between A and A' shown in Figure 3A, and Figure 3C is a cross-sectional view between B and B' shown in Figure 3A. The structure of the semiconductor device 100 from the LO port 120 to the antenna conductor 140 is formed on a semiconductor substrate 130 made of a compound semiconductor such as InP. As shown in Figures 3B and 3C, dielectric layers 131 to 134 are arranged on the semiconductor substrate 130, and conductors are arranged between and on the dielectric layers 131 to 134, forming the semiconductor device 100 from the LO port 120 to the antenna conductor 140. Here, each of the dielectric layers 131 to 134 may be a single layer or may be composed of multiple layers. When composed of multiple layers, each layer may be made of a dielectric of the same composition, or dielectrics of different compositions may be stacked. 【0053】 A GND conductor 116 is placed on the semiconductor substrate 130. The GND conductor 116 is electrically connected to the RTD oscillator 110 and the diode section 125. A Schottky electrode 138, which is a conductor, and a bias pattern 112 are placed on the upper surface of the dielectric layer 131. A dielectric layer 132 is placed on top of the dielectric layer 131, and a resonant conductor 111, which is a conductor of the RTD resonant section 102, is placed on the upper surface of the dielectric layer 132. A dielectric layer 133 is placed on the upper surface of the dielectric layer 132, and microstrip lines that constitute the LO port 120, LO filter 122, RF filter 124, and antenna 104 are placed on the upper surface of the dielectric layer 133. In addition, in the RTD resonant section 102, vias 136 are placed in a vertical structure that penetrates the dielectric layer 132 in order to electrically connect the RTD oscillator 110 to the upper resonant conductor 111. Similarly, vias 113 for bias supply are arranged in a vertical structure penetrating the dielectric layer 132 in order to electrically connect the bias pattern 112 to the upper resonant conductor 111. In addition, in the diode section 125, vias 137 are arranged in a vertical structure penetrating the dielectric layers 131 and 132 in order to electrically connect the Schottky electrode 138 to the conductor of the upper microstrip line. 【0054】The GND conductor 116, RTD oscillator 110, and diode section 125 are composed of carrier-doped semiconductors on an InP semiconductor substrate 130. The GND conductor 116 is composed of a carrier-doped layer on the InP substrate, and the RTD oscillator 110 functions as a resonant tunnel diode by having a carrier-doped barrier layer 135 inside. The diode section 125 is Schottky connected by a relatively highly conductive semiconductor diode layer 139 located directly beneath the barrier layer 135 and a Schottky electrode 138, and functions as a Schottky barrier diode. 【0055】 A portion of the microstrip line, which is a wiring pattern, overlaps the resonant conductor 111, sandwiching the dielectric layer 133. This is called a Metal Insulator Metal (MIM) structure, and it acts as a capacitance, electrically connecting the AC components through AC coupling. 【0056】 The resonant conductor 111 that constitutes the resonant structure of the RTD resonant section 102 is a microstrip line resonator. The resonant structure of the RTD resonant section 102, which is an injection-synchronous oscillator, using a microstrip line, is composed of a broad lower conductor, the GND conductor 116, a line-shaped upper conductor, the resonant conductor 111, and a dielectric 141 disposed between the lower and upper conductors. The dielectric 141 is composed of a dielectric layer 131 and a dielectric layer 132. 【0057】 Antenna 104 comprises a GND conductor 116, an antenna conductor 140, and a dielectric 142 disposed between the GND conductor 116 and the antenna conductor 140. Antenna 104 is a planar patch antenna that radiates terahertz waves vertically above the GND conductor. Dielectric 142 comprises dielectric layer 131, dielectric layer 132, and dielectric layer 133. 【0058】Figures 4A to 4F show examples of filter circuits used in this embodiment. Figure 4A is an example of a bandpass filter whose frequency band can be designed by adjusting the width of the microstrip line, and can be used, for example, as the IF filter 123 described above. The width of the line between the input port 410 and the output port 411 changes. The wider sections of the line 412, 414, and 416 can be considered as capacitances inserted between GND. The narrower sections of the line 413 and 415 can be considered as inductances connected in series with the input and output ports. The arrows 421a to 421e shown in Figure 4A represent the line lengths of each line, and λ ｆ２ It is set to / 8. λ ｆ２ This represents the effective wavelength of the data signal 105 that passes through this filter. 【0059】 Figure 4B is an equivalent circuit diagram of the filter shown in Figure 4A. In this embodiment, the resonant structure functions as an IF filter 123. In this case, the impedance of the transmission line is set so that the capacitance C1 = 25 fF, C2 = 30 fF, and inductance L1 = 80 pH. These values ​​can be configured to maintain a low impedance in the bandwidth of the IF signal while attenuating the signal to less than 1 / 100th (-20 dB) at terahertz frequencies. 【0060】 Furthermore, in order to form such a filter on the same semiconductor substrate as the RTD resonant section 102, the size of the filter needs to be such that the inductance and capacitance are within a size that can be formed by the semiconductor process. In addition, the length of the transmission line should be adjusted to match the frequency of the reference signal 108, which is the LO signal to be blocked, and the length of the transmission line should be such that the effective wavelength of the reference signal 108 is λ. ｆ１ It may also be set to / 8. With this setting, the IF filter 123 functions as an LPF, allowing the IF signal, which has a low frequency of several GHz to tens of GHz, to pass through, while blocking the high frequency (reference signal 108). This prevents deterioration of signal quality in the IF signal circuit and an increase in the loss of the reference signal 108. 【0061】Figure 4C shows an example configuration of the LO filter 122 shown in Figure 3A. Between the input port 425 and the output port 426, there is a capacitive coupling section 427, two capacitance sections 428 and 430, and an inductance section 429 positioned between the capacitance sections 428 and 430. The capacitive coupling section 427 is capacitively coupled (AC coupling) by overlapping the conductor and insulator of the microstrip line. This DC-isolates the input port 425 and the output port 426. 【0062】 The capacitive coupling unit 427 needs to allow the reference signal 108, which is the LO signal, to pass through, while blocking the data signal 105, which is the IF signal, which is in the lower frequency band. The IF signal that reaches the reference signal oscillator 101 can adversely affect the oscillation stabilization of the reference signal oscillator 101. Therefore, the strength of the IF signal should be lower than the strength of the oscillation signal oscillated by the reference signal oscillator 101, for example, it may be less than 1 / 100th. Accordingly, a configuration is used in which the capacitive coupling unit 427 and the LO filter can reduce the impedance in the frequency band of the IF signal (several GHz to tens of GHz) to about 1 / 10th, or even less than 1 / 100th, of the frequency band in which the reference signal oscillator 101 oscillates. Specifically, the capacitance of the capacitive coupling unit 427 is about 20 fF, and the capacitive coupling unit 427 can be configured to be considered conductive in the frequency band of the LO signal oscillated by the reference signal oscillator 101. Furthermore, the capacitive coupling section 427 can also be provided at the connection point between the resonant conductor 111 and the LO filter 122. 【0063】The capacitance section 428, inductance section 429, and capacitance section 430 of the LO filter 122, together with the capacitive coupling section 427, constitute a band-pass filter (BPF) using LC resonance. Primarily, this filter is configured as a narrowband filter to allow the reference signal 108, which is the LO signal, to pass through at a frequency f1 of the reference signal. By narrowing the bandwidth of the LO filter 122, it reduces the amount of data signal 105 (IF signal) and signals of multiple frequencies generated by the mixer 103 that return to the RTD resonance section 102. Here, by determining the line width so that the capacitance components of capacitance sections 428 and 430 are approximately 50 to 100 fF each, and the inductance component of inductance section 429 is approximately a few pH, an MPF ​​matched to the frequency f1 of the reference signal 108 can be formed. 【0064】Figure 4D is a BPF, an example configuration of the RF filter 124 shown in Figure 3A. Capacitive coupling sections 433 and 435 are connected in series between the input port 431 and the output port 432. Short stubs 434 and 436 are placed between these two capacitive coupling sections 433 and 435, and between the capacitive coupling section 435 and the output port 432. The capacitive coupling sections 433 and 435 are configured such that the capacitance component due to capacitive coupling and the inductance component of the transmission line are connected in series. Furthermore, by adjusting the length indicated by arrow 117, the short stubs 434 and 436 have an inductance component between them and the GND conductor 116, and can therefore be used as inductance components. The RF filter 124 transmits the RF signal output by the mixer 103 to the antenna 104, while simultaneously cutting out unwanted signals incident from the antenna 104 and removing unwanted harmonic components such as the intermodulation product generated in the mixer 103. Furthermore, transmission loss is reduced by adjusting the impedance between the preceding mixer 103 and the subsequent antenna 104. The capacitance components of the capacitive coupling sections 433 and 435 are approximately 1 to 30 fF, and the inductance components of the signal flowing through the capacitive coupling sections 433 and 435 are approximately several fF to several tens of fF. The inductance components of the short stubs 434 and 436 are approximately 10 to 30 pF. By using these values, the design is made to resonate in the RF frequency band. By connecting multiple similar configurations, a band-pass filter with a wide passband of several tens of GHz or more can be constructed. 【0065】Furthermore, a filter configuration such as the one shown in Figure 4E is also possible. Coupling of the inductance component L and the capacitance component C can be considered at various points in the T-shaped portion 418 extending from the transmission line 417. Figure 4F shows a filter configuration using an open stub 419 and a radial stub 420. In the microstrip transmission line from the input port 410 to the output port 411, filtering can be performed according to the frequency band of the signal being passed by appropriately setting the length, size, and position of the stubs on the transmission line. In this embodiment as well, by appropriately changing the width of the stubs and patterns, and their positions on the transmission line, it is possible to adjust the impedance for a specific frequency and also adjust the impedance before and after the filter. 【0066】 The equivalent circuit of the semiconductor device 100 in this embodiment will be explained with reference to Figure 5. The RTD resonant section 102 is connected to a bias power supply in the bias section 500, and a drive voltage (bias voltage) is applied to the RTD resonant section 102. The RTD resonant section 102 consists of an RTD oscillator 110, a resonant conductor 111 represented by impedance Z2, and the capacitive impedance of the RTD, all connected in parallel. In Figure 5, a reference signal oscillator 101 is located on the far left. The reference signal oscillator 101 is connected to a transmission line 501. The transmission line 501 is an equivalent circuit of a microstrip line that constitutes a connection section that transmits a reference signal 108 to synchronize the LO port 120 and the RTD resonant section 102. The transmission line 501 has series components of parasitic inductances L1, L2 and the impedance Z1 of the transmission line, and parallel components of parasitic capacitances C1, C2 and parasitic resistances R1, R2. 【0067】The transmission line 501 is electrically connected to the mixer 103. The internal configuration of the mixer 103 will now be described. In the mixer 103, the transmission line 501 is connected to the diode section 125 via an LO filter 122 having impedance Z4. Also, the IF connection port 121 is connected to the diode section 125 via an IF filter 123 having impedance Z5. The diode section 125 is electrically connected to the RTD resonant section 102 via an RF filter 124, indicated by impedance Z6. Impedances Z4 and Z5 are circuits that serve the roles of filtering and impedance matching, arranged on the signal line connecting the transmission line 501 and the mixer 103. These are arranged to efficiently input and output the data signal 105, which is an IF signal, from the IF connection port 121, and the reference signal 108, which is an LO signal, input from the transmission line 501, to the mixer 103. 【0068】 The mixer 103 and the RTD resonant section 102 are connected via a capacitance C3. The capacitance C3 is electrically connected by AC coupling to allow the RF signal 106 at frequency f3 in the terahertz band, which is a mixture of the reference signal 108 at frequency f1 and the data signal 105 at frequency f2, to pass through. Specifically, it may have a capacitance of several tens to several hundred fF to act as a filter to allow the signal output from the mixer 103 in the range of 100 GHz to several hundred GHz to pass through. 【0069】The RF signal 106, which is a terahertz frequency f3 obtained by mixing a reference signal 108 with frequency f1 and a data signal 105 with frequency f2, is set to half the frequency f2 of the THz signal 107 output from the RTD resonant section 102. The capacitance C3 may be configured to pass the RF signal 106 at frequency f3 and block the THz signal 107 at frequency f4. As described above, the capacitance C3 may have a capacitance of several tens to several hundred [fF] as a filter for passing signals of several hundred to several hundred GHz from the mixer 103. For example, the capacitance C3 can use AC coupling with an insulator sandwiched between the resonant conductor 111 of the RTD resonant section 102 and the microstrip line from the mixer 103. Various filters are configured to pass desired frequencies and block unwanted frequencies and are connected using filters consisting of capacitance, fan-shaped radial stubs, microstrip lines, etc. Furthermore, a planar antenna such as a patch antenna is used for antenna 104, and antenna 104 (antenna conductor 140) is connected to the transmission line at a position inset from the end of the patch antenna in order to achieve impedance matching with the transmission line. The THz signal 107, modulated by the mixer 103 and upconverted by the RTD resonator 102, forms a standing wave on antenna 104 and is radiated into space. A filter circuit or the like may be appropriately placed on a stripline between antenna 104 and the RTD resonator 102. 【0070】 Furthermore, it is possible to insert filter circuits or the like around the RTD resonant section 102 as appropriate. Here, the wavelength of the RF signal 106 output by the mixer 103 is λ ｆ３ The wavelength of the THz signal 107 generated in the RTD resonant section 102 is λ ｆ４ For example, on the input side of the RTD resonant section 102, a λ commonly used in high-frequency circuits is used. ｆ３A short stub of λ / 4 may be provided. By providing an appropriate short stub on the input side of the RTD resonant section 102, the RF signal 106 at frequency f3 in the terahertz band, which is a mixture of the reference signal 108 at frequency f1 and the data signal 105 at frequency f2, is ignored by the short stub, and only the RF signal 106 can be transmitted. On the other hand, in the THz signal 107 which is twice the RF signal 106 oscillating in the RTD resonant section 102, λ ｆ３ The length of / 4 is λ ｆ４ Since it becomes / 2, it is shorted to GND, which prevents the THz signal 107 from flowing back to the mixer 103. Furthermore, on the output side of the RTD resonant section 102, on the side connected to the antenna 104, λ ｆ４ A λ / 4 open stub may also be provided. In this case, the λ / 4 open stub operates in the opposite way to a short stub, preventing the transmission of the RF signal 106 at frequency f3, and creating a filter that only passes the RTD resonant section 102 at frequency f4, which is twice the frequency. The stub filter configuration can utilize techniques used in microwave and millimeter-wave circuits, and can be appropriately arranged by scaling it to match the terahertz frequency band. 【0071】 The reference signal oscillator 101, mixer 103, antenna 104, and associated peripheral circuits arranged in the semiconductor device 100 described in this embodiment do not necessarily have to be arranged on the same semiconductor substrate. For example, multiple of these components can be integrated into a semiconductor device on separate semiconductor substrates. In that case, the semiconductor device 100 can be realized by connecting multiple semiconductor substrates using wire bonding or flip-chip technology during the semiconductor substrate mounting process. Furthermore, the semiconductor device 100 can be constructed by joining and combining semiconductor substrates using a semiconductor process, or by integrating multiple semiconductor substrates into a single package. In addition, an Antenna-In-Package (AiP) configuration, which integrates the antenna and semiconductor substrates into a single package, may be used for the semiconductor device 100. 【0072】Baluns, conversion circuits, matching circuits, etc., may be appropriately placed at the boundaries when connecting multiple semiconductor substrates. Furthermore, in the above configuration, no active elements are placed in the path between the RTD resonant section 102, which is an injection synchronous oscillator, and the mixer 103, which is a modulation section. However, it is not limited to this. In addition to the components described above, the peripheral circuits may appropriately include amplification circuits such as power amplifiers (PAs) and low-noise amplifiers (LNAs), phase adjustment circuits such as phase shifters, switches, etc. The semiconductor substrate may be made of silicon or a compound semiconductor such as indium phosphide (InP) or gallium arsenide (GaAs). For example, the reference signal oscillator 101 may be formed on a silicon substrate, which is the first semiconductor substrate, and the section from the LO port 120 shown in Figure 3A down (towards the antenna 104) may be formed on a compound semiconductor substrate, which is the second semiconductor substrate. In that case, a silicon substrate on which a reference signal oscillator 101 etc. is formed and a compound semiconductor substrate on which RTD resonant sections 102a, 102b etc. are formed can be stacked to form a semiconductor device 100. The above does not limit the configuration of the semiconductor substrate and mounting form of the semiconductor device 100 to which the present invention can be applied. In any case, it is sufficient that the output signal of the RTD resonant section 102, which functions as an injection synchronous oscillator based on a reference signal supplied from the reference signal oscillator 101, is given a modulated data signal. By doing so, it is possible to configure a semiconductor device 100 to which the present disclosure can be applied, which can realize a transmitting and receiving circuit using the RTD resonant section 102. 【0073】 In this embodiment, the use of the RTD resonant section 102 makes it possible to efficiently generate terahertz wave oscillation signals in a small area. Furthermore, by arranging the RTD resonant section 102 near circuits such as mixers, losses in transmission lines and other components can be minimized. In other words, miniaturization of the semiconductor device 100 and improvement of signal quality through improved high-frequency characteristics are achieved. 【0074】Next, the semiconductor device 600 of the second embodiment of this disclosure will be described using Figures 6A and 6B. Figure 6A is a conceptual diagram showing the wiring of the functional blocks of the semiconductor device 600 of this embodiment. Figure 6B is a diagram showing an overview of the RTD resonant section 102 and the modulation / demodulation section 6100 in the semiconductor device 600 of this embodiment. Unlike the configuration using the mixer 103 described above, the modulation / demodulation section 6100 is connected to the RTD resonant section 102. Furthermore, an FM modulated signal with a center frequency f2 is used as the data signal 602 input to the modulation / demodulation section 6100. The modulation / demodulation section 6100 is a modulation section when used as a transmitter, and a demodulation section when used as a receiver. 【0075】 Similar to the embodiment described above, the reference signal oscillator 101 is an oscillator that outputs a reference signal 108 with frequency f1. The frequency f1 of the reference signal 108 approximately matches the subharmonic frequency of the oscillation frequency of the RTD resonant section 102. The reference signal 108 with frequency f1 and the data signal 602, which is an FM modulated signal with a center frequency f2 via the modulation / demodulation section 6100, are mixed to produce an RF signal 106 in the terahertz band with frequency f3, which is supplied to the RTD resonant section 102. The RTD resonant section 102, upon receiving the RF signal 106, outputs a THz signal 607 with a frequency f4 that has a higher Q factor than self-oscillation, by subharmonic injection synchronization. 【0076】 The modulation / demodulation unit 6100 used in this embodiment will be described with reference to Figure 6B. The modulation / demodulation unit 6100 has a configuration using a varactor diode mixer, and the varactor diode is used as an impedance variable element. The modulation / demodulation unit 6100 receives the effective wavelength λ of the terahertz band THz signal 607 oscillating in the RTD resonant unit 102. ｆ４ λ ｆ４The impedances Z20 and Z21, which are 4 / 4 lines, are connected in series. The terminal of impedance Z20 opposite to the terminal connected to impedance Z21 is open and functions as an open stub. The terminal of impedance Z21 opposite to the terminal connected to impedance Z20 is connected to the resonant conductor 111 via AC coupling (capacitive coupling). An impedance variable circuit 6101 is connected to the connection between impedances Z20 and Z21. The impedance variable circuit 6101 has a capacitor C20, an impedance Z22, and a varactor diode which is a variable capacitance Cv, connected between the node where impedances Z20 and Z21 are connected and GND. To adjust the capacitance of the variable capacitance Cv, a signal line that supplies an FM modulated signal as a data signal 602 passes through inductor L20 and is connected to the node connecting impedance Z22 and variable capacitance Cv. 【0077】 The modulation / demodulation unit 6100 is connected to the RTD resonant unit 102 and can change the impedance of the resonant conductor 111. By changing the impedance of the resonant conductor 111 in accordance with the data signal 602, which is an FM modulated signal, the RTD resonant unit 102 can output a THz signal 607 whose oscillating phase and intensity change. The THz signal 607 is injected and synchronized by the RF signal 106. The RF signal 106 is a signal obtained by combining (superimposing) the reference signal 108 and the data signal 602, which is an FM modulated signal, and the frequency of the RF signal 106 is the same frequency f1 as the reference signal 108. The impedance variable element is not limited to a varactor diode. Variable capacitors, variable resistors, or impedance variable circuits using transistors may be used as impedance variable elements. As a result, the modulation of the RTD resonant unit 102 can be performed not only by amplitude but also by frequency and phase modulation. The circuit configuration of the other semiconductor devices 600 may be the same as in the embodiment described above, so a description is omitted here. Similar to the embodiment described above, a THz signal 607 is input to the antenna 104, and a terahertz wave TW with frequency f4 is radiated into space. 【0078】By applying this embodiment, the RTD resonant section 102 and the modulation section can be integrated, making it possible to reduce the scale of the terahertz wave generation circuit and modulation circuit. In other words, further miniaturization of the semiconductor device 600 can be achieved. 【0079】 Next, a semiconductor device 900 of the third embodiment of this disclosure will be described using Figures 7A and 7B. Figure 7A is a circuit diagram of the semiconductor device 900 in this embodiment, and Figure 7B is a modified example of the circuit diagram shown in Figure 7A. Unlike the embodiments described above, the semiconductor device 900 has two RTD oscillators 110a and 110b arranged in the RTD resonant section 902. Also, unlike the first embodiment, the circuit constituting the mixer 103 is eliminated, and the RTD resonant section 102 and the modulation section are integrated, making it possible to reduce the size of the circuit that generates and modulates terahertz waves. 【0080】 As shown in Figure 7A, in this embodiment, the port 905 to which the reference signal oscillator 101 is supplied is connected to the RTD resonant section 102 by a capacitive coupling section 906. The capacitive coupling section 906 is a DC cut filter that prevents the DC bias from being input to the reference signal oscillator 101. In addition, a DC shunt 907 is connected to suppress unwanted parasitic oscillations that occur in the RTD resonant section 102. The RTD resonant section 902 is composed of two RTD oscillators 110a and 110b, with the DC shunt 907 connected between them. The connection of the DC shunt 907 is not limited to within the RTD resonant section 102. Unlike the embodiments described above, the semiconductor device 900 superimposes the bias and IF signal in the bias pattern 112 shown in Figure 1B and supplies it to the RTD resonant section 102. As a result, the RTD resonant section 102, which is an injection-synchronous oscillator, also functions as a mixer 103, which acts as a modulation or demodulation section. Therefore, it can be said that the THz signal 107 and the RF signal 106 are the same signal. 【0081】Figure 7B shows a modified configuration in which the input terminal and bias input terminal of the data signal 105 are separated, and an AC shunt 908 is connected to suppress parasitic oscillations in the RTD resonant section 102. A capacitive coupling section 906a is provided at the connection between port 905 and the RTD resonant section 102. Similarly, a capacitive coupling section 906b is provided at the connection between the input terminal of the data signal 105 and the RTD resonant section 102. The capacitive coupling sections 906a and 906b are used as DC cut filters. 【0082】 By applying this embodiment, the RTD resonant section 902 and the modulation section can be integrated, making it possible to reduce the scale of the terahertz wave generation circuit and modulation circuit. In other words, miniaturization of the semiconductor device 900 can be achieved. 【0083】 Next, the semiconductor device 1000 of the fourth embodiment of this disclosure will be described using Figures 8A to 8D. Figure 8A is an equivalent circuit diagram of the semiconductor device 1000 of this embodiment. Figure 8B is a top view showing an example of the configuration of the semiconductor device 1000 of this embodiment. Figure 8C is a cross-sectional view between A and A' shown in Figure 8B. Figure 8D is a cross-sectional view between B and B' shown in Figure 8B. The semiconductor device 1000 of this embodiment does not have a mixer circuit. Furthermore, this embodiment uses a patch antenna 1004 as an antenna, and the patch antenna 1004 is connected to the RTD oscillators 110a and 110b as a resonant conductor constituting the resonant structure of the RTD resonant section 1002. In other words, the patch antenna 1004 is an active antenna that emits and radiates terahertz waves, integrated with the RTD resonant section 102, which is an injection-synchronous oscillator. It can also be said that the patch antenna 1004 constitutes a part of the RTD resonant section 102. In the respects described above, this differs from the first embodiment. 【0084】The RTD resonant section 1002 of this embodiment will be described using Figure 8A. The RTD resonant section 1002 is supplied with a reference signal 108 of frequency f1 from a reference signal oscillator 101. In addition, the RTD resonant section 1002 is input with a data signal 105, which is an IF signal of frequency f2. The RTD resonant section 1002 modulates the supplied reference signal 108 and data signal 105 and outputs a THz signal 107 of frequency f4. The RTD resonant section 1002 is composed of a patch antenna 1004, RTD oscillators 110a and 110b, an AC shunt 908, and a capacitive coupling section 906. The RTD resonant section 1002 receives a bias voltage via a bias pattern 112 and oscillates terahertz waves. The RTD resonant section 1002 also receives an IF signal (data signal 105) via a data signal line 1005 and performs modulation, oscillation, and radiation. The functions of the RTD oscillators 110a and 110b, the capacitive coupling unit 906, and the AC shunt 908 are the same as described above. 【0085】 Using Figure 8B, the connection relationships centered on the patch antenna 1004 will be explained. The patch antenna 1004 is connected to the RTD oscillators 110a and 110b. The patch antenna 1004 is also connected to the bias pattern 112 via vias 113a and 113b. Furthermore, the patch antenna 1004 is connected to the conductive pattern 1006 and the data signal line 1005 via capacitive coupling. The AC shunt 908 is connected to the bias pattern 112 via via 1034. The bias pattern 112 is connected to the DC input. The conductive pattern 1006 is connected to the reference signal oscillator 101. The data signal line 1005 is connected to the IF input. In other words, the RTD resonant section 1002 as a whole is a three-terminal device, a semiconductor device that receives inputs from the DC input, the reference signal oscillator 101, and the IF input, and oscillates, modulates, and radiates terahertz waves. 【0086】The resonant structure of the RTD resonator 102 is a microstrip resonator and comprises a broad lower conductor, the GND conductor 116, an upper conductor, the patch antenna 1004, and a dielectric (dielectric layers 131-133) placed between the lower and upper conductors. The GND conductor 116 is also called the reflector layer. RTD oscillators 110a and 110b are electrically connected to the upper and lower conductors. The RTD oscillators 110a and 110b are connected to the lower conductor, the GND conductor 116, and are connected to the upper conductor, the patch antenna 1004, vias 136a and 136b. In this embodiment, the material used for the vias 136a and 136b has a resistivity of 1 × 10⁻⁶. －６ Materials with a density of Ω·m or less may be used. Specific materials used for vias 136a and 136b may include metals and metal compounds such as Ag, Au, Cu, W, Ni, Cr, Ti, Al, AuIn alloys, and TiN. Each component, such as the lower electrode, upper electrode, and vias, is formed from conductive materials including metals or semiconductors with high carrier concentrations, depending on the applicable process. 【0087】 An insulating layer, such as a passivation layer 1033, may be placed on the patch antenna 1004 and the dielectric layer 133 to protect the semiconductor device 1000. A bias pattern 112 is placed between the dielectric layer 132 and the dielectric layer 133, and the bias pattern 112 is connected to the patch antenna 1004 via vias 113a and 113b for bias supply. These structures are formed on the semiconductor substrate 130 using a semiconductor process. 【0088】 The upper conductor patch antenna 1004 has a wire length (resonator length) indicated by arrow 1016 in Figure 8B, which corresponds to the wavelength λ of the signal oscillating within the RTD resonator 1002. ｅｆｆ For λ ｅｆｆ It can be set to be a multiple of / 2. Specifically, the wavelength λ in a vacuum at 0.5 THz. ０ Since it becomes 0.6 mm, the relative permittivity on the resonator is about 2, so λ ｅｆｆThe electrical length of / 2 is 0.15 mm. This wire length may be used as a reference in the design. As a result, the patch antenna 1004 resonates at a frequency in the terahertz band, and a resonant electric field is generated as a standing wave within the patch antenna 1004. The resonant electric field is generated in the direction of arrow 1016, with antinodes (points of maximum amplitude) at both ends of the patch antenna 1004 and nodes (points of zero amplitude) in the center of the patch antenna 1004. Also, λ ｅｆｆ The patch antenna 1004 of the / 2 configuration has two RTD oscillators 110a and 110b positioned opposite each other, centered on the node of the resonant electric field. In this configuration, the two RTD oscillators 110a and 110b oscillate in push-pull mode. Push-pull mode is an oscillation mode in which the two RTD oscillators 110a and 110b, arranged on the resonant conductor 111, oscillate synchronously with their phases inverted (opposite phase). 【0089】 Furthermore, the width of the connection portion of the bias pattern 112 connected to the bias supply vias 113a and 113b is smaller (narrower) than the width of the arrow 1016 where the standing wave of the patch antenna 1004 is generated. This width may also be 1 / 10 or less (λ / 10 or less) of the effective wavelength λ of the terahertz signal standing in the patch antenna 1004. This is because arranging the bias supply vias 113 and bias pattern 112 in a size and position that does not interfere with the resonant electric field within the patch antenna 1004 is suitable for improving the resonance efficiency. In addition, the bias supply vias 113a and 113b are located in the central part, which is a node of the standing wave of the patch antenna 1004. By making the connection point of the bias supply vias 113a and 113b a node of the standing wave within the patch antenna 1004, the impedance of the oscillation signal at the connection point is maximized. Therefore, losses due to leakage of the oscillation signal from the patch antenna 1004 to the bias line vias 113a and 113b for bias supply can be minimized. The bias section is adjusted as appropriate to efficiently synchronize the reference signal oscillator 101 and the RTD resonant section 102. 【0090】Here, the number of RTD oscillators 110 arranged in the RTD resonant section 102 may be one or three or more. A larger number of RTD oscillators 110 allows for stronger oscillation, but it is necessary to address power consumption, variations between elements, and the suppression of parasitic oscillations. Furthermore, the patch antenna 1004 is not limited to a rectangular patch antenna as shown in Figure 8B; slot antennas, dipole antennas, loop antennas, and various planar antennas can be used. 【0091】 The conductive pattern 1006 and data signal line 1005 are conductors called microstrip lines. If the width of these microstrip lines is too large, it can lead to deterioration of the resonant characteristics of the patch antenna 1004 and a decrease in radiation efficiency due to increased parasitic capacitance. Therefore, the width of the microstrip lines should be such that it does not interfere with the resonant electric field, typically the oscillation frequency f that is fixed in the RTD resonant section 1002. ＴＨｚ The terahertz wave can be configured to be less than or equal to 1 / 10 of its effective wavelength λ. Furthermore, the width of the conductive pattern 1006 and the data signal line 1005 can be small enough not to increase the series resistance, and can be reduced to, for example, about twice the skin depth. Considering the need to reduce the series resistance to no more than 1Ω, the width of the conductive pattern 1006 and the data signal line 1005 can typically be in the range of 0.1 μm or more and 20 μm or less. 【0092】 Furthermore, the bias pattern 112 is electrically connected to the AC shunt section 1036. The AC shunt section 1036 uses a MIM structure as shown in Figure 8B. Specifically, the conductor layer of the GND conductor 116 and the MIM conductor pattern 1035 are connected capacitively via the dielectric layer 131. The GND conductor 116 is connected to GND at an unshown position on the semiconductor substrate 130. 【0093】Changes in the AC component of the bias voltage change the amplitude of the signal oscillated by the RTD resonant section 102, which is injected and synchronized from the reference signal of the reference signal oscillator 101. This results in amplitude modulation. Here, the RTD oscillators 110a and 110b within the RTD resonant section 102 have the characteristic that the amplitude of the output signal changes depending on the applied bias voltage. In this embodiment, modulation is performed by the AC shunt section 1036 provided in the RTD resonant section 1002 and the characteristics of the RTD oscillators 110a and 110b, whose signal output, such as amplitude, changes depending on the bias voltage. The RTD resonant section 102 modulates in response to changes in the bias voltage from the bias pattern 112 while being synchronized with the reference signal from the reference signal oscillator 101. Therefore, the injected power of the reference signal from the reference signal oscillator 101 needs to be strong enough so that the lock is not released due to modulation. Specifically, in the graph of Figure 2A, the bias voltage in the locking range was 0.69 [V] to 0.71 [V]. Therefore, the amplitude of the data signal is set so that the amplitude of the bias voltage superimposed on the data signal is 0.7 ± 0.01 [V]. Since this locking range can be controlled by the injected power, it is necessary to appropriately set the locking range based on the reference signal and the amplitude of the data signal. 【0094】The configuration for modulating the oscillation signals of the RTD oscillators 110a and 110b with a data signal is not limited to this embodiment. For example, a general bias T circuit can be formed and connected to the bias pattern 112. The bias T circuit can handle similar modulation by connecting the DC side to the bias power supply, the AC side to the data signal, and the AC / DC output side to the RTD resonant section 1002. The data signal line 1005 can also be electrically connected to the bias layer with AC coupling. In this case, the connection can be made in a position or configuration with capacitance components that results in different impedances in two frequency bands: the frequency band in which the RTD resonant section 102 oscillates and the frequency band of the IF signal of the data signal. Furthermore, modulation is also possible in an impedance adjustment section, such as by using the connected data signal line to change the impedance of the patch antenna 1004. The impedance adjustment section can control and change the impedance from an external data signal by using an impedance variable element such as a varactor diode. 【0095】 By applying this embodiment, the RTD resonant section 1002 and the modulation section can be integrated. Furthermore, the patch antenna 1004 also functions as a resonant conductor for the RTD resonant section 1002. This makes it possible to reduce the scale of the terahertz wave generation circuit and modulation circuit. In other words, miniaturization of the semiconductor device 1000 is achieved. 【0096】Next, the semiconductor devices 700 and 7000 of the fifth embodiment of this disclosure will be described using Figures 9A to 9C and Figures 10A and 10B. Figure 9A is a conceptual diagram showing the wiring of the functional blocks of the semiconductor device 700 of this embodiment. In addition to the configuration of the semiconductor device 100 of the first embodiment described above, the semiconductor device 700 has a phase adjuster 701 for adjusting the phase of the reference signal 710 between the reference signal oscillator 101 and the RTD resonant section 102, which is an injection synchronous oscillator. The semiconductor device 700 of this embodiment can control the phase of the terahertz wave TW radiated from the semiconductor device 700 by using the phase adjuster 701. In this embodiment, a reference signal 710 with frequency f1 is supplied from the reference signal oscillator 101 to the phase adjuster 701. The phase adjuster 701 inputs the phase-controlled reference signal 709 with frequency f1 to the mixer 103. Next, a terahertz-band RF signal 106 with frequency f3, which is obtained by mixing a reference signal 108 with frequency f1 and a data signal 105 with frequency f2 in a mixer 103, is supplied to the RTD resonant section 102. 【0097】 The phase adjuster 701 may have the function of delaying the phase of the reference signal 710 by a set amount. In the configuration shown in Figure 9B, the reference signal 710 output from the reference signal oscillator 101 is used as the phase reference (0°), and it is possible to change the phase by π / 4 (90°). 【0098】 The phase adjuster 701 has three terminals: an input port 703, an output port 704, and an output port 705. For example, the line 707 on the output port 705 side is connected to the line 706 on the output port 704 side, and the effective wavelength λ of the input reference signal is connected to it. ｆ１ For λ ｆ１Design it to be 4 units longer. For example, as shown in Figure 9B, when the reference signal 710 passes through the line 706, it is output from the output port 704 as a reference signal 709 that is 90° behind in phase with reference signal 710. Also, when the reference signal 710 passes through the line 707, it is output from the output port 705 as a reference signal 709 that is 180° behind in phase with reference signal 710. As a result, a reference signal 709 with a phase difference of 90° from the reference signal 710 input from the input port 703 can be obtained. 【0099】 Input port 703 is connected to the reference signal oscillator 101. By switching a switch 708 located on the phase adjuster 701, output port 704 or output port 705 is connected to the mixer 103. The phase adjuster 701 can perform phase control on the reference signal 710 and output the reference signal 709 by providing multiple paths with different amounts of phase change for the reference signal 710. The combination of output ports 704, 705 and lines 706, 707 is not limited to two, and can be arranged on the phase adjuster 701 according to the number of phases to be set, thereby enabling even finer phase control. 【0100】 Furthermore, by using, for example, a transistor in the switch 708 of the phase adjuster 701, inputting a control signal to the transistor, and selecting a transmission line based on the control signal, phase control can be achieved. With the above configuration, it becomes possible to control the phase delay according to the control signal, and the semiconductor device 700 can be controlled from an external source and synchronized with the time between the transmitting and receiving sides. 【0101】The reference signal 709 output from the phase adjuster 701 is input to the mixer 103, which outputs an RF signal 106. The RF signal 106 is a terahertz frequency f3 signal obtained by mixing the reference signal 709 and the data signal 105. The RF signal 106 is supplied to the RTD resonator 102. The RTD resonator 102 outputs a THz signal 107 at a terahertz frequency f4 with a phase difference set by the phase adjuster 701 by injection-synchronous oscillation. The THz signal 107 is radiated into space as a terahertz wave TW from the antenna 104. This terahertz wave TW also has a phase difference set by the phase adjuster 701. 【0102】 By arranging multiple semiconductor devices 700 in parallel and radiating terahertz wave tweeters (TWs) with phase-locked or phase-controlled terahertz wave tweeters from adjacent antennas, it is possible to combine terahertz wave tweeters. The radiation angle of these combined terahertz wave tweeters changes according to the controlled phase. This phenomenon is called beamforming. Here, we will explain the principle of beamforming. As an example, we will explain using the configuration shown in Figure 9C, which includes three semiconductor devices 700. 【0103】 When three antennas 104 with the same distance d between them are made to oscillate with the same phase difference φ = φ1 = φ3, the radiation angle is determined by the following equation (4): sinθ = (λ / d) × φ / 2π ... (4) This radiation angle can also be changed by the antenna shape, antenna type, and array arrangement. 【0104】 In this embodiment, by controlling each phase adjuster 701 provided in multiple semiconductor devices 700 to set the same phase difference, terahertz waves TW with the same phase difference are emitted between the multiple semiconductor devices 700, and their emission angle is controlled. 【0105】Figure 10A shows a semiconductor device 7000 in which multiple active antennas AA are arranged, each including an RTD resonant section 102 which is an injection-synchronous oscillator and a patch antenna (antenna 104). Figure 10A shows block diagrams of three of the n (n is a natural number) active antennas AA. Each active antenna AA comprises an RTD resonant section 102 which also functions as a modulation section and an antenna 104 which is a patch antenna, and its circuit configuration may be the same as that of the RTD resonant section 1002 shown in Figure 8A above. A phase adjuster 701 is arranged between the reference signal oscillator 101 and the RTD resonant section 1002. Active antennas AA1 to AAn each have a similar configuration, and a common data signal line 7005 for supplying data signals 105 is connected to these active antennas AA. 【0106】 The data signal line 7005 is connected to the RTD resonant section 1002. The reference signal 710 supplied from the reference signal oscillator 101 is input to the RTD resonant section 1002 as a phase-adjusted reference signal 709 via the phase adjuster 701. Although not shown in Figure 10A, the multiple active antennas AA may include active antennas AA to which the reference signal 710 is supplied to the RTD resonant section 7002 without going through the phase adjuster 701. Since the active antennas AA1 to AAn are configured to be supplied with the same data signal, they can radiate similar RF signals into space as terahertz waves TW. 【0107】Figure 10B is a top view of the semiconductor device 7000. The semiconductor device 7000 forms an antenna array, with four active antennas AA1 to AA4 arranged in a 2x2 matrix. Each of these active antennas AA1 to AA4 radiates an RF signal represented by a terahertz wave TW into space. Note that the number of active antennas AA is not limited to four; the active antennas may be arranged in a 3x3 matrix, or in a 2x4 matrix. A transmitter to which this disclosure is applied can be configured as long as the number of active antennas is M x N (where M and N are natural numbers). The THz signal radiated from these active antennas is a signal oscillated by an RTD resonant section 7002 synchronized with a reference signal 709 (710) supplied from a reference signal oscillator 101. Therefore, the RF signals in each of the active antennas AA1 to AA4 can be synchronized with high precision. As a result, terahertz waves of higher intensity can be radiated into space than a single active antenna. Therefore, the gain of the antenna array can be improved. 【0108】Furthermore, adjusting the phase with respect to the reference signal 710 is not limited to using the phase adjuster 701. In this embodiment, a common bias pattern 112 was used to supply power to the RTD resonant section 7002 arranged in each active antenna AA. However, a bias voltage may be supplied to the RTD resonant section 7002 arranged in each active antenna AA using individual bias patterns. In that case, the wiring pattern can be avoided by forming through-vias on a semiconductor substrate and stacking them with a CMOS integrated circuit. As shown in Figure 2B, the phase of the RTD resonant section can be changed within the locking range by changing the bias voltage. In other words, the phase of each active antenna AA can be changed according to the frequency difference between the actual oscillation frequency locked by the reference signal of the reference signal oscillator 101 and the self-oscillation frequency determined by the bias voltage of the RTD resonant section 7002. For example, the self-oscillating frequency of the RTD resonant section 7002, which is an injection-synchronous oscillator located in active antenna AA1 (first active antenna), and the self-oscillating frequency of the RTD resonant section 7002, which is an injection-synchronous oscillator located in active antenna AA2 (second active antenna), may be operated to be different from each other. This makes it possible to perform beamforming in active antenna arrays arranged in one-dimensional and two-dimensional planes (in a two-dimensional array). 【0109】 The semiconductor devices 700 and 7000 of this embodiment can be positioned with a positional accuracy of the distance d between the antennas 104 that is sufficiently small in relation to the wavelength in the terahertz band, thereby improving the accuracy of the radiation angle. Furthermore, when the control signal is used as a data signal, it can also have a phase modulation function and a communication function. The arrangement and order of each component in the above-described embodiments are just one way of realizing this disclosure, and the connection configuration to which this disclosure can be applied is not limited to this embodiment. 【0110】 Other Embodiments Although embodiments of the present invention have been described above, the present invention is not limited to these embodiments, and various modifications and changes are possible within the scope of its gist. 【0111】For example, although the above embodiment uses a transmitter in a wireless communication device as an example, the application is not limited to this, and may be a radar device, for example. In the case of a frequency-modulated continuous wave (FMCW) radar, it is necessary to perform a constant frequency sweep, but the output frequency of the reference signal oscillator 101 may be swept using the circuit of this disclosure as shown in Figure 1, and an oscillation signal may be generated and radiated in the RTD resonant section 102. In that case, problems such as the injection synchronization lock being released due to a large frequency sweep width, and the phase of the oscillation signal of the RTD resonant section 102 changing due to the frequency difference with the self-oscillated oscillation, can be addressed. By appropriately adjusting the bias voltage supplied to the RTD resonant section 102 according to the frequency of the signal output by the reference signal oscillator 101, instability in frequency and phase can be suppressed. 【0112】 Furthermore, although a square patch antenna is used as the terahertz wave antenna in the above embodiment, the shape of the antenna is not limited to this. For example, patch conductors such as rectangles, triangles or polygons, circles, and ellipses, as well as planar antennas such as loop antennas, log-periodic antennas, and Vivaldi antennas, and horn antennas may be used as antennas. 【0113】 Furthermore, although we have described the relationship between one RTD resonator and one antenna, connections such as 1 to N, N to 1, or N to N (where N is a natural number) are also possible. In addition, although we have described the structure of the RTD resonator using a multilayer structure, it is not limited to this. That is, the above discussion can also be applied to oscillators that do not use a multilayer structure. 【0114】 Furthermore, any of the following combinations may be used as materials for the RTD oscillator: • GaAs / AlGaAs and GaAs / AlAs, InGaAs / GaAs / AlAs formed on a GaAs substrate • InGaAs / InAlAs, InGaAs / AlAs, InGaAs / AlGaAsSb formed on an InP substrate • InAs / AlAsSb and InAs / AlSb formed on an InAs substrate • SiGe / SiGe formed on a Si substrate The above structures and materials can be appropriately selected according to the desired frequency and other factors. 【0115】 Next, we will describe the case where the semiconductor devices of each embodiment described above are applied to a terahertz camera system (imaging system). The following description will be made with reference to Figure 11A. The terahertz camera system 1100 includes a transmitter 1101 that emits terahertz waves TW and a receiver 1102 that detects the terahertz waves TW. Furthermore, the terahertz camera system 1100 includes a control unit 1103 that controls the operation of the transmitter 1101 and the receiver 1102 based on an external signal, processes an image based on the detected terahertz waves, or outputs it to the outside. The semiconductor device of each embodiment may be the transmitter 1101 or the receiver 1102. The terahertz waves emitted from the transmitter 1101 are reflected by the subject 1105 (object) and detected by the receiver 1102. A camera system having such a transmitter 1101 and receiver 1102 may also be called an active reflection imaging camera system. Furthermore, a camera system comprising a transmitter 1101 and a receiver 1102 can also be applied to a transmission imaging system that observes terahertz waves transmitted through a subject by placing the transmitter 1101 and receiver 1102 opposite each other and the subject in between. Moreover, in a passive camera system without a transmitter 1101, the semiconductor devices of each embodiment described above can be used as the receiver 1102. In addition, by using the semiconductor devices of each embodiment that are capable of beamforming, it is possible to improve the detection sensitivity of the camera system and obtain high-quality images. 【0116】Furthermore, a case in which any of the above-described embodiments are applied to a terahertz communication system (communication device) will be described below with reference to Figure 11B. Semiconductor devices can be used as components of a communication system. As a communication system, simple ASK methods, superheterodyne, and direct conversion methods can be envisioned. A superheterodyne communication system, for example, includes an antenna 1200, an amplifier 1201, a mixer 1202, a filter 1203, a mixer 1204, a converter 1205, a digital baseband modulator / demodulator 1206, and local oscillators 1207 and 1208. In the above-described embodiment, the RTD resonant section 102 can be used as a local oscillator 1207, the antenna 104 as an antenna 1200, and the mixer 103 in Figure 5 as a mixer 1202. In the case of a receiver, the terahertz wave received via antenna 1200 is converted to an intermediate frequency signal by mixer 1202, then converted to a baseband signal by mixer 1204, and the analog waveform is converted to a digital waveform by converter 1205. The digital waveform is then demodulated in the baseband to obtain a communication signal. In the case of a transmitter, after the communication signal is modulated, it is converted from a digital waveform to an analog waveform by converter 1205, then frequency converted via mixers 1204 and 1202, and output as a terahertz wave from antenna 1200. The direct conversion communication system includes antenna 1200, amplifier 1211, mixer 1212, modulator / demodulator 1213, and local oscillator 1214. In the direct conversion method, during reception, the received terahertz wave is directly converted into a baseband signal by mixer 1212, and during transmission, the baseband signal to be transmitted is converted into a terahertz signal by mixer 1212. The other configurations are the same as in the superheterodyne method. The apparatus according to each of the above embodiments can perform beamforming of terahertz waves by electrical control of semiconductor devices. Therefore, it is possible to align radio waves between the transceiver and receiver.Therefore, by using the semiconductor devices of each embodiment that are capable of beamforming, it becomes possible to improve wireless quality such as the signal-to-noise ratio in communication systems and to transmit large amounts of information over a wide coverage area at low cost. Furthermore, if frequency sweeping can be performed using the same local oscillator for both transmission and reception, it becomes possible to measure distance from the delay and phase information of the transmitted signal (radiated wave) and reflected wave as an FMCW radar device. By applying the semiconductor devices of this disclosure, it becomes possible to improve wireless quality such as the signal-to-noise ratio, improve ranging accuracy, and do so at low cost. 【0117】 The technical ideas derived from this disclosure are not limited to the exemplary embodiments disclosed, but are intended to encompass various modifications of the exemplary embodiments, or substitutions with equivalent structures or functions. The scope of the following claims should be interpreted in the broadest way to encompass all such modifications and equivalent structures and functions. 【0118】 This application claims priority based on Japanese Patent Application No. 2024-210771, filed on December 3, 2024, and all of its contents are incorporated herein by reference.

Claims

1. A semiconductor device comprising: a reference signal oscillator that generates a first signal of a first frequency; a modulation unit that generates a third signal of a third frequency from the first signal and a second signal of a second frequency corresponding to a signal supplied from a baseband circuit; and an injection-synchronous oscillator that generates a fourth signal of a fourth frequency in the terahertz band in synchronization with the third signal.

2. The semiconductor device according to claim 1, characterized in that the fourth frequency is a higher frequency than the third frequency, and the third frequency is a frequency equal to or greater than the first frequency.

3. The semiconductor device according to claim 1 or 2, characterized in that the first frequency is a higher frequency than the second frequency.

4. The semiconductor device according to any one of claims 1 to 3, characterized in that the second signal is an IF signal or an FM modulated signal.

5. The semiconductor device according to any one of claims 1 to 4, characterized in that the injection-synchronous oscillator includes a resonant tunneling diode.

6. The semiconductor device according to any one of claims 1 to 5, characterized in that the injection-synchronous oscillator comprises a microstrip line having an intrinsic impedance for oscillating in the terahertz band.

7. The semiconductor device according to claim 6, characterized in that the resonant structure using the microstrip line of the injection-synchronous oscillator includes a resonant conductor, a lower conductor, and a dielectric disposed between the resonant conductor and the lower conductor.

8. The semiconductor device according to claim 7, characterized in that the resonant conductor comprises a λ / 4 open stub.

9. The semiconductor device according to claim 7 or 8, wherein a bias pattern for supplying power to the injection-synchronous oscillator is provided on the dielectric, the bias pattern is connected to the resonant conductor via vias provided on the dielectric, and the width of the connection portion of the bias pattern connected to the via is smaller than the width in the direction in which the standing wave of the resonant conductor is generated.

10. The semiconductor device according to claim 9, characterized in that an AC shunt is connected to the bias pattern.

11. The semiconductor device according to claim 9 or 10, characterized in that the modulation unit includes an impedance variable circuit for changing the impedance of the resonant conductor in accordance with the second signal.

12. The semiconductor device according to any one of claims 1 to 10, characterized in that the modulation unit includes a mixer.

13. The semiconductor device according to claim 12, characterized in that the modulation unit comprises a microstrip line having a unique impedance.

14. The semiconductor device according to claim 12 or 13, characterized in that a signal line for supplying the third signal is electrically connected between the reference signal oscillator and the modulation unit.

15. The semiconductor device according to any one of claims 12 to 14, characterized in that a first filter is provided between the reference signal oscillator and the modulation unit to allow the first signal to pass through and to reduce the passage of the second signal and the third signal.

16. The semiconductor device according to claim 9 or 10, characterized in that the second signal is supplied to the bias pattern, and the second signal superimposed on the bias voltage is supplied from the bias pattern to the injection synchronous oscillator, thereby causing the injection synchronous oscillator to function as the modulation unit.

17. The semiconductor device according to claim 16, characterized in that the second signal is amplitude modulated.

18. The semiconductor device according to any one of claims 1 to 17, further comprising an antenna for radiating the fourth signal into space.

19. The semiconductor device according to 18, wherein a second filter is disposed between the injection-synchronous oscillator and the antenna, a third filter is disposed between the modulation unit and the injection-synchronous oscillator, and the third filter allows signals of a lower frequency than the signals that the second filter allows to pass through.

20. The semiconductor device according to claim 15, further comprising an antenna for radiating the fourth signal into space, a second filter disposed between the injection-synchronous oscillator and the antenna, a third filter disposed between the modulation unit and the injection-synchronous oscillator, the first filter allowing signals of a lower frequency than those allowed to pass through the second and third filters, and the third filter allowing signals of a lower frequency than those allowed to pass through the second filter.

21. The semiconductor device according to any one of claims 18 to 20, characterized in that the antenna includes a patch antenna.

22. The semiconductor device according to claim 21, characterized in that the patch antenna constitutes a part of the injection-synchronous oscillator.

23. The semiconductor device according to any one of claims 18 to 22, characterized in that a plurality of active antennas, each including the injection-synchronous oscillator and the antenna, are arranged in a two-dimensional array.

24. The semiconductor device according to 23, wherein the plurality of active antennas include a first active antenna and a second active antenna, and the self-oscillating frequency of the injection synchronous oscillator arranged in the first active antenna and the self-oscillating frequency of the injection synchronous oscillator arranged in the second active antenna are operated to be different from each other.

25. The semiconductor device according to any one of claims 1 to 24, characterized in that the third frequency is the subharmonic frequency of the fourth frequency.

26. The semiconductor device according to any one of claims 1 to 25, further comprising a first semiconductor substrate on which the reference signal oscillator is disposed, and a second semiconductor substrate on which the injection synchronous oscillator is disposed, wherein at least a portion of the first semiconductor substrate and the second semiconductor substrate are laminated.

27. The semiconductor device according to any one of claims 1 to 26, characterized in that a phase adjuster for adjusting the phase of the first signal is disposed between the reference signal oscillator and the modulation unit.

28. The semiconductor device according to 27, characterized in that the phase adjuster delays the phase of the first signal.

29. The semiconductor device according to claim 27 or 28, characterized in that the phase adjuster comprises a plurality of paths with different amounts of phase change of the first signal.

30. A communication device comprising: a transmitting unit equipped with a semiconductor device according to any one of claims 1 to 29; and a receiving unit for detecting a signal radiated from the transmitting unit.

31. An imaging system comprising a semiconductor device according to any one of claims 1 to 29, and characterized by having a transmitting unit that emits a signal toward a subject, and a receiving unit that detects the signal reflected or transmitted from the subject.

32. A radar device comprising: a transmitting unit equipped with a semiconductor device according to any one of claims 1 to 29; and a receiving unit for detecting reflected waves emitted from the transmitting unit and reflected by an object, wherein the device measures the distance between the emitted waves emitted from the transmitting unit and the reflected waves.

Images