Open Access
Add to BibSonomy
Abstract
THz waves are promising wireless carriers for next-generation wireless communications, where a seamless connection from wireless to optical communication is required. In this study, we demonstrate carrier conversion from THz waves to dual-wavelength NIR light injection-locking to an optical frequency comb using asynchronous nonpolarimetric electro-optic downconversion with an electro-optic polymer modulator. THz wave in the W band was detected as a stable photonic RF beat signal of 1 GHz with a signal-to-noise ratio of 20 dB via the proposed THz-to-NIR carrier conversion. In addition, the results imply the potential of the photonic detection of THz waves for wireless-to-optical seamless communication.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Terahertz (THz) waves are wireless carriers for next-generation mobile communications (6 G, expected carrier freq. > 300 GHz) [1]. Despite wireless electronics being widely used until recent wireless communications (5 G, carrier freq. = 28 GHz or more), they may face technical limitations in 6 G owing to the high frequency such as increased phase noise of wireless carriers and/or increased signal loss in electric transmission lines. One potential method to overcome this technical limitation is to use photonic technology in THz communications [2]. For example, regarding the generation of THz wave, a promising approach for the photonic generation of low-phase-noise THz waves (freq. = fTHz) is the use of an optical frequency comb (OFC) [3–6] owing to the low phase noise of frequency spacing (= frep), where two OFC modes with THz frequency spacing (= mfrep = fTHz) are extracted and used for photomixing with the help of a uni-traveling carrier photodiode (UTC-PD) [7,8]. Particularly, on-chip Kerr microresonator soliton combs with a considerably large frep, namely, the soliton microcomb, have attracted attention for the photonic generation of ultralow-phase-noise THz waves at 300 GHz [9] and 560 GHz [10] because frep is close to fTHz. Thus, two adjacent microcomb modes can be directly used for photomixing, benefiting from the low-phase-noise advantage of frep without the influence of optical frequency multiplication. Furthermore, photonic-generated THz waves have been applied to wireless data transfer [11–13]. Conversely, regarding the detection of THz wave, electronic THz detectors, such as Schottky barrier diodes and mixers have been still used in THz wireless communication regardless of bulky, fragile, complicated, and expensive device.
If carrier conversion from THz waves to near-infrared (NIR) light can be realized by use of photonic technology, existing mature photonic devices for optical communication can be used with minimal modification, enabling photonic THz detection for a seamless connection from wireless communication to optical communication. One potential method for such THz-to-NIR carrier conversion is the use of nonpolarimetric electro-optic (EO) downconversion (NP-EO-DC), which has been applied for the visualization of millimeter-wave electric fields [14–17]. In this method, dual-wavelength NIR light is used as an optical carrier, and millimeter-wave-induced modulation sidebands are generated corresponding to the dual-wavelength NIR light via the EO effect. When the optical frequency spacing of the dual-wavelength NIR light is close to the millimeter-wave frequency, an optical beat signal between one of the dual-wavelength NIR light beams and the modulation sideband of the other appears in the RF region. The heterodyned detection of the optical beat signal enhances the RF beat signal corresponding to the baseband signal. In particular, since the asynchronous NP-EO-DC [16,17] uses two OFC modes with a frequency spacing close to the frequency of electromagnetic waves being extracted from an electro-optic modulator OFC (EOM-OFC) for stable dual-wavelength NIR light, it is applicable for self-oscillating electromagnetic wave and thus has a good affinity with THz-to-NIR carrier conversion. When an optical signal-to-noise ratio (OSNR) is defined as a ratio of optical signal to background noise level, the remaining bottleneck is the limited OSNR of this dual-wavelength NIR light owing to the background of residual unwanted EOM-OFC modes and/or amplified spontaneous emission (ASE). The means to eliminate this bottleneck is injection-locking of high-power CW lasers to EOM-OFC modes, enabling the amplification and phase noise transfer of each EOM-OFC mode while eliminating the residual unwanted EOM-OFC modes and reducing the ASE background. Although such dual-wavelength NIR light injection-locking to EOM-OFC modes have effectively applied for photonic generation of millimeter-wave and THz waves with UTC-PD [18,19], there have been no attempts to apply it for the asynchronous NP-EO-DC.
In this study, for a proof-of-concept experiment prior to the demonstration at the 6 G carrier frequency band, we demonstrate THz-to-NIR carrier conversion in the W-band by the asynchronous NP-EO-DC using high-OSNR, stable, dual-wavelength NIR light injection-locking to a 10-GHz-spacing EOM-OFC [20,21]. We simultaneously use an electro-optic polymer (EOP) modulator [22,23] in place of EO crystals used in previous research, because it has advantages such as high-frequency response, long interaction length, low absorption of THz waves, a high EO coefficient, and good phase matching between THz and optical carriers.
2. Principle of operation
Figure 1 shows a principle of operation for the carrier conversion from a modulated THz wave (wireless carrier, carrier freq. = fTHz, modulation width = ΔfTHz) to dual-wavelength NIR light (optical carrier, opt. freq. = ν1, ν2). The optical frequency spacing between the pair of high-power CW lasers (CWL1, opt. freq. ≈ ν1; CWL2, opt. freq. ≈ ν2) was set close to fTHz. When two OFC modes (opt. freq. = ν1, ν2) with a frequency spacing of mfrep (= ν2 - ν1 ≈ fTHz) are extracted by an optical bandpass filter (BPF) and then are respectively incident into two CW lasers, the injection locking of each CW laser to each OFC mode enables optical amplification and phase noise transfer of two OFC modes while eliminating the residual unwanted EOM-OFC modes and reducing the ASE background. The resulting stable dual-wavelength NIR light (i.e., ν1 carrier and ν2 carrier) was used for the THz-to-NIR carrier conversion. Thereafter, the modulated THz wave is incident onto an antenna of an EOP modulator as an external electric field, whereas ν1 carrier propagates through an optical waveguide in the EOP modulator as optical carriers. The phase modulation in the EOP modulator creates the modulation sidebands of ν1 carrier that are only fTHz apart. ν2 carrier bypasses the EOP modulator for an unmodulated optical carrier. As the frequency difference between ν1 carrier and ν2 one (= ν2 - ν1 = mfrep) is set to be close to fTHz, ν1 sideband appears near ν2 carrier. Importantly, the ν1 and ν2 carriers are phase-locked to each other because of their injection-locking to the EOM-OFC modes, and their fluctuations in optical frequency are common to each other while maintaining a constant frequency spacing of mfrep. Therefore, a pair of ν2 carrier and ν1 sideband is equivalent to a pair of ν1 carrier and ν1 sideband except their frequency difference. An optical beat signal between the ν2 carrier and ν1 sideband was detected as a baseband signal by the photodetector after optical bandpass filtering of the ν1 carrier with an optical bandpass filter. The use of optical heterodyning detection enables us to enhance the weak ν1 sideband signal by optical interfering with the strong ν2 carrier. The resulting baseband signal (freq. = fRF = fTHz - mfrep) corresponded to the modulated THz wave.
Fig. 1. Principle of carrier conversion from a modulated THz wave to dual-wavelength NIR light. EOM-OFC, electro-optic modulator optical frequency comb; CWL1 and CWL2, high-power CW lasers; EOP modulator, electro-optic polymer modulator.
3. Experimental setup
Figure 2 shows a schematic drawing of the experimental setup. We used a pair of 4-mW distributed feedback laser diodes (DFB1,Gooch & Housego, AA1408-193350-100-PM900-FCA-NA, ν1 = 193.39 THz corresponding to a wavelength of 1550.2 nm; DFB2, Gooch & Housego, AA1408-193350-100-PM900-FCA-NA, ν2 = 193.49THz corresponding to a wavelength of 1549.4 nm) for slave lasers and a pair of 2-µW EOM-OFC modes for master lasers, respectively. Two EOM-OFC modes with a frequency spacing of 10frep (EOM-OFC-M1 and EOM-OFC-M2) were extracted from an EOM-OFC (center wavelength = 1550 nm, frep = 10 GHz) [20,21] by tunable ultra-narrowband optical bandpass filters (BPF1 and BPF2; Alnair labs, CVF-300CL-PM-FA, center wavelength = 1550.20 nm and 1549.50 nm, optical passband = 0.05 nm, insertion loss = 5.5 dB). The EOM-OFC-M1 and EOM-OFC-M2 were incident on DFB1 and DFB2, respectively, via optical circulators (OCs). When the optical frequencies of DFB1 and OFB2 were tuned to those of EOM-OFC-M1 and EOM-OFC-M2 within the locking range of DFB1 and DFB2 (typically a few hundred MHz) by current control, injection-locking was achieved. The details of injection locking are given elsewhere [18,19].
Fig. 2. Experimental setup. EOM-OFC, electro-optic modulator optical frequency comb; FCs, fiber coupler; BPF1 and BPF2, tunable ultra-narrowband optical bandpass filters; OCs, optical circulators; DFB1 and DFB2, distributed feedback laser diode; PC, polarization controller; EOP-MOD, electro-optic polymer modulator; L1, spherical THz lens; CL1 and CL2, a crossed pair of cylindrical THz lenses; BPF3, optical bandpass filter; PD, photodiode; ESA, electric spectrum analyzer.
The ν1 carrier injection locking to EOM-OFC-M1 was fed into the optical waveguide of a W-band non-coplanar patch-antenna-type EOP modulator [23] with a figure of merit of 1.3 W−1/2 at 100 GHz estimated according to the literature [24] whereas the ν2 carrier injection-locking to EOM-OFC-M2 bypassed the EOP modulator. The patch antenna array has a horizontally elongated shape along the EOP waveguide [23], so an elliptical focus spot is preferred to efficiently irradiate it without wastage. Therefore, THz wave (fTHz = 101 GHz, power = 4 dBm = 2.5 mW) was irradiated as an elliptic focus on the parch antenna by a combination of a spherical THz lens (L1, diameter = 50 mm, focal length = 100 mm) and a crossed pair of cylindrical THz lenses (CL1, size = 30 mm by 50 mm, focal length = 96 mm; CL2, size = 85 mm by 85 mm, focal length = 50 mm). This results in the generation of an ν1 sideband separated by fTHz from the ν1 carrier. After the ν1 carrier and the ν1 sideband were combined with the ν2 carrier by a fiber coupler (FC), the optical bandpass filter (BPF3) was used to extract the ν1 sideband and the ν2 carrier while eliminating the ν1 carrier. The optical beat signal between the ν1 sideband and the ν2 carrier was generated and then detected by a photodiode (PD, wavelength = 1550 nm, RF bandwidth = 10 GHz). The RF beat signal was acquired by an electric spectrum analyzer (ESA, RIGOL, DSA815, freq. = 9kHz-1.5 GHz) after passing through an electric amplifier (AMP, Mini-Circuits, ZX60-V62+, freq. = 0.05-6 GHz, gain = 15.5 dB).
4. Results
First, we evaluated the basic performance of injection locking of the DFB to the EOM-OFC modes. The red plot in Fig. 3(a) shows the optical spectrum of DFB1(optical power = 8 mW) injection-locking to EOM-OFC-M1 corresponding to the ν1 carrier. For comparison, the blue plot shows the optical spectrum of EOM-OFC-M1 (optical power = 2 µW) extracted from EOM-OFC. The background noise of the EOM-OFC-M1 represents the noise floor of the optical spectrum analyzer (resolution bandwidth or RBW = 0.02 nm); conversely, the background noise of the ν1 carrier exhibits a higher level compared to the noise floor of the optical spectrum analyzer due to the weak luminescence background light from the DFB1. The wavelength-magnified optical spectrum of them is shown in Fig. 3(b). The wavelength of the ν1 carrier was the same as that of the extracted EOM-OFC-M1, indicating successful injection locking. The reason for no difference in the linewidth of the DFB laser before and after injection locking is the insufficient RBW in the optical spectrum analyzer. The reduced linewidth by the injection locking can be confirmed in the RF spectrum of the optical beat signal [19]. Despite ultrafine spectrally modulated luminescence owing to the internal grating structure characteristic of the DFB, which exists in both tails of the ν1 carrier, its intensity is sufficiently lower than that of the the ν1 carrier, and the resulting OSNR was achieved to 70 dB. Conversely, the OSNR of the extracted EOM-OFC-M1 remained at approximately 50 dB because of the limited optical power of each EOM-OFC mode and the noise background. Thus, we obtained a sufficient OSNR in the ν1 carrier for the generation of modulation sidebands using the EOP modulator.
Fig. 3. (a) Optical spectra of the DFB1 injection locked to EOM-OFC-M1 corresponding to the ν1 carrier (red plot) and EOM-OFC-M1 extracted from EOM-OFC (blue plot) and (b) their magnified spectra. Resolution bandwidth of optical spectrum is 0.02 nm.
Subsequently, the modulation sideband of the ν1 carrier was observed by irradiating the EOP modulator with THz waves. Figures 4(a) and 4(b) compare the optical spectra at the output port of the optical waveguide of the EOP modulator without and with THz-wave irradiation (RBW = 0.02 nm). Both the ν1 carrier and its modulation sidebands are observed under THz irradiation, whereas only the ν1 carrier is observed without THz irradiation. Two ν1 modulation sidebands appeared exactly ± fTHz (= 101 GHz) away from the ν1 carrier. When a carrier-to-sideband ratio (CSR) is defined as a ratio of optical carrier to its modulation sideband, it is 38 dB for a pair of ν1 carrier and ν1 modulation sideband. The improved CSR results from shaping the THz focus as an oval to match the shape of the patch antenna. The OSNR of the ν1 modulation sideband was achieved to be 35 dB, owing to the enhanced OSNR of the ν1 carrier by injection locking. From the OSNR of the ν1 modulation sideband and the irradiated THz power (= 4 dBm = 2.5 mW), the minimum detectable THz power can be estimated as −29 dBm or 1.3 µW. Similar CSR and OSNR should be maintained for a pair of ν2 carrier and ν1 modulation sideband that are close to each other.
Fig. 4. (a) Optical spectra of the ν1 carrier without irradiation of THz wave. (b) Optical spectra of the ν1 carrier and its modulation sideband with irradiation of THz wave. Resolution bandwidth of optical spectrum is 0.02 nm.
Finally, we measured the RF spectrum of the optical beat signal between the ν2 carrier and ν1 modulation sideband. Figure 5 shows the RF spectrum of the beat signal (resolution bandwidth or RBW = video bandwidth or VBW =1 MHz). Because the frequency difference between fTHz (= 101 GHz) and 10frep (= 100 GHz) was 1 GHz, the RF beat signal between them was observed at 1 GHz, achieving an SNR of 20 dB. The pronounced elevation in the spectrum at frequencies below 0.5 GHz is due to the noise signal of optical intensity, PD, and AMP. We consider that SNR of the RF signal will show a linear dependence on the input THz power because the detection of the 1-GHz RF signal is based on the optical heterodyning process between the ν1 modulation sideband and the ν2 carrier. Possible factors that could lead to nonlinear behavior include the nonlinearity of electro-optic Pockels effect in the EOP modulator; however, at the current THz power levels, this is expected to be negligible. Although experimental validation of the input THz power dependence for the RF signal is important, we believe that it would be more practically significant when using modulated THz signals (for example, On-Off-Keying modulation), rather than the unmodulated THz signal used in the present study. This is something we intend to address in our future work.
Fig. 5. RF spectra of optical beat signal between the ν1 modulation sideband and the ν2 carrier. Resolution bandwidth and video bandwidth of RF spectrum are 1 MHz. The corresponding movie is given in Visualization 1.
To confirm the effectiveness of the stable dual-wavelength NIR light injection-locking to the two EOM-OFC modes in the present THz-to-NIR carrier conversion, we continuously monitored the behavior of the RF spectrum before and after the injection locking was lost, as shown in Visualization 1. When dual-wavelength NIR light was injection-locked to two EOM-OFC modes, the RF beat signal remained fixed at 1 GHz. Because the RF beat signal is sufficiently stable in frequency and phase owing to injection locking to the EOM-OFC, it can be used for THz wireless communication. However, once the injection locking was off, dual-wavelength NIR light is in free-running state; in this case, the RF beat signal fluctuated significantly and/or multiple RF beat signals appeared, making it difficult to apply the RF beat signal to THz wireless communication.
5. Discussion
To highlight the effectiveness of the proposed method in THz communication, we compare the following three scenarios. Firstly, considering the use of single-wavelength NIR light (v1 carrier): In this case, two ν1 modulation sidebands are observed in addition to the ν1 carrier [see Fig. 4(b)]. If only one ν1 modulation sideband is extracted using a bandpass filter and detected by a photodetector, a baseband signal appears near DC, yielding only amplitude information; importantly, its amplitude is too weak to use it as a baseband signal. Secondly, considering the use of dual-wavelength NIR light without injection-locking to the EOM-OFC: In this scenario, the ν2 carrier is present near the ν1 modulation sideband. Extracting both signals with a bandpass filter results in the generation of an optical beat signal due to optical heterodyning between them, producing a beat frequency in the RF range. This allows the generation of a baseband signal in the RF range (1 GHz in this paper), and the weak ν1 modulation sideband is enhanced by interference with the strong ν2 carrier, leading to improved SNR. However, since the ν1 carrier and ν2 carrier are not phase-locked to each other, there are large relative fluctuations of frequency and phase between the ν1 modulation sideband and the ν2 carrier, causing significant fluctuations in the RF beat frequency (see Visualization 1, injection locking: OFF), which might degrade the quality of data transfer in THz communication. Lastly, considering the use of dual-wavelength NIR light with injection-locking to the EOM-OFC: In this scenario, besides the RF baseband signal generation and SNR enhancement due to optical heterodyning, their injection-locking to EOM-OFC modes leads to phase synchronization between the ν1 carrier and the ν2 carrier. In other words, the ν1 modulation sideband and ν2 carrier are indirectly phase-locked to each other. As a result, relative frequency and phase fluctuations are suppressed, leading to stable RF beat frequency (see Visualization 1, injection locking: ON). This high stability of frequency and phase is crucial for THz communication utilizing modulation formats that exploit both amplitude and phase (e.g., quadrature amplitude modulation or QAM). Therefore, this approach offers benefits in THz communication.
We used dual-wavelength NIR light injection-locking to the EOM-OFC modes as an optical carrier in the asynchronous NP-EO-DC using the EOP modulator. We here discuss the effectiveness of the injection locking in the asynchronous NP-EO-DC by comparing optical spectra of ν1 carrier and its modulation sidebands when using the DFB1 injection-locking to EOM-OFC-M1 and the extracted EOM-OFC-M1 as a dual-wavelength NIR light, respectively. We set the optical power of them fed into the EOP modulator to be 3.3 mW, which was limited by the maximum optical power of the extracted EOM-OFC-M1. Figure 6(a) compares optical spectra of the DFB1 injection-locking to EOM-OFC-M1 with and without irradiation of THz wave (RBW = 0.02 nm). Two ν1 modulation sidebands clearly appeared without interference of residual unwanted EOM-OFC modes due to the injection locking. Conversely, Fig. 6(b) shows the comparison of optical spectra of the extracted EOM-OFC-M1 with and without irradiation of THz wave. Even though two ν1 modulation sidebands are generated by unmodulated THz carrier wave, residual unwanted EOM-OFC modes significantly interfered their modulation sidebands, limiting the OSNR of the modulation sideband signal. If the modulation bandwidth of THz wave becomes wide by coding the modulation signal into THz carrier wave, such interference effect will become large that cannot be ignored. We also compare RF spectra of the optical beat signal generated by the ν2 carrier and ν1 modulation sideband between the DFB1 injection-locking to EOM-OFC-M1 and the extracted EOM-OFC-M1. Figures 6(c) and 6(d) show the RF spectra of the optical beat signal (RBW = VBW =1 MHz) when using the DFB1 injection-locking to EOM-OFC-M1 (optical power = 3.3 mW) and the extracted EOM-OFC-M1 (optical power = 3.3 mW) as a dual-wavelength NIR light, respectively. In this case, while their frequency stability is identical to each other, OSNR is different between them. Although the RF beat signal was confirmed at 1 GHz in both RF spectra, the SNR was enhanced in the former. In this way, we confirmed the effectiveness of the injection locking in the asynchronous NP-EO-DC.
Fig. 6. (a) Optical spectra of the DFB1 injection-locking to EOM-OFC-M1 (optical power = 3.3 mW) without and with irradiation of THz wave. (b) Optical spectra of the EOM-OFC-M1 (optical power = 3.3 mW) without and with irradiation of THz wave. Resolution bandwidth of optical spectrum is 0.02 nm. RF beat spectra obtained by (c) DFB1 injection-locking to EOM-OFC-M1 (optical power = 3.3 mW) and (d) EOM-OFC-M1 (optical power = 3.3 mW). Resolution bandwidth and video bandwidth of RF spectrum are 1 MHz.
The remaining technical challenge is further decreasing the relative phase noise in the injection-locking dual-wavelength NIR light and further increasing the measured fTHz. Despite such injection-locking dual-wavelength NIR light being more stable in frequency and phase than free-running dual-wavelength NIR light, there is still room for a further decrease in its relative phase noise. Particularly, the use of two extracted EOM-OFC modes with a large frequency separation (= mfrep) for injection-locking spoils the low phase noise of frep characteristic in the OFC because it is equivalent to the high-order optical frequency multiplication of frep; hence, it increases the relative phase noise of the resulting dual-wavelength NIR light. One promising method for no optical frequency multiplication of frep is the use of the soliton microcomb instead of the EOM-OFC because it significantly increases frep up to a few hundreds of GHz to a few THz, which is much larger than frep in the EOM-OFC, while achieving stable soliton mode-locking oscillation with low phase noise. This soliton microcomb is a good reference for injection locking without the need for optical frequency multiplication [25,26]. Also, since the soliton microcomb has the larger ASE background than the EOM-OFC, the injection locking is more powerful to reduce it. In addition, the use of the soliton microcomb in the asynchronous NP-EO-DC provides a margin to further increase the measured fTHz toward the 6 G carrier frequency band. Soliton-microcomb-based, asynchronous NP-EO-DC has a good affinity for soliton-microcomb-based photonic THz generation and wireless data transfer at 6 G carrier frequency bands of 300 GHz and 560 GHz [9–13]. The anticipated challenge to use the soliton microcomb would likely involve the high-frequency compatibility of the EOP modulator. The EOP modulator used in this study was optimized for the W-band [23], but tuning it for higher frequency bands would be necessary. Additionally, if there is a decrease in conversion efficiency in frequency bands used for 6 G wireless carrier, there would be a need for technical innovations to address this issue.
6. Conclusions
We demonstrated THz-to-NIR carrier conversion by combining optical-heterodyne-enhanced detection of the asynchronous NP-EO-DC and high-OSNR, phase-locked dual-wavelength NIR light injection-locking to the EOM-OFC together with a high-sensitivity EOP modulator. The resulting RF beat signal, corresponding to the THz wireless carrier, achieved a SNR of 20 dB and high frequency stability. Despite an unmodulated THz wave being used as the wireless carrier, the achieved SNR and stability of the RF beat signal imply the possibility of extending this method to on-off keying amplitude modulation or multilevel modulation using amplitude and/or phase in the W-band. Although the center frequency of baseband signal was 1 GHz in the present setup, it can be selected within the range of spectral width in EOM-OFC (several THz) through the selection of injection-locked EOM-OFC modes. However, in practical terms, it is constrained to be several tens of GHz by the frequency bandwidth of the photodetector and subsequent amplifier. In the future, a combination of this method with the soliton microcomb will open the door of photonic THz detection for wireless data transfer in 6 G carrier frequency band. The proposed method is a powerful tool for photonic THz detection in THz communication.
Funding
Cabinet Office, Government of Japan (Promotion of Regional Industries and Universities); Japan Science and Technology Agency (JPMJPR1905); Ministry of Internal Affairs and Communications (JPJ000254); Tokushima Prefecture, Japan (Creation and Application of Next-Generation Photonics); Research Clusters program of Tokushima University (2201001).
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. S. Dang, O. Amin, B. Shihada, and M.-S. Alouini, “What should 6 G be?” Nat. Electron. 3(1), 20–29 (2020). [CrossRef]
2. T. Nagatsuma, G. Ducournau, and C. C. Renaud, “Advances in terahertz communications accelerated by photonics,” Nat. Photonics 10(6), 371–379 (2016). [CrossRef]
3. T. Udem, J. Reichert, R. Holzwarth, and T. W. Hänsch, “Accurate measurement of large optical frequency differences with a mode-locked laser,” Opt. Lett. 24(13), 881–883 (1999). [CrossRef]
4. M. Niering, R. Holzwarth, J. Reichert, P. Pokasov, T. Udem, M. Weitz, T. W. Hänsch, P. Lemonde, G. Santarelli, M. Abgrall, P. Laurent, C. Salomon, and A. Clairon, “Measurement of the hydrogen 1S-2S transition frequency by phase coherent comparison with a microwave cesium fountain clock,” Phys. Rev. Lett. 84(24), 5496–5499 (2000). [CrossRef]
5. D. J. Jones, S. A. Diddams, J. K. Ranka, A. Stentz, R. S. Windeler, J. L. Hall, and S. T. Cundiff, “Carrier-envelope phase control of femtosecond mode-locked lasers and direct optical frequency synthesis,” Science 288(5466), 635–639 (2000). [CrossRef]
6. T. Udem, R. Holzwarth, and T. W. Hänsch, “Optical frequency metrology,” Nature 416(6877), 233–237 (2002). [CrossRef]
7. T. Ishibashi and H. Ito, “Uni-traveling-carrier photodiodes,” J. Appl. Phys. 127(3), 031101 (2020). [CrossRef]
8. D. Fukuoka, K. Muro, and K. Noda, “Coherent THz light source based on photo-mixing with a UTC-PD and ASE-free tunable diode laser,” Proc. SPIE 9747, 974717 (2016). [CrossRef]
9. T. Tetsumoto, T. Nagatsuma, M. E. Fermann, G. Navickaite, M. Geiselmann, and A. Rolland, “Optically referenced 300 GHz millimetre-wave oscillator,” Nat. Photonics 15(7), 516–522 (2021). [CrossRef]
10. N. Kuse, K. Nishimoto, Y. Tokizane, S. Okada, G. Navickaite, M. Geiselmann, K. Minoshima, and T. Yasui, “Low phase noise THz generation from a fiber-referenced Kerr microresonator soliton comb,” Commun. Phys. 5(1), 312 (2022). [CrossRef]
11. Y. Tokizane, S. Okada, K. Nishimoto, Y. Okamura, H. Kishikawa, T. Minamikawa, E. Hase, J. Fujikata, M. Haraguchi, A. Kanno, S. Hisatake, N. Kuse, and T. Yasui, “Terahertz wireless communication in a 560-GHz band using a Kerr micro-resonator soliton comb,” Opt. Continuum 2(5), 1267–1275 (2023). [CrossRef]
12. B. M. Heffernan, Y. Kawamoto, K. Maekawa, J. Greenberg, R. Amin, T. Hori, T. Tanigawa, T. Nagatsuma, and A. Rolland, “60 Gbps real-time wireless communications at 300 GHz carrier using a Kerr microcomb-based source,” APL Photonics 8(6), 066106 (2023). [CrossRef]
13. Y. Tokizane, S. Okada, T. Kikuhara, H. Kishikawa, Y. Okamura, Y. Makimoto, K. Nishimoto, T. Minamikawa, E. Hase, J. Fujikata, M. Haraguchi, A. Kanno, S. Hisatake, N. Kuse, and T. Yasui, “Wireless data transmission in a 560-GHz band using low-phase-noise terahertz wave generated by photomixing of a pair of distributed feedback lasers injection-locking to Kerr micro-resonator soliton comb,” arXiv, arXiv:2308.02999 (2023). [CrossRef]
14. S. Hisatake, H. H. N. Pham, and T. Nagatsuma, “Visualization of the spatial–temporal evolution of continuous electromagnetic waves in the terahertz range based on photonics technology,” Optica 1(6), 365–371 (2014). [CrossRef]
15. S. Hisatake and T. Nagatsuma, “Nonpolarimetric technique for homodyne-type electrooptic field detection,” Appl. Phys. Express 5(1), 012701 (2012). [CrossRef]
16. S. Hisatake, H. Nakajima, H. H. N. Pham, H. Uchida, M. Tojyo, Y. Oikawa, K. Miyaji, and T. Nagatsuma, “Mapping of electromagnetic waves generated by free-running self-oscillating devices,” Sci. Rep. 7(1), 9203 (2017). [CrossRef]
17. S. Hisatake, K. Yamaguchi, H. Uchida, M. Tojyo, Y. Oikawa, K. Miyaji, and T. Nagatsuma, “Visualization of frequency-modulated electric field based on photonic frequency tracking in asynchronous electro-optic measurement system,” Appl. Phys. Express 11(4), 046601 (2018). [CrossRef]
18. S. Hisatake, G. Carpintero, Y. Yoshimizu, Y. Minamikata, K. Oogimoto, Y. Yasuda, F. van Dijk, T. Tekin, and T. Nagatsuma, “W-band coherent wireless link using injection-locked laser diodes,” IEEE Photonics. Tech. Lett. 27(14), 1565–1568 (2015). [CrossRef]
19. S. Hisatake, G. Carpintero, Y. Yoshimizu, Y. Minamikata, K. Oogimoto, Y. Yasuda, F. van Dijk, T. Tekin, and T. Nagatsuma, “Photonic millimeter wave transmitter for a real-time coherent wireless link based on injection locking of integrated laser diodes,” IEICE Trans. Electron. 98(12), 1105–1111 (2015). [CrossRef]
20. T. Sakamoto, T. Kawanishi, and M. Izutsu, “Asymptotic formalism for ultraflat optical frequency comb generation using a Mach-Zehnder modulator,” Opt. Lett. 32(11), 1515–1517 (2007). [CrossRef]
21. I. Morohashi, T. Sakamoto, H. Sotobayashi, T. Kawanishi, I. Hosako, and M. Tsuchiya, “Widely repetition-tunable 200 fs pulse source using a Mach–Zehnder-modulator-based flat comb generator and dispersion-flattened dispersion-decreasing fiber,” Opt. Lett. 33(11), 1192–1194 (2008). [CrossRef]
22. T. Kaji, I. Morohashi, Y. Tominari, N. Sekine, T. Yamada, and A. Otomo, “W-band optical modulators using electro-optic polymer waveguides and patch antenna arrays,” Opt. Express 29(19), 29604–29614 (2021). [CrossRef]
23. T. Kaji, I. Morohashi, Y. Tominari, M. Ohara, T. Yamada, and A. Otomo, “D-band optical modulators using electro-optic polymer waveguides and non-coplanar patch antennas,” Opt. Express 31(11), 17112–17121 (2023). [CrossRef]
24. D. Park, V. R. Pagán, P. S. Cho, J. Luo, A. K.-Y. Jen, and P. V. Petruzzi, “RF photonic downconversion of vector modulated signals based on a millimeter-wave coupled electrooptic nonlinear polymer phase-modulator,” Opt. Express 25(24), 29885–29895 (2017). [CrossRef]
25. Q. Zhang, B. Liu, Q. Wen, J. Qin, Y. Geng, Q. Zhou, G. Deng, K. Qiu, and H. Zhou, “Low-noise amplification of dissipative Kerr soliton microcomb lines via optical injection locking lasers,” Chin. Opt. Lett. 19(12), 121401 (2021). [CrossRef]
26. N. Kuse and K. Minoshima, “Amplification and phase noise transfer of a Kerr microresonator soliton comb for low phase noise THz generation with a high signal-to-noise ratio,” Opt. Express 30(1), 318–325 (2022). [CrossRef]
