Flat Millimeter-Wave Noise Generation Based on Mixing of Multiple Optical Spectra

被引:0
|
作者
Chen Y. [1 ]
Liu W. [1 ]
Sun Y. [1 ]
Romain Z. [1 ]
Wang Y. [1 ]
机构
[1] Guangdong Provincial Key Laboratory of Information Photonics Technology, School of Information Engineering, Guangdong University of Technology, Guangdong, Guangzhou
来源
关键词
flatness; millimeter-wave noise; mixing of multiple optical spectra; optical devices; opto-electrical conversion;
D O I
10.3788/CJL202249.0714001
中图分类号
学科分类号
摘要
Objective In this paper, we propose a novel photonic method for generating flat millimeter-wave (mm-wave) noises. The principle of our method consists in reshaping the broadband amplified spontaneous emission (ASE) spectrum into a multiple Gaussian optical noise spectrum with different central wavelengths before beating on a wideband photo-detector (P D). For a better comparison of the quality of the generated noise, the relative flatness (η) defined as the ratio of generated noise power fluctuation to its average power in a certain frequency range is put forward. The experimental results show that the relative flatness with the value of 0.46 [@ (35 ± 15) GHz] is obtained by the proposed method. Benefitting from the low frequency-independent loss due to photonics devices, similar results can be achieved in higher frequency bands such as F-band (90-140 GHz) and G-band (140-220 GHz). This point is numerically confirmed. Methods The aim of this paper is to improve the radio frequency (R F) spectral flatness of electrical noises generated by optically mixing multiple wavelength-sliced ASE lights. To do so, a broadband ASE spectrum from a superluminescent diode (SLD) source is reshaped into m narrowband Gaussian optical spectra using optical tunable filters with different central wavelengths (λ1, λ2, ..., λ m ). Among them, the wavelength difference of λ1 and λ2 is adjusted to correspond to a desirable central frequency f1 in the RF spectrum, and the difference between λ1 and λm is adjusted to decide the expected central frequency fm-1. At the same t i m e, except λ1 the optical wavelength difference (i. e., λ 3 - λ 2, λ 4 - λ 3, ..., λm - λm - 1) between two consecutive spectra needs to be less than or equal to the 3 dB bandwidth of the optical tunable filter. In these conditions, the generation of the flat mm-wave noise in a frequency range of f1 to fm-1 can be achieved. In addition, beating between the optical spectra centered at λ2 to λm produces the electrical noise near the zero frequency region. This electrical noise component is neglected in this paper, because it is out of the target frequency range. Results and Discussions According to Eqs. (3) and (4), it is demonstrated that the relative flatness of the noise signal is relevant to the number of optical spectra and full width at half maximum (FWHM) in F-band. The numerical simulation is performed assuming that the input optical power is maintained constant (Fig. 3). The optimal value of relative flatness is obtained when using six optical spectra and 0.1 nm FWHM. Due to the bandwidth limitation of the electrical devices such as the high-speed PD and electrical spectrum analyzer (ESA), we experimentally demonstrate the generation of the flat mm-wave noise in frequency range of 20 GHz to 50 GHz with a relative flatness of 0.46 [@(35 ± 15)GHz] using the mixing of multiple optical spectra [Fig. 4 (b)]. In theory, the optical wavelength differences of λ1 - λ2 and λ 1 - λ 3 can produce two Gaussian electrical spectra with center frequencies of f1 = 25 GHz and f2 = 37.5 GHz. The RF spectral flatness improvement of the mm-wave noise is obtained due to the superimposition of these two Gaussian electrical noise spectra. For a better comparison, we also investigate the relative flatness in the mixing conditions of two optical spectra with central wavelengths of λ1 and λ3 [Fig. 4 (d)]. It is noticed that the relative flatness is reduced to 0.65 [@(35 ± 15)GHz]. We can deduce that our proposed method enables a relative flatness improvement of about 0.19 [@(35 ± 15)GHz]. The simulation results relative to F-band noise generation using the proposed method are shown in Fig. 5. First, six optical noise spectra are applied to produce the mm-wave noise according to the Eqs. (3) and (4) with center wavelengths of λ1 = 1550.0 nm, λ2 = 1550. 7 nm, λ3 = 1550.8 nm, λ4 = 1550. 9 nm, λ5 = 1551. 0 nm, and λ6 = 1551. 1 nm, respectively [Fig. 5 (a)]. Except for λ1, the differences between two consecutive optical spectra with λ 3 - λ 2, λ 4 - λ 3, λ 5 - λ 4, and λ 6 - λ 5 are all equal to the FWHM of the optical spectra. The mixing of six optical noise spectra generates Gaussian electrical noise spectra with central frequency of f1 = 87.5 GHz, f2 = 100.0 GHz, f3 = 112.5.0 GHz, f4 = 125.0 GHz, and f5 = 137. 5 GHz. Their superimposition yields a broadband electrical noise covering the frequency band of 90-140 GHz. It is obtained a relative flatness of 0.18 [@(115 ± 25)GHz] in F-band [Fig. 5 (b)]. At the same time, the case of two optical spectra mixing with center wavelengths of λ± and λ4 is investigated [Fig. 5 (c)]. The produced uneven electrical noise spectrum has a relative flatness of 1.73 [@(115 ± 25)GHz] [Fig. 5 (d)]. In summary, our proposed method for optical mixing of multiple wavelength-sliced ASE signals can generate a flatter F-band mm-wave noise in comparison to that for optical mixing of two noise spectra. Conclusions We theoretically and experimentally demonstrate the generation of flat millimeter-wave noises by optically mixing multiple Gaussian wavelength-sliced ASE lights. The proof-of-principle experiment has shown the effectiveness of our proposed method with a satisfactory relative flatness as low as 0. 46 [@ (35 ± 15)GHz]. Considering the non-dependency of fibers loss on frequency and the use of a photodetector with faster response such as uni-traveling-carrier photodiode (UTC-PD), the experimental implementation of our method in a higher frequency band can be expected. Our proposed method exhibits potential features for the noise parameter measurement of millimeter-wave devices. © 2022 Science Press. All rights reserved.
引用
下载
收藏
相关论文
共 21 条
  • [11] Duan G H, Georgiev E., Non-white photodetection noise at the output of an optical amplifier: theory and experiment, IEEE Journal of Quantum Electronics, 37, 8, pp. 1008-1014, (2001)
  • [12] Song H J, Shimizu N, Furuta T, Et al., Subterahertz noise signal generation using a photodetector and wavelength-sliced optical noise signals for spectroscopic measurements, Applied Physics Letters, 93, 24, (2008)
  • [13] Shi J W, Kuo F M, Chiueh T, Et al., Photonic generation of millimeter-wave white-light at W-band using a very broadband and high-power photonic emitter, IEEE Photonics Technology Letters, 22, 11, pp. 847-849, (2010)
  • [14] Song H J, Shimizu N, Kukutsu N, Et al., Microwave photonic noise source from microwave to sub-terahertz wave bands and its applications to noise characterization, IEEE Transactions on Microwave Theory and Techniques, 56, 12, pp. 2989-2997, (2008)
  • [15] Sun J, Wu Y D, Wu W F, Et al., Optimization of polarization-dependent loss of arrayed waveguide grating demultiplexer [J], Chinese Journal of Lasers, 47, 1, (2020)
  • [16] Wang T, Chen H M, Jia H M, Et al., Performance research and fabrication of 1310 nm superluminescent diodes with high power [J], Acta Photonica Sinica, 50, 6, (2021)
  • [17] Derickson D., Fiber optic test and measurement, pp. 169-220, (1998)
  • [18] Nazarathy M, Sorin W V, Baney D M, Et al., Spectral analysis of optical mixing measurements, Journal of Lightwave Technology, 7, 7, pp. 1083-1096, (1989)
  • [19] Vidal B., Broadband photonic microwave noise sources, IEEE Photonics Technology Letters, 32, 10, pp. 592-594, (2020)
  • [20] Qiao L J, Lii T S, Xu Y, Et al., Generation of flat wideband chaos based on mutual injection of semiconductor lasers [J], Optics Letters, 44, 22, pp. 5394-5397, (2019)