Open Access
Add to BibSonomyAbstract
We present the first down-conversion intensity modulated-direct detection (IM-DD) RF photonic link that achieves frequency down-conversion using the nonlinear optical phase modulation inside a Mach-Zehnder (MZ) modulator. The nonlinear phase modulation is very sensitive and it can enable high RF-to-IF conversion efficiency. Furthermore, the link linearity is enhanced by canceling the nonlinear distortions from the nonlinear phase modulation and the MZ interferometer. Proof-of-concept measurement was performed. The down-conversion IM-DD link demonstrated 28dB improvement in distortion levels over that of a conventional IM-DD link using a LiNbO3 MZ modulator.
© 2016 Optical Society of America
1. Introduction
An RF photonic link can be used to connect a remote antenna and a signal processor for advanced radar front end applications. The fiber optic link provides notable benefits, including low weight, small size, wideband, and immunity to electro-magnetic interference. With recent developments in high dynamic RF photonic links [1], the optic link also promises to bring the much needed dynamic range improvement to a radar front end. Separately, a radar front end should also down-convert antenna signal to intermediate frequency (IF) or baseband before it can be digitized and processed. Usually the frequency down-conversion function is performed by electronic mixers. But they often add conversion loss and limit the dynamic range of the entire front end. Several optical techniques for frequency down-conversion have been proposed, including dual MZ modulator [2], optical heterodyning [3,4], optical sampling [5,6], coherent I/Q demodulation [7] and nonlinear optical phase modulation [8]. Among these methods, the nonlinear optical phase modulation has demonstrated the best down-conversion performance with 116dB∙Hz2/3 spurious-free dynamic range (SFDR) and 2dB conversion loss. The performance should further scale up with longer modulator length. However, this is a phase modulated coherent system that is sensitive to environmental perturbations. Although mitigation is possible, mitigation approaches often add complexity and cost. Therefore, for many front end systems that only require moderate SFDR, an efficient frequency down-conversion method based on a simple Intensity Modulated-Direct Detection (IM-DD) RF photonic link is still appealing.
In this paper we propose a new down-conversion IM-DD RF photonic link that achieves frequency down-conversion by nonlinear optical phase modulation. As depicted in Fig. 1, the RF input from an antenna is combined with an optically fed LO signal, and then applied to a multiple quantum well (MQW) nonlinear optical phase modulation (NPM) section of a Mach-Zehnder (MZ) intensity modulator. Due to the second order nonlinearity, the output optical phase of the NPM section contains the mixing product between the RF and LO signals. Therefore, at the output of the MZ modulator, the RF signal is down-converted to IF and coded to the optical intensity. Via a fiber optic link, the IF modulated optical signal is delivered to a photodetector inside a receiver.
Fig. 1 Down-conversion IM-DD RF photonic link using nonlinear phase modulation inside an MZ modulator.
In this new approach the system linearity is enhanced by canceling the nonlinear distortions from the Mach-Zehnder interferometer and the nonlinear optical phase modulation section. This new approach demonstrated 28dB improvement in nonlinear distortion level over that of a conventional IM-DD link with a LiNbO3 MZ modulator.
2. Theoretical analysis
The output of the down-conversion IM-DD RF photonic link (see Fig. 1) is given by:
where IPD is the photodetector DC photocurrent, ZPD is its termination resistance, and ∆θ is the nonlinear optical phase modulation inside the MQW nonlinear phase modulation section. The nonlinear optical phase modulation inside the MQW modulator is caused by band-filling [9]. When biased near forward turn-on, the phase modulation ∆θ can be expressed as:where VT is the thermal voltage. ARF and ALO are the RF and LO voltage magnitudes, respectively. ωRF and ωRF are their angular frequencies. η is a constant related to the MQW modulator forward current-voltage characteristics [10]. α is a coefficient that is related to the MQW modulator bias voltage, quantum well design and the modulator length. It should be noted that the exponential response (2) contains an unwanted linear phase modulation term. However, the modulator has a low-pass response due to its band-filling mechanism. Therefore, the linear modulation term is suppressed as they are in the high RF or LO frequencies. In addition, the linear modulator term can be further significantly suppressed by employing a balanced configuration similar to that of a balanced electronic mixer. From Eq. (2), we can obtain the MQW modulator’s effective down-conversion Vπ, which is defined as the RF input amplitude required to achieve π phase shift in the IF band:For deriving (3), we considered up to the 6th order term of the exponential function in (2) and we also assume that the RF signal is small. Upon substituting (2) into (1) and assuming the small angle approximation (i.e. sin(θ) = θ), we found that the RF to IF conversion voltage gain is:The nonlinearity of the down-conversion IM-DD link output arises from the nonlinear responses of the phase modulation (Eq. (2)) and the MZ interferometer (Eq. (1)). In presence of two-tone RF input (with frequencies of ωRF1 and ωRF2), the in-band third order intermodulation distortions (IMD3) generated by the nonlinear phase modulation and the MZ interferometer are respectively given by:
where ω1 and ω2 are the frequencies of the in-band IMD3: and . It is important to note from Eq. (5) that the distortions from the nonlinear phase modulation and the MZ interferometer are of opposite polarities. Therefore, they always cancel each other and perfect IMD3 cancellation may be achieved.by controlling the bias voltage applied to the nonlinear phase modulation section and the LO power.3. Experiment
Proof-of-concept experiment was performed using an MQW MZ modulator. As shown in Fig. 2, the MQW modulator employs a p-i-n device configuration. Its intrinsic region contains 13 periods of QWs. The MQW modulator uses a deep-etched ridge optical waveguide. The MQW MZ modulator consists of four 1.5-mm long phase modulation sections and two MMI couplers.
Fig. 2 MQW MZ modulator used in measurement. (a) Modulator waveguide structure; (b) MZ modulator layout.
The MQW MZ modulator was fabricated using Harvard CNS facilities. The MQW base epitaxial wafer is provided by Landmark, Inc. In device fabrication, deep ridge waveguides (see Fig. 3(a)) were first etched by inductively coupled plasma reactive-ion etching (ICP-RIE) using a SiO2 hard mask. Then, wet etch was performed to expose the n-contact layer. Next, the n-contact metal (5nm Ni/14.5nm Ge/25.5nm Au/14.5nm Ge/25.5nm Au/20nm Ni/200nm Au) was deposited and annealed. In order to electric-isolate different modulation sections, the InGaAs p-contact layer was selectively removed followed by selective proton implantation. To reduce the parasitic capacitance and improve the modulator bandwidth, the n-contact and the n-doping InP layers were also selectively removed. Then, p-contact metal (20nm Ti/40nm Pt/300nm Au) was deposited on top of the waveguide ridge (see Fig. 3(b)) and annealed. In order to further reduce the parasitic capacitance, a thick low stress dielectric film (~1 µm) was deposited by plasma-enhanced chemical vapor deposition (PECVD). Then, the vias for n-contact metal and p-contact metal were opened and thick interconnect metal (20nm Ti/1500nm Au) was deposited. At last, the wafer was thinned by mechanical lapping, and devices were then diced out and mounted on AlN carriers. Figure 3(c) shows a microscope image of the fabricated MQW MZ modulator devices.
Fig. 3 MQW MZ modulator fabrication. (a) SEM image of the deep-ridge waveguide; (b) SEM image of the waveguide after p-contact metal deposition; (c) the fabricated MQW MZ modulator.
The complete down-conversion IM-DD link experimental setup is shown in Fig. 4. The amplified output of a polarization maintaining (PM) single frequency fiber laser was coupled in and out of the MQW MZ modulator using two PM lens-fibers. The lens-fibers have a focused beam spot diameter of 2.5 microns. The RF and LO signals were combined using power combiners and applied simultaneously to the two 1.5 mm long modulation sections on the upper arms of the MQW MZ modulator. Therefore, the total phase modulation length is 3mm. A DC voltage was applied to one phase modulation section on the other arm of the MZ modulator to set its phase offset at quadrature. The modulator output is detected by a photodetector. An RF spectrum analyzer was used to analyze the link output.
As suggested by the theoretical analysis, the nonlinear distortion can be significantly cancelled by controlling the bias voltage applied to the nonlinear phase modulation section and the LO power. Figure 5(a) shows an example of the photodetector output IF spectrum when the bias voltage is set at 0.73volt. Prior to power combining, the RF (2GHz) and LO (1.99GHz) powers were −7dBm and 10dBm, respectively. But because of the modulator’s low impedance (2.8-j4.25 Ω, measured by a network analyzer) at 2GHz and the loss of the power combiners, the RF and LO voltage applied to the modulator were only 0.014V and 0.147V, respectively. The measured output IF power was ~-19dBm (0.035V) per tone. Thus, the RF-to-IF voltage gain was found to be ~8dB.
Fig. 5 IMD3 measurement at IF output. (a) Down-conversion IM-DD link; (b) LiNbO3 MZM link under similar modulation index.
The measured IF output power corresponds to a phase modulation index of 0.35 since the DC photocurrent of the photodetector was 2mA. From the input RF voltage and the phase modulation index, we can determine the MQW modulator’s effective down-conversion Vπ to be:
Even with the added down-conversion function, this Vπ value is far superior to that of existing state-of-the-art MZ modulator devices.
In addition, Fig. 5(b) shows the output spectrum of a commercial LiNbO3 MZ modulator under an identical modulation index (0.35). We can see that the nonlinear distortion level of the down-conversion IM-DD link is 28dB better than that of the non-down-conversion IM-DD link with the commercial LiNbO3 MZ modulator.
Figure 6(a) shows the IM-DD link OIP3 measurement. The OIP3 is also measured with different RF frequencies. It was found to be near 10dBm from 500MHz to 4GHz RF frequencies (see Fig. 6(b)). The measurement was limited to 4GHz due to equipment limitation. It should also be noticed that the nonlinear distortion cancellation is highly dependent on the bias voltage. For field deployment of the proposed link, both temperature and bias voltage control should be implemented to assure optimum distortion cancellation.
4. Conclusion
We presented a new down-conversion IM-DD RF photonic link that employs nonlinear optical phase modulation inside an MQW phase modulator. Since the nonlinear distortions from the phase modulation and the MZ interferometer has opposite polarity, the link linearity is enhanced by canceling these unwanted nonlinear distortions. The concept was verified using a custom MQW MZ modulator. By adjusting the phase modulator’s bias and LO power, the cancellation of the nonlinear distortions were observed. The proposed link exhibited 28dB improvement in distortion levels comparing with a conventional IM-DD link without down-conversion. The OIP3 of the link was measured to be ~10dBm over a RF input frequency range from 500MHz to 4GHz.
Acknowledgments
This work was supported by the Air Force Office of Scientific Research (AFOSR) (grant FA9550-12-1-0194).
References and links
1. S. Jin, L. Xu, P. Herczfeld, A. Bhardwaj, and Y. Li, “Recent progress in attenuation counter-propagating optical phase-locked loops for high-dynamic-range radio frequency photonic links,” Photon. Res. 2(4), B45–B53 (2014). [CrossRef]
2. G. K. Gopalakrishnan, W. K. Burns, and C. H. Bulmer, “Microwave-optical mixing in LiNbO3 modulators,” IEEE Trans. Microw. Theory Tech. 41(12), 2383–2391 (1993). [CrossRef]
3. Y. Li, P. Herczfeld, A. Rosen, M. Bystrom, and T. Berceli, “Optical domain down-conversion of microwave signals for high dynamic range microwave fiber optics links,” in IEEE International Topical Meeting on Microwave Photonics, (IEEE 2006), pp. 1–4. [CrossRef]
4. K. Y. Tu, M. S. Rasras, D.M. Gill, S.S. Patel, Y.K. Chen, A.E. White, A Pomerene, D. Carothers, J. Beattie, M. Beals, J. Michel, and L.C. Kimerling, “Silicon RF-Photonic filter and down-converter,” J. Lightwave Technol. 28(20), 3019–3028 (2010).
5. D. Zibar, L. A. Johansson, H. F. Chou, A. Ramaswamy, M. J. W. Rodwell, and J. E. Bowers, “Investigation of a novel optical phase demodulator based on a sampling phase-locked loop,” IEEE International Topical Meeting on Microwave Photonics, (IEEE 2006), pp. 1–4. [CrossRef]
6. A. Ramaswamy, L. A. Johansson, J. Klamkin, D. Zibar, L. A. Coldren, M. J. Rodwell, and J. E. Bowers, “Optical phase demodulation of a 10GHz RF signal using optical sampling,” in Coherent Optical Technologies and Applications, (COTA 2008), CtuC3.
7. T. R. Clark, S. R. O’Connor, and M. L. Dennis, “A phase-modulation I/Q-demodulation microwave-to-digital photonic link,” IEEE Trans. Microw. Theory Techn. 58(11), 3039–3058 (2010). [CrossRef]
8. Y. F. Li, R. Y. Wang, P. Herczfeld, J. Klamkin, L. Johansson, and J. Bowers, “RF frequency down-conversion with quadratic electro-optic effect,” in IEEE MTTS Int Microw Symp., (IEEE 2009), pp.153–156. [CrossRef]
9. B. R. Bennett, R. A. Soref, and J. A. Alamo, “Carrier-induced change in refractive index of InP, GaAs and InGaAsP,” IEEE J. Quantum Electron. 26(1), 113–122 (1990). [CrossRef]
10. B. L. Sharma and R. K. Purohit, Semiconductor Heterojunctions (Pergamon, 1974), Chap. 1.
