Jpn. J. Appl. Phys. Vol. 40 (2001) pp. 2728–2730 Part 1, No. 4B, April 2001 c 2001 The Japan Society of Applied Physics
Optical Binary-Phase-Shift-Keying Modulation of 10 GHz Using an Integrated Hybrid Device Myunghun S HIN∗ , Jiyoun L IM, Jeha K IM1 , Jeong-Soo K IM1 , Kwang-Eui P YUN1 and Songcheol H ONG Department of Electrical Engineering, Korea Advanced Institute of Science and Technology (KAIST), 373-1 Kusong-Dong, Yusong-Gu, Taejon 305-701, South Korea 1 Telecommunication Basic Research Laboratory, Electronics Telecommunication Research Laboratory (ETRI), Yusong P.O. Box 106, Taejon 305-600, South Korea (Received August 29, 2000; accepted for publication November 28, 2000)
An integrated optical/RF hybrid device was designed and fabricated for binary-phase-shift-keying (BPSK) subcarrier modulation. The device is composed of two identical high-speed multiple quantum well electroabsorption modulators branched with two multimode interference couplers. Using the device, we demonstrated that a RF subcarrier of 10 GHz is directly modulated in a BPSK scheme with an independently applied RF carrier and digital signal. KEYWORDS: BPSK, MQW, electroabsorption, modulator, MMI
1. Introduction RF resources of microwaves and millimeter waves (MMWs) are receiving increasing attention since they are expected to provide wide-band channels required for both mobile communication systems and wireless local area networks.1) To realize such systems, an optical and RF hybridization approach is proposed as a promising method. Hybrid fiber/coax networks allow robust transmission of RF subcarrier signals via a single mode fiber, so that significant reduction in a number of cascade RF amplifiers is achieved. Digital transmission systems with microwaves and MMWs using quadrature-phase-shift-keying (QPSK) or binary-phase-shift-keying (BPSK) modulation have been reported.2, 3) Although MMW bands have the capacity higher than 1 Gb/s, previous studies reported data transmission of only a few hundred Mb/s.4, 5) This is due to the lack of a proper modulation scheme for the hybrid approach. In general, up-conversion schemes in the transmitter use an intermediate frequency to convert a baseband signal to a MMW band signal, thus the data rate cannot exceed the intermediate frequency. High-frequency amplifiers and mixers, which are very difficult to realize, are also needed if one wants to utilize the full capacity of MMW carriers. We proposed a binary-phase-shift method using a single integrated hybrid device. The device is composed of two identical high-speed multiple quantum well (MQW) electroabsorption (EA) modulators branched with two multimode interference (MMI) couplers. The phase of the RF carrier was changed by both the optical amplitude modulation of modulators and the interference resulted from the phase delay at MMI couplers. The results of the modulation operation of the device were presented at the low-frequency region.6) In this paper, the BPSK modulation of 10 GHz is reported. The f 3dB bandwidth of the device and the frequency spectrum of the modulated optical signal were also measured. 2. Device and Fabrication The device structure is illustrated in Fig. 1. The input 3 dB MMI coupler divides and splits an incident light into two identical MQW EA modulators. The MMI couplers are de∗ E-mail
Fig. 1. An integrated optical/RF hybrid device. The input optical signal undergoes a phase change of π when it passes through two MMIs.
signed in such a way that it has a mirror image at an output due to paired interference.7) Thus, two optical outputs of the MMI coupler have the same amplitude but have a π /2 phase difference. Since two identical MMI couplers are placed at both input and output ports as shown in Fig. 1, the optical fields from two modulators interfere with each other destructively so that the output of the device produces an out-of-phase image. On the other hand, the amplitude of the signal in each branch is changed when it passes through the EA modulator (M1 and M2 ). When M1 is turned off, the optical amplitude at the output is directly proportional to that of M2 . When M1 is turned on, the output amplitude is dependent on the difference of the optical amplitudes through both M1 and M2 . Assume that a digital on-and-off signal (0 or 1) is applied to M1 while a sinusoidal signal of amplitude m 2 and frequency ω is applied to M2 , then the optical amplitude at the output can be written experimentally as follows. A0 [1 − m 2 sin(ωt)] if M1 is turned ON ∼ Pout-of-phase = A0 [1 + m 2 sin(ωt)] if M1 is turned OFF = A0 [1 + m 2 sin(ωt + π(M1 ))],
(1)
where A0 is the maximum optical power at the output port. Note that the sign of amplitude m 2 is changed. Thus, the phase of the sinusoidal signal generated at M2 is shifted by
address:
[email protected] 2728
Jpn. J. Appl. Phys. Vol. 40 (2001) Pt. 1, No. 4B
π in accordance with the state of M1 . Since the subcarrier and the digital data are introduced independently to the modulators, the subcarrier frequency and the data rate are limited only by the bandwidth of high-speed EA modulators. Therefore, the direct up-conversion from baseband to MMW band can be achieved when modulators with a 3 dB bandwidth of a few ten GHz are used. In addition, the synchronization of a local oscillator for RF subcarrier generation and a digital signal is not necessary in this scheme. This is because this modulation mechanism makes the signal phase change at an arbitrary time by optical interference. The device was fabricated in the following process. The epitaxial layers of the modulator region were prepared on an n+ -InP substrate by metalorganic chemical vapor deposition (MOCVD). The MQW absorption layers of 250 nm thickness were grown between graded-index clads of 70 nm thickness. The MQW layers consisted of 10 pairs of 1.52Q-InGaAsP wells (λg = 1.52 µm) with 10 nm thickness and 1.2Q-InGaAsP barriers (λg = 1.2 µm) with 6.5 nm thickness. The low-loss passive region was grown by butt-coupled regrowth. Deep dry etching was applied to form the waveguide structures including MMI couplers and MQW EA modulators. A 3-µm-thick polyimide layer was employed to block the leakage current and to reduce the pad capacitance.6)
M. S HIN et al.
2729
plied. When a voltage of −4 V was applied to M1 (OFF state), the out-of-phase output was decreased as the reverse voltage was increased to M2 . On the contrary, when M1 was biased to 0 V (ON state), the output was increased with the reverse voltage at M2 . This is due to the fact that when M1 is on, the difference in the optical output between the two modulators increases at the output port as M2 becomes opaque. We measured the overall modulation characteristics using an optical power detector. Figure 4 shows that the bias voltage of M1 changes the sign of the optical signal modulated with the reverse voltage of M2 . The phase of the output signal was then changed by π in comparison with that of the input signal. The 3 dB bandwidth of the device was 18 GHz, as shown in Fig. 5. An RF signal of 10 GHz with 5 dBm to M2 was driven and a 4 V 10 MHz rectangular pulse train to M1 was driven directly as a baseband digital signal. We have reported a low-frequency operation,6) but the light wavelength is slightly changed (below 1 nm) by high-speed amplitude modulation so that the interference between the RF carrier light and digitally modulated light could be different from that of low-frequency operation. We confirmed its high-frequency operation via frequency spectrum analysis since in this way the frequency performance of the modulated signal can be easily analyzed even without using a high-speed optical oscil-
3. Measurement Results The characteristics of MMI couplers and EA modulators strongly depend on the wavelength and polarization of light. We used an HP8168F tunable laser and controlled the polarization with a three-paddle fiber rotating polarizer. The adjustments were monitored using an IR-Vidicon camera. As for the EA modulator, we measured the optical and electrical properties of the EA modulators in this device. The on-off ratio at −3 V was 16 dB for the TE mode and 13 dB for the TM mode at 1.53 µm wavelength as in Fig. 2. The divide optical power from the MMI couplers was measured as the different launched wavelength in Fig. 3. Thus, a 1.53 µm light of the TE mode was selected to operate the MMI couplers under the optimum condition and to have a large on-off ratio of the modulators. The incident power of the light was set at 5 mW. Each modulator was turned off when a reverse voltage of 4 V is ap-
Fig. 2. Extinction ratio vs wavelength. Solid square and open circle represent TE and TM polarization modes, respectively.
Fig. 3. Divided output power at the output of MMI couplers vs wavelength for the TE mode.
Fig. 4. Normalized output power vs applied carrier signal voltage. –●– VM1 = 0 V –– VM1 = −4 V.
2730
Jpn. J. Appl. Phys. Vol. 40 (2001) Pt. 1, No. 4B
M. S HIN et al.
(carrier) frequency. As expected, the power output at the center frequency f c of 10 GHz appeared to have decreased. This result implied that the phase of the carrier signal is shifted by π periodically in accordance with the pulse train. This shows that a high-frequency operation yields the same results as previous results of a low-frequency operation.6) This is because the operation wavelength range of the MMI couplers (Fig. 3) is wider than the wavelength changes brought about by presence of an RF carrier. 4. Conclusions
Fig. 5. Small signal frequency response of the MQW EA modulator at the bias voltage of −1 V for the optical wavelength λ and the length of the modulator L. (a) λ = 1530 nm, L = 100 µm (b) λ = 1530 nm, L = 150 µm (c) λ = 1550 nm, L = 150 µm.
We demonstrated that a baseband signal has been up-converted to a 10 GHz RF subcarrier using the optical/RF hybrid device. The device was composed of two identical MQW EA modulators of 18 GHz bandwidth and two MMI couplers that are mirror images. Since the working frequency and the data rate are limited only by the bandwidth of the modulator, one can readily extend this operation principle to the MMW subcarrier and the data rate to its full bandwidth capacity. Acknowledgements This study was partially supported by the Ministry of Information and Communications of ETRI.
Fig. 6. Spectrum response of the modulated output near the center frequency.
loscope. The frequency behavior of the modulated signal was observed using a frequency spectrum analyzer as in Fig. 6. It was similar to that of the rectangular pulse train, a unique form of a sinc-function-like spectrum except at the center
1) D. Novak, G. H. Smith, C. Lim, H. F. Liu and R. B. Waterhouse: Tech. Dig. Opt. Fiber Commun. Conf. OFC’98, San Jose, CA, Feb. 1998, paper TuC1. 2) L. Noel, D. Wake, D. G. Moodie, D. D. Marcenac, L. D. Westbrook and D. Nesset: IEEE Trans. Microwave Theory & Techniq. 45 (1997) 1416. 3) T. K. Woodward, S. Hunsche, A. J. Ritger and J. B. Stark: IEEE Photon. Technol. Lett. 11 (1999) 382. 4) C. Cox III, E. Acherman, R. Helkey and G. E. Betts: IEEE Trans. Microwave Theory & Techniq. 45 (1997) 1375. 5) K. Kitayama, T. Kuri and Y. Ogawa: Tech. Dig, Opt. Fiber Commu. Conf. OFC’98, San Jose, CA, Feb. 1998, paper TuC2. 6) M. Shin, J. Lim, C. Y. Park, J. Kim, J. S. Kim, K. E. Pyun and S. Hong: IEEE Photon. Technol. Lett. 12 (2000) 516. 7) L. B. Soldano and E. C. M. Pennings: J. Lightwave Technol. 13 (1995) 615.
