Widely Tunable Short-Pulse Generation With Ultralong Semiconductor ...

754

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 28, NO. 5, MARCH 1, 2010

Widely Tunable Short-Pulse Generation With Ultralong Semiconductor Optical Amplifiers Patrick Runge, Christian-Alexander Bunge, Member, IEEE, Klaus Petermann, Fellow, IEEE, Michael Schlak, Walter Brinker, and Bernd Sartorius

Abstract—Ultralong bulk semiconductor optical amplifiers (bulk UL-SOAs) have tremendous four-wave mixing (FWM) efficiencies so that due to two continuous wave (CW) input signals, a broad mode comb at the output of the UL-SOA can be obtained. When using tunable single mode lasers as input signals, the generated FWM mode comb is tunable in the mode spacing and the wavelength. As a result in time domain, widely tunable pulses in the repetition rate and the carrier frequency are obtained. Since the phase signature of the FWM modes is not linear over the whole mode comb, different filtering techniques can be applied for filtering modes with linear phase in order to efficiently generate short pulses. With the optimal filtering technique, pulses with full width at half maximum of less than 2 ps and a pulse repetition rate of 20 GHz can be obtained. Index Terms—Four-wave mixing, picosecond pulse generation, supercontinuum, ultralong semiconductor optical amplifier.

I. INTRODUCTION

W

ITH increasing demand for higher transmission rates in optical communication systems, the transmission capacity has to be optimised. In [1] a combination of wavelength division multiplexing (WDM) and time division multiplexing (TDM) has been applied to transmit 2.56 Tb/s over a 160 km fiber link. For the TDM part, add-drop multiplexing plays a key role, where short pulses as a clock or for switching are needed. Moreover, a trend towards optical packet switched networks can be observed where again short pulses are needed for all-optical signal processing. Furthermore, in packet switched networks the components should be agile, in order to fulfil operations at different wavelengths and different data rates [2]. In general, the Kerr effect in highly nonlinear fibers (HNLFs) is used for generating a supercontinuum [3] which in turn can be used for the generation of short pulses [4], [5]. Typically, the Kerr effect creates self-phase modulation (SPM) due to its dependence of the input signal’s optical power. The additional modulation in

Manuscript received August 17, 2009; revised November 05, 2009 and November 08, 2009. First published December 01, 2009; current version published February 24, 2010. This work was supported by the Deutsche Forschungs-gemeinschaft (DFG). P. Runge and K. Petermann are with the Fachgebiet Hochfrequenztechnik, Technische Universität Berlin, D-10587 Berlin, Germany (e-mail: [email protected]). C.-A. Bunge is with the Fachgebiet Photonik, Hochschule für Telekommunikation Leipzig, D-04277 Leipzig, Germany (e-mail: [email protected]). M. Schlak, W. Brinker, and B. Sartorius are with the Fraunhofer-Institut für Nachrichtentechnik, Heinrich-Hertz-Institut, D-10587 Berlin, Germany (e-mail: [email protected]). Color versions of one or more of the figures in this article are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/JLT.2009.2037522

combination with a dispersion managing component creates a broader spectra so that in time domain short pulses can be obtained. But due to the HNLFs, the source cannot be easily integrated. Another device that creates short pulses and can be integrated, is the mode-locked laser, but its possibilities for tuning the repetition rate or the carrier frequency of the pulses are very limited, because these properties are defined by the material and the geometry [6]. Opposite to the passively mode-locked laser, the monolithic fundamental actively mode locked laser can be tuned in the repetition rate of the pulses, but needs a complex electronic driving stage in order to obtain high repetition rates [7]. Furthermore, when reaching with the pulse duration the terahertz regime (0.6–6 THz) also non-telecommunication applications in the field of biomedical or imaging applications are possible [8], [9]. Recently, the capability of ultralong semiconductor optical amplifiers (UL-SOAs) for pulse compression was presented in [10]. In this article, the capability of UL-SOAs for short pulse generation is investigated numerically and experientially. Opposite to the previously discussed methods of generating short pulses, the single-pass scheme presented here is widely tunable and can be integrated. Since UL-SOAs have a tremendous fourwave mixing (FWM) efficiency, they can create broad mode combs at their output from two continuous wave (CW) input signals. The article is structured as follows. In Section II the mechanisms in UL-SOAs for the supercontinuum generation are briefly discussed and some results are presented. With these basics, the results of the pulse generation with UL-SOAs are discussed in Section III. An outlook for the UL-SOA’s potential of pulse generation is given in Section IV. II. FWM MODE COMB OF UL-SOAS A. Principle of Operation Due to the device length, the main part of UL-SOAs is deeply saturated by the amplified input signals and the amplified spontaneous emission (ASE) noise. Typically, the saturation starts for standard bulk SOAs after approximately 1 mm of propagation. Because of the high optical power in the saturated section, the carrier density is fixed to the net transparency level and only the fast nonlinear intraband effects like carrier heating (CH) and spectral hole burning (SHB) influence the signals. While for the intraband effects the carrier density distribution inside of a band changes, the overall number of carriers of the band remains unchanged. For this reason, the response time of these effects is about 1000 times shorter compared to the slow interband effects. On the other hand, their impact is also weaker so that UL-SOAs

0733-8724/$26.00 © 2010 IEEE

RUNGE et al.: WIDELY TUNABLE SHORT-PULSE GENERATION WITH ULTRALONG SOAS

755

B. Results

Fig. 1. Conceptual simulation and measurement setup; TSML = tunable single mode laser OSA = optical spectrum analyser and OSO = optical sampling oscilloscope. ;

have a length of several mm in order to benefit from the fast intraband effects. The slow interband effects should be suppressed as far as possible in order to provide a stable operation point. When injecting multiple co-polarized signals at the same time into an UL-SOA, in the saturated section of the UL-SOA the beating of the signals creates a dynamic gain grating due to the fast intraband effects. In active semiconductors a gain change is also related to a phase change due to the -factor also causing a dynamic index grating. The interaction of the signals with these dynamic gratings generates FWM. Due to the short relaxation times of the intraband effects, FWM can take place over several nm. Furthermore, the FWM products again interact with their neighbour signals so that cascaded FWM processes take place and proceeds multiple times. As a result, a broad spectral FWM mode comb can be obtained at the end of the UL-SOA. Moreover, a FWM related effect, called the Bogatov-like effect, occurs in the saturated section. In [11], Bogatov has demonstrated that due to the dynamic gain and index grating in nonlinear semiconductor media, the weaker signals amplification is dependent on the stronger signals wavelength detuning. Due to the -factor causing the asymmetry, the weaker signal is less amplified on the shorter wavelength side than on the longer wavelength side. With decreasing -factor the amplification asymmetry becomes more and more symmetric [12]. Different from [11], in UL-SOAs the effect is also caused by the fast intraband effects so that the amplification asymmetry can be observed over several nanometres. Similar to the amplification asymmetry, the phase of the weaker signal is also affected in dependence of the detuning. The device and the simulation tool, used in the investigations, are the same as in [13]. The simulation tool is time-domain SOA model [14]. The SOA is divided into segments corresponding to the sampling time of the input signals. The wavelength dependent gain of each these SOA segments is modelled with adaptive finite-impulse response filter. With the help of rate equations, the model also accounts for nonlinearities like SHB and CH. For the simulation, typical parameters for a 1550 nm InGaAsP bulk SOA had been taken from the literature. The setup of the presented scheme is depicted in Fig. 1. As default parameters for simulations and measurements an 8-mmlong bulk UL-SOA was pumped with a current of 300 mA/mm. The co-polarized CW input signals were injected with an input power of 8.5 dBm each into the UL-SOA. When two tunable single mode lasers (TSMLs) are used as input signal sources of the UL-SOA, the mode spacing and the center wavelength can be easily tuned. As a result, the generated output mode comb can also be tuned with respect to the mode spacing and the center wavelength.

Fig. 2 shows spectra at the output of the UL-SOA for two CW signals at three different input wavelengths. Comparing measurements and simulations an excellent match can be observed. A saddle point close by the high optical power modes can be observed on the shorter wavelength side of the output spectra. No matter where the FWM comb is located in the gain spectrum, the saddle point is always on the shorter wavelength side of the FWM comb. Similarly, the Bogatov-like effect attenuates modes on the shorter wavelength side of a stronger signals. For this reason the saddle point can be ascribed to the attenuation caused by the Bogatov-like effect. Furthermore, for the low power modes the feedback of the FWM products at the parent modes can be neglected and a strictly decaying smooth shape due to cascaded FWM processes can be observed. Opposite to the low power modes, the power of the high power FWM modes around the wavelength of the input modes fluctuates over several dBm compared to the neighbor modes. In the high power mode region also the feedback of the FWM products on the parent modes has to be considered. As a result, irregularities caused by the Bogatov-like effect can be carried on by different FWM processes (degenerated FWM, non-degenerated FWM and multi-wave mixing) operating in opposite direction. For this reason, these power fluctuations can be also ascribed to the Bogatov-like effect in combination with different FWM processes interacting in opposite direction. Since the Bogatov-like effect only becomes a mentionable influence for high optical power, the saddle point and the modes power fluctuation are close by the high power modes. Moreover, the investigation shows that a broad mode comb can be obtained over a range of input wavelengths of 15 nm so that in later applications the carrier frequency of the short pulses will also be tunable over a similar range. III. GENERATION OF SHORT PULSES A. Phase Relation of the FWM Modes In order to create short pulses, the phase relation between the FWM modes has to be coherent. The simulations in Fig. 3 show that the phase of the FWM modes is not linear for the regime with high optical power FWM modes while on the shorter wavelength side a region with approximately 40 modes with linear phase relation can be found. Again, the distortion can be ascribed to the Bogatov-like effect since it also affects the phase of the modes as mentioned in Section II-A. For this reason, it is of advantage to make use of the low power FWM modes from the linear phase regime for generating short pulses even if the output power is reduced. If one wants to make use of all FWM modes, an additional element with anomalous chromatic dispersion is needed in order to correct the phase. Since the phase signature of the FWM modes cannot be fitted with a second-order polynomial, dispersion compensation fibers cannot be used for the phase correction. An individual phase correction of each FWM mode can be done with distributed Bragg gratings, but this device would dramatically limit the tunability of the presented scheme. In order to improve the tunability, novel devices like a double AWG with a thermal lens could resolve the problem [15].

756


Fig. 2. Simulated and measured output spectra of an UL-SOA, generated from two CW input signals: (a) simulations and (b) measurements; the two copolarized CW input signals for the three different spectra were injected with a mode spacing 0.2 nm around 1550 nm, 1560 nm and 1565 nm, respectively.

a MZI driven in the quadrature point while the simulation are done with two CW input signals. C. Results

Fig. 3. Simulation of the FWM modes phase relation for two CW input signals with a mode spacing of 0.16 nm injected at 1563 nm; the figure shows the phase and the power only at a FWM modes over a bandwidth of 9 nm resulting in approximately 60 modes.

B. Locking of the Generated Pulses In order to investigate the short pulses generated by the UL-SOA with an optical sampling oscilloscope, it needs to be triggered to the repetition rate of the pulses. For first investigations it is advantaging, to generate the input signal of the UL-SOA from the trigger signal. For stable triggering, the trigger signal should be an electrically generated sine, while the input signal of the UL-SOA has to be an optical signal only consisting of two spectral modes. In order to synchronise both signals to each other, the optical signal has to be generated from the electrical signal. For this reason the electrical signal drives a Mach–Zehnder interferometer (MZI) in the quadrature point and generates a carrier suppressed sine modulated signal [16]. Opposite to simulations, the triggering is only a problem for measurements since in simulations the phase relation of the CW lasers can be easily set to be fixed. For this reason, the input signal of all further measurements in this article is created with

Regarding the time-domain signal for the whole FWM comb from Fig. 3, the simulations and the measurements show a spike followed by an inverted spike, both with a constant power offset [Fig. 4(a) and (b)]. The repetition rate is 20 GHz corresponding to the input mode spacing of 0.16 nm. As predicted in Section III-A, no nice pulses can be obtained due to the phase distortion of some FWM modes. In frequency-domain the differences between simulation and measurements can be ascribed to the slightly different input signals as mentioned in Section III-B since for the measurements the carrier signal is not totally suppressed and also contributes to the FWM processes. As a result an increased noise floor can be observed in the spectra. Moreover, the measured output signal is amplified with an EDFA and transmitted with a fiber to the optical sampling oscilloscope creating additional phase distortion due to chromatic dispersion while the simulation shows the output signal directly after the UL-SOA resulting in additional differences in the time-domain. When filtering the FWM modes in the linear phase regime (4 nm from the combs shorter wavelength side), mainly FWM modes with a linear phase signature are used for the pulse generation. As a result, the time-domain signal has a clear slightly nonsymmetric pulse shape and pulses with full width at half maximum (FWHM) of approximately 5 ps and a repetition rate of 20 GHz could be obtained from simulations and measurements [Fig. 4(c) and (d)]. In order to further shorten the generated pulses, the influence of the FWM modes from the linear phase regime has to be improved. For the results in Fig. 4(a)–(d), the high power FWM modes with a nonlinear phase signature dominated or had a nonnegligible contribution to the pulse generation. In the next approach the center wavelength of the filter is located at the very short wavelength side of the FWM mode comb, so that the filter’s slope is used to equalize the amplitudes of the FWM


757

Fig. 4. Output signal of the UL-SOA and the bandpass filter in time and frequency domain for different filtering techniques (left column—simulation results and right column—measurement results); filtering techniques (red line—filter characteristic): first row—all FWM modes (without filtering), second row—only taking FWM modes with linear phase and third row—equalisation of the FWM modes with the help of the filter slope.

modes. Regarding the spectra in Fig. 4(e) and (f), the FWM modes from the linear phase regime became more dominant due to the filtering, so that the main part of the optical power

is now located in the linear phase regime. Regarding the output pulses in the time-domain, pulses with only slight distortions and a FWHM of less than 2 ps can be observed.

758


Fig. 5. Simulated (a) pulsewidth and (b) time-bandwidth product of the generated pulses for different filtering techniques in dependence of the input CWs’ mode spacing.

IV. POTENTIAL OF THE UL-SOA FOR PULSE GENERATION A. Repetition Rate of the Pulses As mentioned in Section II-A, the presented scheme should also provide a tunability in the repetition rate of the pulses depending on the modes spacing of the input CWs. Furthermore, the bandwidth of the generated FWM mode comb strongly depends on the mode spacing of the CW input signals. For this reason, the pulsewidth of the generated pulses also depends on the mode spacing of the input CWs [Fig. 5(a)]. The simulations show that for all filtering techniques, the pulsewidth decreases with increasing mode spacing. Hence, with increasing repetition rate of the pulses, shorter pulses can be obtained. Moreover, assuming that an ideal phase correction as discussed in Section III-A limits the bound for generating short pulses with this UL-SOA device, the investigation shows that the equalising filtering technique creates nearly optimal results. Another interesting point from the investigation is that the generated pulsewidth is limited to about 400 fs. For the equalising filtering technique and the case with ideal phase correction, the pulsewidth saturates for a mode spacings greater than 400 GHz due to the gain bandwidth of the UL-SOA (400 fs corresponding to approximately 20 nm). Furthermore, the figure shows that around 200 GHz of mode spacing, the pulsewidth for the case with all FWM modes becomes shorter than for the case where only modes of the linear phase regime are used for the pulse generation. The reason for this effect can be ascribed to the Bogatov-like effects decaying with increasing mode spacing. As a result, the phase distortion cannot be carried over multiple modes. Therefore, the distortion only affects one mode for bigger modes spacings. In addition, the power of the mode with the distorted phase is reduced by the Bogatov-like effect. Hence, its contribution to the pulse generation is reduced and also the few modes with linear phase on the long wavelength side efficiently contribute to the pulse generation. In order to comment the efficiency of the generated pulses with the here presented scheme, the time-bandwidth product can be used being defined as the product of the pulsewidth and the spectral width (both measured at FWHM). Regarding

Fig. 5(b), the spectral efficiency of the generated pulses with ideal phase recovery is optimal since the value of the -pulses. time-bandwidth product corresponds the value of The value of the time-bandwidth product for the case where the filter slope is used to equalize the FWM modes is close to the value of sinc-pulses for small mode spacings underlining the equalisation process because sinc functions in time-domain corresponds to rectangular functions in frequency-domain. It also has to be mentioned that chromatic dispersion is not considered in the simulations. First estimations showed that up to approximately 250 GHz of mode spacing, the influence of the input CW signals’ phase walk-off for an 8 mm long UL-SOA can be neglected and for this reason chromatic dispersion should not dramatically influence the results in Fig. 5 for high mode spacings. B. Possible Further Improvements In order to further improve the efficiency of creating short pulses with this UL-SOA device, one could think of using modulated input signals as a FWM seed, already having a linear mode comb in frequency-domain. Opposite to HNLFs where typically modulated input signals are used and the SPM due to the Kerr effect causes the pulsewidth reduction, in UL-SOAs the pulse shortening is due to dynamic gain and index gratings creating FWM products. The most efficient FWM mode creation can be obtained from modes with nearly equal power (2 CW input signals) while for amplitude modulated signals always the carrier mode dominates and so the efficiency of creating FWM modes is reduced. Furthermore, when using short pulses as FWM seed having a very broad mode comb in frequency-domain and the time between the two pulses is in the range of the carrier lifetime, the UL-SOA is not driven in a stable operation point and slow interband effects distort the pulse generation. Also the usage of phase modulated input signals does not improve the performance. Although phase modulated signals have a broad spectral mode comb, they cannot create the dynamic gratings for the FWM processes because the nonlinear effects causing the dynamic gratings are dependent on the photon density being constant for phase modulated signals.


Another possibility for improving the generation of short pulses could be device optimization. In [17] it has been shown, that the mode confinement plays an important factor for the creation of the FWM mode comb. For this reason, bulk devices should have a clear advantage compared to multi quantum well (MQW) devices. On the other hand, MQW devices have a slightly smaller -factor resulting in slightly less phase distortion due to the Bogatov-like effect. Nevertheless, the phase distortion only has a minor effect on the pulse generation due to the equalising filtering technique. Hence, bulk devices with a high mode confinement should be the number one choice for pulse generation. V. CONCLUSION A concept for an integrated short pulse generator, widely tunable in the repetition rate and the carrier frequency of the pulses has been presented. The simple scheme is based on two tunable CW laser sources and an UL-SOA. Due to the FWM efficiency of the UL-SOA, broad FWM combs can be obtained at the output tunable over a range of 15 nm. According to calculations, the total mode spectrum has due to complex FWM interactions, a nonlinear phase signature in the regime with high optical power. Therefore, no proper pulses can be obtained at the output of the UL-SOA. When mainly filtering the FWM modes with a linear phase signature from the combs shorter wavelength side, short pulses with FWHM of approximately 5 ps and a pulse repetition rate of 20 GHz have been generated. Moreover, even shorter pulses with FWHM less than 2 ps and a pulse repetition rate of 20 GHz can be obtained from simulations and measurements when the filter’s slope is used to equalize the FWM modes so that the modes from the linear phase regime dominate. Furthermore, full numerical simulation showed a strong dependence of the pulsewidth on the CW mode spacing. With increasing mode spacing not only the repetition rate of the pulses increases, but also the pulsewidth shortens and pulses in the sub-picosecond regime can be obtained. Also an optimization of the device should further improve the performance. REFERENCES [1] H.-G. Weber, R. Ludwig, S. Ferber, C. Schmidt-Langhorst, M. Kroh, V. Marembert, C. Boerner, and C. Schubert, “Ultrahigh-speed OTDM-transmission technology,” J. Lightw. Technol., vol. 24, no. 12, pp. 4616–4627, Dec. 2006. [2] A. Viglienzoni, “Evolution of products and enabling technologies for optical networks,” in Proc. Photonics in Switching’08, Sapporo, Japan, 2008, Plenary 2. [3] H. Takara, T. Ohara, K. Mori, K. Sato, E. Yamada, Y. Inoue, T. Shibata, M. Abe, T. Morioka, and K.-I- Sato, “More than 1000 channel optical frequency chain generation from single supercontinuum source with 12.5 GHz channel spacing,” Electron. Lett., vol. 36, no. 25, pp. 2089–2090, 2000. [4] T. Inoue, N. Kumano, M. Takahashi, T. Yagi, and M. Sakano, “Generation of 80 nm wavelength-tunable 100 fs pulse based on comblike profiled fiber comprised of HNLF and zero dispersion-slope NZDSF,” J. Lightw. Technol., vol. 25, no. 1, pp. 165–169, Jan. 2007.

759

[5] H. A. Haus, K. Tamura, L. E. Nelson, and E. P. Ippen, “Stretched-pulse additive pulse mode-locking in fiber ring lasers: Theory and experiment,” IEEE Quantum Electron., vol. 31, no. 3, pp. 591–598, Mar. 1995. [6] R. Kaiser, B. Hüttl, H. Heidrich, S. Fidorra, W. Rehbein, H. Stolpe, R. Stenzel, W. Ebert, and G. Sahin, “Tunable monolithic mode-locked lasers on InP with low timing jitter,” IEEE Photon. Technol. Lett., vol. 15, no. 5, pp. 634–636, May 2003. [7] K. Sato, K. Wakita, L. Kotaka, Y. Kondo, and M. Yamamoto, “Monolithic strained-InGaAsP multiple-quantum-well lasers with integrated electroabsorption modulators for active mode locking,” Appl. Phys. Lett., vol. 65, no. 1, pp. 1–3, 1994. [8] R. M. Woodward, B. E. Cole, V. P. Wallace, R. J. Pye, D. D. Arnone, E. H. Linfield, and M. Pepper, “Terahertz pulse imaging in reflection geometry of human skin cancer and skin tissue,” J. Phys. Med. Biol., vol. 47, pp. 3853–3863, 2002. [9] D. M. Mittleman, R. H. Jacobsen, and M. C. Nuss, “T-ray imaging,” IEEE J. Sel. Top. Quantum Electron., vol. 2, no. 3, pp. 679–692, May/ Jun. 1996. [10] C. Bornholdt, J. Slovak, B. Sartorius, M. Schlak, and Ch. Schmidt, “Optical comb generator using pulse compression in ultra-long semiconductor amplifiers,” in Proc. ECOC’05, Glasgow, 2005, paper Tu.1.5.5. [11] A. Bogatov, P. Eliseev, and B. Sverdlov, “Anomalous interaction of spectral modes in a semiconductor laser,” IEEE J. Quantum Electron., vol. QE-11, no. 7, pp. 510–515, Jul. 1975. [12] P. Runge, R. Elschner, C.-A. Bunge, and K. Petermann, “Extinction ratio improvement in ultralong semiconductor optical amplifiers,” IEEE J. Quantum Electron., vol. 45, no. 6, pp. 578–585, Jun. 2009. [13] P. Runge, R. Elschner, C.-A. Bunge, K. Petermann, M. Schlak, W. Brinker, and B. Sartorius, “Operational conditions for the extinction ratio improvement in ultralong SOAs,” IEEE Photon. Technol. Lett., vol. 21, no. 2, pp. 106–108, Jan. 2009. [14] P. Runge, R. Elschner, and K. Petermann, “Time-domain modelling of ultralong semiconductor optical amplifiers,” IEEE J. Quantum Electron., to be published. [15] F. Kerbstadt and K. Petermann, “Analysis of adaptive dispersion compensators with double-AWG structures,” J. Lightw. Technol., vol. 23, no. 3, pp. 1468–1477, Mar. 2005. [16] Y. Miyamoto, A. Hirano, K. Yonenaga, A. Sano, H. Toba, K. Murata, and O. Mitomi, “320 Gbit/s (8 40 Gbit/s) WDM transmission over 367 km with 120 km repeater spacing using carrier-suppressed return-to-zero format,” Electron. Lett., vol. 35, no. 23, pp. 2041–2042, 1999. [17] P. Runge, R. Elschner, and K. Petermann, “Optimising four-wave mixing in ultralong SOAs,” in Proc. NUSOD’09, Gwangju, South Korea, 2009, paper TuB2.

2

Patrick Runge was born in Berlin, Germany, in 1979. He received the Dipl.Ing. degree in computer science from the Technische Universität Berlin, Berlin, Germany, in 2005. From 2005 to 2007, he was with Hymite GmbH, where he was involved in the design and measurement of optoelectronic packages for optical communication. In 2007, he returned to the Technische Universität Berlin to pursue the Ph.D. degree, where he is currently researching the physics and applications of ultralong semiconductor optical amplifiers.

Christian-Alexander Bunge (S’01–A’02–M’03) was born in Berlin, Germany, in 1973. He received the Dipl.-Ing. degree in electrical engineering and the Ph.D. degree from the Technische Universität Berlin, Berlin, Germany, in 1999 and 2002, respectively. From 2002 to 2004, he was with the Polymer Optical Fiber Application Center (POF-AC) in Nuremberg, where he was responsible for fiber modeling and short-haul systems design. In 2004, he joined the TU Berlin as a Senior Scientist, working on high-speed optical transmission systems, studying nonlinear optics, and modelling of optical components. In 2009, he became a Professor with Deutsche Telekom’s University for Telecommunication, Leipzig, where his main research interests are short-haul and access transmission techniques, multimode glass and polymer-fiber links, and signal processing.

760

Klaus Petermann (M’76–SM’85–F’09) was born in Mannheim, Germany in 1951. He received the Dipl.-Ing. and Dr.-Ing. degrees from the Technische Universität Braunschweig, Braunschweig, Germany, in 1974 and 1976, respectively, both in electrical engineering. From 1974 to 1976, he was a Research Associate with the Institut für Hochfrequenztechnik, Technische Universität Braunschweig, where he worked on optical waveguide theory. From 1977 to 1983, he was with AEG-Telefunken, Forschungsinstitut, Ulm, Germany, where he was engaged in research work on semiconductor lasers, optical fibers, and optical fiber sensors. In 1983, he became a Full Professor at the Technische Universität Berlin, Germany. His current research interests include optical fiber communications and integrated optics. Mr. Petermann is a member of the BerlinBrandenburg Academy of Science. He was the recipient of the Leibniz Award from the Deutsche Forschungsgemeinschaft in 1993, and the “Distinguished Lecturer Award” from the Laser and Electro-Optics Society within the IEEE in 1999/2000. From 1999 to 2004, he was an Associate Editor of the IEEE PHOTONICS TECHNOLOGY LETTERS. From 1996 to 2004, he was a member of the board of the Verband Der Elektrotechnik Elektronik Informationstechnik (VDE). From 2004 to 2006, he was the Vice President for research at the Technische Universität Berlin.

Michael Schlak was born in Berlin, Germany, in 1950. He received the M.S. degree in physics from the Technische Universitt Berlin, Berlin, Germany, in 1979. Since 1980, he has been working as Member of Research Staff, HeinrichHertz-Institut, Berlin. He became head of a project group on monolithically integrated all optical demultiplexers in 1998. He has diverse publications on epitaxy and integrated optics. He is experienced in crystal growth/epitaxy in the InP-system, InP-technology and measurement technologies, InP integration technology, design and development of InPbased devices, measurement technologies for opto-electronic devices.


Walter Brinker was born in Vechta, Germany, in 1955. He received the diploma in physics from the Westfalian Wilhelms-Universitiy Münster, Germany, in 1984. In 1985, he joined Heinrich-Hertz-Institut fr Nachrichtentechnik (HHI), Berlin, where he first worked on projection display technologies. Since 1995 he has been engaged in the optical characterization of semiconductor based photonic devices. One focus of his scientific interest was the all-optical signal processing based on monolithically integrated active interferometers. His current research activities include the characterization of nonlinear optical effects in ultralong semiconductor optical amplifiers and the hybrid integration of photonic semiconductor devices with polymeric waveguide structures.

Bernd Sartorius received the Ph.D. degree in physics from the Technische Universität Berlin, Berlin. Germany in 1982. He then joined the Heinrich-Hertz-Institute for Telecommunications, Berlin, where he first worked on optical techniques for characterization of semiconductors. In 1991 he became head of projects developing multi-section lasers and optical amplifiers for applications in all-optical signal processing. The focus of the joint work with Alcatel was particularly on high-speed 3R signal regeneration. Since 2006, he has been leading a project team developing InP-based components for applications in Terahertz sensor systems.