Mode-locked semiconductor lasers.
In microwave applications, broadband electrical amplifiers are devices with tens of GHz of bandwidth using MESFET or HEMT transistors. In optical applications, broadband optical amplifiers are semiconductor optical amplifiers with thousands of GHz of bandwidth. This paper describes new methods that take advantage of the THz bandwidth in semiconductor optical amplifiers using mode-locking techniques. Mode-locked semiconductor lasers are devices that produce short optical pulses in which the pulse width is limited only by the optical amplifier's bandwidth. Optical pulse widths as short as a few picoseconds and with pulse repetition rates from several hundred MHz to several hundred GHz are achieved in mode-locked semiconductor lasers. This paper explains the principles of operation, gives typical performance characteristics and describes how mode-locked semiconductor lasers can be used in microwave applications.
Mode-Locked Semiconductor Laser Operation
Figure 1a shows a scanning electron micrograph (SEM) of an entire mode-locked semiconductor laser. The chip has a continuous active waveguide along the device length. Figure 1b shows a SEM of the active waveguide's cross section. The optical amplifier uses a PIN heterostructure diode. In this example, a 1550 nm wavelength bandgap InGaAsP active region is surrounded by InP material. When the diode is forward biased, the lower bandgap active region material has holes injected from the p-side and electrons injected from the n-side in order to produce a carrier inversion and therefore, optical gain. The active region may consist of a single composition (bulk structure) or contain a sandwich of thin quantum well layers. The active region has a higher index of refraction than the surrounding higher bandgap regions, forming an optical waveguide. The mirrors for the laser are the cleaved facets on the edge of the chip. The top metalization of the semiconductor laser chip is broken up into several segments, allowing for nonuniform electrical pumping of the structure. With proper biasing of these contacts, short optical pulses can be obtained.
Figure 2 shows the operation by which short optical pulses can be generated in multisegment laser structures by mode-locking. The idea of mode locking is to initiate an optical pulse circulating back and forth inside the laser cavity. The repetition rate of the pulses are equal to the round trip time in the laser cavity.
The three important segments in a mode-locked semiconductor laser are the gain segment, the gain modulation segment and the saturable absorber. The gain segment is forward biased (DC) to provide the optical gain necessary to overcome the waveguide and mirror losses in order to obtain lasing. The gain modulation segment is pumped by an electrical current pulse providing a time dependent gain function. The saturable absorber segment provides a time dependent loss function. The time dependent gain and loss functions are used to force the laser to emit pulses of light instead of a continuous light stream.
Gain Section Characteristics
To understand the characteristics of the mode-locked laser gain segment, an understanding of the gain modulation and saturable absorber segments is needed. Figure 3 shows the calculated gain characteristics of an optical amplifier with an InGaAs quantum well active region. The gain coefficient g is plotted vs. wavelength for several values of the carrier density in the active region. The actual gain is calculated by
Gain = |e.sup.|gamma~~|g.sup.l~ (1)
L = amplifier length
|gamma~ = fraction of the optical mode overlapping the gain region
When the gain section is not forward biased by an electrical current, the device is highly attenuating. By pumping the amplifier with a DC electrical current, a large carrier density and therefore, a large optical gain can be achieved. The optical bandwidth shown in Figure 3 can be as wide as 100 nm, which corresponds to a 12 THz bandwidth at a center frequency of 193 THz. Optical amplifiers are relatively narrowband in terms of percentage bandwidth. However, the absolute bandwidth is extremely large. Gains of 30 dB can be achieved with large pumping currents and high carrier densities in semiconductor optical amplifiers. Gains greater than 30 dB are difficult to achieve due to gain saturation caused by amplification of spontaneously emitted light along the length of the amplifier.
The gain saturation characteristics of a semiconductor optical amplifier for a pulsed optical input are shown in Figure 4. The energy gain is plotted vs. the input pulse energy to the amplifier, normalized to a characteristic saturation energy. A semiconductor optical amplifier's characteristic saturation energy is proportional to the cross-sectional area of the optical mode in the waveguide. The saturation energy is inversely proportional to the slope of the gain coefficient vs. carrier density function. The saturation energy is a measure of the output energy capability of an optical amplifier and has a typical value of 1 to 10 pJ. As the input energy to the amplifier approaches the saturation energy of the amplifier, the achievable energy gain from the laser amplifier rapidly decreases.
Active Mode-Locking and Gain Modulation
Figure 5a shows the function of the active gain modulation segment. A repetitive electrical current pulse with a period nearly equal to the round trip time in the laser cavity is applied to the gain modulation segment. Figure 5b shows the electrical, current pulse increasing the carrier density and thus the gain in the segment. In steady state mode-locked laser operation, an optical pulse arrives at the gain modulation segment slightly after the peak of the current modulation pulse. Figure 5c shows the pulse narrowing that is accomplished in the active gain modulation segment. As the optical pulse enters the gain modulation segment, the leading portion of the pulse receives a small amount of amplification since the electrical pulse has just started increasing the segment gain. The middle portion of the pulse receives the largest amount of gain. The trailing edge of the optical pulse receives less amplification than the middle of the optical pulse because the earlier portions of the pulse saturate the gain of the amplifier. The output optical pulse from the gain modulation segment is slightly shorter in width than the input pulse. Over many round trips in the laser cavity, the optical pulse width becomes much narrower than the electrical current modulation pulse width. In order to get very short optical pulses from an active gain modulation segment, the electrical modulation pulses should be short and of large amplitude. The laser gain modulation segment must also have small electrical parasitics so that the high frequency content of the electrical modulation signal is not attenuated. The final optical pulse width value is reached when the gain modulation segment's pulse shortening ability is equaled by the pulse broadening effects of the gain segment.
Passive Mode-locking and Loss Modulation
The saturable absorber segment is effective in narrowing optical pulses by modulating the loss of the segment. A saturable absorber is formed when a short segment of the laser diode is reverse biased. Since there is no pumping current in the diode, the carrier density is small and the reverse-biased section is highly attenuating. Figure 6 shows the pulse shortening operation of a saturable absorber segment. When an optical pulse enters the reverse-biased segment, the leading edge of the optical pulse is absorbed. For each absorbed photon, an electronhole pair is created in the saturable absorber active waveguide, which increases the carrier density in the segment. Later portions of the optical pulse receive less attenuation than the leading edge. For the high input energy pulses found in mode-locked lasers, the carrier density increases in the saturable absorber segment until the transmission through the segment becomes unity for later portions of the input pulse. The saturable absorber removes the leading edge of the optical pulse and causes pulse width narrowing of 10 to 25 percent on each pass.
The saturable absorber also functions as an integrated photodetector. A reverse-biased optical amplifier segment acts as a PIN photodetector. The large electric field that exists in the intrinsic region of the diode sweeps the photogenerated carriers out of the saturable absorber into an external electrical termination. Figure 6a shows that for every optical pulse in a saturable absorber, an electrical current pulse is also formed. The removal of the photogenerated carriers is also important for proper mode-locked laser operation. The carriers are removed from the optical waveguide in less than 10 ps by the electric field. This carrier removal resets the absorption in the saturable absorber to a high value for the next optical pulse circulating in the mode-locked laser cavity.
Cavity Configurations and Mode-locking Techniques
Using the gain, gain modulation and saturable absorber segments, three different mode-locking techniques are possible, including active, passive and hybrid. In addition, there are two types of cavity configurations, monolithic and external cavities. Figure 7 shows devices that use external optical cavities to produce mode-locked laser operation. External cavities are useful for making very long optical delays for lower repetition rates. One facet of the semiconductor laser is antireflection coated and the light is coupled with a lens into an external optical cavity. The light is reflected back into the laser with an external mirror. The light can be collimated or brought to a focus at the external mirror. The simplest method for producing mode-locked laser pulses is with a single-segment external cavity arrangement shown in Figure 7a. This configuration uses a single-segment laser diode that is readily available from many manufacturers. The device is actively mode-locked with the functions of gain and gain modulation combined in the same segment. This is accomplished by adding a DC current bias onto the modulation current pulses. Narrower width optical pulses can be obtained by separating the functions of gain and gain modulation into separate contacts, as shown in Figure 7b. In Figure 7c, the laser is mode-locked using only passive gain modulation with a saturable absorber. A repetitive optical pulse stream is formed in the laser by applying two DC biases to the device (no RF excitation is required). With passive mode-locking, a microwave oscillator is also formed, since electrical pulses are produced from the saturable absorber at the same time as the optical pulses. Figure 7d shows a configuration that combines the active gain modulation with the passive gain modulation. This combination of active and passive gain modulation is called hybrid mode-locking. Hybrid mode-locking usually forms the shortest width optical pulses because the two pulse shortening techniques are being used together.
Figure 8 shows the monolithic cavity configurations in which the external cavity is replaced with a semiconductor waveguide. In these configurations, all of the mode-locking functions are integrated in a single semiconductor chip.
There are many other mode-locked laser sources, such as dye and solid-state lasers, that are capable of producing short optical pulses. Semiconductor lasers are smaller in size, can be electrically pumped, are less power hungry and are easier to use than other short pulse sources. Mode-locked semiconductor lasers can perform at much higher repetition rates and interface well with electronics. An alternative short optical pulse generation technique is gain switching using semiconductor lasers. Gain-switched semiconductor lasers typically have wider optical pulse widths and do not operate well at repetition rates above 10 GHz. Gain-switched semiconductor lasers have the advantage of being able to operate at a repetition rate independent of the round trip time in the laser cavity. The major disadvantage of semiconductor lasers is the moderate output power of the devices. Mode-locked semiconductor lasers are replacing dye and solid-state lasers in many applications, and have generated new applications for short optical pulses.
Semiconductor lasers are important sources for sending analog and digital signals at microwave and mm-wave modulation frequencies over low loss fiber-optic cables, as shown in Figure 9a. Directly current modulated semiconductor lasers can operate at modulation frequencies up to 30 GHz. Mode-locked semiconductor laser techniques can modulate the light at much higher frequencies than are obtainable by directly modulated semiconductor lasers. Mode-locked semiconductor lasers offer efficient light modulation at frequencies nearly equal to one over the round trip time in the mode-locked laser cavity. High frequency fiber-optic systems using mode-locked semiconductor lasers are useful for signal distribution and antenna remoting where metal waveguides are replaced by fiber optics.
The optical signals from mode-locked lasers can be used to interact with photosensitive electronic devices. Optical pulse streams have been used to reduce the phase noise in microwave and mm-wave oscillators. The optical signal can also be photo-mixed with electronic signals in photosensitive devices. In the described passive and hybrid mode-locked semiconductor lasers, microwave and mm-wave electronic signals are generated as a by-product of the optical signal generation process.
Optical pulse sources are useful in optoelectronic measurement systems. Impulse response testing, of optical-to-electrical converters and detectors with short optical pulses allows the magnitude and phase response of the system to be measured. The electrical impulses generated from photoreceivers have been used to trigger photoconductive switches to measure the response of electrical networks. Electro-optic sampling of electronic signals, shown in Figure 9b, has been demonstrated using semiconductor lasers instead of more complicated and expensive mode-locked Nd:YAG laser sources with fiber and grating pulse compressors. Electrical input analog-to-digital A/D converters with picosecond time resolution have been demonstrated using optical sampling pulses from mode-locked semiconductor lasers, as shown in Figure 9c.
Mode-locked Laser Performance Characteristics
Table 1 lists performance characteristics of mode-locked semiconductor laser designs. The characteristics of pulse width, optical spectral width, optical pulse energy, repetition rate and pulse jitter are listed for various cavity configurations and mode-locking techniques.
On each round trip of the optical pulse through the optical cavity, there are mechanisms that widen and narrow the optical pulse width. The final pulse width is reached when an equilibrium occurs between the pulse narrowing and the pulse widening mechanisms. The two pulse narrowing mechanisms are active gain modulation and saturable absorption. The sources of pulse broadening are the gain-bandwidth of the optical amplifier, pulse broadening during gain saturation, dispersion in the cavity and self-phase modulation effects. For semiconductor lasers, the equilibrium TABULAR DATA OMITTED pulse width is in the 1 to 50 ps range depending on the cavity configuration and mode-locking technique.
Active mode-locked lasers produce the widest optical pulse width. Active gain modulation is less effective in shortening the optical pulse on each round trip as the input pulse width becomes very narrow. The width of the optical pulse is dependent on the width of the electrical modulation current pulses. Step recovery diode and high frequency sinusoidal driving signals are typically used for driving the active gain modulation segment. The active mode-locked designs in which the functions of gain and gain modulation are separated into two segments are more effective in producing short optical pulses.
Passive mode-locking using a saturable absorber produces a shorter pulse width than active mode-locking, since passive gain modulation is a more effective pulse shortening mechanism than active gain modulation. The saturable absorber produces the same amount of pulse shaping per pass independent of the input pulse width. Pulse widths from 1 to 10 ps have been produced using passive mode-locking techniques. Passive mode-locked semiconductor lasers using quantum well active regions produce shorter optical pulse widths than those using bulk active regions.
Hybrid mode-locking, a combination of active and passive mode-locking, produces the shortest pulse width. Since both active and passive pulse shortening mechanisms are used, the maximum pulse shortening per pass and the shortest width pulse are achieved. Hybrid mode-locked devices also have superior pulse-to-pulse timing stability compared to passive mode-locked lasers.
If the pulse width is multiplied by the spectral width, a time-bandwidth product is calculated. If the electric field zero crossings under the optical pulse modulation envelope are uniformly spaced in time, the time-bandwidth product should be approximately 0.5. The actual time-bandwidth products are found to be much larger. This extra bandwidth indicates that the pulses created in mode-locked semiconductor lasers are highly frequency chirped. The chirping is caused by a coupling between the carrier density and the index of refraction in the laser. The mechanism of chirp formation in the optical pulses is shown in Figure 10. When an optical pulse is applied to a semiconductor optical amplifier, conversion of carriers to photons causes a reduction in the carrier density and therefore, gain in the amplifier. The index of refraction is proportional to the carrier density, causing the phase shift through the amplifier to become time dependent. The time dependent phase delay introduces a chirp onto the optical pulse. For many applications of the mode-locked laser pulses, frequency chirping is not important. When a chirped optical pulse is propagated through long lengths of optical fiber, the extra optical bandwidth caused by the frequency chirp will lead to a premature broadening of the optical pulse width. In applications where lower time-bandwidth product pulses are necessary, wavelength filter elements are often added to the cavity or the power level from the device can be decreased to reduce the self-phase modulation effects.
Mode-locked semiconductor lasers can be used at repetition rates between a few tens of MHz to over 100 GHz. For the lower repetition rates, external cavity lasers are used. The upper bound on the repetition rate of external cavity lasers is imposed by the finite length of the external cavity optical elements and is about 20 GHz. The lower bound on repetition rate is determined by practical lengths for the external cavity. Monolithic cavity devices become practical for repetition rates above 4 GHz. Due to limitations in fabrication technology and yield, semiconductor devices longer than about 1 cm are impractical. Monolithic cavity devices have been shown to operate at upper repetition rates above 100 GHz. The upper repetition rate is reached with passive mode-locking techniques and is limited by the saturable absorber and gain recovery time constants. Devices with quantum well active regions are found to be superior for repetition rates above 20 GHz. Figure 11 shows autocorrelation traces for monolithic cavity devices operating at 36 and 81 GHz repetition rates. Pulse autocorrelation techniques must be used to measure mode-locked laser pulses because they are too narrow to be directly measured with sampling oscilloscopes and high speed photodetectors. As the pulse period decreases, the mode-locked laser output tends to look more sinusoidal in shape and less pulse-like.
In mode-locked lasers, the repetition time period is constrained so that it is approximately the round trip time of the optical cavity. For a fixed cavity length, active and hybrid mode-locked lasers can be slightly tuned from this resonance by varying the active gain modulation frequency with minimal degradation in device performance. Measurements of this tunability show that monolithic cavity devices have tunable ranges up to 5 percent of the center repetition frequency and external cavity devices have up to 1 percent tunability. In passive mode-locking, the device's repetition rate can be electrically tuned by varying the current in the gain region or by varying the reverse bias voltage on the saturable absorber. The repetition rate's electrical tuning is important because it allows passive mode-locked devices to be phase locked to lower frequency electrical sources.
Timing Jitter and Intensity Noise
For use in microwave applications, the phase and intensity noise of the optical pulse envelope at the mode-locked frequency is important. The phase noise of the optical and electrical pulses was measured using spectrum analyzer and heterodyne techniques.|22,23~ The detected electrical phase noise characteristics of a 1.25 mm long monolithic cavity device were measured. The results for passive and hybrid mode-locking at 36 GHz are shown in Figure 12, which shows the L(f) single sideband noise level in a 1 Hz bandwidth offset from the mode-locking repetition rate. The device was hybridly mode-locked with the frequency doubled output of a synthesizer that was amplified to 26 dBm with a traveling-wave tube amplifier. The detected modulation spectrum is measured in a synthesized spectrum analyzer. The phase noise from the mode-locked laser is similar to that of the driving source below 10 kHz, but the mode-locked laser makes a significant contribution above this frequency. When the laser is passively mode-locked using only DC power supplies, a higher level of phase noise is seen. The integrated RMS timing jitter is 0.2 ps (150 Hz to 100 MHz) for the electrical driving source, 0.6 ps (150 Hz to 100 MHz) for the hybrid mode-locked laser and 3.3 ps (100 kHz to 100 MHz) for the passive mode-locked laser. The phase noise of a passive mode-locked laser can also be stabilized using the electronic tuning of repetition rate. The detected output from the saturable absorber serves as the input to phase locking circuitry and the frequency error signal is applied to the saturable absorber bias voltage.|21~ The phase noise of an active mode-locked laser can be extremely small, adding negligible phase noise to that of the driving oscillator.|23~
The energy from a mode-locked laser pulse is limited by the saturation energy of the optical amplifier in the mode-locked laser. The saturation energy of the optical amplifier is defined as
|Mathematical Expression Omitted~
h = Planck's constant
v = optical frequency
A = mode cross-sectional area
|Mathematical Expression Omitted~ = differential gain with respect to carrier density
The saturation energy is a measure of the energy necessary to saturate the gain of a gain segment or the absorption of a saturable absorber segment. In passive and hybrid mode-locked devices, the output pulse energy is operated close to the saturation energy of the gain segment. The saturation energy for semiconductor laser amplifiers may range from 1 pJ in a tightly confined index-guided laser to up to 20 pJ for lightly-confined gain guided edge emitting structures. In order to increase the satuation energy from the device, the device cross-sectional area can be increased or the differential gain value can be decreased. Quantum well laser amplifiers can have a very low differential gain when operated at high carrier densities resulting in a high saturation energy. The area of the mode can be increased by using arrays of coupled optical waveguides.|24~
At very high repetition rates, for example greater than 20 GHz, the optical amplifiers become limited by the saturation power of the amplifier. The saturation power is given as
|P.sub.sat~ = |E.sub.sat~/||tau~.sub.c~ (3)
||tau~.sub.c~ = carrier recombination lifetime
Typical values for the saturation power are 1 to 10 mW. Semiconductor optical amplifiers can be used to increase the output power from mode-locked semiconductor lasers. At 1550 nm, erbium doped fiber amplifiers can also be used.
Mode-locked semiconductor lasers offer a convenient means of generating optical and electrical pulses. Monolithic cavity devices with gain, gain modulation and saturable absorption functions integrated on a single semiconductor chip are easy-to-use. Passive mode-locking offers a simple method to generate short optical pulses, since only DC power supplies are necessary to generate optical and electrical pulse streams. Gain modulation in active and hybrid mode-locking serves to stabilize the optical pulses in time and amplitude. The mode-locked lasers add only a small amount of noise to the driving electrical modulation signal. Pulse repetition rates well into the mm-wave frequencies are possible with these devices. The small size and low power requirements of these devices make them useful in many microwave applications.
1. G.P. Agrawal and N.K. Dutta, Long Wavelength Semiconductor Lasers, New York, Van Nostrand Reinhold, 1986.
2. S.W. Corzine, PhD Dissertation, University of California at Santa Barbara, 1993.
3. G.P. Agrawal and N.A. Olsson, "Self Phase Modulation and Spectral Broadening of Optical Pulses in Semiconductor Laser Amplifiers," IEEE Journal of Quantum Electronics, Vol. 59, 1989, pp. 2297-2306.
4. D.J. Derickson, R.J. Helkey, A. Mar, J.G. Wasserbauer and J.E. Bowers, "Microwave and mm-Wave Signal Generation Using Mode-locked Semiconductor Lasers with Intra-Waveguide Saturable Absorbers," IEEE International Microwave Theory and Technology Symposium, paper V-2, Albuquerque, NM, 1992.
5. D.J. Derickson, R.J. Helkey, A. Mar, J.R. Karin, J.G. Wasserbauer and J.E. Bowers, "Short Pulse Generation Using Multisegment Mode-locked Semiconductor Lasers," IEEE Journal of Quantum Electronics, Vol. 28, 1992, pp. 2186-2202.
6. A. Mar, D.J. Derickson, R.J. Helkey and J.E. Bowers, "1.4 ps Pulses Directly Generated Using Tandem Contact Actively Mode-locked Semiconductor Lasers," IEEE Lasers and Electro-optics Society, 1991 Annual Meeting, paper SDL 14.1, Sen Jose, CA, 1991.
7. J. Werner, H. Melchior and G. Guekos, "Stable Optical Picosecond Pulses from Actively Mode-locked Twin-Section Diode Lasers," Electronics Letters, Vol. 24, 1988, pp. 140-141.
8. W. Kaiser, Ultrashort Laser Pulses, Springer-Verlag, 1988.
9. E. Scholl, et al., "Kinetics of Picosecond Pulse Generation in Gain-switched Semiconductor Lasers," IEEE Journal of Quantum Electronics, Vol. 20, 1984, p. 394.
10. K.Y. Lau, "Narrowband Modulation of Semiconductor Lasers at mm-Wave Frequencies (|is greater than~100 GHz) by Mode-locking," IEEE Journal of Quantum Electronics, Vol. 26, 1990.
11. R. Nagarajan, S. Levy, A. Mar and J.E. Bowers, "Resonantly Enhanced Semiconductor Lasers for Efficient Transmission of mm-Wave Modulated Light," IEEE Photonics Technology Letters, January 1993.
12. C.H. Cox III, "Analog Fiber-Optic Links with Intrinsic Gain," Microwave Journal, September 1992, pp. 90-99.
13. Alwyn Seeds, "Optical Beamforming Techniques for Phased-Array Antennas," Microwave Journal, July 1992, pp. 74-83.
14. Stephen E. Lipsky and Afshin S. Daryoush, "Fiber-Optic Methods for Injection-locked Oscillators," Microwave Journal, January 1992, pp, 80-88.
15. R.T. Hawkins, II, Michael D. Jones, Steven H. Pepper, Jeffrey H. Goll and Mihir K. Ravel, "Vector Characterization of Photodetectors, Photoreceivers and Optical Pulse Sources by Time-Domain Pulse Response Measurements," IEEE Transactions on Instrumentation and Measurement, Vol. 41, August 1992, pp. 467-475.
16. Barry A. Bell, Arnold G. Perry and Robert Sadler, "Gallium Arsenide (GaAs)-based Photoconductive Switches for Pulse Generation and Sampling Applications in the Nanosecond Regime," IEEE Transactions on Instrumentation and Measurement, Vol. 38, February 1989, pp. 92-97.
17. J. Wiesenfeld, "Electro-optic Sampling of High Speed Devices and Integrated Circuits," IBM J. of Research and Development, Vol. 34, 1990, pp. 141-161.
18. R.A. Becker, C.E. Woodward, F.J. Leonberger and R.C. Williamson, "Wide-band Electro-optic Guided Wave Analog-to-Digital Converters," Proceedings of the IEEE, Vol. 72, 1984, pp. 802-819.
19. S. Sanders, L. Eng, J. Paslaski and A. Yariv, "108 GHz Passive Mode-locking of a Multiple Quantum Well Semiconductor Laser with Intracavity Absorber," Applied Physics Letters, Vol. 56, 1990, pp. 310-312.
20. Y.K. Chen, M.C. Wu, T. Tanbun-Ek, R.A. Logan and M.A. Chin, "Subps Monolithic Colliding Pulse Mode-locked Multiple Quantum Well Lasers," Applied Physics Letters, Vol. 58, 1991, pp. 1253-1255.
21. R.J. Helkey, D.J. Derickson, A. Mar, J.G. Wasserbauer, J.E. Bowers and R.L. Thornton, "Repetition Frequency Stabilization of Passively Mode-locked Semiconductor Lasers," Electronics Letters, Vol. 28, 1992, pp. 1920-1922.
22. D.J. Derickson, P.A. Morton, J.E. Bowers and R.L Thornton, "Comparison of Timing Jitter in External and Monolithic Cavity Mode-locked Semiconductor Lasers," Applied Physics Letters, Vol. 59, 1991, pp. 3372-3374.
23. D.J. Derickson, A. Mar and J.E. Bowers, "Residual and Absolute Timing Jitter in Actively Mode-locked Semiconductor Lasers," Electronic Letters, Vol. 26, 1990, pp. 2026-2027.
24. A. Mar, D.J. Derickson, R.J. Helkey, J.E. Bowers and D, Botez, "Mode-locking of High Power Resonant-Optical-Waveguide Diode Laser Arrays," 1992 IEEE Semiconductor Laser Conference Proceedings.
25. D.J. Derickson, R.J. Helkey, A. Mar, J.R. Karin, J.E. Bowers and R.L. Thornton, "Suppression of Multiple Pulse Formation in External Cavity Mode-locked Semiconductor Lasers Using Interwaveguide Saturable Absorbers," IEEE Photonics Technology Letters, Vol. 4, 1992, pp. 333-335.
This work was sponsored by the Office of Naval Research. R.J. Helkey was supported by a Newport Research Fellowship. The authors thank all of the members of the high speed optoelectronics lab at the University of California.
Dennis Derickson received his BS degree in electrical engineering from South Dakota State University in 1981, his MS degree from the University of Wisconsin in 1982 and his PhD degree from the University of California, Santa Barbara (UCSB) in 1992. Currently, he is with the lightwave technology group at Hewlett-Packard Co. From 1983 to 1988, he was a member of the technical staff at HP.
Roger Helkey received his BS degree in engineering from the California Institute of Technology in 1982 and his MS degree in 1988 in electrical engineering from the UCSB, where he is currently working on his PhD. From 1982 to 1984, he was a design engineer with Trimble Navigation. From 1984 to 1986, Helkey was a member of the technical staff at Watkins-Johnson Co.
Alan Mar received his AB degree in physics from Occidental College, Los Angeles, CA in 1985 and his MS degree in 1989 in electrical engineering from the University of California, where he is pursuing his PhD degree.
John Wasserbauer received his BS degree in material science and engineering at Cornell University in 1984 and his DEA degree from the Ecole Centrale de Lyon, Ecully, France in 1986. He is pursuing his PhD at UCSB.
John Bowers received his MS and PhD degrees in applied physics from Stanford University. He is a professor at UCSB and a member of the Optoelectronics Technology center and the NSF Science and Technology Center on Quantized Electronic Structures. In 1979, he worked at Honeywell. From 1982 to 1987, Bowers worked at Bell Laboratories.
|Printer friendly Cite/link Email Feedback|
|Author:||Derickson, Dennis; Helkey, Roger; Mar, Alan; Wasserbauer, John; Bowers, John|
|Date:||Feb 1, 1993|
|Previous Article:||Advanced alumina: a manufacturing medium for microwave oscillators and amplifiers.|
|Next Article:||Predicting post-detection pulse characteristics using the FFT.|