Picosecond Diode Lasers
There are two fundamentally different kinds of diode lasers (lasers based on laser diodes) which are made such that they do not emit light continuously, but rather in the form of ultrashort pulses with pulse durations in the picosecond regime. These laser types are described in the following sections.
Mode-locked Diode Lasers
A suitable kind of diode laser (typically an edge-emitting single-mode laser) can be mode-locked with active, passive or hybrid techniques. In any case, mode-locked operation means that a single ultrashort pulse (or in some cases a couple of equidistant pulses) circulates in the laser resonator. Each time when a pulse hits the output coupler mirror, one obtains a pulse at the output, and this leads to a high repetition rate pulse train. Electrical pumping may be done with a constant or modulated pump current, depending on the type of mode locking.
For more details, see the article on mode-locked diode lasers.
As the laser resonators of diode lasers are typically quite short (except for some external cavity diode lasers), the resulting pulse repetition rates are high – typically in the high gigahertz region (with a relatively short external cavity).
Mode-locked diode lasers can in principle also emit femtosecond pulses, but pulse durations of a couple of picoseconds are more common.
The average output power achievable in mode-locked operation is often substantially lower than in continuous-wave operation, since stable mode locking with well-shaped output pulses may not work at maximum power level. Because of the moderate average output power and the high pulse repetition rate, the pulse energy is normally quite small – at most in the low picojoule region, but frequently even far below 1 pJ.
In most cases, the obtained pulses are synchronized with a supplied electrical signal and exhibit a relatively small timing jitter, compared with gain-switched lasers (see below). The output pulses are mutually coherent, since they are usually derived from a single circulating pulse.
Mode-locked diode lasers are mostly suitable for applications in optical fiber communications, e.g. in the 1.5-μm spectral region. With a fiber amplifier, the obtained average power and pulse energy can be increased substantially, if required.
Other possible applications are in fields like optical metrology, e.g. distance measurements. Here, different spectral regions may also be utilized, e.g. around 0.8 μm or 1 μm. However, not for all wavelength regions mode-locked diode lasers are available; for example, it may be difficult to find such devices for emission in the ultraviolet or in the mid-infrared region.
Gain-switched Laser Diodes
Operation Principle; Limit for the Pulse Energy
A relatively simple technique for obtaining short pulses from diode lasers is gain switching. Here, one applies short current pulses to a diode laser [2, 4] – in this case, a laser diode without an external cavity. Each current pulse (with a duration of e.g. a few hundred picoseconds) brings the carrier density in the active region of the laser diode above the value at the laser threshold, so that laser operation starts. If the conditions are properly chosen, this leads to the generation of a single well-shaped optical pulse with some time delay; this is caused by the need to build up the laser radiation in some number of resonator round-trips, starting from spontaneous emission. Due to that time delay, it is possible to obtain optical pulses which are substantially shorter than the electrical pump pulses: it creates the opportunity to supply additional excitation energy during a time where the optical power is still at a very low level.
Note that the conditions in the laser diode are far from the steady state; there is a substantial variation of the gain (through the carrier density) during the pulse, much in contrast to quasi-continuous-wave operation for the generation of much longer light pulses.
Because no particularly special requirements on the laser diode arise, gain switching can be practiced in a wide range of spectral regions, including for example the ultraviolet.
Such a picosecond laser head can be quite compact, because no optical components beyond the laser diode and perhaps a beam collimator lens or a fiber launch system are required. However, one generally needs to integrate the electronic driver, because such short current pulses cannot be transmitted over cables with a substantial length.
If the electrical pump pulses bring the laser only slightly above the laser threshold, one quite certainly obtains clean pulses, but with a quite low pulse energy and a not yet ideally short duration. For stronger current pulses, the pulse energy increases and the pulse duration decreases, because the round-trip gain gets higher, leading to a faster build-up of the pulse. That implies that the optical peak power increases substantially. However, if the current pulses become too strong, the peak of the pulse is reached before the end of the current pulse (since the gain is high), and that can easily lead to a second pulse or a more complicated temporal structure after the main pulse. Also, one may get some power in higher-order modes, if the waveguide of the laser diode is not single-mode; that would also lead to a reduced beam quality. Such distortions of the temporal and/or spatial shape can be detrimental for applications. Some suppliers may tolerate larger distortions of the pulse shape in order to present a high pulse energy and peak power.
In conclusion, there is a maximum pulse energy, usually limited by the mentioned distortions for too strong current pulses. Typically, one obtains pulses with a peak power of the order of 1 W and a pulse duration of a few tens of picoseconds, which implies a pulse energy of the order of some tens of picojoules. One may try to optimize the performance through the choice of an ideally suited laser diode in conjunction with an electronic circuit for realizing the optimum driving conditions. Often, one applies the current pulses together with a constant bias current.
Even when such a laser is driven with a relatively high pulse repetition rate (e.g. 100 MHz), the average output power will be much lower than for continuous-wave operation, because the peak power is not much higher than the possible continuous-wave power, while the duty cycle is rather small.
The achievable pulse repetition rate is fundamentally limited only by the need to have near-complete recombination of the excited carriers to the point in time where a pump pulse starts. Otherwise, the pulse build-up could occur prematurely. However, in practice the limits are often set by the used driver electronics, e.g. to 50 or 100 MHz.
The achievable pulse duration is often of the order of a few tens of picoseconds. Although it might in principle help to use a laser diode with particularly short resonator, reducing the build-up time of the laser pulse, one would then also require very short current pump pulses. Therefore, some not too small resonator round-trip time will usually be better in practice, and there is probably no realistic potential for achieving much shorter pulse durations.
However, with additional pulse compression outside the laser, one can reduce the pulse durations further to a few picoseconds .
During operation, one can to some extent tune the pulse duration together with the pulse energy by varying the energy of the current pulses.
Coherence and Timing Jitter
As each pulse is generated starting from spontaneous emission, there is no fixed phase relationship between subsequent pulses. That implies a lack of temporal coherence, and that even a highly regular pulse train generated with gain switching will generally not be associated with a frequency comb. This is essentially because laser operation completely disappears between two pulses, so that the phase information is lost. Only in cases with extremely high repetition rate (and correspondingly higher duty cycle) this might be different.
Also, the relative timing between the used current pulse and the generated optical powers is subject to some fluctuations (timing jitter). Fluctuations of the electrical pulse parameters are converted into timing fluctuations, and even for a virtually noiseless electrical pump system there are some timing fluctuations arising from quantum noise of the optical field.
Apart from the simplicity of the method of pulse generation, a main attraction of gain switching is that pulses can be generated on demand, simply based on a suitable electronic driver. For example, one can generate regular pulse trains at variable pulse repetition rates, or bursts of pulses, even with the pulse energy evolving as desired within a bunch of pulses. That creates high flexibility for certain applications.
For example, one may combine such a flexible pulse source with a multi-stage fiber amplifier for obtaining sufficiently energetic pulses for laser micromachining, e.g. for fine laser cutting. In order to account for gain saturation in the amplifier, one may use pulse energies which rise from pulse to pulse during each burst. Note, however, that the amplification from the picojoule to the microjoule domain requires a quite large amplifier gain, so that a single-stage fiber amplifier is usually not sufficient, and proper measures have to be applied between the amplifier stages in order to avoid problems with amplified spontaneous emission. Also, such a high-gain system can be highly sensitive to back-reflections, even when a Faraday isolator is used.
In that respect, an amplifier system using with a different kind of seed laser – for example, a passively mode-locked solid-state laser and a pulse picker – could be substantially simpler, because with a higher seed pulse energy one requires less amplifier gain. There is thus a trade-off between complexity of the seed laser and the amplifier system.
Gain-switched laser diodes may also be used as seed lasers for optical parametric amplifiers. Compared to an optical parametric generator, one can then obtain higher output pulse energies and substantially more consistent pulse parameters. Due to the simplicity of such a gain-switched seed source, the overall cost may not be increased substantially. However, the wavelength tuning range will then be limited by the diode laser, so that the technique is not suitable for broadly wavelength-tunable sources.
Pulses from a gain-switched laser diode may also be directly used for certain measurements, for example for fluorescence lifetime measurements and imaging, spectroscopy, optical tomography and optical time-domain reflectometry.
The RP Photonics Buyer's Guide contains 16 suppliers for picosecond diode lasers.
Questions and Comments from Users
Here you can submit questions and comments. As far as they get accepted by the author, they will appear above this paragraph together with the author’s answer. The author will decide on acceptance based on certain criteria. Essentially, the issue must be of sufficiently broad interest.
Please do not enter personal data here; we would otherwise delete it soon. (See also our privacy declaration.) If you wish to receive personal feedback or consultancy from the author, please contact him e.g. via e-mail.
By submitting the information, you give your consent to the potential publication of your inputs on our website according to our rules. (If you later retract your consent, we will delete those inputs.) As your inputs are first reviewed by the author, they may be published with some delay.
|||P.-T. Ho et al., “Picosecond pulse generation with a cw GaAlAs laser diode”, Appl. Phys. Lett. 33 (3), 241 (1978), doi:10.1063/1.90312|
|||C. Lin et al., “Simple picosecond pulse generation scheme for injection lasers”, Electron. Lett. 16, 600 (1980), doi:10.1049/el:19800416|
|||J. AuYeung, “Picosecond optical pulse generation at gigahertz rates by direct modulation of a semiconductor laser”, Appl. Phys. Lett. 38 (5), 308 (1981), doi:10.1063/1.92351|
|||P. Paulus et al., “Generation and optimum control of picosecond optical pulses from gain-switched semiconductor lasers”, IEEE J. Quantum Electron. 24 (8), 1519 (1988), doi:10.1109/3.7078|
|||J. E. Bowers et al., “Actively mode-locked semiconductor lasers”, IEEE J. Quantum Electron. 25 (6), 1426 (1989), doi:10.1109/3.29278|
|||P. J. Delfyett et al., “High-power ultrafast laser diodes”, IEEE J. Quantum Electron. 28 (10), 2203 (1992), doi:10.1109/3.159528|
|||P. P. Vasil'ev, “Ultrashort pulse generation in diode lasers”, Opt. Quantum Electron. 24, 801 (1992), doi:10.1007/BF00620194|
|||D. J. Derickson et al., “Short pulse generation using multisegment mode-locked semiconductor lasers”, IEEE J. Quantum Electron. 28 (10), 2186 (1992), doi:10.1109/3.159527|
|||L. Chusseau and C. Kazmierski, “Optimum linear pulse compression of a gain-switched 1.5 μm DFB laser”, IEEE Photon. Technol. Lett. 6 (1), 24 (1994), doi:10.1109/68.265878|
|||S. Arahira et al., “Mode-locking at very high repetition rates more than terahertz in passively mode-locked distributed-Bragg-reflector laser diodes”, J. Sel. Top. Quantum Electron. 32 (7), 1211 (1996), doi:10.1109/3.517021|
|||I. Ogura et al., “Precise operation-frequency control of monolithic mode-locked laser diodes for high-speed optical communication and all-optical signal processing”, Opt. Quantum Electron. 33, 709 (2001), doi:10.1023/A:1017511214111|
|||K. Sato, “Semiconductor light sources for 40-Gb/s transmission systems”, J. Lightwave Technol. 20 (12), 2035 (2002)|
|||L. A. Jiang et al., “Semiconductor mode-locked lasers as pulse sources for high bit rate data transmission”, J. Opt. Fiber Commun. Rep. 2, 1 (2005), doi:10.1007/s10297-004-0022-0|