Open Access
Add to BibSonomyAbstract
Lasers-thyristors with a narrow (20 μm) mesa stripe contact have been studied. It was shown that the laser peak power reaches a value of 2.5 W in the long-pulse mode at a pulse width of 13 ns. It was demonstrated that generation of a controlled train of laser pulses with peak power of 1.6 W and width of 90 ps is possible in the short-pulse mode. The maximum value of the pulse repetition frequency was 470 kHz at the following working characteristics of the laser-thyristor: blocking voltage 5.8 V, control current pulse 25 mA. A number of specific features were observed in the short-pulse mode. It was found that the blocking voltage and amplitude of the control current pulse affect the lasing process. We observed that in short pulse mode the lasing spectra have a two-band structure and the lateral near field may degenerate into a single spot with size substantially smaller than the mesa stripe width. It was shown that the main reason for the observed specific features of lasing is the clearly pronounced effect of the spatial localization of the current.
© 2016 Optical Society of America
1. Introduction
Recently, the interest in high-power pulsed semiconductor lasers has been increasing. This is due to the wide opportunities for practical application of these sources of light and, in particular, that for free-space communications, metrology, materials processing, pumping sources for semiconductor optical amplifiers, spectroscopy, LIDAR, and frequency doubling. Each application has its own specific requirements to such laser pulse characteristics as width, peak power, far-field structure, and lasing spectrum. Passing from the continuous-wave mode to the pulsed lasing mode makes it possible to substantially raise the peak power [1,2]. However, a significant increase in the pulsed current gives rise to a number of nonthermal factors that restrict the maximum peak power of semiconductor lasers [2–5]. These are the finite energy relaxation time [5] and the increase in the internal optical loss due to the accumulation of free carriers in the waveguide layers [2–4]. In this case, the simplest way to raise the peak power at increased pumping current amplitude is to make wider the emitting mesa stripe. A disadvantage of this approach is the poor quality of the mode structure of laser light in the parallel plane, which results in a fall of brightness and lower efficiency of coupling with an optical fiber. Another problem of the multimode laser emission is related to the low dynamic stability of emitted laser pulses. This is due to the effect of mode hopping within the multi-mode Fabry-Perot cavity. These effects are the most clearly pronounced in the nanosecond range. One of ways to tackle with this problem is to use a linear array of single emitters with a narrow mesa stripe. Making the mesa-stripe contact narrower enables transition to a single-mode or quasi-single-mode lasing [6]. As a consequence, passing to the design based on a linear array with narrow mesa stripe may be an effective way to generate high-power laser pulses with high stability and high-quality mode structure. Another aspect of generation of high-power pulsed emission is associated with the requirement that an effective pulsed current pumping should be formed. There exist several approaches that provide generation of nanosecond and subnanosecond laser pulses in semiconductor heterostructures. A part of these are based on pulsed current pumping. Here, the research avenue based on the pumping of an optical amplifier by a low-power master laser can be distinguished. In this case, a high spatial and spectral quality of the mode structure can be achieved by using a single-mode master oscillator with spectrally selective elements (DFB or DBR) and an optical amplifier providing an effective power buildup, with the main mode and spectral characteristics preserved. As a result, a multiwatt output power and a 4-8 ns pulse width were demonstrated [7,8]. A different way to achieve a short-pulse generation is the gain-switching mode. In this case, pulses with durations of ~100 ps can be generated under direct pumping with current pulses having a width of several ns [9–12]. The main achievements in this field are associated with the development of a laser heterostructure having a thicker active region at a low optical confinement factor, and also with the development of pulsed current sources. Commercial current sources are available both as separate devices [9] and as integrated transistor-based switches [10–12]. The other way for generation of high-power laser pulses is based on using a multijunction N-p-N-i-P laser-thyristor heterostructure [13–19]. The suggested heterostructure can be represented as an optocouple constituted by a phototransistor (N-p-N) and a semiconductor laser (N-i-P). The optocouple can combine the functions of a current switch and a semiconductor laser. The high control efficiency and the monolithic integration simplify the electrical circuit and diminish the number of parasitic couplings, which is particularly important for generation of nanosecond current pulses. Experimental studies confirmed the possibility of generating laser pulses with power of 43 W and width of 100 ns for lasers-thyristors with output aperture of 200 μm. Simulations [14] demonstrate that laser and current pulses can be, in principle, generated in the nanosecond range of durations. Therefore, the goal of our study was to examine the dynamics of lasers-thyristors with narrow mesa stripe contact in the mode of generation of nanosecond and subnanosecond laser pulses. Adherence to the specifications listed in this style guide is essential for efficient review and publication of submissions.
2. Experimental samples
The experimental multijunction heterostructure was grown by MOCVD epitaxy. The transistor part of the multijunction heterostructure included a 0.5-μm-thick wide bandgap N-AlGaAs emitter (x = 0.15), 2-μm-thick narrow-gap p-GaAs base, and 2-μm-thick wide-bandgap N-AlGaAs collector (x = 0.35) (Fig. 1). The laser part of the multijunction heterostructure included wide-bandgap AlGaAs-based N- and P-type emitters, each 2 μm thick, between which a 0.4-μm-thick AlGaAs waveguide layer (x = 0.3) was situated, with a 10-nm-thick InGaAs-based quantum well at the center of this layer (Fig. 1). The composition of the active region provided the maximum electroluminescence (EL) intensity at a wavelength of 900 nm. The active region was sandwiched between GaAs spacers, which made it possible to improve the optical activation efficiency of the transistor part [15]. The experimental laser-thyristor samples had a mesa-stripe design of the laser part, with a stripe width of 20 μm. The control contact was formed in the wide-bandgap N-AlGaAs collector (x = 0.35), which enabled effective pumping of the laser part by control current pulses. The laser-thyristor crystals were mounted substrate-down onto copper heatsinks with indium solder. A study of the optical characteristics of the laser part demonstrated that, for the heterostructure design with a narrow waveguide, the internal optical loss is 1.7 cm—1 [13]. In this case, the optimal cavity length providing the high emission efficiency of the laser part is 600–1000 μm. It was shown that lasers-thyristors with cavity lengths within this range have close power-related and dynamic characteristics [13,18], and, therefore, experimental samples with cavity length of 760 μm were chosen for study.
The dynamic characteristics of the lasers-thyristors were studied with a circuit that included an external dc voltage source, capacitor connected in parallel with the laser-thyristor, and a generator of control current pulses (Fig. 1). The working cycle of the laser-thyristor includes two stages. In the first stage the laser-thyristor is in the OFF state and has a high resistance. In this case, the voltage from the external source drops on the reverse-biased collector junction of the transistor part (or, in other words, the laser-thyristor blocks the voltage). This enables charging of the external capacitance. The energy stored in the external capacitance is determined by the voltage being blocked and by the capacitance, which, in turn, sets the amplitude and width of the current pulse pumping the laser part. In the second stage, the control pulse switches the laser-thyristor to the ON state with low resistance (the collector junction is unbent), which provides an effective discharge of the external capacitance through the multijunction heterostructure and pumping of the laser part. A controlled turn-on was realized by a low amplitude control current pulse (an amplitude of control current is defined as ICONTR) (Fig. 1). The control pulse had a negative polarity, which provided the pumping of the laser part when the laser-thyristor is in the OFF state. It is important to note that the laser-thyristor may be in the OFF state for an arbitrarily long time if the blocking voltage is lower than the maximum value (~20 V in our case) and only pumping by a control pulse switches the laser-thyristor to the ON state. Also noteworthy is the high control efficiency of the laser-thyristor, which is manifested in that the power of the control pulse is substantially lower than the power that the laser-thyristor can emit as a laser [16]. Another specific feature of the circuit is associated with the turn-off process of the laser-thyristor. In the variant under consideration, we use no additional control pulses to turn-off the laser-thyristor. The turn-off process occurs automatically when the external capacitance discharges to the critical voltage [17].
3. Experimental studies of lasing dynamics
Making narrower the mesa-stripe contact automatically reduces the maximum peak current at which the operation of the laser structure remains effective, which is due to the temperature-unrelated roll-off of the light-current characteristic [1–5]. In the case of the laser-thyristor, the maximum peak pumping current can be reduced by diminishing the external capacitance. Therefore, we used in the first part of the study a 10-nF external capacitor, which provided a laser pulse width of about 15 ns. Typical dependences of the long-pulse mode lasing dynamics are shown in Fig. 2. The maximum peak power reached a value of 2.5 W at a blocking voltage of 12 V. The peak power was determined from the experimentally measured waveform of the photoresponse, obtained using a photodetector with a bandwidth of 1 GHz (200-μm photosensitive surface area) and an oscilloscope with bandwidth of 6 GHz (emitted light was focused onto the photodetector pad with two aspheric lenses having focal distances of 5 mm), the average power was measured with an Newport 842-PE/918D-SL-OD3 optical power meter. Further increase in voltage failed to provide any significant rise in the peak power. Our results are comparable with those previously obtained for lasers-thyristors with wide aperture [13,18]. However, a subnanosecond front of lasing turn-on is present in the pulse. This front has not been resolved previously [13,18]. A subnanosecond front of lasing turn-on can be used for generation of nanosecond and subnanosecond laser pulses. To determine whether short pulses can be generated in the structures of this kind, the external capacitance was reduced to 0.5 nF. The lasing dynamics of the short-pulse mode was measured with subnanosecond resolution by using a photodetector having a 25-GHz bandwidth and an Agilent 86100 oscilloscope with an 86117A unit. As a result of reducing the external capacitance, the width of the current pulse generated in the laser-thyristor circuit decreased to 3-5 ns. The resulting lasing dynamics of the short-pulse mode is shown in Fig. 3. It can be seen that passing to the short-pulse mode markedly changed the nature of lasing. It was found that the blocking voltage and control current amplitude (ICONTR) strongly affect the shape and width of the laser pulse. A comparison of the current and lasing dynamics demonstrated that a “dead time” appears in the short-pulse mode, during which there is a current through the structure, but no lasing (in Fig. 4(b), the turn-on instant of the laser-thyristor corresponds to the beginning of the voltage decay).
Fig. 2 Dynamics of the output optical power of a laser-thyristor in the long-pulse mode for various blocking voltages (L = 760 μm, C = 10 nF).
Fig. 3 Dynamics of the output optical power of a laser-thyristor in the short-pulse mode for various blocking voltages and control currents (L = 760 μm, C = 0.5 nF).
Fig. 4 (a) Diagram of lasing modes of a laser-thyristor in the long- and short-pulse modes. (b) Dynamics of voltage on the external capacitance and of lasing in the short-pulse mode for U = 10 V and control current pulse amplitudes of 10 and 46 mA (insert: the voltage dynamics for these pulses) (L = 760 μm, C = 0.5 nF).
That is, the laser pulse is narrower than the current pulse (Fig. 3, control currents 10 and 16 mA). As shown below, this is due to the spatial localization of the current, at which only a part of the Fabry-Perot cavity is pumped. It is important to note that the amplitude of the current generated in the laser-thyristor circuit substantially exceeds the threshold current (in our case, the threshold current is 90 mA). This is well seen in the laser pulses obtained at high control currents, the peak power of which is hundreds of mW (see Fig. 3, ICONTR = 46 mA). This can be seen for a set of pulses obtained for, e.g., 10 V at control currents of 10, 16, and 46 mA, Fig. 3.
For working modes close to UMIN and UMAX (Fig. 4(a)), the pulse width is at a minimum, which is determined by the duration of the first relaxation peak in the gain-switching mode. On the UMIN side, the optical flux at the beginning of a current pulse is insufficient for the effective bleaching of the unpumped part of the Fabry-Perot cavity, which results in that the switching-on of lasing is delayed and the laser pulse becomes shorter (see Fig. 3, UMIN, ICONTR = 10 and 16 mA). A possible reason for the decrease in the pulse width on approaching UMAX (see Fig. 3, UMAX, ICONTR = 10 and 16 mA) is that the pumped region becomes smaller. For control currents close to the lasing threshold, when there is no need to bleach the unpumped part of the Fabry—Perot (because the control current pumps the whole Fabry-Perot cavity), the laser pulse duration is close to that of the current pulse generated in the laser-thyristor circuit (see Fig. 3, ICONTR = 46 mA) for the whole range of blocking voltages. The extreme case is the mode in which there is absolutely no lasing. The diagram of the working modes (amplitudes of control current pulses and blocking voltages) in which lasing is observed in the short- and long- pulse mode is shown in Fig. 4(a). In this diagram, and also in Fig. 3 and Fig. 5, UMIN and UMAX correspond to the minimum and maximum blocking voltages at which lasing can be observed for a particular control current amplitude (ICONTR). There is no lasing at voltages lower than UMIN and higher than UMAX. It can be seen that lowering the amplitude of the control current short-pulse mode makes narrower the range of voltages in which lasing is observed. As a result, the range of working modes in the short-pulse mode (nanosecond duration) is noticeably narrower than that in the long-pulse mode (>10 ns durations). (6) To achieve the maximum efficiency of the circuit, we removed all reference resistors that can furnish information about the current through the laser-thyristor. In this case, the current pulse duration can be judged from the voltage dynamics, measured across the capacitor C. Then, the voltage decay duration will demonstrate the current pulse width (see the inset of Fig. 4(b)). Figure 4(b) shows laser pulses obtained for a blocking voltage of 10 V at control currents of 10 and 46 mA, which provide a change in the laser pulse duration from 1 to 5 ns, respectively. The voltage dynamics for these pulses is shown in the inset of Fig. 4(b). It can be seen that the voltage decay duration and, accordingly, the pulse width of the current through the laser-thyristor is 5 and 7 ns at control currents of 10 and 46 mA, respectively. Thus, it can be seen that making lower the control current results in that the laser pulse becomes shorter at comparable durations of the current pulse through the laser-thyristor.
Fig. 5 Averaged lasing spectra of a laser-thyristor in the short-pulse mode for various blocking voltages and control currents (L = 760 μm, C = 0.5 nF).
To understand the reason for the observed specific features, we studied the near fields and the lasing spectra. The time-averaged spectra are shown in Fig. 5. The characteristic feature of the spectra is associated with the presence of two lines. The main lasing channel is via the long-wavelength line at limiting blocking voltages (near UMIN and UMAX), being the most pronounced for low control current amplitudes. The reason is that the unpumped part of the Fabry-Perot cavity has an additional optical loss associated with the interband absorption in the active region. Because the threshold conditions of lasing are expressed in that the modal gain is equal to the optical loss, the spectral dependence of the optical loss may lead to a shift of the lasing line relative to the peak of the gain spectrum. As a result, the lasing begins with the long-wavelength line, which provides a lower interband absorption in the active region in the unpumped part of the Fabry–Perot cavity. Raising the control current at a fixed voltage gives rise to a short-wavelength line in the lasing spectrum (see, e.g., Fig. 5, U = 10 V, ICONTR = 10 and 16 mA) and is simultaneously accompanied by the broadening of the laser pulse (Fig. 3, U = 10 V, ICONTR = 10 and 16 mA), with the long-wavelength line preserved. This indicates that raising the control current leads to a decrease in the loss in the unpumped part of the Fabry-Perot cavity. As a result, the spectrum is rearranged in a natural way into the short-wavelength region corresponding to the maximum gain. Thus, the appearance of the long-wavelength line for the limiting voltages (UMIN and UMAX) is due to the contribution of the absorption spectrum of the active region in the unpumped part of the Fabry-Perot cavity. The transformation of the spectrum via the inclusion of the long-wavelength line was first observed in [20] and attributed to the appearance of a new high-Q mode structure, i.e., the closed mode that covers the whole volume of the crystal, including the passive regions and characterized by nearly zero output optical losses. In our case, appearance of high-Q mode structures with low mirror loss is unlikely because of the inclined walls of the mesa stripe. It can also be noted that lasing only at the long-wavelength line is characteristic of the short-pulse mode with pulse widths shorter than 1 ns. However, the efficiency of this lasing is not high because the characteristic peak power levels are 300-400 mW, which is substantially lower than the peak power for the comparable blocking voltages in the long-pulse mode (~15 ns). As already noted, such a low power may be due to the poor differential efficiency because of the high optical loss in the unpumped region.
The time-averaged near field in the plane parallel to the p–n junction is shown in Fig. 6. The following characteristic features can be distinguished in the behavior of the near field. In generation of laser pulses with widths shorter than 1 ns (this mode is typical of low control currents and limiting blocking voltages Fig. 3), the laser mode is localized into single spots. The size of a single spot is ~5 μm, which is substantially smaller than the lateral size of the emitting region (20 μm). In this case, the near-field mode structure of the second harmonic reproduces the field structure of the main laser line. As shown in [20], the intensity of the emerging second-harmonic emission is determined by the intensity of the IR laser emission at the output mirror on the side of the Fabry-Perot cavity, irrespective of the mode incidence angle onto the output face. At the same time, the intensity of the output IR emission is determined only by the laser mode directed toward the output face at angles smaller than the total-internal-reflection angle. In addition, the Fabry-Perot modes and closed modes are characterized by different field distributions in the cavity. In this case, the contribution of the closed mode can be found by comparison of the fields distributions at the output mirror for IR laser emission of the Fabry-Perot modes and the second-harmonic emission. Therefore similarity of the second harmonic and the main laser line near-field mode structure also confirms that there are no high-Q mode structures. Raising the width of a laser pulse to several nanoseconds is accompanied by the filling of the whole lateral waveguide by laser emission. This behavior is the most characteristic of the mode with the maximum control current, when the mode structure fills the whole lateral waveguide in the whole range of working voltages (Fig. 6, ICONTR = 46 mA).
Fig. 6 Averaged intensity distribution in the near field of a laser-thyristor operating in the short-pulse mode for various blocking voltages and control currents (L = 760 μm, C = 0.5 nF).
All the results presented above characterized the lasing dynamics. However, a spatial current dynamics is possible in a laser-thyristor into which the function of a current switch is integrated. The first experimental results characterizing the spatial dynamics separately in a low-voltage phototransistor were published in [21]. It was shown that the current localization occurs in low-voltage phototransistors, and the size of the current filament may be substantially smaller (~10 μm) than the size of the optical activation region [21]. To examine the effect of current localization in the laser-thyristor, we analyzed distributions of the spontaneous emission intensity along the optical axis of the cavity in the laser part. In experiments, we recorded the spontaneous emission emerging from the side walls of a mesa stripe in the direction perpendicular to the layers of the heterostructure (Fig. 7). This made it possible to eliminate the contribution from the scattered laser emission. In this case, the emission intensity being recorded is determined by the carrier concentration in the active region of the laser part of the heterostructure and, consequently, it must depend on the lasing threshold and be stabilized at currents above the lasing threshold. The distributions obtained are shown in Fig. 7(a). It can be seen that, for the minimum control current at which the laser-thyristor is turned-on without lasing, the high-intensity spontaneous emission is localized near the mirror of the Fabry-Perot cavity. In this case, there is no emission from the rest of the cavity, which indicates that the current is localized. The size of the current-localization region cannot exceed that of the EL region, the minimum size of which is ~110 μm (Fig. 7(b)). Raising the control current amplitude is accompanied by expansion of the current-localization region to 200 μm and by a certain decrease in the EL intensity from the current-localization region (Fig. 7(b)). It can also be noted that traces of EL appear in the remaining part of the cavity as the control current amplitude increases (Fig. 7(a)). In this case, the current localization is preserved for all the working modes, irrespective of the control current amplitude and blocking voltage.
Fig. 7 (a) Intensity distribution of the spontaneous emission along the cavity axis for a laser-thyristor operating in the short-pulse mode at various control current amplitudes. (b) Dependence of the width and peak intensity of the spontaneous emission from the current-localization region on the control current amplitude. (c) Schematic of the laser-thyristor crystal and of the recording direction of spontaneous emission (L = 760 μm, C = 0.5 nF).
4. Analysis of experimental results
The experimental results we obtained enable a rather full description of the processes occurring when lasers-thyristors generate short pulses. A control current pulse turns-on the laser-thyristor; however, the current pumps only a part of the mesa stripe because of the localization effect. Then, the laser-thyristor crystal can be regarded as an analog of a laser with a double-section design. In our case, the “gain section” is the current-localization region, and the “absorption section” is the remaining part of the crystal. The characteristics of the “absorption section” are changed by the optical pumping from the “gain section” and by the current pumping by the control pulse, which uniformly pumps the whole mesa stripe. The “gain section” determines the threshold condition (modal gain compensates for the optical loss). However, the optical confinement factor in the “gain section” may be several times smaller than that in ordinary semiconductor lasers due to the localization of the current pumping region. This requires a substantial increase in the material gain (and, as a consequence, in the threshold concentration) in the pumping region for the same optical loss to be compensated. At low control currents and low blocking voltages, the modal gain in the current pumping region is insufficient for the loss compensation. As a result, there is no lasing when a current pumping the laser-thyristor. In addition, the size of the current filament may depend on the blocking voltage. At a high current and blocking voltage the threshold cannot be approached too. This is due to the saturation of the material gain at high carrier concentrations in the active region [22]. In addition, the time necessary for the threshold concentration to accumulate in the “gain section” and for bleaching of the “absorption section” determines the lasing turn on delay. Operation in the mode of a quasi-double-section design also accounts for the appearance of the long-wavelength line in the lasing spectrum. In this case, the optical pumping and control current are insufficient for bleaching of the “absorption section” in the energy range corresponding to the peak in the gain spectrum. It results in that the threshold condition is satisfied in the long-wavelength part of the gain spectrum, corresponding to the lower absorption (bleaching) [23].
The current localization and the quasi-double-section cavity structure can be effectively used for operation in the gain-switching mode. The experimental results demonstrate that we can realize lasing modes, in which only the first relaxation peak is observed. These lasing modes were already demonstrated above (Fig. 3), but the low peak power and the operation at the long-wavelength line show that the chosen conditions are not optimal. An example of optimal lasing modes for generation of a single subnanosecond pulse is shown in Fig. 8. The optimization of the control current (25 mA) and blocking voltage (5.8 V) made it possible to obtain a single pulse with peak power of 1.6 W and width of 90 ps (Fig. 8(a)). In this case, the maximum generation frequency of a controlled train of such pulses was 470 kHz and was limited by the effect of laser-thyristor “sticking” in the ON state [17]. The voltage dynamics in the given mode, which reflects the repetition frequency of the current pulses is shown in Fig. 8(c). In addition, there is no continuous zero line in the optical pulse being recorded. This means that there is no optical pulse skipping relative to the electric pulses being generated. It can be seen that raising the control current provides that a part of the cavity that is not pumped by the main current pulse due to the localization effect is bleached. As a result, the lasing spectrum includes only the main short-wavelength line (Fig. 8(d)). The near field also includes a single spot (see the inset of Fig. 8(d)). A study of the spatial dynamics of the near field shows that the central part of the spot is characterized by the maximum intensity and minimum width of 50 ps, whereas the side satellites of the spot have a slower dynamics (Fig. 8(b)). This may indicate that there are several mode structures. Selection of these mode structures may be of key importance for obtaining the minimum duration of the first relaxation peak in the gain-switching mode.
Fig. 8 (a) Output dynamics of the optical power of the laser-thyristor, demonstrating the operation at only the first relaxation peak in the gain-switching mode; (b) lasing dynamics at various points of the near field; (c) voltage dynamics of the laser-thyristor; (d) lasing spectrum for the mode with U = 5.8 V and ICONTR = 25 mA (insert: distribution of the time-averaged near field along the mesa stripe) (L = 760 μm, C = 0.5 nF).
A study of lasers-thyristors with narrow (20 μm) stripe contact demonstrated that varying the cavity length within the range 600-1000 μm has no effect on the frequency characteristics. A similar result was obtained for laser-thyristors with wide stripe contact (200 μm). As shown in [17], the main factors determining the maximum lasing frequency are (i) holding current (minimum current at which a laser-thyristor in the ON state) and (ii) differential resistances of the heterostructure in the ON state. It was found that, for the cavity lengths in the range under study, these characteristics are rather close. The most effective way to raise the generation frequency is by making higher the holding current because an increase in the differential resistance results in that the energy efficiency of the device decreases [17]. This can be done by mismatching the absorption spectrum of the p-base and that of the spontaneous emission from the active region [15], or by introducing additional absorption regions into the heavily doped collector [13]. It was shown in [17] that raising the holding current to 100 mA in high-power lasers-thyristors with a 200-μm mesa stripe makes it possible to raise the working generation frequencies of laser pulses to 12 MHz.
As shown above, the change in the pulse width (for the given switch-on conditions: blocking voltage, capacitance) is due to the influence exerted by the control current amplitude on the loss-gain of the unpumping and the size of the pumping region in the Fabry-Perot cavity. It can be seen for lasers-thyristors with a narrow stripe contact that raising the drive current favors broadening of the laser pulse. As the stripe width increases, the effect of an inhomogeneous spreading of the control current is manifested, associated with the design of the control electrode (Fig. 1). As a result, the influence exerted by the control current on the loss-gain and localization of the pumping region in the Fabry-Perot cavity may be different for lasers with a wide stripe contact. An experimental analysis of these factors requires an additional study.
5. Conclusion
The study demonstrated that the developed low-voltage lasers-thyristors with a narrow mesa stripe contact provide generation of Watt-level nanosecond and subnanosecond laser pulses. The amplitude of the control current did not exceed 25 mA, which is substantially lower than the amplitudes of current pulses across a laser-thyristor. An important advantage of the suggested structure is that an effective current switch and a laser emitter are integrated. As a result, there is no need to use external pulsed sources with high current amplitude.
It was shown that the main factor that enables the mode in which subnanosecond pulses are generated is the local pumping of the Fabry-Perot cavity. In this case, the unpumped part of the Fabry-Perot cavity serves as a saturable absorber, which enables operation in the gain-switching mode under pumping with current pulses having widths of several nanoseconds. It was demonstrated that the characteristics of the lasing modes in which short pulses are generated (duration, power, spectrum) are determined by the operation modes (blocking voltage, control current). Two factors can be distinguished, which affect the laser pulse dynamics in opposite ways. The first of these is the control current providing that the localization region expands and the optical loss decreases. The second is the blocking voltage, which can enhance the current localization. The current localization effect can serve as a fundamental basis for development of high-power integrated subnanosecond generators of laser pulses.
Funding
The study was carried out by the grant from Russian Science Foundation (project Nº 14-19-01560).
References and Links
1. D. A. Veselov, V. A. Kapitonov, N. A. Pikhtin, A. V. Lyutetskiy, D. N. Nikolaev, S. O. Slipchenko, Z. N. Sokolova, V. V. Shamakhov, I. S. Shashkin, and I. S. Tarasov, “Saturation of light –current characteristics of high-power lasers (l = 1.0 –1.1 mm) in pulsed regime,” Quantum Electron. 44(11), 993–996 (2014). [CrossRef]
2. X. Wang, P. Crump, H. Wenzel, A. Liero, T. Hoffmann, A. Pietrzak, C. M. Schultz, A. Klehr, A. Ginolas, S. Einfeldt, F. Bugge, G. Erbert, and G. Trankle, “Root-cause analysis of peak power saturation in pulse-pumped 1100 nm broad area single emitter diode lasers,” IEEE J. Quantum Electron. 46(5), 658–665 (2010). [CrossRef]
3. D. A. Veselov, N. A. Pikhtin, A. V. Lyutetskiy, D. N. Nikolaev, S. O. Slipchenko, Z. N. Sokolova, V. V. Shamakhov, I. S. Shashkin, N. V. Voronkova, and I. S. Tarasov, “Study of the absorption coefficient in layers of a semiconductor laser heterostructure,” Quantum Electron. 45(7), 604–606 (2015). [CrossRef]
4. B. S. Ryvkin and E. A. Avrutin, “Asymmetric, nonbroadened large optical cavity waveguide structures for high-power long-wavelength semiconductor lasers,” J. Appl. Phys. 97(12), 123103 (2005). [CrossRef]
5. L. V. Asryan and Z. N. Sokolova, “Optical power of semiconductor lasers with a low-dimensional active region,” J. Appl. Phys. 115(2), 023107 (2014). [CrossRef]
6. A. Y. Leshko, A. V. Lyutetskii, N. A. Pikhtin, S. O. Slipchenko, Z. N. Sokolova, N. V. Fetisova, E. G. Golikova, Y. A. Ryaboshtan, and I. S. Tarasov, “High power single-mode (l = 1.3–1.6 mm) laser diodes based on quantum well InGaAsP/InP heterostructures,” Semiconductors 36(11), 1308–1314 (2002). [CrossRef]
7. A. Klehr, H. Wenzel, J. Fricke, F. Bugge, A. Liero, T. Hoffmann, G. Erbert, and G. Trankle, “Generation of spectrally stable continuous-wave emission and ns pulses at 800 nm and 975 nm with a peak power of 4 W using a distributed Bragg reflector laser and a ridge-waveguide power amplifier,” Proc. SPIE 9382, 93821I (2015). [CrossRef]
8. B. Sumpf, A. Klehr, T. N. Vu, G. Erbert, and G. Trankle, “975 nm high-peak power ns-diode laser based MOPA system suitable for water vapor DIAL applications,” Proc. SPIE 9382, 93821K (2015). [CrossRef]
9. S. Riecke, K. Posilovic, T. Kettler, D. Seidlitz, V. A. Shchukin, N. N. Ledentsov, K. Lauritsen, and D. Bimberg, “10.7W peak power picosecond pulses from high-brightness photonic band crystal laser diode,” Electron. Lett. 46(20), 1393–1394 (2010). [CrossRef]
10. B. Lanz, B. S. Ryvkin, E. A. Avrutin, and J. T. Kostamovaara, “Performance improvement by a saturable absorber in gain-switched asymmetric-waveguide laser diodes,” Opt. Express 21(24), 29780–29791 (2013). [CrossRef] [PubMed]
11. B. Lanz, S. N. Vainshtein, and J. T. Kostamovaara, “High power gain-switched laser diode using a superfast GaAs avalanche transistor for pumping,” Appl. Phys. Lett. 89(8), 081122 (2006). [CrossRef]
12. J. M. T. Huikari, E. A. Avrutin, B. S. Ryvkin, J. J. Nissinen, and J. T. Kostamovaara, “High-energy picosecond pulse generation by gain switching in asymmetric waveguide structure multiple quantum well lasers,” IEEE J. Sel. Top. Quantum. Electron. 21(6), 1501206 (2015). [CrossRef]
13. S. O. Slipchenko, A. A. Podoskin, A. V. Rozhkov, N. A. Pikhtin, and I. S. Tarasov, “High-power laser-thyristors with high injection efficiency (λ=890-910 nm),” IEEE Photonics Technol. Lett. 27(3), 307–310 (2015). [CrossRef]
14. S. O. Slipchenko, V. S. Yuferev, A. A. Podoskin, O. S. Soboleva, N. A. Pikhtin, and I. S. Tarasov, “Specific features of the injection processes dynamics in high-power laser thyristor,” IEEE Trans. Electron Dev. 62(12), 4091–4096 (2015). [CrossRef]
15. A. A. Podoskin, O. S. Soboleva, M. S. Zakharov, D. A. Veselov, V. V. Zolotarev, N. A. Pikhtin, I. S. Tarasov, T. A. Bagaev, M. A. Ladugin, A. A. Marmalyuk, V. A. Simakov, and S. O. Slipchenko, “Optical feedback in 905 nm power laser-thyristors based on AlGaAs/GaAs heterostructures,” Semicond. Sci. Technol. 30(12), 125011 (2015). [CrossRef]
16. S. O. Slipchenko, A. A. Podoskin, V. V. Vasil’eva, N. A. Pikhtin, A. V. Rozhkov, A. V. Gorbatyuk, V. V. Zolotarev, D. A. Veselov, A. V. Jabotinskii, A. A. Petukhov, I. S. Tarasov, T. A. Bagaev, M. V. Zverkov, V. P. Konyaev, Y. V. Kurniavko, M. A. Ladugin, A. V. Lobintsov, A. A. Marmalyuk, A. A. Padalitsa, and V. A. Simakov, “On the control efficiency of a high-power laser thyristor emitting in the 890–910 nm spectral range,” Semiconductors 48(5), 697–699 (2014). [CrossRef]
17. S. O. Slipchenko, A. A. Podoskin, O. S. Soboleva, D. A. Veselov, V. V. Zolotarev, N. A. Pikhtin, T. A. Bagaev, M. A. Ladugin, A. A. Marmalyuk, V. A. Simakov, and I. S. Tarasov, “Generation of laser pulses in the MHz range of repetition frequencies by low-voltage AlGaAs/GaAs lasers-thyristors,” IEEE Trans. Electron Dev. (to be published).
18. S. Slipchenko, A. Podoskin, O. Soboleva, M. S. Zakharov, K. Bakhvalov, D. Romanovich, N. Pikhtin, I. Tarasov, T. Bagaev, M. Ladugin, A. Marmalyuk, and V. Simakov, “High-efficiency and compact semiconductor lasers with monolithically integrated switches for generation of high-power nanosecond pulses in time-of-flight (TOF) systems,” Proc. SPIE 9751, 97510P (2016). [CrossRef]
19. S. O. Slipchenko, A. A. Podoskin, A. V. Rozhkov, N. A. Pikhtin, I. S. Tarasov, T. A. Bagaev, V. P. Konyaev, M. A. Ladugin, A. A. Marmalyuk, A. A. Padalitsa, and V. A. Simakov, “A study of nonlinear lasing dynamics of an InGaAs/AlGaAs/GaAs heterostructure power laser-thyristor emitting at 905 nm,” J. Appl. Phys. 116(8), 084503 (2014). [CrossRef]
20. S. O. Slipchenko, A. A. Podoskin, N. A. Pikhtin, and I. S. Tarasov, “Fundamental aspects of closed optical mode formation in Fabry–Perot semiconductor lasers based on AlGaAs/GaAs (905nm) asymmetric heterostructures,ˮ,” Laser Phys. 24(10), 105001 (2014). [CrossRef]
21. S. O. Slipchenko, A. A. Podoskin, O. S. Soboleva, N. A. Pikhtin, T. A. Bagaev, M. A. Ladugin, A. A. Marmalyuk, V. A. Simakov, and I. S. Tarasov, “Spatial dynamics of high current turn-on in low-voltage AlGaAs/GaAs phototransistors,ˮ,” J. Appl. Phys. 119(12), 124513 (2016). [CrossRef]
22. S. O. Slipchenko, A. A. Podoskin, N. A. Pikhtin, A. L. Stankevich, N. A. Rudova, A. Y. Leshko, and I. S. Tarasov, “Electroluminescence and absorption spectra of low-optical-loss semiconductor lasers based on InGaAs/AlGaAs/GaAs QW heterostructures,” Semiconductors 45(5), 673–678 (2011). [CrossRef]
23. S. O. Slipchenko, A. A. Podoskin, N. A. Pikhtin, V. V. Zolotarev, I. S. Shashkin, A. Y. Leshko, A. V. Lutetskyi, M. G. Rastegaeva, I. S. Tarasov, and P. S. Kop’ev, “Multi-wavelength integrated optical-laser emission modulator based on semiconductor heterostructures,” Semiconductors 48(5), 691–696 (2014). [CrossRef]
