Characteristics of GaAs lasers near room temperature

The efficiency and threshold of GaAs lasers can be affected by a delay occurring between the application of a current pulse and the laser output. The delays may be some tenths of a microsecond in certain diffused junction lasers when operated near room temperature, e.g., when the donor substrate dopant is selenium with a concentration of the order of 10¹⁸atoms per cm³. No such delays are observed in diodes in which the substrate dopant is silicon. In the case of selenium, the delay depends on the doping level, decreasing as the concentration increases. The delay is also very sensitive to injection current, decreasing rapidly with increasing current from its maximum at threshold. During the pulse, after lasing commences, the output continues to increase. A similar effect has also been observed with spontaneous emission. Operation of these devices below room temperature shows that the delay is also dependent on temperature. The delay decreases over a fairly narrow temperature range and in all cases is no longer observed ( < 20 n/s) below -70°C. These observations are explained by considering the effect of impurity trapping levels which may be associated with the     

Spiking in light pulses from GaAs Q-switched junction lasers

Using a detection system with 150-ps resolution, a spiking behavior has been detected in theQ-switched light output from GaAs junction lasers. At the lower currents in theQ-switching region, a single light spike, whose width is about 300 ps, is observed. At higher currents, additional light spikes appear whose widths and spacings decrease as the current is increased. At the highest pumping levels, only the initial spike is clearly resolved and its width has decreased to less than 200 ps. Qualitative agreement is obtained from a simple theory based on the standard rate equations. The necessary modifications to the theory are discussed and results of computer calculations are presented which predict that the width of the initial spike can be much less than 100 ps at sufficiently high pumping levels.     

Characteristics of bistable CW GaAs junction lasers operating above the delay-transition temperature

The conditions for bistable operation of CW GaAs junction lasers are developed in terms of the previously published double-acceptor trap theory. The experimental CW operation of such devices is shown to agree well with the theoretical results. In addition the fabrication of these bistable lasers is described and several pulsed experiments are reported that indicate a significant increase in the number of trapping centers in the vicinity of the junction.     

Internal Q-switching in GaAs junction lasers

InternallyQ-switched light pulses have been obtained from junction lasers. It is believed that this is a completely new observation for semiconductor lasers. Using specially fabricated diodes, narrow bursts of light were detected immediately after the termination of the injection current pulse. The effect persists for a wide variation in the length of the current pulse, from less than 2 ns to several μs. The width of theQ-switched light pulse itself is less than 0.4 ns, this value being the resolution of our detection system. Its energy increases rapidly with the amplitude of the injection current. The occurrence of stimulated emission after the end of the injection pulse indicatesQ-switching due to a reduction of the internal absorption. This reduction allows those injected carriers that have not yet spontaneously recombined to produce the narrow burst of stimulated light. TheQ-switching is observed over a current amplitude range that is a very strong function of temperature. This range can be relatively large. One diode at 150°K showedQ-switching after current pulses from 1.2 to 5.0 amperes; for amplitudes greater than 5.0 amperes, normal stimulated emission occurred during the current pulse. As the temperature is increased, theQ-switched pulse is first observed near Tt, the so-called "transition temperature" where trapping effects first start to cause long delays between the application of the current pulse and the onset of stimulated emission. Therefore, it is believed that the same traps are involved in both the long delays and theQ-switching. The latter is observed only in diodes with low Ttvalues where the absorption due to traps accounts for a large proportion of the total losses of the laser. A model explaining these effects will be presented.     

Resonant modes of GaAs junction lasers

The resonant modes of GaAs junction lasers are obtained from a proposed model and compared with experimental results. Theoretical results are based on an assumed laser medium whose dielectric constant varies both along and perpendicular to the junction plane. The frequency separations of the transverse modes are found to be in very good agreement with presented high-resolution spectral measurements of stripe-geometry laser radiation. Furthermore, the theoretical field distributions are also in good agreement with observed transverse field variations reported previously. The laser output spectrum usually shows a number of "satellites" located adjacent to each longitudinal (Fabry-Perot) resonance. It is shown that each satellite represents the frequency of a transverse resonance having a different mode number along the junction plane and a corresponding Hermite-Gaussian intensity profile along the plane. Theory shows that the frequency separation of two adjacent satellites is related to the falloff rate of the dielectric constant along the junction plane. This focusing is measured from the intensity profile of a far-field pattern. From this information, the theoretical frequency separation is calculated and found to agree well with the measured separation of 6.4 GHz (0.15 Å).     

Continuous microwave oscillations in GaAs junction lasers

Microwave oscillations produced by continuously operating GaAs stripe geometry junction lasers are reported here for the first time. These oscillations have been measured both in the light output of the diodes and in the dc current applied to the laser connections. The frequency of the oscillation lies in the 0.5- to 3.0-GHz range and depends strongly on current and temperature. The frequency typically increases with increasing current at a rate of 15 to 20 MHz/mA and decreases with increasing temperature at 100 to 150 MHz/°K. Introduction of an external microwave signal locks the frequency of the oscillation and reduces the oscillation width from 10 MHz to the external signal width. The frequency of oscillation agrees with a theory of intensity fluctuations in lasers based on the rate equations.     

External resonator turn-on time-delay effects with GaAs injection lasers at room temperature

Stimulated emission has been observed at the end of the current pulse for a GaAs injection laser operated with an external resonator at room temperature. It has been observed in the peak injection current range of 65- 95 A and for current pulse half-widths in the investigated range of 250-500 ns. This effect was found only when a diode was operated with the external resonator. Stimulated emission both during the current pulse and at the end of the current pulse also has been observed.     

Time-resolved spectral output of pulsed GaAs lasers

When pulse operated, all injection lasers show a laser line shift which is linear with time during the pulse length. This is evident for time-resolved spectra obtained by a scanning technique. The data present time-resolved spectra for pulse-operated gallium arsenide laser diodes. At high peak current values, joule heating results and the time-resolved laser spectra are seen to shift to longer wavelengths. A line-shift coefficient is defined for which the units are angstroms per (ampere)²per microsecond. For high peak current and long pulse length, the laser lines may shift one hundred angstroms or more.     

Dependence of the total stimulated light power and the gain factor in GaAs junction lasers on laser length

The use of an auxiliary loop aerial in conjunction with a dipole is considered and it is shown to lead to a resonant frequency of operation which may be greater than, equal to, or less than that of the same dipole in isolation.     

Noise characteristics of stripe-geometry double-heterostructure junction lasers operating continuously--I. Intensity noise at room temperature

Measurements of the noise fluctuations present in the output intensity from stripe-geometry double-heterostructure junction lasers operating continuously at room temperature are reported. In some but not all the lasers studied, the low-frequency (50-MHz) fluctuations exhibit the quieting expected of an amplitude-stabilized oscillator operating above threshold. The intensity noise in these lasers becomes shot noise limited at currents about 10 percent above threshold, even when many longitudinal modes are oscillating. Additional measurements demonstrate explicitly the effective elimination of the wave-interaction, or excess, noise during the transit of the threshold region. However, in other lasers which are nominally similar, only a partial reduction of the excess noise occurs above threshold, resulting in a noise level which can be more than ten times the shot-noise limit. In addition, we find in all lasers studied thus far no quieting at high frequencies (4 GHz), in apparent contradiction to the behavior expected of a well-stabilized oscillator.     

Subnanosecond light pulses from GaAs injection lasers

Self-induced regular pulsations of the light intensity of GaAs injection lasers have been observed in pulse-driven diodes at liquid-nitrogen temperatures. The pulses occur at a repetition frequency of 500 MHz to 1 GHz, and the pulsewidth is less than 400 ps. These results are examined in the light of two theories that have been proposed to explain this kind of behavior in the GaAs laser. The basic differences between these two models are briefly described.     

Optimum stimulated light powers from GaAs injection lasers

The total stimulated light power obtainable from GaAs injection lasers is optimized by varying reflection coefficient, laser length, and width. The laser performance is assumed to be limited by heating, and a given temperature rise is taken as a constant parameter of the optimization procedure. Two cases are considered: 1) operation with short pulses at low duty cycle, and 2) continuous operation. For both cases optimum reflection coefficients are obtained. Case 2) leads to optimum values for laser length and width, case 1) to favorable length values only.     

Directional light output from photonic-wire microcavity semiconductor lasers

We have obtained directional light output from a recently realized InGaAsP photonic-wire microcavity ring lasers. The output was achieved by fabricating a 0.45-/spl mu/m-wide U-shape waveguide next to a 10-/spl mu/m diameter microcavity ring laser. The laser has a threshold pump power of around 124 /spl mu/W when optically pumped at 514 nm. It is comparable to the former structure without output coupling. The output coupling efficiency can be controlled carefully by choosing the spacing between the laser cavity and the waveguide.     

The phase shift of the light output in sinusoidally modulated semiconductor lasers

The dependence of the phase shift of the light output from sinusoidally modulated semiconductor lasers was investigated as a function of the modulation current. This measurement is effective in accurately determining the short damping time constant associated with the relaxation oscillation. The frequency half width of this phase shiftDelta fwas found to be inversely proportional to the damping time constant. For narrow stripe lasers, the phase shift occurs more gradually, which corresponds to the fact that the narrow stripe lasers have shorter damping time constants. To analyze the narrow stripe effect, the recently developed time-dependent self-consistent theory was applied, considering the transverse distribution of both optical field and carrier density and including the carrier diffusion term. This theory can explain the shorter damping time constant for narrow stripe lasers compared with broader stripe lasers.     

Recombination enhanced annealing effect in AlGaAs/GaAs remote junction heterostructure lasers

New structure lasers, the remote junction heterostructure (RJH) lasers, are made to obtain information about slow degradation of AlGaAs/GaAs DH lasers. The RJH laser is characterized by the presence of a thin clad layer between the active layer and the p-n junction. During the LD and LED mode aging process, the RJH lasers showed a marked reduction of threshold current. This reduction was accompanied by increased spontaneous lifetime and pileup of defects at the p-n junction. From these observations, a model was proposed in which point defect generation in the active layer and defect motion toward the p-n junction during the aging are assumed. The rate equation was derived for concentration of the point defect, and the solution of this equation was compared with the experimental results with reasonable agreement. The parameters relating to the slow degradation were determined, and the ultimate life of conventional DH lasers was discussed using these parameters.     

Transverse light coupling from GaAs lasers into optical fibres

We report a method for transverse light coupling from GaAs lasers into optical fibres. The fibre end is positioned perpendicular to the laser-beam direction and is cut at the end under 45° to reflect the laser radiation. We measured 10% coupling efficiency for a plastics fibre which was simply cut by a blade. We used this method for an automatic-gain-control circuit and picked up reference light from the back side of a d.h. laser.     

Surface aspects of the thermal degradation of GaAs p-n junction lasers and tunnel diodes

Owing to the similarity of the junction widths and doping levels in epitaxially formed GaAs lasers and conventional tunnel diodes, the gradual degradation of the latter under use might be expected to be intensified in the case of lasers operating under high duty cycle or CW conditions. It was found that the current-voltage characteristics of both the lasers and tunnel diodes underwent considerable thermal degradation even when no bias was applied, and that the degradation was limited to surface or "junction edge" effects. The degradation of the current-voltage characteristic was correlated with degradation in the light output under forward bias. It was concluded that degradation under forward bias was also a surface effect induced by heating rather than by deterioration of the junction in the bulk, as commonly believed. Methods of controlling the degradations are discussed.