B. W. Hakki

Gain spectra in GaAs double−heterostructure injection lasers

Gain spectra for GaAs double−heterostructure junction lasers have been obtained with high resolution. This is accomplished by using an automated data aquisition system to analyze the Fabry−Perot resonance modulation in the spontaneous emission spectra. For active regions doped with Ge at a level of 4×1017 cm−3, the gain in the TE polarization at a fixed wavelength increases linearly with current, below lasing threshold. However, the peak gain (at a variable wavelength) increases slightly faster than linearly with current. The photon energy at which gain is a maximum increases logarithmically with current. Gain in the TM polarization depicts the same general behavior as that for the TE case, except that it is slightly less than the TE gain. It is concluded that for this particular doping the spectral gain characteristics are intermediate between those for undoped and heavily doped active regions. Above the threshold for lasing in the TE mode the TE gain spectra are well saturated, with new fine details rev...

cw degradation at 300°K of GaAs double‐heterostructure junction lasers. II. Electronic gain

The rapid degradation at 300°K in the cw regenerative output of stripe‐geometry GaAs double‐heterostructure junction lasers is shown to be a result of the formation of a local optical absorber in the laser cavity. Gain measurements performed on diodes before and after degradation show that the optical loss within the cavity increases during degradation. By observing the (predominantly) spontaneous emission from the active region directly through the n‐GaAs substrate, it is confirmed that the increased loss is localized in a region where little or no spontaneous emission takes place at lasing energies. In such diodes, the internal radiative efficiency of the undegraded portion of the optical cavity shows a relatively small decrease compared to the external differential quantum efficiency. When the local absorber extends over a sufficient length of the cavity the electronic gain in the undegraded section is insufficient to overcome the loss and the device ceases to act as a regenerative optical oscillator. ...

Carrier and gain spatial profiles in GaAs stripe geometry lasers

The spontaneous emission generated in the active region of stripe geometry GaAs double‐heterostructure (DH) laser diodes is observed through the n‐GaAs substrate. The spatial variation of the spontaneous intensity along the junction plane is measured in a direction normal to the stripe axis at current levels below and above lasing threshold. It is observed that the carrier concentration profile is consistent with a model in which carriers out diffuse along the junction plane away from the active region. No evidence has been found of a local depletion of carrier concentration due to regenerative optical oscillations. Analytical expressions for the concentration profiles are obtained from a diffusion model. A comparison of the analytical expressions with experiment yields the carrier diffusion lengths inside and outside of the active region. The computed profiles are in excellent agreement with measurement. From the computed carrier concentration profiles and measurements of mode gain dependence on current ...

Catastrophic failure in GaAs double-heterostructure injection lasers

Catastrophic failure in GaAs double‐heterostructure (DH) injection lasers is shown to manifest itself in two ways concurrently: rupture of one or both laser mirrors, and internal damage to the laser active region. When the electroluminescent emission is observed in a direction normal to the plane of the junction, internal damage takes the form of a dark line which starts from the damaged mirror and extends normal to it into the active region. The increase in lasing threshold and decrease of external efficiency of a catastrophically damaged laser are due to (i) decrease in reflectivity of the damaged mirror, (ii) reduction of electronic gain (and possible conversion into loss) of the active region immediately adjacent to the dark line, and (iii) optical scattering due to internal crystal damage associated with the dark line. It is shown that catastrophic damage in GaAs is consistent with a model whereby failure occurs when the optical electric field intensity at the GaAs mirror exceeds a certain critical v...

GaAs double heterostructure lasing behavior along the junction plane

Properties along the junction plane of GaAs double heterostructure (DH) injection lasers are studied in a self−consistent analytical model which takes into account the effect of lasing action in the active region. Solutions are obtained for the distribution of current under the stripe, and profiles of carrier concentration and gain in the active region, all as functions of lasing intensity. The results of the analysis are applied to practical situations in which the stripe width is varied from 10 to 20 μm, and the distance between the active region and the metallic stripe is varied between 0.5 and 3 μm. For these situations it is found that lasing action in the active region can exert a strong influence on the spatial distribution of the gain component of junction current. This influence increases as a function of increase in stripe width, increase in resistivity in the layer separating the active region from the stripe contact, and reduction of carrier diffusion length. In particular, it is found that as...

Band Structure of InGaP from Pressure Experiments

Positive identification of the Γ and X conduction‐band minima in InGaP has been made by performing hydrostatic‐pressure experiments on forward‐biased p‐n junction diodes. The Γ and X valleys are coincident in energy at a composition of In0.37Ga0.63P, and the corresponding bandgap is (2.17±0.02) eV at 300°K. The indirect bandgap EX in InP is inferred from the measurements to be 2.0 eV at 300°K. In addition, the pressure coefficients of the direct and indirect bandgaps, ∂EΓ/∂P and ∂EX/∂P, respectively, have been measured at various In(1−x)GaxP compositions. For InP, ∂EΓ/∂P is 8.7×10−3 eV/kbar and this coefficient increases to 13×10−3 eV/kbar for compositions close to In0.5Ga0.5P. On the other hand, ∂EX/∂P = −1.25×10−3 eV/kbar for GaP and shows little change for compositions in the range 0.4<x<1. The deformation potential is 5.7 eV for the Γ valley in InP, and increases to about 9 eV for direct bandgap ternary compositions.

Polarization mode dispersion in a single mode fiber

The Jones matrix method is used to measure the polarization mode dispersion (PMD) of a large variety of single mode fibers in the 1500 nm range. The dependence of PMD on wavelength, time, and temperature are studied in two different regimes: adiabatic and isothermal. In the adiabatic regime, time dependent stresses are introduced in the fiber by subjecting it to large and rapid changes in temperature. In this regime it is shown that the probability density function of PMD, obtained as a function of temperature and wavelength, fits very closely the theoretically predicted Maxwellian function. In the isothermal regime, the temperature of the fiber is held constant and the stresses are allowed to relax to their long term steady state conditions. In this regime the PMD exhibits a strong dependence on wavelength but otherwise is a bounded function which is nearly stationary with time. Test and analysis of the deterministic PMD in a specially constructed polarization maintaining fiber are used to study the dependence of PMD on temperature and wavelength. Finally, the system implications of this PMD study are described briefly.

Striped GaAs lasers: Mode size and efficiency

The stimulated efficiency of the fundamental mode and its beamwidth along the junction plane are calculated for GaAs double‐heterostructure (DH) stripe‐geometry lasers in a self‐consistent analytical model. Carrier diffusion along the junction plane creates a gain profile which acts as the guiding mechanism for the propagating mode. The size of the mode and the diffusion‐governed gain profile are the two interdependent parameters that enter into the characteristic equation whose solution determines the lasing threshold. The analytical results obtained indicate that the beamwidth is almost linearly related to the carrier diffusion length and stripe width. The minimum beamwidth obtainable for zero stripe width is ?1.3 times a diffusion length. The calculated waveguide loss, due to absorption in the unpumped region, increases with reduced stripe width. This mode loss is 5–7 cm−1 for 10–12‐μm‐wide stripes, in agreement with experiment. Due to the increased waveguide loss the predicted external efficiency decr...

Amplification in Two‐Valley Semiconductors

Microwave amplification and negative conductance are discussed theoretically and experimentally for those semiconductors that exhibit the Gunn effect, e.g., GaAs, InP, and CdTe. Field‐induced transfer of carriers, from the high‐mobility, main conduction band minimum to energetically higher, low‐mobility valleys, gives the medium a negative differential bulk conductivity. The resultant growing space‐charge waves impart to the semiconductor structure a calculated negative conductance which is in good agreement with experiment. The regime of stable amplification in n‐GaAs is found to be (2×107/l)<−(1+γ)n1l<0.65×1011 cm−2, where n1 is the carrier concentration in the lower valley, l is the sample length, and γ is the intervalley field rate of transfer of carriers. Static field distortion imposes a lower limit at n1l=1010 cm−2 on the above relation. Amplification under conditions of strong field distortion is also discussed, as well as other modes of amplification that exist in the bulk semiconductor.

Evaluation of transmission characteristics of chirped DFB lasers in dispersive optical fiber

Experimental and analytical results are presented for pulse distortion and receiver penalty associated with wavelength chirp of 1.5 mu m distributed feedback (DFB) lasers in the presence of fiber dispersion. An approximate analytical model for dispersion penalty due to chirp is given and shown to provide reasonable agreement with the experimental results obtained for digital transmission at 1.7 Gb/s. The model captures the important features of chirp that lead to dispersion penalty, such as the maximum wavelength excursion, the chirping time, and the energy content at the leading edge of the pulse during chirp. The utility of the model is that the parameters used in it can be readily obtained from single experimental measurements. The limitations of the model are also discussed. >