<p>A detailed study is presented of power and spectral measurement methodology commonly used to characterize and optimize the fundamental continuous-wave properties of semiconductor lasers. These properties include: efficiency, optical loss, the temperature sensitivity of threshold, gain, and spectral linewidth. The techniques studied are found to often yield erroneous or misleading results. The conditions under which errors can occur are investigated and precautionary measures necessary to avoid these problems are outlined. The effect of well number, length, and temperature on efficiency and optical loss is investigated. A phenomenological model based on inter-valence band absorption (IVBA) for the description of the results is developed. The model provides evidence of the importance of IVBA in determining laser characteristics, including the failure to lase. The operating regime where IVBA dominates is found to be clearly identifiable. It is also found that, in the presence of IVBA, there exists a poptential for misleading measurements of the internal efficiency and the optical loss. Using the high temperature sensitivity of the gain coefficient, it is possible to obtain an indication as to whether IVBA is affecting the experimental results. In order to study the temperature dependence of the threshold current, a derivation of the empirical Tmax relation is given which provides meaning to the fitting parameters. The experimentally determined Tmax is shown to correlate with the Tmax predicted from data obtained by the IVBA model for a range of lengths and well numbers. The conditions under which the Tmax relation is valid are determined. The gain coefficient for lasers having five quantum wells is calculated theoretically and compared to the gain coefficient obtained experimentally both from the IVBA model for efficiency and from the length dependence of the threshold current density. The efficiency method agrees with theory to within experimental error. However, the threshold method yields a value which is approximately one-half of the theoretical value, a phenomenon which has been previously observed in the literature but for which no explanation has been found. Direct measurement of the below-threshold output spectrum of the lasers demonstrates that the cause of this discrepancy is an unexpected length dependence of the gain coefficient. Non-uniform carrier injection into the quantum wells is suggested as a possible physical mechanism. A potential for gross inaccuracies in the measurement of laser linewidth using the delayed self-homodyne (DSH) technique is discovered. It is found that the accuracy of the DSH technique depends on the amount of noise in the laser bias current source to a previously unappreciated degree. This is due to a combination of the high FM sensitivity of semiconductor lasers and the long optical delay lines required by the DSH technique. The errors behave in a manner usually associated with intrinsic 1/f noise, and can cause all of the associated properties of residual linewidth, floor, non-Lorentzian lineshape, and premature re-broadening of the laser line. Guidelines for proper filtering of the current source are given to ensure accurate measurements.</p> / Doctor of Philosophy (PhD)
| Identifer | oai:union.ndltd.org:mcmaster.ca/oai:macsphere.mcmaster.ca:11375/5751 |
| Date | January 1998 |
| Creators | Prosyk, Kelvin |
| Contributors | Simmons, J. G., Engineering Physics |
| Source Sets | McMaster University |
| Detected Language | English |
| Type | thesis |
Page generated in 0.028 seconds