Introduction:
Back when KES-400A's were all the rage, in testing them with the P-I-V autoramper I noticed that at certain currents their output power slowly declined a minute amount. In normal pointer operation, this decline is barely noticeable, but it is there. Now is the time to revisit this effect, and to see if we can make use of it when testing very expensive diodes, to prevent the necessity of killing diodes and to extract the most information out of each test.
Hypothesis:
The theory goes like this: if one is able to slowly level up the laser output and measure the output sag at each level, one should be able to construct an aging curve for the laser
before it dies! The caveat is that the laser must age, not instantaneously die, to get useable data.
Protocol:
To test this hypothesis, the following procedure was used. A single ramp test was performed as usual on a PHR-803T blu ray diode. Next the diode was driven for 2 hours at 60mA while being shut off for one second at 5 minute intervals. During this time, power, voltage, and temperature data were recorded. After 2 hours, the current was raised 5mA and the cycle repeated. This test 2 hours, raise 5mA, test 2 hours procedure was performed up to 200mA (designed to eventually kill the diode in this case). After data collection, the aging of the diode during each 2 hour time span was calculated and plotted.
Results:
The diode had a typical PHR P-I-V curve:
and exhibited marked aging at output levels over 100mW
Unexpectedly, at 97mW, there was a 'reverse' aging process taking place, and between 100 and 120mW the aging varied between 0.2mW/Hr and 0. This area corresponds with the first 'knee' of the P-I-V plot.Above the first 'knee' the aging increased to 1mW/Hr. Unexpectedly, aging again ceased at 148mW after the second 'knee' but quickly increased again shortly before final failure.
Discussion:
While between 40 and 90mW the aging is negligible and probably the result of system noise, the reverse aging point at 97mW is not an anomaly. The diode did increase in output power very slightly during this time. One possible explanation is that the aging process on the die or output coupling mirror was changing the cavity properties in such a way as to push the first 'knee' further up the diode current slope, resulting in a temporarily increased lasing efficiency at that point. Above the 'knee' at 120mA, no further reverse aging levels were recorded.
Without this aging test (and our experimental knowledge gained through the sacrifice of many diodes), it might not be immediately clear that operation of the diode above the knee is detrimental to its life. However with this single test, this becomes clear. It remains to be seen if aging occurs significantly in 'kneeless' blu ray diodes before their death, but if it does, or if a knee develops spontaneously at high currents after a period of time, this aging test will find it and report it before the diode is completely destroyed.
The testing methodology utilized here, while longer running, may be superior to that of running a lifetime test at a fixed current. While the latter methodology will give results for that specific current/output power, it is possible to determine the output power at which significant aging begins with the former test technique.
