In industrial process monitoring environments, it is important to maximize plant productivity by having distributed sensing components of high reliability. Fiber lasers for moisture sensing leverage high reliability platforms developed for the telecommunications and optical data storage industries in order to project mean times to failure (MTTF’s) in excess of 100,000 hours.

Test Conditions

Fiber laser operation is enabled by a high power “pump” semiconductor diode laser. The reliability of this component is key to the reliability of the fiber laser system. Accelerated aging studies have been performed to evaluate the reliability of the pump laser component. A total of 28 pump lasers were evaluated in this exercise. The lasers were operated at 70^oC at a constant power which resulted in an approximate 50% overstress condition relative to normal operating current for the pump lasers. End of life for the pump lasers was defined very conservatively as a 20% increase in the operating current required in order to maintain this constant power condition. The actual operating current is monitored periodically during the course of the test.

Accelerated Aging Results

During 1000 hours of operation under these conditions, slight degradation of the performance characteristics is observed, but no failures were observed during this time. Linear regression is therefore used to extrapolate the degradation slope for each device to project the time to achieve 20% increase in operating current. These extrapolated times to failure are tabulated and subjected to statistical analysis.

The statistical analysis demonstrates that the failure rate in time fits very well to  the lognormal distribution, which is generally associated with wear-out failure mechanisms. From the Lognormal distribution analysis, we determine an MTTF at the elevated stress condition of 24,299 hours.

Extrapolation to Normal Operating Temperature

In order to translate this measure into failure rates under expected operating conditions, we must implement a model for the temperature dependence of the reliability.  It is well known, that these lasers commonly exhibit temperature dependence to their reliability that is described well by an Arrhenius model,[1] where failure statistics are exponentially dependent on temperature. In order to extract the “activation energy” which determines the temperature dependence of these devices, it is necessary to evaluate the aging characteristics at multiple temperatures. Since the present study has only used a single temperature, we will rely on literature reported values of the activation energy to make our projection[2]. The literature shows us different degradation processes having associated with them different activation energies.

With our present uncertainty as to the relative importance of the different degradation mechanisms in these devices, we select the lowest of these activation energies to further our analysis. This will result in the most conservative reliability projection. Taking this approach and using the 0.3eV activation energy, we project a room temperature MTTF for these lasers of 112,353 hours, or somewhat in excess of 12 years, under continuous operation. We are thus quite encouraged that fiber lasers based on these pump lasers will meet the reliability requirements for a variety of demanding industrial applications.

Conclusions:

In summary, accelerated aging studies have been used to extract failure rates for fiber laser modules based on the failure rates for the diode laser pump components. We have obtained an MTTF of 112,353 hours for our pump lasers. We emphasize that measures have been taken to maximize the likelihood that this is a conservative estimate relative to the actual MTTF. Firstly, the test is conducted under an electrical current stress condition of 50% greater starting current than normal operating current. Secondly, we have set the failure criterion as a 20% increase in operating current. With appropriate control circuitry, operating current increases of twice this value can be readily accommodated. In our model, this would translate directly into a doubling of the MTTF. Finally, in extrapolating the results to operating temperature, we have assumed the lowest physically plausible value of the activation energy. It is therefore reasonable to expect that actual field performance of these devices will exceed the estimate that we have projected from this study.

Note:   This Application Note is available for comment and discussion.

References:

[1]   J. Spencer, “Calculating laser diode reliability,” Global Telecommunications Conference, 1988, and Exhibition. 'Communications for the Information Age.' Conference Record, GLOBECOM '88., IEEE, 1988, pp. 63-69 vol.1.

[2] T. Kim, D. Kim, K. Kim, K. Jang, G. Moon, J. Park, S. Lee, M. Kim, and J. Koh, “Continuous wave operation of a 250 mW AlGaAs laser diode,” Solid-State Electronics,  vol. 49, Oct. 2005, pp. 1674-1677.