[an error occurred while processing this directive] [an error occurred while processing this directive]

PTB >> Improving the Performance of High-Power Diode Lasers

High-power, 808-nm diode lasers are used to pump solid-state lasers, which are employed in many scientific, microelectronics, materials processing, biomedicine, and metrology applications. Achieving improved performance, extended reliability, and lower cost of ownership for these solidstate lasers often translates into developing diode lasers with higher output powers and longer lifetimes. This article reviews several recent technical innovations that enable improved diode laser performance and reduced cost per watthour of pump diodes.

Characterizing Performance
The compact nature of diode lasers is extremely attractive from an applications standpoint, but also presents a number of design and manufacturing challenges in terms of device reliability. While virtually every laser creates thermal, electrical, mechanical, and optical stress, the compact size of diode lasers results in a much higher power/size ratio than any other laser type, making these stress issues more severe. Examples include facet damage due to high photon density and junction degradation from high current density and temperatures, as well as subsequent mechanical stress caused by thermal mismatch of materials.

These issues have largely been dealt with for the lower power diode lasers used in data storage and telecom applications; however, development is ongoing for high-power diodes at 808 nm. Tremendous reliability improvements have already been achieved through the use of aluminum-free active area (AAA) construction, which first enabled the industry-standard, 1-cm room temperature, conduction-cooled diode laser bar to go above the 40-watt, 10,000-hour lifetime mark. Current research and development efforts in 808-nm lasers involve advancing several different aspects of diode design, fabrication, and packaging in order to further reduce the cost per watt-hour of these diode laser bars.

To understand the significance of these efforts, it is first necessary to understand how performance for 808-nm pump diode lasers is typically defined. These lasers are produced as single emitters, linear bars (a monolithic chip with multiple emitters), or two-dimensional arrays (stacks of bars). Output power (in watts) is typically the most common specification for diode laser devices, but specifying power alone is insufficient. Bar power, for example, can be increased simply by adding more emitters and making the device longer. Thus, power is usually specified in conjunction with brightness — namely power as a function of the size and number of emitters. In many applications it is advantageous to maximize brightness, as this enables refocusing of the output to the highest possible power density.

In addition to power and brightness, it is also critical to specify the operating lifetime of the diode laser. This mean time to failure is defined-with a 90% confidence level-as the mean time that passes before the power drops by 20%. Because the major degradation mechanisms (e.g., facet oxidation) that reduce lifetime are power dependent, there is a trade-off between these parameters. Consequently, lowering the output power of a given device extends its lifetime significantly. For this reason many laser manufacturers operate the pump diodes at reduced power levels in order to extend their lifetimes well beyond 20,000 hours.

It is also necessary to specify the operating temperature at which these performance specifications are being measured or certified. Because facet damage and the spread of flaws, such as lattice defects, are all highly temperature dependent, a 10°C rise in operating temperature typically results in a 50% drop in lifetime given the same operating power and brightness.

Damage to the output facet has been an important performance-limiting factor in 808-nm diode lasers. In the diode laser manufacturing process, the facet is produced by cleaving the wafer and then applying a multilayer coating. This coating
provides the appropriate level of reflectivity and some oxidation protection for the cleaved surface; however, some oxidation can still occur with conventional multilayer coatings. With highphoton fluxes, this oxidation can spread back from the surface, causing performance degradation and ultimately device failure due to catastrophic optical mirror damage (COMD).

The COMD threshold power is a very useful parameter associated with the power-lifetime product (watt-hours) at a given brightness and operating temperature. COMD can be quickly measured and induced at any time by raising the device current and hence, the output power. It is typically referenced after a specified burn-in period (e.g., 200 hours) because COMD decreases with time. As a rule of thumb, a safe operating power level for long-term operation of diode lasers is some fraction of the COMD threshold value.

Expanded Mode Designs
Figure 1 shows a schematic cross-section of a typical, double hetero-junction diode architecture. The laser light is created and amplified in the thin junction and propagates through the junction and waveguide layers. In traditional 808-nm designs, this action produces a single spatial mode in the vertical direction, with a beam cross-section on the order of a micron. With high-power diodes now producing over 2 watts from each emitter, this output corresponds to very high intensity both within the chip and at the output facet.

A large optical cavity (LOC) design has been one approach used to reduce the intensity in high-power laser diode devices. This design utilizes thicker waveguide layers in order to increase vertical mode size. Recent investigations have successfully employed an alternate device structure device to further expand the optical mode and reduce the power intensity in high-power laser diodes. Typical data indicates an increase of over 30% in COMD threshold for an “expanded mode” device relative to a typical control design group.

The expanded mode design offers many benefits over other laser diode designs. Because the expanded mode design lowers photon density at the output facet and in the cavity, it improves the trade-off between lifetime and output power. This allows expanded mode devices to either be operated for much longer lifetimes or at higher powers than diode lasers based on more conventional architecture, as shown in Figure 2. This data demonstrates the ability to achieve 270 watts from a 1-cm bar. The expanded mode design also reduces beam divergence, simplifying the optics required for fiber-coupling or other beam re-imaging tasks.

NAM: Reducing COMD
Another way to extend facet lifetime is to lower the absorption in the facet and its immediate vicinity. Several techniques have been developed for this purpose and are often collectively referred to as non-absorbing mirror (NAM) technology.

In the vicinity of the laser facet, photon re-absorption and subsequent localized heating occurs, setting up a positive feedback loop for increased photon re-absorption. The goal of NAM is to shift the band-gap of the junction at and behind the vulnerable output facet, so that resonant absorption cannot occur in this region. Quantumwell disordering is used to change the band-gap, which is achieved either by ion implantation and/or impurity diffusion. The end result is a blue shift of the facet region relative to the junction band-gap, making the facet region transparent at the lasing wavelength. This effect is also sometimes referred to as Impurity-Induced Layer Disordering (IILD).

Diode laser manufacturers are also investigating several novel facet coating and treatment protocols. The goal of these efforts is to reduce the effects of laser damage and oxidation at these critical facets. Several of these proprietary techniques already show great promise.

Lower Thermal Resistance
Because many damage mechanisms are highly dependent on temperature, it is important to minimize a laser’s operating temperature. One way of accomplishing this is by lowering the thermal resistance (Rth) of the device. Thermal resistance is the increase in temperature as a function of output power, and is measured in units of °C/W.

A physically longer cavity can reduce thermal resistance (e.g., a 1.5-mm-long cavity instead of the typical value of 1 mm). Modeling results and empirical data both show that Rth decreases with increasing cavity length.

As with the expanded mode designs, this approach allows the device to be operated at increased power for the same overall lifetime, or at the same power for a longer lifetime. However, the cavity cannot be lengthened indefinitely. Wafer economics, for instance, dictate that as the device size increases, fewer units can be produced on a single wafer, which increases cost. Cavity losses also place an upper limit on useful device length. This is because the defect density is never zero; a longer cavity will contain more defects and decrease the overall manufacturing yield.

Performance Milestones

The most common 808-nm diode laser form is the 1-cm bar, which typically consists of 19 to 60 separate emitters. DPSS laser applications require diode laser manufacturers to deliver these bars at a rated power level with a minimum lifetime of 10,000 hours. As recently as 1998, this allowed a maximum power rating of only 20-watts/bar. Since then, the AAA material system has enabled the commercialization of the 40-watt bar, and other incremental technical improvements have resulted in higher powers. As the technology described here is implemented for 808-nm diodes, a steady increase in the power-lifetime rating will continue. As shown in Figure 3, the state of the art is the 100-watt power level, with QCW power levels predicted to pass the 200-watt milestone in 2005.

In conclusion, high-power laser diode technology has made tremendous performance strides over the past ten years, and already delivers higher overall efficiency than any other laser type. Further improvements in this technology will enable enhanced output from diode-pumped, solid-state systems at reduced cost per watt-hour.

This article was written by Mark Mondry, Paul Rudy, and Hailong Zhou for Coherent Inc., Santa Clara, CA. For more information contact Paul Rudy at paul.rudy@coherent.com. Visit Coherent online at www.coherent.com.


home | about | subscribe | advertise | editorial | contact us