Thermal Management of Laser Diodes

High-power laser diodes can generate a great deal of heat. Even for laser diodes operating with 70% or higher efficiency, a large amount of applied energy is converted to heat. Since the performance and optical efficiency of a laser diode is dependent upon operating temperature, maintaining a stable temperature close to room temperature (about +25°C) is instrumental in achieving high reliability.

Liquid-cooled components such as microchannel coolers (MCCs) can provide dependable thermal management and contribute to long operating lifetimes for high-power laser diodes. But what are the expected operating lifetimes of the MCCs themselves? By performing lifetime testing on MCCs without laser diodes attached, it is possible to determine useful lifetimes for these components, of 20,000 hours or more. The quality of the cooling water in the MCCs has a great deal to do with the duration.

In a typical high-power laser diode array, liquid-cooling components help dissipate heat generated by the diodes. An MCC is one such component, a miniature heat exchanger that is also used for electrical connections to the laser diodes as busbars. It is formed with microchannels only a few hundred microns in size and comprised of thermally conductive material, such as high-quality copper. High-power laser diodes are mounted on MCCs with the semiconductor’s p-side facing down (Figure 1). In an array, the MCCs may be stacked in horizontal or vertical configurations. An electrical circuit is assembled by mounting an MCC with an isolator and a top lid for mounting wire bonds for the diodes’ n-contacts. Inlet and outlet holes in the MCCs enable the flow of cooling water through the microchannels to help dissipate heat.

1. A high-power laser diode busbar is mounted on an MCC for proper thermal management.

Since the cooling water is surrounded by such an effective electrical conductor, deionized (DI) water (H2O), with its low electrical conductivity (and high electrical resistance), is typically used for MCC cooling purposes. Not only can the DI H2O minimize thermal conduction resistance, it also helps reduce the possibility of the cooling water conducting current and causing a short circuit. Although copper and DI water combine for excellent cooling properties in MCCs, water also causes electrochemical reactions with the copper that can result in corrosion and degradation of the metal’s mechanical and electrical properties.

Lifetime MCC Testing

By performing lifetime testing on MCCs, it is possible to better understand these breakdown mechanisms. Lifetime testing was focused on the material properties of the MCCs, without connection to high-power laser diodes and without considering the impact of thermal energy from the diodes on the MCCs. Testing was concerned with the different types of corrosion effects on copper over an operating lifetime and how the cooling water through the MCCs affects their structural integrity and useful operating lifetimes.

2. Typical set-up for a 5 diode laser vertical MCC stack

Oxidation effects from water can be particularly corrosive to copper, especially as copper wears down over time and forms copper solutions with the water within an MCC. If the electrode potentials of these solutions are low, the corrosion will be minimal. But as the electrode potential increases, reactions with or without applied electrical potential can increase, and increase the rate of corrosion. One way to combat this is the use of different materials as coatings for the copper microchannels, to act as electrical barriers between the copper surfaces and the DI H2O without sacrificing thermal conductivity.

Copper degradation can occur as an electrochemical reaction with water and applied electrical energy. But even without applied potential, water alone can cause copper erosion. As copper dissolves in an aqueous solution over time, the dissolved copper will be deposited on different surfaces without an MCC, with the possibility of obstructing the passages through the cooler and slowing the flow of cooling water. Atmospheric carbon dioxide (CO2) can also become dissolved in the cooling water in MCCs, decreasing the pH of the water solution and contributing to complex corrosion and degradation effects. Thus, the quality of the cooling water is a critical element in minimizing the breakdown of MCCs.

3. A total of 80 coolers with three different MCC configurations were evaluated as part of the lifetime testing.

While they contribute to corrosion, electrochemical reactions between copper and water are also acting to protect the copper from excessive oxidation. As part of a passivation process, solid copper from copper oxides and copper hydroxides formed in the reactions of the copper and water also cover the internal copper surfaces, serving to block the effects of further passivation processes. Unfortunately, the passivation layer is not as corrosion resistant as the bare copper and it can be removed at high fluid flow rates (1 to 2 L/minute). This allows the bare copper to come in contact with electrolytic, oxygen-containing water to form another oxidation layer on the copper, in a process that constantly repeats. But by employing optimal liquid flow rates, the passivation process can be controlled and the operating lifetime of an MCC can be extended.

To better understand the lifetime mechanisms of microchannel coolers, three different microstructures (CUR 3/1, CUR 5/1, and CUR 7/1) were evaluated at different flow rates: 0.2, 0.3, and 0.5 L/minute in cooling water with constant conductivity of 5 μS/cm. The cooling water was maintained at a constant temperature of about room temperature (+25°C), at typical conduction for electronic testing. The analysis was performed to learn more about the lifetime effects of the microcoolers; the heating effects of devices such as laser diodes were not considered as part of the MCC lifetime testing.

4. The three different MCC configurations evaluated for lifetime testing were the CUR 3/1, CUR 5/1, and CUR 7/1 coolers with 5, 6, and 5 layers, respectively.

A test system was assembled with a single loop and bypass line (Figure 2). A total of 80 different MCCs were evaluated as part of the lifetime testing, using a total of 300 MCCs (Figure 3). The single loop includes a pump, piping to the laser diode stacks, the stacks and associated fittings and fixtures, the return piping from the stacks, the chiller, a 15-μm filter, and a DI water reservoir. The bypass consists of a control valve and an ion-exchange bed.

The system was filled with distilled water with conductivity of 5 μS/cm, automatically controlled by a mixed-bed ion exchanger. The water temperature was maintained at +25°C (room temperature). Constant flow rates of 0.2, 0.3, and 0.5 L/min were achieved for stacks with 5, 8, and 12 MCCs for three different MCC designs (Figure 4). Testing was performed to reveal the effects of as much as 15,000 hours under different cooling water conditions.

The testing allowed the initial conditions of the three different MCCs (Figure 5) to be compared to the same MCCs after 5,000 (Figure 6) and 15,000 hours of service (Figure 7). As the images of the MCCs reveal, thicker oxidation layers form on the surfaces of the microchannels with decreasing water flow velocity and increasing time. In areas where the microchannels have been narrowed to achieve higher fluid flow velocities, lower amounts of passivation and greater areas of polished copper surfaces are evident.

5. This is how the laser diode bar area appeared prior to lifetime testing (o h test time).

In the CUR 3/1 cooler, deposits are visible in the transition between the microchannel region and the wider reflux channel. Under higher magnification than shown here, deposited grains exhibited a red color, indicating a copper oxide (Cu2O) composition. These deposits are formed from copper that has been removed from various areas within the MCCs by high water flow velocity. These high-flow-velocity areas are typical for the CUR 3/1 MCCs, with the lowest thermal resistance of the three experimental MCCs. The copper deposits and the associated removal of copper from other areas within the CUR 3/1 MCCs are the greatest at the highest flow rates, 0.5 L/min N which has the lowest thermal resistance. As expected, these deposits and the corresponding copper removal are highly pronounced at the highest flow rates of 0.5 L/min.

6. The reference photos compare the three MCC configurations after 5,000 hours for three different water flow rates.

In comparison, the CUR 5/1 MCC shows no deposits are visible in the transition between the microchannel area and the wider reflux channel. This MCC has a larger channel cross section than the CUR 3/1 MCC, which results in lower water flow velocity for the CUR 5/1 MCC than the CUR 3/1 design. As with the CUR 3/1 MCCs, there were no visible geometrical variations in the CUR 5/1 microchannel structure. For the CUR 7/1 coolers, the only visible changes were in the formation of the copper coating, caused by this MCC’s flow-optimized structure. As with the other two MCC designs, there were no changes in the geometric structure.

Controlling Water Parameters

This lifetime testing was conducted to learn more about the effects of the cooling water on the metal surfaces and structures of MCCs for cooling laser diodes. It did not consider the effects of applied potential or temperature on the operating lifetimes of these MCCs. While the effects of electrical potential used to drive laser diodes may be negligible on the microchannel coolers, the effects of applied heat from the laser diodes can shorten the useful lifetimes of microchannel coolers if that heat is not effectively dispersed.

7. The reference photos compare the three MCC configurations after 15,000 hours for three different water flow rates.

Still, as the results demonstrated even after 15,000 hours, pure copper coolers can be sufficiently reliable under standard conditions for water flow rate, temperature (room temperature), and conductivity. When comparing the initial appearance of the three different MCCs with their conditions after 5000 and 15,000 hours, no changes are visible in the microchannel structures. It is safe to assume that all three MCCs will function consistently over that time, with no degradation in cooling performance. Given the lack of changes in the MCCs, and provided the water flow rate, temperature, and conductivity are also properly maintained, these three MCC designs are likely usable for an additional 15,000 hours with no degradation in performance.

In summary, if the water parameters are properly controlled, copper microchannel coolers can have a long lifetime (greater than 20,000 h) over a range of flow velocities and cooler designs. In contrast, copper microchannel coolers with less than ideal cooling water and flow conditions can suffer catastrophic failure after only hundreds of hours of operation.

This article was written by Manfred Goetz, Product Marketing Manager, and Andreas Meyer, Product Development Manager, Rogers Germany GmbH Power Electronics Solutions (PES) (Eschenbach, Germany). For more information, Click Here .