IBM Uses Supercomputer Cooling Technology on Concentrated Solar PV

IBM’s Zurich laboratory is incorporating cooling technology used in supercomputers into its “Sunflower” concentrated solar PV technology.

Concentrated solar PV (CPV) generates electricity from sunlight by using lenses and mirrors to focus sunlight onto high efficiency solar cells. One of the traditional challenges with CPV is that the modules get very hot and requires cooling. As the temperature of the CPV module rises, its efficiency decreases.




CPV systems are usually segmented by how much they concentrate the sunlight – measured in “suns”. High concentration systems are usually rated above 1,000 suns. As you can imagine, this level of sunlight concentration produces a lot of waste heat.

Image Credit: www.airlightenergy.com / Sunflower System

The IBM CPV system can safely utilize a 2,000 sun magnification thanks to the same cooling technology that is implemented in supercomputers. This cooling system uses a network of capillaries carrying water in order to keep the CPV system cool.

Image Credit: www.engadget.com / IBM Aquasar Supercomputer Cooling System

The magic of this cooling system is that it brings the cooling water extremely close (a few tenths of a micrometer) to the hot components without the water actually touching the electronics. This close proximity to the heat source allows for very effective heat transfer. Based on IBM estimates, the cooling system can keep the CPV system safe up to a light concentration of 5,000 suns.




With this advanced cooling technology, the CPV system can capture up to 80% of the incoming solar energy. However, a good portion of this captured energy is in the form of hot water. The IBM team foresees the hot water getting used for activities such as desalination. If used for desalination, it is estimated that each square meter of CPV module area can produce 30-40 liters of drinkable water each day.

Image Credit: www.airlightenergy.com / Desalination Install Location

The team’s ultimate goal is to reduce the cost to under $250 per square meter. IBM hopes to have this technology commercially available by the end of 2017.

Leave a Reply

Your email address will not be published.