Convert exhaust heat from data centers into power

By Dr. Neil Canter, Contributing Editor | TLT Tech Beat May 2026

A two-phase refrigerant R601 (n-pentane) was found to theoretically exhibit the best balance of efficiency and environmental compatibility. 
HIGHLIGHTS
A detailed thermoeconomic model was produced that shows how placement of solar cells on a data center can facilitate the production of electricity by increasing the temperature of the waste heat generated.
A small scale, organic Rankine cycle (ORC)-based waste recovery system was developed to theoretically evaluate the efficacy of solar energy thermal boosting.
Two theoretical case studies conducted on two data centers located in the U.S. demonstrated the potential by finding that the amount of electricity generated increased when solar cells were used.
 
Heat remains a source of energy that is wasted in such processes as the generation of friction. Reducing the magnitude of heat generated is assuming a larger role in tribological processes due to the increase in electrification in machinery such as battery electric vehicles. For a lithium-ion battery to demonstrate a long-operating life, heat must be dissipated efficiently.

With the dramatic growth in employing artificial intelligence, data center construction is leading to a significant increase in demand for power, which is also leading to higher operational costs, significant heat production and carbon emissions. The need for efficient cooling has emerged as a major challenge in managing data centers. It is estimated that cooling systems account for 30%-50% of the total energy consumed in data centers. 

One option for handling the high degree of heat produced during data center operations is to figure out an effective waste heat recovery method. Dr. Laura Schaefer, Burton J. and Ann M. McMurtry Chair of Mechanical Engineering at Rice University in Houston, Texas, says, “Removing heat generated from a data center is particularly daunting because of the ultra-low temperatures involved for chip operation. The typical temperature range for waste heat is in the 40°C-60°C range. Evaporative cooling through a heat exchanger is the reason for the low temperatures.”

The organic Rankine cycle (ORC) potentially can be used to recover the heat and convert it to power. Schaefer says, “The Rankine cycle is a basic power cycle used at the majority of power plants with water as the working fluid. For the ORC, an organic fluid (such as a hydrocarbon or a refrigerant) is utilized and cycles through evaporation, expansion, condensation and pressurization.”

Even with the potential of having an ORC recover waste heat, the low temperature of the waste heat leaving a data center is difficult to recover efficiently and at a reasonable cost. Schaefer indicates that in using ORC, the base temperature is atmospheric temperature, but this figure is usually only slightly lower than the heat released by a data center.

Solar energy is becoming more cost effective as newer materials incorporated into solar cells are achieving higher efficiencies. In a previous TLT article,1 researchers developed a potential new solar material, zintl phosphide, that was found to exhibit comparable performance to the well-known solar absorber but also is very stable in ambient air. 

Schaefer speculated that increasing the temperature of the waste heat may produce a more favorable environment for using ORC to generate power. She says, “Our objective is to utilize solar power to increase the temperature of the heat generated by data centers to increase ORC efficiency and the output of electricity. We know that this technique has worked in the past when solar cells were installed on the roof of a brewery and shoemaker.”



Solar thermal-boosted ORC
Schaefer and her colleague, Kashif Liaqat, a graduate student in mechanical engineering at Rice University, produced a detailed thermoeconomic model to demonstrate the potential for placing solar cells on two of the biggest data centers located in the U.S. in very different climates. 

The researchers prepared a model with an array of flat plate collector solar cells positioned on the roof of a data center. A heat transfer circulating fluid system was designed to initially pick up heat from the solar cells (when the sun is available) and then pass through the ORC evaporator at a higher temperature due to the additional thermal boost. A simple ORC configuration was devised that included an evaporator, expander, condenser and pump to ascertain the feasibility of solar thermal-boosted ORC. 

Even though the solar thermal-boosted ORC could not be directly validated, the researchers conducted a comparative analysis of existing literature and felt comfortable that their approach is feasible. As part of this process a specific heat transfer fluid needed to be selected to be used in the model. Schaefer says, “For the ORC, we felt that the best fluid to include in our approach was a two-phase refrigerant. We calculated thermal efficiencies at a source temperature of 95°C and a sink temperature of 25°C. The fluid with the best balance of efficiency and environmental compatibility was R601 (n-pentane).”

The researchers used a small scale, ORC-based waste heat recovery system to determine that the heat transfer fluid should enter the data center servers at a temperature of 25°C and exit at 53°C. These represent typical values for modern liquid cooled server racks.

Two theoretical case studies were conducted on data centers in Ashburn, Va., and Los Angeles, Calif., to evaluate this approach. Schaefer says, “Ashburn and Los Angeles are hubs in the two regions that contain the highest number of data centers. They display different climates, which should provide an indication about whether solar thermal-boosted ORC can be effective under different operating conditions.”

Results from the case studies indicate that solar thermal-boosted ORC produced greater efficiency and higher power output. In Ashburn, the amount of electricity generated increased by 60% while in Los Angeles the increase was 81%. One parameter that impacted the results is the number of solar hours experienced by both locations annually. Los Angeles has 400 more solar hours than Ashburn. 

Schaefer points out that if the waste heat temperature increases to 75°C, an inflection point is reached where solar thermal-boosted ORC becomes less economical than conventional ORC. She notes that this is unlikely to occur because the temperature of waste heat generated by data centers is no higher than 60°C. 

The efficiency of solar thermal-boosted ORC may be increased by using parabolic trough solar collectors that can operate at temperatures above 200°C. 

Additional information on this research can be found in a recent paper2 or by contacting Schaefer at Laura.Schaefer@rice.edu. 

REFERENCES
1. Canter, N. (2024), “Potential new solar cell material,” TLT, 80 (7), pp. 12-13. Available at www.stle.org/files/TLTArchives/2024/07_July/Tech_Beat_I.aspx.
2. Liaqat, K. and Schaefer, L., (2025), “Techno-economic analysis of a solar thermal-boosted organic Rankine cycle system for data center heat recovery,” Solar Energy, 300, 113893.
 
Neil Canter heads his own consulting company, Chemical Solutions, in Willow Grove, Pa. Ideas for Tech Beat can be submitted to him at neilcanter@comcast.net.