Through aggregated analysis of waste heat data from the literature, ARPA-E found that most waste heat (~75%) is low-grade (≤230oC). This temperature regime is not easily converted to usable work as its exergy is roughly a third of the total heat generated (Figure 2); a Carnot analysis yields a maximum efficiency of only ~40% (e.g. 25°C cold-side). A majority of the higher grade waste heat resides in the 230°C to 400°C range. This can be seen in Figure 2, which shows the cumulative percentage of total waste heat as a function of temperature differential. Figure 2 also illustrates the cumulative percent of the total maximum work potential at each temperature difference. The maximum work potential is defined here as the amount of waste heat available at any temperature multiplied by the Carnot efficiency at that temperature. From Figure 2, it can be seen that approximately 85% of work potential from waste heat sources across all sectors in the United States comes from waste heat sources at or below 400°C. Thus, ARPA-E is keenly interested in waste heat conversion in this temperature range.
Several technologies exist to realize the opportunity of lower-grade waste heat recovery, and are typically either mechanical, solid state, or hybrid systems. Examples of mechanical systems include the Organic Rankine cycle, and Kalina cycle, while examples of solid-state devices include thermoelectric generators, piezoelectrics, and multiferroics among others. Mechanical systems are often limited by their complexity, large footprint (e.g. size/mass), and parasitic power requirements. These are particularly challenging limitations for waste heat recovery in the transportation or mobile sectors, where a majority of the opportunity lies (Table 2). Solid-state devices have advantages in mobile applications due to their small footprint and lack of complexity and parasitic power requirement. Unfortunately, existing solid-state technologies have low efficiency and high cost. However, there may exist an opportunity to greatly improve most solid-state technologies.
For example, one might seek to improve the performance per unit cost of a thermoelectric generator (TEG). To date, TEG devices remain very inefficient