Recent quantum dots are three times better than thermoelectric devices from the mid-1990's. There are quantum wells that get 4.5ZT.
Vehicle efficiency technology gets about 176 million per year The new thermoelectronics are not magic technology. They are part of one of the highest value US government research programs. The advanced versions should be ready in 2014. (Actually the first versions will be in BMW car in 2010. Crude versions are already used for beer refridgerators and are used for car seat warmers).
It can be seen on this chart that 4.5 ZT gets 38% efficiency for 500 degrees and 54% efficiency for 1000 degrees
Recent quantum well samples have achieved 4.5 ZT.
The 4.5 ZT would have 17-25% efficiency under more common temperature ranges
Nanostructured material could achieve ZT of 10-15.
The 2014 target is a ZT of 10 which would make thermoelectric conversion of 35% for temperature ranges typically found in cars. The goal is raise diesel engine efficiency from 38% now to 60%. More than 50% better without adjust the weight of the vehicle.
Powerchips claim 70-80% carnot efficiency.
Quantum well thermoelectronics 28 page presentation
A 55 page presentation on using thermoelectronics for cars and other applications. It also discusses scaling up the production processes.
Another set of slides on making diesels more efficiency
A presentation on a near term target of 10% efficient thermoelectric system.
2006 advanced engine project status review.
The target for engine efficiency increase was 10%– 13% which was met by achieving 10.5% efficiency improvement when the engine was operating over a steady-state cycle. However, the achievement in the road cycle was 8.2% increase in efficiency. When the engine was coupled to a CVT the efficiency increase was only 5%.
A comprehensive study showed that with a 5% increased efficiency and the fuel price of $2.50/gal, it takes a little less than 1.5 years to compensate the initial charges added to the engine due to the efficiency recovery units. This study has been done for a 10,000unit production volume per year. With increase in annual production numbers and considering a learning factor of 0.87 the pay back time will reduce significantly.
A thermoelectric addon device costing a few thousand dollars would pay for itself in 9-18 months. Better systems will perform even better, cost less and would become economic as add-ons for smaller trucks and cars.