Not everything improves with time.

The typical industrial power plant in the U.S. is only about half as energy efficient as those used in 1910, according to Sean Casten, CEO of Recycled Energy Development (RED).

In fact, the ones Thomas Edison designed were more efficient. The problem? Waste heat. Edison's plants weren't actually very efficient as making electric power, Casten noted, but he sold the heat generated during operations, which boosted the overall. A full two-thirds of the fuel burned to generate power in today's power plants – which for the most part were built in the mid-1960s with 1850s technology – gets lost he asserts.

"Edison's were 50 percent efficient," he said.

Although it can be tricky and expensive to harness, waste heat is getting increased focus as a source of power in both the U.S. and China, mostly because of the quantities of heat out there. A study conducted by Lawrence Berkeley National Labs estimated in 2005 that the U.S. alone has 100 gigawatts of untapped electrical capacity in the form of waste heat that annually could produce 742 terawatt hours of power. That's bigger than the solar fleet, which gets measured in megawatts. UC Berkeley's Arun Majumdar estimates that the U.S. consumes 100 quads (100 quadrillion BTUs) of energy a year and 55 to 60 percent of it gets dissipated as waste heat (see Tapping America's Secret Power Source).

Generating that heat, naturally, also means excess greenhouse gases. Approximately 42 percent of carbon dioxide emissions come from power plants, said Casten. If power plants are truly only 33 percent efficient, that means that 28 percent of the carbon dioxide output in the U.S. could be eliminated without crimping the national lifestyle. Cars only account for 19 percent.

Some companies, such as Cypress Semiconductor and GMZ Energy, are trying to develop thermoelectric materials. These are semiconductors that, wrapped around a steam pipe, could convert ambient heat to electricity.

Companies like Israel's Ormat and Westmont Ill.-based RED – which raised a $1.5 billion fund with Denham Capital Management to take on waste heat projects – are largely focusing on the more traditional techniques. Namely, exploit excess steam pressure and heat to turn a turbine, power heating systems or boil more water. It all depends on the circumstances on the ground. Waste fuels can also be harvested.

"Everything is custom," Casten said.

While the bulk of waste heat is generated in large plants, there are also smaller pockets. Natural gas pipelines are equipped with booster stations which maintain the pressure inside the pipeline as the gas travels from one point to another. Each one on average requires 10 megawatts of power but gives off about 3 megawatts worth of waste heat. 

"There are opportunities all over the place – silicon manufacturers, cement, steel," he said. "They have high volumes of fairly high quality waste heat."

One of the company's more dramatic projects will go online in 2010. West Virginia Alloys, a silicon manufacturer, will install a waste heat recovery system that will generate 45 megawatts of electrical power. The company only uses 120 megawatts right now. (Put another way, the company only really needs 75 megawatts for its operations and is currently burning off 45 megawatts.)

To date, the big challenge has been cost. Most industrial-scale waste heat projects cost between $5 and $50 million. That's too high for most to pay out of capital budgets and too low for a public financing project.

"There's a huge Goldilocks problem," he said. To get around this, RED pays for any waste heat recovery system it installs and then gets paid for energy savings under long-term contracts.

"It is becoming mostly easier for a whole lot of reasons," Casten said.