Refrigeration, Thermoelectricity. Schematic Drawing of a Simple Thermoelectric Device

The scientific principles underlying thermoelectric refrigeration were understood by the mid-nineteenth century. In 1822, German scientist Thomas Johann Seebeck discovered that a needle would move when held near the junction of two dissimilar metals maintained at different temperatures. Seebeck misidentified the effect as magnetic, but Hans Christian Oersted, the father of electromagnetism, and James Cumming, a Cambridge chemist, correctly categorized Seebeck’s discovery, known as the Seebeck effect, as an electrical phenomenon.

A Parisian clockmaker, Jean Charles Athanese Peltier, made the second important discovery in the field of thermoelectricity in 1834 while performing an experiment to measure the conductivity of bismuth and antimony. As he had predicted, the temperature at the junction of the two conductors changed with the application of an electrical current. He also discovered that the temperature of the metals differed at their ends and that the current absorbed heat at one end and released it at the other.

Like Seebeck, however, Peltier misinterpreted his results. With an ingeniously simple experiment—placing a drop of water at the junction of the two conductors and watching it freeze and melt depending on the direction of the current—Emil Lenz first demonstrated and correctly interpreted the Peltier effect in 1838.

From the time Lord Kelvin clarified the relationship between the Peltier and Seebeck effects in 1854 until the 1950s, research on thermoelectricity moved along at a languid pace. Bold efforts to develop practical devices based on thermoelectric principles met with little success. Known materials allowed for efficiencies of just 1 percent, far too low to justify any serious development efforts.

The most important contribution to the study of thermoelectricity in the early twentieth century came from E. Altenkirch, a German scientist, who determined that progress in thermoelectricity depended on finding materials that exhibited three characteristics — (1) high electrical conductivity, (2) high voltage capacity, and (3) low thermal conductivity. Since he knew of no such materials, Altenkirch abandoned his search.

Thermoelectric researchers developed a greater understanding of the possibilities of semiconductors in the 1930s, and the positive-negative (p-n) junction, a crucial component of thermoelectric devices, was developed in 1942.

However, not until the 1950s, after researchers at the Soviet Institute of Semiconductors declared the inevitability of a thermoelectric breakthrough and H.J. Goldsmid of General Electric’s London laboratory provided a rationale for studying the heaviest semiconductor compounds such as bismuth telluride and lead telluride, was there a concerted worldwide effort to overcome the technical barriers to the development of thermoelectric devices.

At the core of technologies based on thermoelectric principles is a simple solid-state device that facilitates the exchange of thermal and electrical energy through the movement of electrons and holes (Figure 27). An electrical current passing through the device will draw heat from one side of the p-n junction and release it on the other side.

Figure 27. Schematic drawing of a simple thermoelectric device

If the current is reversed, so too is the heat transfer process. Employed in this way, as a solid-state heat pump, the thermoelectric device can be used for refrigeration, air conditioning, and heating. The same device also can be used to exploit temperature differentials of opposite sides of the p-n junction to produce electricity. In thermoelectric generators, thermal energy is transformed into electrical energy as it passes from one side of the device to the other.

The 1950s witnessed an explosion of interest in thermoelectric research, especially for refrigeration. Optimistic about the potential of thermoelectricity, researchers in corporate laboratories convinced electrical industry executives and government research agencies to devote substantial resources to thermoelectric research and development for military and commercial applications.

Researchers worldwide published twice as many papers on thermoelectricity in the four years between 1956 and 1960 as they had during the previous 130 years. In the U.S. alone, over 75 organizations, including research universities, government laboratories, private research institutes, and major corporations, maintained substantial thermoelectric research programs.

Major electrical appliance manufacturers, such as General Electric, RCA, Westinghouse, and Whirlpool, pinned their hopes on the large, growing, and highly profitable market for standardized household refrigeration units in the late 1950s. Thermoelectric refrigerators, however, could not compete in price or efficiency with existing mechanical refrigerators. Furthermore, manufacturers were unwilling to adjust their expectations and to shift their focus to smaller markets for specialty devices in which thermoelectric units might have been competitive.

As the promise of big profits in the household refrigeration market evaporated by the mid-1960s, all the major appliance manufacturers and all but a handful of specialty firms abandoned thermoelectric research and development for the consumer market. Research on military applications of thermoelectricity continued nonetheless and proved productive over the long term. Thermoelectric devices have been used to power electronic equipment in spacecraft, to cool submarines, and to provide a number of other military applications in which cost and efficiency are of limited concern.

Interest in commercial applications of thermoelectricity refrigeration reemerged in the 1990s. Niche markets developed for portable coolers, and at least one automobile manufacturer embedded thermoelectric cooling units in its car seats.

Manufacturers in the U.S., Europe, and Japan began to experiment with using thermoelectric chips to cool computers and medical devices such as blood analyzers. Most of these devices used bismuth telluride alloys in simple p-n junctions, which were still too inefficient to employ in larger units such as household refrigerators.

The technical breakthrough that researchers had been seeking for 40 years appeared to have arrived in 2002, through U.S. government-funded research begun in 1993 at the Research Triangle Institute (RTI) in North Carolina. RTI researchers used a thin film deposition process to lay down thermoelectric alloys in alternating layers.

The microscopic structure, known as a superlattice, enhances electron flow while hindering heat transfer, resulting in efficiencies two-and-a-half times greater than allowed with simple p-n junctions. Whether devices based on superlattice structures will finally bring about the long-promised thermoelectric revolution in refrigeration and cooling remains to be seen.