We consider new possibilities for cooling by vacuum tunneling. We examine a nanogeometry and find that large cooling currents can be obtained by a combination of energy selective tunneling of electrons and thermionic emission. The energy selective tunneling is a result of the special form of a potential barrier which has wider gap for low energy electrons, which results in electrons above the Fermi level being the principal tunneling component. Numerical calculations show that available material with work functions about 1.0 eV are useful for cooling. For gaps of 5-15 nm, which are well within the present state of the art, only a small external voltage ͑1-3 V͒ is required to create large currents and a useful Peltier coefficient of about 0.3, and cooling power of 100 W/cm 2 .

The thermoelectric figure of merit is calculated for a compound material comprising thin semiconductor and wider metallic layers. The layers are perpendicular to the direction of current. The semiconductor barriers exclude electrons with energies ε<μ from the current. This exclusion increases thermopower. One may obtain a material with a very high ZT if the distance between the barriers is on the order of the energy relaxation length. This material should have the resistivity characteristic of a metal and the thermopower characteristic of a semiconductor. An additional significant rise in ZT can be achieved by increasing the contact area at the metal–semiconductor interface.

A new concept for converting heat energy to electrical energy using thermionic energy converters (TECs) is proposed. It has potential for increasing the Carnot efficiency to an unprecedented 80% when the TEC is combined as a topping cycle with a conventional external combustion engine. The optimal electrical power density, 5–20 W cm−2, is satisfactory for many applications including stationary power generators and propulsion drives for vehicles. But unfortunately the substantial losses, ∼50% of the output power, that have been needed to compensate the space charge have prevented the TEC from realizing its thermodynamic potential. We present two ways of overcoming this limitation. Both utilize a triode configuration (rather than a simple diode) with a longitudinal magnetic field. In the first method the magnetic field together with the grid separate the relatively few hot electrons required for volume generation of Cs ions used for the space charge compensation from the majority of the thermal electrons which constitute the current to the collector. In the second method the Cs ions are generated on the surface of the grid wires and injected into the space between the electrodes. Grid wires with high work functions are required for this.

We have completed an investigation of cooling at room temperature by thermionic emission. The use of a small nm-sized gap lowered the vacuum barrier between the electrodes, enabling emission from surfaces with work functions of ∼1 eV at room temperature. We utilized a microfabricated cantilever with a cesiated metal coating on the tip, and an integrated thermometer to initiate and control an emission current of 1–10 nA, and to detect the resulting temperature changes. Using a lock-in technique, temperature changes of 0.1–1.0 mK were observed, corresponding to cooling power of 1–10 nW. The amplitude of this signal and its dependence on emission current and bias voltage are in good agreement with our model. Possible applications for cooling and energy conversion are discussed.

No abstract